100% found this document useful (10 votes)

14K views801 pages

JOHN ZINK Combustion Handbook

Uploaded by

Baher Elsheikh

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

100% found this document useful (10 votes)

14K views801 pages

JOHN ZINK Combustion Handbook

Uploaded by

Baher Elsheikh

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 801

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
C M Y CM MY CY CMY K

COMBUSTIONHANDBOOK
THE J OHN Z INK

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

INDUSTRIAL COMBUSTION SERIES

Edited by Charles E. Baukal, Jr.

PUBLISHED TITLES
Oxygen-Enhanced Combustion
Charles E. Baukal, Jr.

Heat Transfer in Industrial Combustion

Charles E. Baukal, Jr.

Computational Fluid Dynamics in Industrial Combustion

Charles E. Baukal, Jr., Vladimir Y. Gershtein, and Xianming Li

The John Zink Combustion Handbook

Charles E. Baukal, Jr.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

COMBUSTION
HANDBOOK
THE J OHN Z INK

CHARLES E. BAUKAL, JR., PH.D., P.E.

Editor
ROBERT E. SCHWARTZ, P.E.
Associate Editor

John Zink Company, LLC

Tulsa, Oklahoma

CRC Press
Boca Raton London New York Washington, D.C.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Library of Congress Cataloging-in-Publication Data

The John Zink combustion handbook / Charles E. Baukal, editor ; Robert Schwartz, associate editor.
p. cm. — (Industrial combustion series)
Includes bibliographical references and index.
ISBN 0-8493-2337-1 (alk. paper)
1. Combustion engineering—Handbooks, manuals, etc. I. Baukal, Charles E.
II. Schwartz, Robert (Robert E.) III. John Zink Company. IV. Series.

TJ254.5 .J63 2000

621.502′3—dc21 00-049357
CIP

This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with
permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish
reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials
or for the consequences of their use.

Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, microfilming, and recording, or by any information storage or retrieval system, without prior
permission in writing from the publisher.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

All rights reserved. Authorization to photocopy items for internal or personal use, or the personal or internal use of specific
clients, may be granted by CRC Press LLC, provided that $.50 per page photocopied is paid directly to Copyright clearance
Center, 222 Rosewood Drive, Danvers, MA 01923 USA. The fee code for users of the Transactional Reporting Service is
ISBN 0-8493-2337-1/00/$0.00+$.50. The fee is subject to change without notice. For organizations that have been granted
a photocopy license by the CCC, a separate system of payment has been arranged.

The consent of CRC Press LLC does not extend to copying for general distribution, for promotion, for creating new works,
or for resale. Specific permission must be obtained in writing from CRC Press LLC for such copying.

Direct all inquiries to CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation, without intent to infringe.

© 2001 by CRC Press LLC

No claim to original U.S. Government works

International Standard Book Number 0-8493-2337-1
Library of Congress Card Number 00-049357
Printed in the United States of America 4 5 6 7 8 9 0
Printed on acid-free paper
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page v Thursday, April 7, 2005 4:00 PM

Dedication to David H. Koch

The staff at John Zink dedicates this book and gives special chemical, petrochemical, and power generation industries.
thanks to David H. Koch, president of John Zink Company He aided in developing a world-class research and develop-
from January 1998 to March 2000. David made significant ment team, built a state-of-the-art testing center, and estab-
contributions to this book from its inception to its comple- lished alliances with the top combustion researchers in the
tion. Without David’s passion and excitement about science world. David’s technological leadership and strong support
and technology, we could not have completed this book. of the John Zink technical staff made this book possible.
David recognized the need for a comprehensive handbook David, we thank you for your respect and dedication to the
regarding combustion that would address applications for the John Zink Company.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

v
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page vi Thursday, April 7, 2005 4:00 PM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Foreword
As we enter the twenty-first century, the importance of tool in combustion research, development, analysis, and
energy for industry, transportation, and electricity generation equipment design.
in our daily lives is profound. Combustion of fossil fuels is Today’s diagnostic tools and instrumentation — with capa-
by far the predominant source of energy today and will likely bilities unimaginable just a few years ago — allow engineers
remain that way for many years to come. and scientists to gather detailed information in hostile com-
Combustion has played major roles in human civilization, bustion environments at both microscopic and macroscopic
including both practical and mystical ones. Since man dis- levels. Lasers, spectroscopy, advanced infrared, and ultra-
covered how to create fire, we have relied on combustion to violet camera systems are used to nonintrusively gather quan-
perform a variety of tasks. Fire was first used for heating and titative and qualitative information, including combustion
cooking, and later to manufacture tools and weapons. For all temperature, velocity, species concentration, flow visuali-
practical purposes, it was not until the onset of the Industrial zation, particle size, and loading. Advanced diagnostic sys-
Revolution in the nineteenth century that man started to har- tems and instrumentation are being transferred beyond the
ness power from combustion. We have made rapid progress laboratory to implementation in practical field applications.
in the application of combustion systems since then, and The information obtained with these systems has consider-
many industries have come into existence as a direct result ably advanced our knowledge of combustion equipment and
of this achievement. has been an indispensable source of CFD model validation.
Demands placed on combustion systems change continu-
Oil refining, chemical process, and power generation are
ously with time and are becoming more stringent. The safety
energy-intensive industries with combustion applications in
of combustion systems has always been essential, but empha-
burners, process heaters, boilers, and cogeneration systems,
sis on effective heat transfer, temperature uniformity, equip-
as well as flares and thermal oxidizers. Combustion for these
ment scale-up, efficiency, controls, and — more recently —
industries presents unique challenges related to the variety of
environmental emissions and combustion-generated noise
fuel compositions encountered. Combustion equipment must
has evolved over time. Such demands create tremendous
be flexible to be able to operate in a safe, reliable, efficient,
challenges for combustion engineers. These challenges have
and environmentally responsible manner under a wide array
been successfully met in most applications by combining
of fuel compositions and conditions.
experience and sound engineering practices with creative and
innovative problem-solving. Combustion is an exciting and intellectually challenging
Understanding combustion requires knowledge of the fun- field containing plenty of opportunities to enhance fundamen-
damentals: turbulent mixing, heat transfer, and chemical tal and practical knowledge that will ultimately lead to devel-
kinetics. The complex nature of practical combustion sys- opment of new products with improved performance.
tems, combined with the lack of reliable analytical models in This book represents the tireless efforts of many John Zink
the past, encouraged researchers to rely heavily on empirical engineers willing to share their unique knowledge and expe-
methods to predict performance and to develop new products. rience with other combustion engineers, researchers, opera-
Fortunately, the combustion field has gained considerable tors of combustion equipment, and college students. We have
scientific knowledge in the last few decades, which is now tried to include insightful and helpful information on com-
utilized in industry by engineers to evaluate and design combustion fundamentals, combustion noise, CFD design, exper-
bustion systems in a more rigorous manner. This progress is imental techniques, equipment, controls, maintenance, and
the result of efforts in academia, government laboratories, troubleshooting. We hope our readers will agree that we have
private labs, and companies like John Zink. done so.
The advent of ever-faster and more powerful computers
has had a profound impact on the manner in which engineers
model combustion systems. Computational Fluid Dynamics David H. Koch
(CFD) was born from these developments. Combined with Executive Vice President
validation by experimental techniques, CFD is an essential Koch Industries

vii --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Preface

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Combustion is described as “the rapid oxidation of a fuel processes so that theoretical correlations and predictions
resulting in the release of usable heat and production of a vis- can be made on the basis of secure and realistic models.7
ible flame.” 1 Combustion is used to generate 90% of the
world’s power.2 Regarding the science of combustion, Liñán Despite the length of time it has been around, despite its
and Williams write, importance to man, and despite vast amounts of research,
combustion is still far from being completely understood. One
Although combustion has a long history and great eco- of the purposes of this book is to improve that understanding,
nomic and technical importance, its scientific investigation particularly in industrial combustion applications in the pro-
is of relatively recent origin. Combustion science can be cess and power generation industries.
defined as the science of exothermic chemical reactions in
flows with heat and mass transfer. As such, it involves
This book is generally organized in two parts. The first part
thermodynamics, chemical kinetics, fluid mechanics, and deals with the basic theory of some of the important disci-
transport processes. Since the foundations of the second plines (combustion, heat transfer, fluid flow, etc.) important
and last of these subjects were not laid until the middle of for the understanding of any combustion process and covers
the nineteenth century, combustion did not emerge as a Chapters 1 through 13. While these topics have been satis-
science until the beginning of the twentieth century.3 factorily covered in many combustion textbooks, this book
treats them from the context of the process and power gener-
Chomiak writes, “In spite of their fundamental importance
ation industries. The second part of the book deals with spe-
and practical applications, combustion processes are far from
cific equipment design issues and applications in the process
being fully understood.” 4 Strahle writes, “combustion is a
and power generation industries.
difficult subject, being truly interdisciplinary and requiring the
merging of knowledge in several fields.”5 It involves the study
of chemistry, kinetics, thermodynamics, electromagnetic radi-
ation, aerodynamics and fluid mechanics including multi-
phase flow and turbulence, heat and mass transfer, and quan- REFERENCES
tum mechanics to name a few. Regarding combustion research,
1. Industrial Heating Equipment Association, Combustion
The pioneering experiments in combustion research, some Technology Manual, Fifth Edition Arlington, VA, Com-
600,000 years ago, were concerned with flame propagation bustion Division of the Industrial Heating Equipment
rather than ignition. The initial ignition source was pro- Association, 1994, 1.
vided by Mother Nature in the form of the electrical dis- 2. N. Chigier, Energy, Combustion, and Environment
charge plasma of a thunderstorm or as volcanic lava, McGraw-Hill, New York, 1981, ix.
depending on location. … Thus, in the beginning, Nature
provided an arc-augmented diffusion flame and the first of 3. A. Liñán and F.A. Williams, Fundamental Aspects of
man’s combustion experiments established that the heat of Combustion Oxford University Press, Oxford, 1993, 3.
combustion was very much greater than the activation
4. Chomiak, Combustion: A Study in Theory, Fact and
energy — i.e., that quite a small flame on a stick would
Application, 1.
spontaneously propagate itself into a very large fire, given
a sufficient supply of fuel.6 5. W.C. Strahle, An Introduction to Combustion Gordon
& Breach, Longhorne, PA, 1993, ix.
In one of the classic books on combustion, Lewis and von
6. F.J. Weinberg, “The First Half-Million Years of Com-
Elbe write,
bustion Research and Today’s Burning Problems,” in
the Fifteenth Symposium (International) on Combus-
Substantial progress has been made in establishing a com- tion, The Combustion Institute, Pittsburgh, PA, 1974, 1.
mon understanding of combustion phenomena. However,
this process of consolidation of the scientific approach to 7. B. Lewis and G. von Elbe, Combustion, Flames and
the subject is not yet complete. Much remains to be done Explosions of Gases, Third Edition, Academic Press,
to advance the phenomenological understanding of flame New York, 1987, xv.

ix
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page x Thursday, April 7, 2005 4:00 PM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Acknowledgments
The authors would like to collectively thank the John Zink he has been involved with the art and science of burning
Company, LLC for the consistent help and support provided wastes. Thanks also to all those who directly or indirectly
during the preparation of this book. Many of our colleagues helped him gain the experience and knowledge to contribute
have helped in the ideas, the writing, and the preparation of to his chapter. It is a continuing education process. Robert
figures and tables. We would like to especially thank the John Schwartz thanks his wife, Stella, for her patience throughout
Zink management including Steve Pirnat (President), Earl the years. In addition, he would like to acknowledge the work
Schnell (Vice President of Burners), Roberto Ruiz (Vice
of each of the authors and thank them for their contributions.
President of Technology and Commercial Development), Jim
A very special thank you to David Koch for the opportunity
Goodman (Vice President of the Systems Group), and Andy
to work on this book and to Chuck Baukal and Dave Fitzgerald
Barrieau (General Manager of Todd Combustion) for their
interest and attention in this project and for providing the for their untiring efforts. Prem Singh thanks his wife and
resources to complete it. The authors would like to thank Dr. daughters for their constant encouragement and inquiry about
David Fitzgerald who spent countless hours formatting, this book that prompted him to prepare his chapters in time.
drawing figures, editing style, getting permissions, and col- Joseph Smith would like to thank the Lord for His continued
lecting information for this book. The project would certainly grace in his life. He gives all credit to Him for anything
have taken much longer without his help. The authors would noteworthy that he’s done. Tim Webster would like to thank
also like to thank Kevin Hardison who drew some of the fig- his parents, Lee and Marilyn Webster, for their continued
ures used in several chapters. support and encouragement, which has made all his personal
Chuck Baukal would like to thank his wife Beth and his and professional accomplishments possible. Jeff White thanks
daughters Christine, Caitlyn, and Courtney for their patience
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

his wife and family who supported him even when he was
and help during the writing of this book. He would also like very late at the office. He also thanks the John Zink Flare
to thank the good Lord above, without whom this would not
Group who continue to find new situations from which we all
have been possible. Larry Berg would like to thank his wife,
can learn something. Roger Witte would like to thank all his
Betty, who has always encouraged him to be his very best at
colleagues at John Zink Company for the wisdom and the
work. Without her support, much of his career would not have
happened. He would also like to thank God who has blessed knowledge they shared with him in writing this book. He
him with the talent and ability to work as a research engineer would especially like to thank Dr. Robert R. Reed, Herschel
and participate in the creation of this book. Joe Colannino Goodnight, Don Iverson, Harold Koons, and Bob Schwartz
gives special thanks to his wife, Judy, for never complaining for the knowledge and wisdom they have shared through the
(not even once) regarding the myriad of evenings and week- years and their patience in letting him make the mistakes that
ends he devoted to this project. He knows that God is the giver one makes in learning the combustion business. He would also
of every good gift whenever he hears “Hi, honey, I’m happy like to thank his wife, Nancy, for putting up with the long
you’re home.” Joe Gifford gives many thanks to his wife hours of working at home in writing his chapters.
Barbara for typing, colleague Jim Heinlein, for suggestions
and review, Kevin Hardison, for preparing figures, and to The authors and especially the editor would like to give a
Charles Baukal and David Fitzgerald for guidance and editing. special acknowledgement to Andrea Demby at CRC Press.
Bob Hayes would like to thank God, his family, and his wife, This project would not have been nearly as successful without
for all of their love and encouragement. Paul Melton thanks her tireless efforts working days, nights, and weekends on
his wife, Toni, (and daughter, Angela, before she left for col- this project for many months. We salute her and thank her
lege) for her patience and understanding during the many years for her unending patience with our numerous revisions!

xi
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xii Thursday, April 7, 2005 4:00 PM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Tables and Figures1

CHAPTER 1
TABLE 1.1 Major Petroleum Refining Processes.
TABLE 1.2 Average Burner Configuration by Heater Type.
TABLE 1.3 Major Refinery Processes Requiring a Fired Heater.
TABLE 1.4 Major Fired Heater Applications in the Chemical Industry.

CHAPTER 2
TABLE 2.1 Alphabetical List of Atomic Weights for Common Elements. (From IUPAC Commission on Atomic Weights and
Isotopic Abundances, Atomic Weights of Elements, 1995. Pure Appl. Chem., 68, 2339, 1996. With permission.)
TABLE 2.2 Molar Ratios for Some Combustion Reactions and Products.
TABLE 2.3 Molecular Weights and Stoichiometric Coefficients for Common Gaseous Fuels.
TABLE 2.4 Combustion Data for Hydrocarbons.
TABLE 2.5 Adiabatic Flame Temperature.

CHAPTER 3
TABLE 3.1 Thermal Conductivity of Common Materials. (From S.C. Stultz and J.B. Kitto, Eds., Steam: Its Generation and Use,
40th ed., The Babcock & Wilcox Company, Barberton, OH, 1992. With permission.)
TABLE 3.2 Properties of Various Substances at Room Temperature.
TABLE 3.3 Properties of Selected Gases at 14.696 psi. (From F.P. Incropera and D.P. DeWitt, Fundamentals of Heat and Mass
Transfer, 4th ed. Copyright© 1996. Reprinted by permission of John Wiley & Sons, Inc.)
TABLE 3.4 One-dimensional, Steady-State Solutions to the Heat Equation with No Generation30.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
TABLE 3.5 Typical Convective Heat Transfer Coefficients.
TABLE 3.6 Summary of Convection Correlations for Flow in a Circular Tube.29 (From F.P. Incropera and D.P. DeWitt, Funda-
mentals of Heat and Mass Transfer, 4th ed. Copyright© 1996. Reprinted by permission of John Wiley & Sons, Inc.)
TABLE 3.7 Constants of Equation (3.80) for a Circular Cylinder in Cross Flow.
TABLE 3.8 Constants of Equation (3.85) for the Tube Bank in Cross Flow. (From F.P. Incropera and D.P. DeWitt, Fundamentals
of Heat and Mass Transfer, 4th ed. Copyright© 1996. Reprinted by permission of John Wiley & Sons, Inc.)
TABLE 3.9 Spectrum of Electromagnetic Radiation.
TABLE 3.10 View Factors for Two-dimensional Geometries.
TABLE 3.11 View Factors for Three-dimensional Geometries.
TABLE 3.12 Normal Emissivities, ∑, for Various Surfaces.
TABLE 3.13 Mean Beam Lengths Le for Various Gas Geometries.

CHAPTER 4
TABLE 4.1 Viscosity Conversion Table.
TABLE 4.2 Properties of U.S. Standard Atmosphere at Sea Level.
TABLE 4.3 Equivalent Roughness for New Pipes.
TABLE 4.4 Loss Coefficients for Various Fittings.

1 Permission and source lines are in addition to or replacement of citations in the text.

xiii
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xiv Thursday, April 7, 2005 4:00 PM

xiv The John Zink Combustion Handbook

CHAPTER 5
TABLE 5.1 Example Pipeline Quality Natural Gas. (From Gas Processors and Suppliers Association, GPSA Engineering Data
Book, Vol. 1, 10th ed., Tulsa, OK, 1987, pp. 2–3. With permission.)
TABLE 5.2 Commercial Natural Gas Components and Typical Ranges of Composition.
TABLE 5.3 Composition of a Typical Refinery Gas.
TABLE 5.4 Typical Composition of Steam Reforming/PSA Tail Gas.
TABLE 5.5 Typical Composition of Flexicoking Waste Gas.
TABLE 5.6 Volumetric Analysis of Typical Gaseous Fuel Mixtures.
TABLE 5.7 Physical Constants of Typical Gaseous Fuel Mixtures.
TABLE 5.8 Physical Constants of Typical Gaseous Fuel Mixture Components.
TABLE 5.9 Quantitative Listing of Products Made by the U.S. Petroleum Industry. (From Gas Processors and Suppliers Associ-
ation, GPSA Engineering Data Book, Vol. 1, 10th ed., Tulsa, OK, 1987, p. 6. With permission.)
TABLE 5.10 General Fraction Boiling Points.
TABLE 5.11 Requirements for Fuel Oils (per ASTEM D 396).
TABLE 5.12 Typical Analysis of Different Fuel Oils.
TABLE 5.13 Naphtha Elemental Analysis.
TABLE 5.14 Viscosity Conversion Chart.

CHAPTER 6
TABLE 6.1 Combustion Emission Factors (lb/106 Btu) by Fuel Type.
TABLE 6.2 Uncontrolled NOx Emission Factors for Typical Process Heaters.
TABLE 6.3 Reduction Efficiencies for NOx Control Techniques.
TABLE 6.4 NOx Control Technologies in Process Heaters.
TABLE 6.5 NOx Reductions for Different Low-NO Burner Types. (From A. Garg, Chemical Engineering Progress, 90, 1, 46–49.
Reproduced with permission of the American Institute of Chemical Engineers. Copyright© 1994 AIChe. All rights
reserved.)

CHAPTER 7
TABLE 7.1 The Ten Octave Bands.
TABLE 7.2 Octave and One-Third Octave Bands.
TABLE 7.3 Addition Rules.
TABLE 7.4 Sound Levels of Various Sources.
TABLE 7.5 OSHA Permissible Noise Exposures.
TABLE 7.6 Overall Sound Pressure Lever from Combustion.
TABLE 7.7 The Overall Sound Pressure Level (OASPL) Determined Experimentally and Using the Mathematical Model.

CHAPTER 8
TABLE 8.1 Exit Mach Number, Area Ratio, Driving Force Ratio, and Driving Force Percentage Increase for Various Gas Pressures
(γ = 1.33 and Pa = 14.3 psia+).
TABLE 8.2 Values for a 90o Mitered Elbow.

CHAPTER 9
TALBE 9.1 Current CFD Applications n the Chemical Process Industry.
TABLE 9.2 Universal “Empirical” Constants Used in k-∑ Turbulence Model.
TABLE 9.3 Cartesian Differential Equation Set.
TABLE 9.4 Cylindrical Differential Equation Set.
TABLE 9.5 Composition of Acid Gas Used in CFD Study.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

TABLE 9.6 Limiting Cases Considering During RCl Combustion Study.

Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xv Thursday, April 7, 2005 4:00 PM

Tables and Figures xv

CHAPTER 10
TABLE 10.1 Flammability Limits for Common Fuels at Standard Temperature and Pressure. (Courtesy of R.J. Reed, North American
Combustion Handbook, Vol. I, 3rd ed., North American Manufacturing Company, Cleveland, OH, 1986.)
TABLE 10.2 Minimum Ignition Temperatures for Common Fuels at Standard Temperature and Pressure. (Courtesy of R.J. Reed,
North American Combustion Handbook, Vol. I, 3rd ed., North American Manufacturing Company, Cleveland, OH,
1986.)
TABLE 10.3 Flammability and Ignition Characteristics of Liquids and Gases. (Adapted from D.R. Lide, Ed., CRC Handbook of
Chemistry and Physics, 80th ed., CRC Press, Boca Raton, FL, 1999. With permission.)
TABLE 10.4 Ignition Sources of Major Fires. (Adapted from the Accident Prevention Manual for Industrial Operations, National
Safety Council, Itasca, IL, 1974.)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
TABLE 10.5 Minimum Ignition Energies Required for Common Fuels.
TABLE 10.6 Benefits of a Successful Process Knowledge and Documentation Program. (Adapted from the GPSA Engineering
Data Book, Vol. II, 10th ed., Gas Processors and Suppliers Association, Tulsa, OK, 1994.)

NO TABLES FOR CHAPTER 11

CHAPTER 12
TABLE 12.1 Gas Valve Data.
TABLE 12.2 Data for Characterizer.

CHAPTER 13
TABLE 13.1 Some Potential Factors Affecting NOx Response from a Burner.
TABLE 13.2 NOx as a Function of Burner Geometry and Operation.
TABLE 13.3 Transforms for Table 13.4.
TABLE 13.4 Transformed Data for Fuel-staged Burner.
TABLE 13.5 Generic ANOVA Table.
TABLE 13.6 F-Distribution, 99%, 95%, and 90% Confidence.
TABLE 13.7 ANOVA Table for Equation 13.2 Applied to Data of Table 13.4.
TABLE 13.8 ANOVA for Table 13.7 with Separate Effects.
TABLE 13.9 ANOVA for Table 13.7 with Pooled Effects.
TABLE 13.10 Factorial Design with Replicate Centerpoints.
TABLE 13.11 Generic ANOVA for Factorial Design with Replicates.
TABLE 13.12 ANOVA for Table 13.10 and Equation.
TABLE 13.13 ANOVA for Factorial Design with Centerpoint Replicates Case 2 of Table 13.11.
TABLE 13.14 ∫ Fractional Factorial [FF(3,-1,0)].
TABLE 13.15 FF(3,0,4) in Two Blocks.
TABLE 13.16 Experimental Design with Categorial Factors.
TABLE 13.17 ANOVA for Table 13.16.
TABLE 13.18 CC(3,0,6) Design.
TABLE 13.19 ~CC(3,0,3) Design.
TABLE 13.20 ANOVA for Table 13.19 and Equation (13.38).
TABLE 13.21 Example of an Orthogonal Subspace for a q = 3 Simplex.
TABLE 13.22 FF(7,-2,0) Design Generating a Combined Mixture-Factorial in Five Factors – Two at Four Levels and Three at Two
Levels.

CHAPTER 14
TABLE 14.1 Tulsa Natural Gas (TNG) Composition and Properties.
TABLE 14.2 Example Refinery Gas.
TABLE 14.3 Comparison of Refinery Gas to Test Blend.
TABLE 14.4 Test Procedure Gas Specification Sheet.
TABLE 14.5 Example Test Procedure.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xvi Thursday, April 7, 2005 4:00 PM

xvi The John Zink Combustion Handbook

CHAPTER 15
TABLE 15.1 Burner Throat Area for Different Tile Dimensions.

CHAPTER 16
TABLE 16.1 Static Draft Effect per Foot of Height.
TABLE 16.2 Typical Flame Dimensions for Different Burner Types.
TABLE 16.3 Typical Excess Air Values for Gas Burners.
TABLE 16.4 Typical Excess Air Values for Liquid Fuel Firing.

CHAPTER 17
TABLE 17.1 Ratio of Upper and Lower Explosive Limits and Flashback Probability in Premix Burners for Various Fuels.
TABLE 17.2 Troubleshooting for Gas Burners.
TABLE 17.3 Troubleshooting for Oil Burners.

CHAPTER 18
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

TABLE 18.1 Typical NOx and CO Emissions from Duct Burners.

NO TABLES FOR CHAPTERS 19 AND 20

CHAPTER 21
TABLE 21.1 Typical Thermal Oxidizer Operating Conditions.
TABLE 21.2 Relative Characteristics of Centrifugal Blowers.

CHAPTER 1
FIGURE 1.1 Typical petroleum refinery.
FIGURE 1.2 Typical refinery flow diagram.
FIGURE 1.3 Offshore oil rig flare.
FIGURE 1.4 Duct burner flame.
FIGURE 1.5 Duct burner in large duct.
FIGURE 1.6 Front of boiler burner.
FIGURE 1.7 Thermal oxidizer.
FIGURE 1.8 Side- (a) and top-fired (b) reformers (elevation view).
FIGURE 1.9 Downfired burner commonly used in top fired reformers.
FIGURE 1.10 Elevation view of a terrace firing furnace.
FIGURE 1.11 Schematic of a process heater.
FIGURE 1.12 Typical process heater.
FIGURE 1.13 Fixed heater size distribution.
FIGURE 1.14 Sketch (elevation view) of center or target wall firing configuration.
FIGURE 1.15 Horizontal floor-fired burners.
FIGURE 1.16 Wall fired burner (side view).
FIGURE 1.17 Sketch (elevation view) of a horizontally mounted, vertically fired burner configuration.
FIGURE 1.18 Examples of process heaters.
FIGURE 1.19 Typical heater types.
FIGURE 1.20 Cabin heater.
FIGURE 1.21 Crude unit burners.
FIGURE 1.22 Typical burner arrangements (elevation view).
FIGURE 1.23 Process heater heat balance. (From Philip Conisbee, Georges de La Tour and His World. National Gallery of Art,
Washington, D.C., 1996, 110. With permission.)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xvii Thursday, April 7, 2005 4:00 PM

Tables and Figures xvii

FIGURE 1.24 Burner (B) arrangement (plan view) in the floor of vertical cylindrical furnaces: (a) small-diameter furnace with a
single centered burner and (b) larger diameter furnace with four burners symmetrically arranged at a radius from the
center.
FIGURE 1.25 Burner (B) arrangement (plan view) in the floor of rectangular cabin heaters: (a) single row of burners in a narrower
heater, (b) two rows of staggered burners in a slightly wider heater, and (c) two rows of aligned burners in an even
wider heater.
FIGURE 1.26 Adiabatic equilibrium NO and CO as a function of the equivalence ratio for an air/CH4 flame.
FIGURE 1.27 Typical combination oil and gas burner.
FIGURE 1.28 Schematic of flue gas recirculation.
FIGURE 1.29 Cartoon of a premixed burner.
FIGURE 1.30 Typical premixed gas burner. (From API Publication 535: Burner for Fire Heaters in General Refinery Services, 1st
ed., American Petroleum Institute, Washington, D.C., July, 1995. With permission.)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 1.31 Painting of a diffusion flame.

FIGURE 1.32 Cartoon of a diffusion burner.
FIGURE 1.33 Cartoon of a partially-premixed burner.
FIGURE 1.34 Cartoon of a staged-air burner.
FIGURE 1.35 Schematic of a typical staged air combination oil and gas burner.
FIGURE 1.36 Cartoon of a staged-fuel burner.
FIGURE 1.37 Schematic of a typical staged-fuel gas burner.
FIGURE 1.38 Typical natural-draft gas burner.
FIGURE 1.39 Natural draft burner.
FIGURE 1.40 Flames impinging on tubes in a cabin heater.
FIGURE 1.41 Flames pulled toward the wall.
FIGURE 1.42 Gas burners needing service.

CHAPTER 2
FIGURE 2.1 Typical cabin-style process heater.
FIGURE 2.2 Species concentration vs. excess air for the following fuels: (a) CH4, (b) natural gas, (c) simulated refinery gas (25%
H2, 50% CH4, 25% C3H8), (d) C3H8, (e) No. 2 oil, and (f) No. 6 oil.
FIGURE 2.2 (b) Natural gas
FIGURE 2.2 (c) Simulated refinery gas (25% H2, 50% CH4, 25% C3H8).
FIGURE 2.2 (d) Propane.
FIGURE 2.2 (e) Fuel oil #2.
FIGURE 2.2 (f) Fuel oil #6.
FIGURE 2.3 Species concentration vs. stoichiometric ratio for the following fuels: (a) CH4, (b) natural gas, (c) simulated refinery
gas (25% H2, 50% CH4, 25% C3H8), (d) C3H8, (e) No. 2 oil, and (f) No. 6 oil.
FIGURE 2.3 (b) Natural gas
FIGURE 2.3 (c) Simulated refinery gas (25% H2, 50% CH4, 25% C3H8).
FIGURE 2.3 (d) Propane.
FIGURE 2.3 (e) Fuel oil #2.
FIGURE 2.3 (f) Fuel oil #6.
FIGURE 2.4 Adiabatic equilibrium reaction process.
FIGURE 2.5 Adiabatic equilibrium calculations for the predicted gas composition as a function of the O2:CH4 stoichiometry for
air/CH4 flames, where the air and CH4 are at ambient temperature and pressure.
FIGURE 2.6 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the major species as a function
of the air preheat temperature for air/CH4 flames, where the CH4 is at ambient temperature and pressure.
FIGURE 2.7 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the minor species as a function
of the air preheat temperature for air/CH4 flames, where the CH4 is at ambient temperature and pressure.
FIGURE 2.8 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the major species as a function
of the fuel preheat temperature for air/CH4 flames, where the air is at ambient temperature and pressure.
FIGURE 2.9 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the minor species as a function
of the fuel preheat temperature for air/CH4 flames, where the air is at ambient temperature and pressure.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xviii Thursday, April 7, 2005 4:00 PM

xviii The John Zink Combustion Handbook

FIGURE 2.10 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the major species as a function
of the fuel blend (H2 + CH4) composition for air/fuel flames, where the air and fuel are at ambient temperature and
pressure.
FIGURE 2.11 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the minor species as a function
of the fuel blend (H2 + CH4) composition for air/fuel flames, where the air and fuel are at ambient temperature and
pressure.
FIGURE 2.12 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the major species as a function
of the fuel blend (N2 + CH4) composition for air/fuel flames, where the air and fuel are at ambient temperature and
pressure.
FIGURE 2.13 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the minor species as a function
of the fuel blend (N2 + CH4) composition for air/fuel flames, where the air and fuel are at ambient temperature and
pressure.
FIGURE 2.14 Equilibrium calculations for the predicted gas composition of the major species as a function of the combustion
product temperature for air/CH4 flames, where the air and fuel are at ambient temperature and pressure.
FIGURE 2.15 Equilibrium calculations for the predicted gas composition of the minor species as a function of the combustion
product temperature for air/CH4 flames, where the air and fuel are at ambient temperature and pressure.
FIGURE 2.16 Adiabatic flame temperature vs. equivalence ratio for air/H2, air/CH4, and air/C3H8 flames, where the air and fuel are
at ambient temperature and pressure.
FIGURE 2.17 Adiabatic flame temperature vs. air preheat temperature for stoichiometric air/H2, air/CH4, and air/C3H8 flames, where
the air and fuel are at ambient temperature and pressure.
FIGURE 2.18 Adiabatic flame temperature vs. fuel preheat temperature for stoichiometric air/H2, air/CH4, and air/C3H8 flames,
where the air is at ambient temperature and pressure.
FIGURE 2.19 Adiabatic flame temperature vs. fuel blend (CH4/H2 and CH4/N2) composition for stoichiometric air/fuel flames, where
the air and fuel are at ambient temperature and pressure.
FIGURE 2.20 Adiabatic flame temperature vs. fuel blend (CH4/H2) composition and air preheat temperature for stoichiometric
air/fuel flames, where the fuel is at ambient temperature and pressure.
FIGURE 2.21 Sample Sankey diagram showing distribution of energy in a combustion system.
FIGURE 2.22 Available heat vs. gas temperature for stoichiometric air/H2, air/CH4, and air/C3H8 flames, where the air and fuel are
at ambient temperature and pressure.
FIGURE 2.23 Available heat vs. air preheat temperature for stoichiometric air/H2, air/CH4, and air/C3H8 flames at an exhaust gas
temperature of 2000oF (1100oC), where the fuel is at ambient temperature and pressure.
FIGURE 2.24 Available heat vs. fuel preheat temperature for stoichiometric air/H2, air/CH4, and air/C3H8 flames at an exhaust gas
temperature of 2000oF (1100oC), where the air is at ambient temperature and pressure.
FIGURE 2.25 Graphical representation of ignition and heat release.

CHAPTER 3
FIGURE 3.1 A typical fired heater.
FIGURE 3.2 Heat transfer through a plane wall: (a) temperature distribution, and (b) equivalent thermal circuit.
FIGURE 3.3 Equivalent thermal circuit for a series composite wall.
FIGURE 3.4 Temperature drop due to thermal contact resistance.
FIGURE 3.5 Temperature distribution for a composite cylindrical wall.
FIGURE 3.6 Transient conduction through a solid.
FIGURE 3.7 Thermal conductivity of (a) some commonly used steels and alloys and (b) some refractory materials.
FIGURE 3.8 Temperature-thickness relationships corresponding to different thermal conductivities.
FIGURE 3.9 Thermal boundary layer development in a heated circular tube.
FIGURE 3.10 Orthogonal oscillations of electric and magnetic waves in the oscillations in electromagnetic waves.
FIGURE 3.11 Spectrum of electromagnetic radiation.
FIGURE 3.12 Spectral blackbody emissive power.
FIGURE 3.13 Radiation transfer between two surfaces approximated as blackbodies.
FIGURE 3.14 Network representation of radiative exchange between surface i and the remaining surfaces of an enclosure.
FIGURE 3.15 View factor of radiation exchange between faces of area dAi and dAj.
FIGURE 3.16 View factor for aligned parallel rectangles.
FIGURE 3.17 View factor for coaxial parallel disks.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Tables and Figures xix

FIGURE 3.18 View factor for perpendicular rectangles with a common edge.
FIGURE 3.19 Infrared thermal image of a flame in a furnace.
FIGURE 3.20 Emission bands of (a) CO2 and (b) H2O.
FIGURE 3.21 Emissivity of water vapor in a mixture with nonradiating gases at 1-atm total pressure and of hemispherical shape.
FIGURE 3.22 Emissivity of carbon dioxide in a mixture with nonradiating gases at 1-atm total pressure and of hemispherical shape.
FIGURE 3.23 Radiation heat transfer correction factor for mixtures of water vapor and carbon dioxide.
FIGURE 3.24 Photographic view of a luminous flame.
FIGURE 3.25 Photographic view of a nonluminous flame.
FIGURE 3.26 Photographic view of a radiant wall burner.
FIGURE 3.27 Vertical heat flux distribution for oil and gas firing in a vertical tube furnace.
FIGURE 3.28 Distribution of dimensionless average radiant flux density at the tube surfaces for various flame lengths (Lf= flame
length, L = heater height, Z = height).

CHAPTER 4
FIGURE 4.1 Variation in measured density with length scale.
FIGURE 4.2 Velocity profile of a fluid flowing along a solid surface.
FIGURE 4.3 Absolute viscosity vs. temperature for various fluids.
FIGURE 4.4 Temperature vs. viscosity for various hydrocarbons. (Courtesy of J.B. Maxwell, Data Book on Hydrocarbons, D.
Van Nostrand, Princeton, NJ, 1950, 174.)
FIGURE 4.5 Viscosity of mid-continent oils. (Courtesy of J.B. Maxwell, Data Book on Hydrocarbons, D. Van Nostrand, Princeton,
NJ, 1950, 164.)
FIGURE 4.6 Compressibility factor Z as a function of reduced pressure and reduced temperature for different gases.
FIGURE 4.7 U-tube manometer.
FIGURE 4.8 Inclined manometer .
FIGURE 4.9 Helium balloon attached to the ground.
FIGURE 4.10 A small packet of fluid from point A to B along an arbitrary path.
FIGURE 4.11 Pressure relief vessel venting to a flare.
FIGURE 4.12 An idealization of a small “differential” control volume.
FIGURE 4.13 Mass flow into and out of a volume in the X-direction.
FIGURE 4.14 Gravitational body force.
FIGURE 4.15 Normal or pressure forces.
FIGURE 4.16 Effect of shear stress on X-direction face.
FIGURE 4.17 Smoke from incense. (Courtesy of The Visualization Society of Japan, Fantasy of Flow, Tokyo, 1993, 93.)
FIGURE 4.18 Water exiting a faucet at low velocity. (Courtesy of The Visualization Society of Japan, Fantasy of Flow, Tokyo, 1993, 97.)
FIGURE 4.19 Leonardo daVinci's view of turbulence.
FIGURE 4.20 Osborn Reynolds’ experimental apparatus used to study the transition from laminar to turbulent flow.
FIGURE 4.21 Water from faucet showing transition. (Courtesy of The Visualization Society of Japan, Fantasy of Flow, Tokyo, 1993, 97.)
FIGURE 4.22 Wake area showing mixing vortices. (Courtesy of The Visualization Society of Japan, Fantasy of Flow, Tokyo, 1993, 3.)
FIGURE 4.23 Laminar flow of smoke over a rectangular obstruction. (Courtesy of M. Van Dyke, An Album of Fluid Motion, The
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Parabolic Press, Stanford, CA, 1982, 10.)

FIGURE 4.24 Free jet structure. (Courtesy of M. Van Dyke, An Album of Fluid Motion, The Parabolic Press, Stanford, CA, 1982, 97.)
FIGURE 4.25 Free jet entrainment. (Courtesy of M. Van Dyke, An Album of Fluid Motion, The Parabolic Press, Stanford, CA, 1982, 99.)
FIGURE 4.26 Flow around the air intake of a jet engine in supersonic flow. (Courtesy of The Visualization Society of Japan, Fantasy
of Flow, Tokyo, 1993, 23.)
FIGURE 4.27 Shock waves from a supersonic fighter. (Courtesy of The Visualization Society of Japan, Fantasy of Flow, Tokyo,
1993, 59.)
FIGURE 4.28 Choked flow test rig.
FIGURE 4.29 Types of flow of a fluid passing through a small orifice.
FIGURE 4.30 Photograph showing slow-moving streaks near a wall using the hydrogen bubble technique.
FIGURE 4.31 Moody diagram.
FIGURE 4.32 Factors that can influence the discharge coefficient.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xx Thursday, April 7, 2005 4:00 PM

xx The John Zink Combustion Handbook

CHAPTER 5
FIGURE 5.1 Simplified process flow diagram for hydrogen reforming/pressure swing absorption. (From R.A. Meyers, Handbook
of Petroleum Refining Processes, 2nd ed., McGraw-Hill, New York, 1997, p. 6.27. With permission.)
FIGURE 5.2 Simplified process flow diagram for flexicoking. (From R.A. Meyers, Handbook of Petroleum Refining Processes, 2nd
ed., McGraw-Hill, New York, 1997, p. 12.5. With permission.)
FIGURE 5.3 100% Tulsa natural gas flame.
FIGURE 5.4 90% Tulsa natural gas/10% nitrogen flame.
FIGURE 5.5 80% Tulsa natural gas/20% nitrogen flame. (From R.A. Meyers, Handbook of Petroleum Refining Processes, 2nd ed.,
McGraw-Hill, New York, 1997, p. 12.11. With permission.)
FIGURE 5.6 90% Tulsa natural gas/10% hydrogen flame.
FIGURE 5.7 75% Tulsa natural gas/25% hydrogen flame.
FIGURE 5.8 50% Tulsa natural gas/50% hydrogen flame.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 5.9 25% Tulsa natural gas/75% hydrogen flame.
FIGURE 5.10 100% hydrogen flame.
FIGURE 5.11 50% Tulsa natural gas/25% hydrogen/25% C3H8 flame.
FIGURE 5.12 50% Tulsa natural gas/50% C3H8 flame.
FIGURE 5.13 100% C3H8 flame.
FIGURE 5.14 100% C4H10 flame.
FIGURE 5.15 Simulated cracked gas flame.
FIGURE 5.16 Simulated coking gas flame.
FIGURE 5.17 Simulated FCC gas flame.
FIGURE 5.18 Simulated refoming gas flame.
FIGURE 5.19 100% Tulsa natural gas flame.
FIGURE 5.20 100% hydrogen.
FIGURE 5.21 100% propane.
FIGURE 5.22 50% hydrogen 50% Propane.
FIGURE 5.23 50% hydrogen 50% Tulsa natural gas.
FIGURE 5.24 50% propane 50% Tulsa natural gas.
FIGURE 5.25 25% hydrogen 75% Propane.
FIGURE 5.26 75% hydrogen 25% Propane.
FIGURE 5.27 25% hydrogen 75% Tulsa natural gas.
FIGURE 5.28 75% hydrogen 25% Tulsa natural gas.
FIGURE 5.29 25% propane 75% Tulsa natural gas.
FIGURE 5.30 75% propane 25% Tulsa natural gas.
FIGURE 5.31 25% hydrogen 25% propane 50% Tulsa natural gas.
FIGURE 5.32 25% hydrogen 50% propane 25% Tulsa natural gas.
FIGURE 5.33 50% hydrogen 25% propane 25% Tulsa natural gas.
FIGURE 5.34 Viewing oil flame through burner plenum.
FIGURE 5.35 Oil derrick, circa 1900.
FIGURE 5.36 Capping a burning oil well.
FIGURE 5.37 Refinery flow diagram.
FIGURE 5.38 Flow diagram of UOP fluid catalytic cracking complex.
FIGURE 5.39 Burner firing heavy oil (1).
FIGURE 5.40 Burner firing heavy oil (2).
FIGURE 5.41 Naphtha distillation curve.
FIGURE 5.42 Crude oil distillation curve.
FIGURE 5.43 Viscosity of fuel oils.

CHAPTER 6
FIGURE 6.1 Cartoon of NO exiting a stack and combining with O2 to form NO2.
FIGURE 6.2 Cartoon of acid rain.
FIGURE 6.3 Cartoon of photochemical smog formation.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxi Thursday, April 7, 2005 4:00 PM

Tables and Figures xxi

FIGURE 6.4 Schematic of fuel NOx formation pathways.

FIGURE 6.5 Adiabatic equilibrium NO as a function of equivalence ratio for air/fuel flames.
FIGURE 6.6 Adiabatic equilibrium NO as a function of gas temperature for stoichiometric air/fuel flames.
FIGURE 6.7 Adiabatic equilibrium NO as a function of air preheat temperature for stoichiometric air/fuel flames.
FIGURE 6.8 Adiabatic equilibrium NO as a function of fuel preheat temperature for a stoichiometric air/CH4 flames.
FIGURE 6.9 Adiabatic equilibrium NO as a function of fuel composition (CH4/H2) for a stoichiometric air/fuel flame.
FIGURE 6.10 Adiabatic equilibrium NO as a function of fuel composition (CH4/N2) for a stoichiometric air/fuel flame.
FIGURE 6.11 Sampling system schematic as recommended by the U.S. EPA.
FIGURE 6.12 Schematic of furnace gas recirculation.
FIGURE 6.13 Adiabatic equilibrium NO as a function of the fuel blend composition for H2/CH4 blends combusted with 15% excess air
where both the fuel and the air are at ambient temperature and pressure.
FIGURE 6.14 Adiabatic equilibrium NO as a function of the fuel blend composition for C3H8/CH4 blends combusted with 15%
excess air where both the fuel and the air are at ambient temperature and pressure.
FIGURE 6.15 Adiabatic equilibrium NO as a function of the fuel blend composition for H2/C3H8 blends combusted with 15% excess
air where both the fuel and the air are at ambient temperature and pressure.
FIGURE 6.16 Ternary plot of adiabatic equilibrium NO (fraction of the maximum value) as a function of the fuel blend composition
for H2/CH4/C3H8 blends combusted with 15% excess air where both the fuel and the air are at ambient temperature
and pressure.
FIGURE 6.17 Raw gas (VYD) burner.
FIGURE 6.18 Test furnace.
FIGURE 6.19 Measured NOx (percent of the maximum ppmv value) as a function of the fuel blend composition for H2/Tulsa natural
gas blends combusted with 15% excess air where both the fuel and the air are at ambient temperature and pressure.
FIGURE 6.20 Measured NOx (percent of the maximum ppmv value) as a function of the fuel blend composition for C3H8/Tulsa natural
gas blends combusted with 15% excess air where both the fuel and the air are at ambient temperature and pressure.
FIGURE 6.21 Measured NOx (percent of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel blend
composition for H2/C3H8 blends combusted with 15% excess air where both the fuel and the air are at ambient
temperature and pressure.
FIGURE 6.22 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel blend composition
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

for Tulsa natural gas/H2/C3H8 blends combusted with 15% excess air where both the fuel and the air are at ambient
temperature and pressure for gas tip #2.
FIGURE 6.23 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel blend composition
for Tulsa natural gas/H2/C3H8 blends combusted with 15% excess air where both the fuel and the air are at ambient
temperature and pressure for gas tip #4.
FIGURE 6.24 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel blend composition
for Tulsa natural gas/H2/C3H8 blends combusted with 15% excess air where both the fuel and the air are at ambient
temperature and pressure for gas tip #6.
FIGURE 6.25 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel blend
composition for Tulsa natural gas/H2/C3H8 blends combusted with 15% excess air where both the fuel and the air are
at ambient temperature and pressure for a constant fuel gas pressure of 21 psig.
FIGURE 6.26 Measured NOx (fraction of the maximum value in ppmvd) as a function of the fuel pressure for all 15 different Tulsa
natural gas/H2/C3H8 blends (A through O) combusted with 15% excess air where both the fuel and the air are at
ambient temperature and pressure.
FIGURE 6.27 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel blend
composition, fuel gas pressure, and calculated adiabatic flame temperature for for Tulsa natural gas/H2/C3H8 blends
combusted with 15% excess air where both the fuel and the air are at ambient temperature and pressure.
FIGURE 6.28 Adiabatic equilibrium CO as a function of equivalence ratio for air/fuel flames.
FIGURE 6.29 Adiabatic equilibrium CO as a function of gas temperature for stoichiometric air/fuel flames.
FIGURE 6.30 Adiabatic equilibrium CO as a function of air preheat temperature for stoichiometric air/fuel flames.
FIGURE 6.31 Adiabatic equilibrium CO as a function of fuel preheat temperature for stoichiometric air/CH4 flames.
FIGURE 6.32 Adiabatic equilibrium CO as a function of fuel composition (CH4/H2) for a stoichiometric air/fuel flame.
FIGURE 6.33 Adiabatic equilibrium CO as a function of fuel composition (CH4/N2) for a stoichiometric air/fuel flame.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxii Thursday, April 7, 2005 4:00 PM

xxii The John Zink Combustion Handbook

CHAPTER 7
FIGURE 7.1 Pressure peaks and troughs.
FIGURE 7.2 Cross-section of the human ear.
FIGURE 7.3 Relationship of decibels to watts.
FIGURE 7.4 Sound pressure level at a distance r.
FIGURE 7.5 Threshold of hearing in humans.
FIGURE 7.6 Threshold of hearing and threshold of pain in humans.
FIGURE 7.7 A-weighted scale for human hearing threshold.
FIGURE 7.8 A-weighted burner noise curve.
FIGURE 7.9 Weighting curves A, B, C, and D.
FIGURE 7.10 Block diagram of a sound level meter.
FIGURE 7.11 Same sound spectrum on three different intervals.
FIGURE 7.12 Typical burner noise curve.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 7.13 Same sound spectrum on 3 different intervals.

FIGURE 7.14 Typical noise signature emitted from a Flare.
FIGURE 7.15 Engineer measuring flare noise level.
FIGURE 7.16 Shadow photograph of a burning butane lighter.
FIGURE 7.17 A steam-assisted flare under normal and over-steamed conditions.
FIGURE 7.18 Sound pressure level burner with instability.
FIGURE 7.19 Development of orderly wave patterns.
FIGURE 7.20 Region of maximum jet mixing noise.
FIGURE 7.21 Shock waves downstream of an air jet.
FIGURE 7.22 Location of screech tone emissions.
FIGURE 7.23 Noise radiating from a valve.
FIGURE 7.24 Two enclosed flares.
FIGURE 7.25 A steam-assisted flare with a muffler.
FIGURE 7.26 Steam jet noise emitted with and without a muffler.
FIGURE 7.27 Noise spectrum from high pressure flare with and without water injection.
FIGURE 7.28 Sound pressure vs frequency with and without a muffler.
FIGURE 7.29 Burner noise example.
FIGURE 7.30 The sound-pressure-level spectrum of a high-pressure flare.
FIGURE 7.31 The noise contributions separately based on the mathematical model.
FIGURE 7.32 Effect of distance on flare noise.

CHAPTER 8
FIGURE 8.1 A typical eductor system.
FIGURE 8.2 Example eductor system.
FIGURE 8.3 Experimental apparatus that has been successfully used to determine the flow coefficients and eduction performance
of various flare and burner eduction processes.
FIGURE 8.4 First and second generation steam flares.
FIGURE 8.5 Typical third generation steam flare tube layout.
FIGURE 8.6 Normalized plot showing the sonic-supersonic eduction performance in a single steam tube used in a typical steam-
assisted flare.
FIGURE 8.7 Normalized eduction performance of a flare pilot operating on Tulsa natural gas with two different motive gas orifice
diameters.
FIGURE 8.8 Experimental and theoretical results of the eduction performance of a particular radiant wall burner firing with two
different orifice sizes and fuel gas compositions.
FIGURE 8.9 Simplified reactor modeling of a staged fuel burner.
FIGURE 8.10 Picture of a radiant wall burner.
FIGURE 8.11 Picture of a thermal oxidizer.
FIGURE 8.12 Sample results of simplified modeling for a premixed burner.
FIGURE 8.13 Sample results of simplified modeling for a thermal oxidizer.
FIGURE 8.14 Capacity curves that many burner manufactures use for sizing burners.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxiii Thursday, April 7, 2005 4:00 PM

Tables and Figures xxiii

FIGURE 8.15 Burner designs typically consist of a muffler, damper, plenum, throat, and tile section.
FIGURE 8.16 Cold flow furnace of a test chamber (8 x 8 x 8) supported on legs approximately 7 feet above the ground.
FIGURE 8.17 Smoking and non-smoking flares.
FIGURE 8.18 Flame of a flare is divided into two parts: 1) the main body of the flame, and 2) the region near the tip.
FIGURE 8.19 Illustration showing the experimental setup utilized to obtain calibration and validation data.
FIGURE 8.20 Prediction of diluent flow for smokeless operation (a) Propylene (b) Propane.
FIGURE 8.21 Figure 8.20 with data points added (a) Propylene (b) Propane.
FIGURE 8.22 Steam-to-hydrocarbon ratios (per Leite).
FIGURE 8.23 Steam-to-hydrocarbon (large diameter flares).
FIGURE 8.24 One smokeless steam flare.
FIGURE 8.25 Typical air flare.
FIGURE 8.26 Effect of high velocity air: (a) blower off, (b) commence blower, and (c) blower on.
FIGURE 8.27 Aeration rate was determined and plotted against the scaling function.
FIGURE 8.28 Annular air-assisted flare (190,000 lb/hr propane).
FIGURE 8.29 Comparison of three different air flares to prediction.
FIGURE 8.30 Typical process heater oil flame.
FIGURE 8.31 Standard John Zink oil gun.
FIGURE 8.32 Schematic of a typical oil gun.
FIGURE 8.33 Comparison of predicted vs. actual oil and steam flow rates.
FIGURE 8.34 John Zink Co. LLC. (Tulsa, OK) Spray Research Laboratory.
FIGURE 8.35 Droplet size comparison between a standard and a newer oil gun.
FIGURE 8.36 High-efficiency new oil gun.
FIGURE 8.37 Heat transfer in a packed bed between the ceramic material and the air stream.
FIGURE 8.38 Installing small, type K thermocouple pairs into numerous ceramic saddles.
FIGURE 8.39 Summary of saddle data.
FIGURE 8.40 Rock temp distribution with time.
FIGURE 8.41 John Zink RTO test unit.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

CHAPTER 9
FIGURE 9.1 Elements in CFD modeling.
FIGURE 9.2 Plot of the β-pdf for several values of <f> and <f ’2> .
FIGURE 9.3 Point measurement of scalar in a turbulent flow.
FIGURE 9.4 Rendered view of a CFD model of a John Zink Co. burner. This view illustrates the complex geometry that necessitates
a variety of cell types. This mesh consists of hexahedral, pyramidal, and tetrahedral cell types.
FIGURE 9.5 Close-up view of primary tip. This view reveals the five fuel jets (indicated by the arrows on the image) issuing from
the primary tip.
FIGURE 9.6 Rendered view inside an ethylene pyrolysis furnace showing flow patterns near the premixed radiant wall burners.
FIGURE 9.7 CFD model of an ethylene pyrolysis furnace. There are six burners shown in each row at the bottom of the furnace,
and the tubes are approximately 35 feet long. The endwalls are not shown in this image.
FIGURE 9.8 Plot showing heat flux to the process tubes in the modeled ethylene furnace as a function of height above the furnace
floor.
FIGURE 9.9 Geometry of a xylene reboiler. This view shows half (sliced vertically) of the furnace. Only three of the six burners
are shown at the bottom of the image.
FIGURE 9.10 This view shows half of the furnace with unmodified burners. The “blob” in the furnace is the 50-ppm OH mole
fraction iso-surface. This surface is colored according to its temperature (oF)
FIGURE 9.11 This view shows half of the furnace with modified burners firing. The 50-ppm OH mole fraction iso-surface is shown
as an indicator of the flame shape. This surface is colored according to its temperature (oF)
FIGURE 9.12 Exterior geometry of the furnace is included in the model. The surface mesh is also shown.
FIGURE 9.13 Burner geometry. The acid swirl vanes are shown in red; the air swirl vanes are shown in green; and the start-up
fuel tip is shown in purple.
FIGURE 9.14 Oxygen mass fractions viewed from above the furnace. The contour scale is logarithmic. The mass fractions are
contoured on the mid-plane of the furnace.
FIGURE 9.15 H2S mole fractions contoured on the midplane of the furnace.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxiv Thursday, April 7, 2005 4:00 PM

xxiv The John Zink Combustion Handbook

FIGURE 9.16 Stoichiometric iso-surface colored by temperature (oC) for the initial burner design.
FIGURE 9.17 Stoichiometric iso-surface colored by temperature (oC) for the final burner design.
FIGURE 9.18 Midplane of geometry colored by temperature (oC). This view shows the burner quarl and the mixing regions of acid
gas and air.
FIGURE 9.19 Temperature profiles (oC) exiting the reaction furnace for the initial (left) and final (right) burner geometries.
FIGURE 9.20 Geometric information describing the thermal oxidizer examined during this study.
FIGURE 9.21 Predicted centerline profiles for excess air case: (a) axial velocity (m/s), and (b) gas temperature (K) for the furnace
section of the thermal oxidizer shown in Figure 9.20. Two distinct combustion zones are illustrated, with an exit
temperature of about 1600 K (2400oF).
FIGURE 9.22 Predicted centerline profiles for excess air case: (a) methane concentration (ppmv), and (b) carbon monoxide
concentration (ppmv) for the furnace section of the thermal oxidizer shown in Figure 9.20. These predictions depict
the CO formation and oxidation zones common to most combustion processes.
FIGURE 9.23 Predicted centerline profiles for excess air case: (a) HCl concentration (ppmv), and (b) Cl2 concentration (ppmv) for
the furnace section of the thermal oxidizer shown in Figure 9.20. The predicted maximum Cl2 concentration, nearly
3200 ppmv, occurs in the cooler reactor regions, while an exit Cl2 concentration of about 100 ppmv is predicted.
FIGURE 9.24 Predicted centerline profiles for the stoichiometric case: (a) axial velocity (m/s), and (b) gas temperature (K) for the
furnace section of the thermal oxidizer shown in Figure 9.20. A single combustion zone is indicated, with the local
maximum temperature of 1450 K (2150oF) and an exit temperature of about 1350 K (1970oF)
FIGURE 9.25 Predicted centerline profiles for the stoichiometric case: (a) methane concentration (ppmv), and (b) carbon monoxide
concentration (ppmv) for the furnace section of the thermal oxidizer shown in Figure 9.20. Here, the post-flame CO
oxidation zone, shown in the first prediction, is not present; this results in a predicted exit CO concentration of 9000 ppmv.
FIGURE 9.26 Predicted centerline profiles for the stoichiometric case: (a) HCl concentration (ppmv), and (b) Cl2 concentration
(ppmv) for the furnace section of the thermal oxidizer shown in Figure 9.20. Dramatically less Cl2 formation is
predicted (local maximum of 7 ppmv and exit concentrations less than 1 ppmv) in this case due to excess H+ radical
present from the increased fuel gas.
FIGURE 9.27 Entrainment of flue gas.
FIGURE 9.28 Re-circulation region in the eductor throat.
FIGURE 9.29 Re-circulation zone starting to occur in eductor throat.
FIGURE 9.30 Re-circulation zone developing in eductor throat.
FIGURE 9.31 Contours of stream function with increasing backpressure.

CHAPTER 10
FIGURE 10.1 Fire tetrahedron.
FIGURE 10.2 Tube rupture in a fired heater. (Courtesy of R.E. Sanders, Chemical Process Safety: Learning from Case Histories,
Butterworth–Heinemann, Woburn, MA, 1999.)
FIGURE 10.3 Trapped steam in a dead-end that can freeze and cause pipe failure.
FIGURE 10.4 CO detector: (a) permanent, (b) portable.
FIGURE 10.5 Flarestack explosion due to improper purging. (Courtesy of T. Kletz, What Went Wrong: Case Histories of Process
Plant Disasters, 4th ed., Gulf Publishing, Houston, TX, 1998.)
FIGURE 10.6 Vapor pressures for light hydrocarbons. (Courtesy of M.G. Zabetakis, AIChE-Inst. Chem. Engr. Symp., Ser. 2, Chem.
Engr. Extreme Cond. Proc. Symp. American Institute of Chemical Engineers, New York, 1965, 99–104.)
FIGURE 10.7 Ethylene oxide plant explosion caused by autoignition. (Courtesy of T. Kletz, What Went Wrong: Case Histories of
Process Plant Disasters, 4th ed., Gulf Publishing, Houston, TX, 1998.)
FIGURE 10.8 Safety documentation feedback flow chart. (Courtesy of GPSA Engineering Data Book, Vol. II, 10th ed., Gas
Processors and Suppliers Association, Tulsa, OK, 1994.)
FIGURE 10.9 Refinery damaged due to improper maintenance procedures. (Courtesy of R.E. Sanders, Chemical Process Safety:
Learning from Case Histories, Butterworth–Heinemann, Woburn, MA, 1999.)

CHAPTER 11
FIGURE 11.1 Graph of sustainable combustion for methane.
FIGURE 11.2 Typical raw gas burner tips.
FIGURE 11.3 Typical premix metering orifice spud and air mixer assembly.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 11.4 Typical gas fuel capacity curve.

Tables and Figures xxv

FIGURE 11.5 Typical liquid fuel atomizer/spray tip configurations.

FIGURE 11.6 Typical liquid fuel capacity curve.
FIGURE 11.7 Typical throat of a raw gas burner.
FIGURE 11.8 Ledge in the burner tile.
FIGURE 11.9 Flame stabilizer or flame holder.
FIGURE 11.10 Swirler.
FIGURE 11.11 Round-shaped flame.
FIGURE 11.12 Flat-shaped flame.
FIGURE 11.13 Internal mix twin fluid atomizer.
FIGURE 11.14 Port mix twin fluid atomizer.
FIGURE 11.15 Regen tile and swirler.
FIGURE 11.16 Typical conventional raw gas burner.
FIGURE 11.17 Typical premix gas burner.
FIGURE 11.18 Typical round flame combination burner.
FIGURE 11.19 Typical round flame, high-intensity combination burner.
FIGURE 11.20 Typical staged-fuel flat flame burner.
FIGURE 11.21 Typical radiant wall burner.

CHAPTER 12
FIGURE 12.1 Programmable logic controller.
FIGURE 12.2 Touch screen.
FIGURE 12.3 Simplified flow diagram of a standard burner light-off sequence.
FIGURE 12.4 Simple analog loop.
FIGURE 12.5 Feedforward loop.
FIGURE 12.6 Double-block-and-bleed system.
FIGURE 12.7 Failsafe input to programmable logic controller.
FIGURE 12.8 Shutdown string.
FIGURE 12.9 Typical pipe rack.
FIGURE 12.10a Large control panel.
FIGURE 12.10b Small control panel.
FIGURE 12.11 Inside the control panel.
FIGURE 12.12a Pressure switch.
FIGURE 12.12b Pressure switch.
FIGURE 12.13 Pneumatic control valve.
FIGURE 12.14 Control valve characteristics.
FIGURE 12.15 Thermocouple.
FIGURE 12.16 Thermowell and thermocouple.
FIGURE 12.17 Velocity thermocouple.
FIGURE 12.18 Pressure transmitter (left) and pressure gauge (right).

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 12.19 Mechanically linked parallel positioning.
FIGURE 12.20 Electronically linked parallel positioning.
FIGURE 12.21 A variation of parallel positioning.
FIGURE 12.22 Fuel flow rate vs. control signal.
FIGURE 12.23 Typical butterfly-type valve calculation.
FIGURE 12.24 The required shape of the air valve characterizer.
FIGURE 12.25 Fully metered control scheme.
FIGURE 12.26 Fully metered control scheme with cross limiting.
FIGURE 12.27 O2 trim of air flow rate.
FIGURE 12.28 O2 trim of air setpoint.
FIGURE 12.29 Multiple fuels and O2 sources.
FIGURE 12.30 Controller.
FIGURE 12.31 Analog controller with manual reset.
FIGURE 12.32 Analog controller with automatic reset.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxvi Thursday, April 7, 2005 4:00 PM

xxvi The John Zink Combustion Handbook

CHAPTER 13
FIGURE 13.1 Contrast of classical experimentation and SED methods.
FIGURE 13.2 NOx contours for furnace temperature and oxygen concentration based on Equation 13.1.
FIGURE 13.3 A fuel-staged burner.
FIGURE 13.4 Municipal solid waste boiler using ammonia injection to control NOx.
FIGURE 13.5 Method of steepest ascent.
FIGURE 13.6 A combination burner capable of firing either oil or gas or both simultaneously.
FIGURE 13.7 Simplex design for q = 3.
FIGURE 13.8 Mixture factors, a transformation, and a combined mixture-factorial.
FIGURE 13.9 Flowchart showing a general sequential experimental strategy.
FIGURE 13.10 Some orthogonal designs for f = 3 arranged in a sequential strategy.

CHAPTER 14
FIGURE 14.1 John Zink Co., LLC, Research and Development Test Center, Tulsa, Oklahoma
FIGURE 14.2 Test furnace for simulation of ethylene furnace.
FIGURE 14.3 Test furnace for simulation of down-fired tests.
FIGURE 14.4 Test furnace for simulation of up-fired tests.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 14.5 Test furnace for simulation of terrace wall reformers.

FIGURE 14.6 Test fuel storage tanks.
FIGURE 14.7 Forced-draft air preheater.
FIGURE 14.8 Heat flux probe schematic.
FIGURE 14.9 Ellipsoidal radiometer schematic.

CHAPTER 15
FIGURE 15.1 Heater cutout and burner bolt circle on a new heater.
FIGURE 15.2 Warped steel on the shell of a heater.
FIGURE 15.3 Burner improperly installed at an angle due to a warped shell.
FIGURE 15.4 Donut ring for leveling burner mounting onto the warped shell of a heater.
FIGURE 15.5 Typical burner drawing.
FIGURE 15.6 Burner mounted on the floor of a heater.
FIGURE 15.7 Burner mounted on the side of a heater.
FIGURE 15.8 Burner mounted on the top of heater.
FIGURE 15.9 Burner mounted in a common plenum.
FIGURE 15.10 Burner in a plenum box mounted to a heater.
FIGURE 15.11 Piping improperly loaded on the burner inlet.
FIGURE 15.12 Picture of a burner tile showing multiple tile pieces.
FIGURE 15.13 Sketch showing a round tile measured in 3 different diameters.
FIGURE 15.14 Sketch showing a square tile measured at different lengths and widths.
FIGURE 15.15 Oil tip in combination burner showing oil tip locations.
FIGURE 15.16 Welding rods in an oil tip.
FIGURE 15.17 VYD burner gas tip in a diffuser with a pilot.
FIGURE 15.18 VYD drawing showing the diffuser cone and the pilot tip.
FIGURE 15.19 Example of an air register.
FIGURE 15.20 Typical fuel gas piping system.
FIGURE 15.21 Typical heavy fuel oil piping system.
FIGURE 15.22 Typical light fuel oil piping system.
FIGURE 15.23 A pat-826 gas tip.
FIGURE 15.24 Example oil gun atomizer.
FIGURE 15.25 Catatlyst deposit within an oil burner tile.
FIGURE 15.26 Typical diffuser cone.
FIGURE 15.27 Typical spin diffuser.
FIGURE 15.28 Example of a damaged stabilizer.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxvii Thursday, April 7, 2005 4:00 PM

Tables and Figures xxvii

FIGURE 15.29 Example of a damaged pilot tip.

FIGURE 15.30 New ST-1S pilot tip without an electronic ignitor.
FIGURE 15.31 New ST-1SE pilot with an electronic ignitor.

CHAPTER 16
FIGURE 16.1 Typical draft measurement points.
FIGURE 16.2 Inclined manometer.
FIGURE 16.3 Excess air indication by oxygen content.
FIGURE 16.4 Location for measuring excess oxygen.
FIGURE 16.5 Oxygen analyzer.
FIGURE 16.6 Cost of operating with higher excess oxygen levels (natural gas).
FIGURE 16.7 Cost of operating with higher excess oxygen levels (#6 oil).
FIGURE 16.8 Fuel gas pressure measurement.
FIGURE 16.9 Graph of fuel pressure vs. heat release.
FIGURE 16.10 Viscosity vs temperature for a range of hydrocarbons.
FIGURE 16.11 Velocity thermocouple.
FIGURE 16.12 Air control device schematic.
FIGURE 16.13 Picture of air control device.
FIGURE 16.14 Primary air door.
FIGURE 16.15 Burner ignition ledge.
FIGURE 16.16 Gas tips.
FIGURE 16.17 Oil tips.
FIGURE 16.18 Long narrow and short bushy flames.
FIGURE 16.19 Typical flame envelope with x-y-z axes.
FIGURE 16.20 Sodium ions in the flame.
FIGURE 16.21 Example of a good flame within the firebox.
FIGURE 16.22 Example of a very bad flame pattern in a fire box.
FIGURE 16.23 Typical draft profile in a natural draft heater.
FIGURE 16.24 Logic diagram for tuning a natural draft heater.
FIGURE 16.25 Logic diagram for tuning a balanced draft heater.
FIGURE 16.26 Unstable flame.
FIGURE 16.27 Broken burner tile.
FIGURE 16.28 Dark line or black streaks on hot refractory surface indicating air leaks.

CHAPTER 17
FIGURE 17.1 Coke deposit causes tube thinning.
FIGURE 17.2 Cracked gas tip causing an irregular flame pattern.
FIGURE 17.3 Damaged diffuser cone.
FIGURE 17.4 Effect of excess O2 on NOx in raw gas burners.
FIGURE 17.5 Effect of combustion air temperature on NOx.
FIGURE 17.6 Effect of firebox temperature on NOx.
FIGURE 17.7 Effect of bound nitrogen in the liquid fuel on NOx.
FIGURE 17.8 Effect of burner model on NOx.
FIGURE 17.9 Staged air burner.
FIGURE 17.10 Staged fuel burner.
FIGURE 17.11 Ultra low NOx burner.

CHAPTER 18
FIGURE 18.1
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Typical plant schematic.
FIGURE 18.2 Cogeneration at Teesside, England. Courtesy of Nooter/Eriksen. St. Louis, MO. With permission.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxviii Thursday, April 7, 2005 4:00 PM

xxviiiThe John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 18.3 Combination (oil and gas) fire duct burners at Dahbol, India. (Courtesy of Ms. Martha Butala, Dabhol Power Company,
Bombay, India, as published in Power Magazine.)
FIGURE 18.4 Typical location of duct burners in an HRSG. (Courtesy of Deltak, Minneapolis, MN. With permission.)
FIGURE 18.5 Schematic of HRSG at Teesside, England. (Courtesy of Nooter/Eriksen, St. Louis, MO. With permission.)
FIGURE 18.6 Fluidized bed startup duct burner.
FIGURE 18.7 An inline burner.
FIGURE 18.8 Linear burner elements.
FIGURE 18.9 Gas flame from a grid burner.
FIGURE 18.10 Oil flame from a side-fired oil gun.
FIGURE 18.11 Approximate requirement for augmenting air.
FIGURE 18.12 Duct burner arrangement.
FIGURE 18.13 Comparison of flow variation with and without straightening device.
FIGURE 18.14 Physical model of burner.
FIGURE 18.15 Sample result of CFD modeling performed on an HRSG inlet duct.
FIGURE 18.16 Drilled pipe duct burner.
FIGURE 18.17 Low emission duct burner.
FIGURE 18.18 Flow patterns around flame stabilizer.
FIGURE 18.19 Effect of conditions on CO formation.
FIGURE 18.20 Typical main gas fuel train: single element or multiple elements firing simultaneously.
FIGURE 18.21 Typical main gas fuel train: multiple elements with individual firing capability.
FIGURE 18.22 Typical pilot gas train: single element or multiple elements firing simultaneously.
FIGURE 18.23 Typical pilot gas train: multiple elements with individual firing capability.
FIGURE 18.24 Typical main oil fuel train: single element.
FIGURE 18.25 Typical main oil fuel train: multiple elements.
FIGURE 18.26 Typical pilot oil train: single element.
FIGURE 18.27 Typical pilot oil train: multiple elements.

CHAPTER 19
FIGURE 19.1 Typical utility boilers. (Courtesy of Florida Power & Light.)
FIGURE 19.2 Typical single-burner industrial boiler. (Courtesy of North Carolina Baptist Hospital.)
FIGURE 19.3 Swirl burner.
FIGURE 19.4 Average flame length as a function of burner heat input.
FIGURE 19.5 A typical low-NOx burner, venturi-style.
FIGURE 19.6 A typical low-NOx burner, a venturi-style (second example).
FIGURE 19.7 A strong flame front established within a maximum of 0.5 diffuser diameters of the face of the diffuser.
FIGURE 19.8 The effects of boiler design on NOx.
FIGURE 19.9 The NOx of various boilers included in the database on oil and gas, respectively.
FIGURE 19.10 NOx generation for natural gas and No. 6 oil (0.5% Nf) vs. adiabatic flame temperature.
FIGURE 19.11 NOx vs. excess O2 with FGR implementation.
FIGURE 19.12 NOx vs. excess O2 (The TFM-94 boiler equipped with nine boilers, at ~94% load).
FIGURE 19.13 NOx vs. relative steam flow at the TGM-94 boilers (natural gas, O2 = 1.2–1.6%).
FIGURE 19.14 NOx vs. relative steam flow at the TGM-94 boilers (No. 6 oil, O2 = 1.2–1.6%).
FIGURE 19.15 NOx vs. relative steam flow with firing natural gas on the TGME-206 boiler equipped with Todd Combustion low-
NOx Dynaswirl burners at O2 = 0.8–1.0%.
FIGURE 19.16 NOx vs. load with firing natural gas on utility burners.
FIGURE 19.17 Degree of the power function NOx = f (load) vs. bounded nitrogen in No. 6 oil.
FIGURE 19.18 Relative NOx vs. relative load on industrial boilers firing natural gas and No. 6 oil with ambient air.
FIGURE 19.19 Relative NOx vs. relative load on industrial boilers firing natural gas and No. 6 oil with preheated air.
FIGURE 19.20 Effect of furnace cleanliness on NOx emissions.
FIGURE 19.21 Effect of HRA cleanliness on NOx emissions.
FIGURE 19.22 Effect of air in-leakage on the burner performance.
FIGURE 19.23 Improvement of mass flow distribution to burners (differences within ±2%).
FIGURE 19.24 Improvement of peripheral air flow distribution to burners (deviations ±10%)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxix Thursday, April 7, 2005 4:00 PM

Tables and Figures xxix

FIGURE 19.25 Improvement of FGR flow distribution to burners.

FIGURE 19.26 A scaled, physical, aerodynamic simulation model.
FIGURE 19.27 Flame-to-flame similarity of appearance.
FIGURE 19.28 Premixing the FGR flow with the combustion air upstream of the windbox.
FIGURE 19.29 Relative NOx concentration vs. FGR flow rate with firing natural gas in the utility boilers at full load.
FIGURE 19.30 Relative NOx concentration vs. FGR flow rate with firing natural gas in the 800 MWe boiler at full load.
FIGURE 19.31 NOx reduction vs. flue gas flow.
FIGURE 19.32 Full load NOx reduction data.
FIGURE 19.33 Overfire air flow rate influence on NOx reduction and CO emission with firing natural gas in the TGM-94 boiler at
~94% load and O2 ~ 1.2%.
FIGURE 19.34 “Horizontal” imbalance intensifies the combustion process and especially burnout while firing oil.
FIGURE 19.35 Data comparing single and three-stage gas combustion.
FIGURE 19.36 NOx generation with firing natural gas and No. 6 oil (0.5% Nf) vs. the theoretical maximal flame temperature.
FIGURE 19.37 The ratio between NOx numbers obtained with gas and oil firing are considered as a function of the furnace space
heat release.
FIGURE 19.38 Schematic for typical low NOx burner.
FIGURE 19.39 NOx vs. heat load with firing natural gas with ambient air in the packaged industrial boiler equipped with the Todd
Combustion low NOx burner.
FIGURE 19.40 CO vs. heat load with firing natural gas with ambient air in the packaged industrial boiler equipped with the Todd Combustion
low NOx burner.
FIGURE 19.41 NOx, O2 and opacity vs. relative heat load with firing No. 6 oil with ambient air in the packaged industrial boiler
equipped with the Todd Combustion low NOx burner.
FIGURE 19.42 Internal FGR impact on NOx with firing No. 6 oil with ambient air in the 30–100% load range.
FIGURE 19.43 NOx and CO emissions vs. injector gas flow rate/total gas flow rate ratio, at the boiler equipped with the Todd
Combustion low-NOx burner firing natural gas with preheated air at full load.
FIGURE 19.44 An ultra-low emissions burner.
FIGURE 19.45 Relationship between the adiabatic flame temperature and thermal NOx formation.
FIGURE 19.46 HCN and NH3 formation at three flame temperatures.
FIGURE 19.47 A nearly uniform fuel/air mixture at the ignition point.
FIGURE 19.48 Swirl vanes also serve as the gas injectors, and provide the burner’s near-perfect fuel/air mixing.
FIGURE 19.49 Outer sleeve contains a second set of gas injector vanes attached to an outer gas reservoir.
FIGURE 19.50 NOx emission from the ultra-low emissions burner firing into the firetube boiler for ambient, 300oF preheat, and
500oF preheat as a function of FGR rate.
FIGURE 19.51 Data from the application of an ultra-low emissions burner on a new 230,000 lb/hr “A” type Nebraska boiler.
FIGURE 19.52 The ultra-low emissions burner can also be used on two-burner applications where NFPA 8501 guidelines are being
followed.
FIGURE 19.53 Atomizers are fitted onto oil “guns”.
FIGURE 19.54 Steam assist, air blast, or other “external mix” atomizers.

CHAPTER 20
FIGURE 20.1 Typical early 1950s flare performance.
FIGURE 20.2 First successful smokeless flare.
FIGURE 20.3 Major flaring event.
FIGURE 20.4 Typical elevated single point flare.
FIGURE 20.5 Typical pit flare installation.
FIGURE 20.6 A grade-mounted, multi-point LRGO flare system.
FIGURE 20.7 Elevated multi-point LRGO flare system.
FIGURE 20.8 Multiple ZTOF installation in an ethylene plant.
FIGURE 20.9 Combination ZTOF and elevated flare system.
FIGURE 20.10 Comparison of the flame produced by burning (a) 25 MW well head natural gas, (b) propane, and (c) propylene.
FIGURE 20.11 Combination elevated LRGO and utility flare system.
FIGURE 20.12 General arrangement of a staged flare system, including a ZTOF and an elevated flare.
FIGURE 20.13 John Zink Co. test facility in Tulsa, Oklahoma.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

xxx The John Zink Combustion Handbook

FIGURE 20.14 Liquid carryover from an elevated flare. (a) Start of flaring event. (b) Liquid fallout and flaming rain from flare
flame. (c) Flaming liquid engulfs flare stack.
FIGURE 20.15 Thermogram of a flare flame.
FIGURE 20.16 API radiation geometry.
FIGURE 20.17 Comparison of stack height and relative cost for various radiation calculation methods.
FIGURE 20.18 Effectiveness of steam in smoke suppression.
FIGURE 20.19 Effectiveness of air in smoke suppression.
FIGURE 20.20 SteamizerTM steam-assisted smokeless flare.
FIGURE 20.21 Typical nonassisted flare.
FIGURE 20.22 Zink double refractory (ZDR) severe service flare tip.
FIGURE 20.23 Simple steam-assisted flare.
FIGURE 20.24 Perimeter:area ratio as a function of tip size.
FIGURE 20.25 Schematic of an advanced steam-assisted flare.
FIGURE 20.26 A comparison of the perimeter:area ratio for simple and advanced steam-assisted flares.
FIGURE 20.27 State-of-the-art SteamizerTM flare burner and muffler.
FIGURE 20.28 Air assisted smokeless flare with two blowers in a refinery.
FIGURE 20.29 Annular air flare. (Courtesy of Shell Canada Ltd.)
FIGURE 20.30 Hydra flare burner in an offshore location.
FIGURE 20.31 LRGO staging sequence during a flaring event from inception (a) to full load (g)
FIGURE 20.32 Multi-point LRGO system with a radiation fence.
FIGURE 20.33 A RIMFIRE endothermic flare.
FIGURE 20.34 OWB liquid flare test firing 150 gpm.
FIGURE 20.35 Forced draft Dragon liquid flare.
FIGURE 20.36 Poseidon flare: water-assisted Hydra.
FIGURE 20.37 Fundamental pilot parts.
FIGURE 20.38 Conventional flame front generator.
FIGURE 20.39 Slip stream flame-front generator.
FIGURE 20.40 Self inspirating flame-front generator.
FIGURE 20.41 SoundProof acoustic pilot monitor.
FIGURE 20.42 Horizontal settling drum at the base of an air assisted flare.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 20.43 Cyclone separator.
FIGURE 20.44 Schematic of a vertical liquid seal.
FIGURE 20.45 “Smoke signals” from a surging liquid seal.
FIGURE 20.46 Various seal head designs.
FIGURE 20.47 AirrestorTM velocity-type purge reduction seal.
FIGURE 20.48 Molecular Seal density-type purge reduction seal.
FIGURE 20.49 Schematic of a ZTOF.
FIGURE 20.50 Self-supported flare.
FIGURE 20.51 Guy wire-supported flare.
FIGURE 20.52 Derrick supported flare.
FIGURE 20.53 Demountable derrick.
FIGURE 20.54 Flare support structure selection guide.
FIGURE 20.55 Steam control valve station.
FIGURE 20.56 Staging control valve assembly.
FIGURE 20.57 Loop seal.
FIGURE 20.58 Purge control station.
FIGURE 20.59 Geometry for dispersion calculations.

CHAPTER 21
FIGURE 21.1 Typical natural-draft burner.
FIGURE 21.2 Typical medium pressure drop burner.
FIGURE 21.3 Typical high pressure drop burner.
FIGURE 21.4 Typical horizontal system with a preheat exchanger.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxxi Thursday, April 7, 2005 4:00 PM

Tables and Figures xxxi

FIGURE 21.5 Watertube boiler.

FIGURE 21.6 Firetube boiler.
FIGURE 21.7 Typical all-welded shell-and-tube heat exchanger.
FIGURE 21.8 Regenerative preheat exchanger.
FIGURE 21.9 Organic fluid transfer system configuration.
FIGURE 21.10 Vertical, down-flow conditioning section.
FIGURE 21.11 Direct spray contact quench.
FIGURE 21.12 Submerged quench.
FIGURE 21.13 Adjustable-plug venturi quench.
FIGURE 21.14 Baghouse.
FIGURE 21.15 Dry electrostatic precipitator.
FIGURE 21.16 Horizontal venturi scrubber.
FIGURE 21.17 Wet electrostatic precipitator.
FIGURE 21.18 Simple packed column.
FIGURE 21.19 Two-stage acid gas removal system.
FIGURE 21.20 Combination quench/two-stage acid removal system.
FIGURE 21.21 Three-stage NOx reduction process.
FIGURE 21.22 Two-stage NOx reduction process.
FIGURE 21.23 Selective noncatalytic reduction system.
FIGURE 21.24 Common catalyst configuration.
FIGURE 21.25a Fan wheel designs.
FIGURE 21.25b Radial blade operating curve for 1780 RPM, 70oF, and 0.075 lb/ft3 density.
FIGURE 21.25c Forward tip blade operating curve for 1780 RPM, 70oF, and 0.075 lb/ft3 density.
FIGURE 21.25d Backward curved blade operating curve for 1780 RPM, 70o F, and 0.075 lb/ft3 density.
FIGURE 21.25e Outlet damper flow control.
FIGURE 21.25f Radial inlet damper/inlet box damper flow control.
FIGURE 21.25g Blower speed control.
FIGURE 21.26 Simple thermal oxidizer.
FIGURE 21.27 Thermal oxidizer system generating steam.
FIGURE 21.28 Heat recovery thermal oxidation system.
FIGURE 21.29 Bypass recuperative system.
FIGURE 21.30 Horizontal thermal oxidizer with firetube boiler and HCl removal system.
FIGURE 21.31 Vertical thermal oxidizer with 180˚ turn quench section.
FIGURE 21.32 Chlorine reaction equilibrium vs. operating temperature.
FIGURE 21.33 Molten salt system.
FIGURE 21.34 Online cleaning with soot blowers.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 21.35 Three-stage NOx system with packed column scrubber.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxxii Thursday, April 7, 2005 4:00 PM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

About the Editor

Charles E. Baukal, Jr., Ph.D., P.E., is the Director of the author of the book Heat Transfer in Industrial Combustion,
John Zink Company, LLC R&D Test Center in Tulsa, OK. He and the general editor of the Industrial Combustion series, all
previously worked for 13 years at Air Products and Chemi- with CRC Press, Boca Raton, FL. He has a Ph.D. in mechani-
cals, Inc., Allentown, PA, in the area of oxygen-enhanced cal engineering from the University of Pennsylvania, is a
combustion. He has more than 20 years of experience in the licensed Professional Engineer in the state of Pennsylvania,
fields of heat transfer and industrial combustion and has has been an adjunct instructor at several colleges, and has
authored more than 50 publications in those fields. He is the eight U.S. patents. He is a member of several Who’s Who
editor of the books Oxygen-Enhanced Combustion and Com- compilations. He is a member of the American Society of
putational Fluid Dynamics in Industrial Combustion, the Mechanical Engineers and The Combustion Institute.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

xxxiii
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxxiv Thursday, April 7, 2005 4:00 PM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Contributors
John Ackland is a Test Engineer at the John Zink Com- John Zink Company in the field of industrial burners and
pany, Tulsa. He has worked in the field of combustion for two combustion equipment since 1974. He has co-authored a
years and has a bachelor’s degree in chemical engineering number of papers and presentations covering combustion,
from the University of Tulsa. combustion equipment, and combustion-generated emissions,
Eugene Barrington is retired as a Senior Staff engineer and is co-holder of several combustion-related patents.
for Shell Oil Company. He had primary responsibility for the Joseph Colannino, P.E., is the Director of Engineering
design, selection, application, and performance improvement and Design at the John Zink Company, LLC, Tulsa, OK, and
of fired equipment. He has taught public courses on this topic is a registered professional chemical engineer with more than
for many years, publishing extensive notes. He has also pub- 15 years of experience regarding combustion and combus-
lished in Hydrocarbon Processing and Chemical Engineering tion-related emissions. He has over 20 publications to his
Progress. Mr. Barrington was heavily involved in writing API credit and is listed in Who’s Who in Science and Engineering,
specifications for both fired and unfired heat transfer equip- Who’s Who in California, and Who’s Who in Finance and
ment. He has an M.S. degree in engineering science from the Industry. He is a member of the American Institute for Chem-
University of California at Berkeley and is a registered Pro- ical Engineers and the American Chemical Society.
fessional Engineer in Texas.
Terry Dark is the Coordinator of Engineering and Tech-
Peter Barry is the Director of Duct Burners for the John
nology Programs at Oklahoma State University’s Tulsa cam-
Zink Company, LLC, Tulsa, OK. He has a B.S. in Mechanical
pus, where he is currently pursuing a Masters of Science in
Engineering from Lehigh University.
Engineering and Technology Management. Terry worked pre-
Lawrence D. (Larry) Berg is a Senior Development Engi-
viously for the John Zink Company as a Combustion Test
neer at the John Zink Company, LLC, Tulsa, OK. He has over
Engineer, focusing on product testing, burner development,
ten years’ experience as a research and product development
and combustion safety. He is a 1998 graduate of Oklahoma
engineer for the company, and has a Masters Degree in
State University’s School of Chemical Engineering.
Mechanical Engineering from MIT. He holds four U.S. patents,
has co-authored five publications, and authored numerous Joe Gifford is a Senior Engineer, Instrumentation and Con-
internal technical documents. trol Systems, at the John Zink Company, LLC, Tulsa, OK.
Wes Bussman is Research and Development Engineer at He has worked in the field of control and facilities design for
the John Zink Company, LLC, Tulsa, OK. He has worked in 40 years and has a B.S. in Physics. For many years, he has
the field of combustion and fluid dynamics and has a Ph.D. conducted company training classes for Control Engineers
in Mechanical Engineering from the University of Tulsa. He and Technicians. He has received numerous awards for inno-
has authored five publications and has two patents. Honors vative control system designs throughout his career, including
achieved include Kappa Mu Epsilon Mathematical Society the General Electric Nuclear Energy Division’s Outstanding
and Sigma Xi Research Society. Engineering award for systems design over a 15-year period.
I-Ping Chung, Ph.D., is a Development Engineer in Tech- Technical society memberships have included the Pacific
nology and Commercial Development Group at the John Zink Association of General Electric Scientists and Engineers
Company, LLC, Tulsa, OK. She has worked in the field of (PAGESE), Instrument Society of America (ISA), American
atomization and sprays, spray combustion, and laser diagno- Society of Mechanical Engineers (ASME), and the National
sis in combustion and has a Ph.D. degree in Mechanical and Fire Protection Association (NFPA).
Aerospace Engineering. She has authored 14 publications and Karl Graham, Ph.D., is currently the Process Engineering
has two patents. She is a registered Professional Engineer of Director at UniField Engineering, Inc., Billings, MT. Karl
Mechanical Engineering in California and Iowa. was formerly the Manager of Flare Design and Development
Michael G. Claxton is a Senior Principal Engineer in the at the John Zink Company, LLC, Tulsa, OK. Karl’s work has
Burner Process Engineering Group of the John Zink Com- been in the design, testing, engineering, and specification of
pany, LLC, Tulsa, OK. He has a B.S. in Mechanical Engi- flare, incineration, and gas cleaning equipment. He has a
neering from the University of Tulsa and has worked for the Ph.D. in chemical engineering from MIT, is a member of the

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
xxxv
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxxvi Thursday, April 7, 2005 4:00 PM

xxxviThe John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
American Institute of Chemical Engineers and is an active waste materials and he has several publications in that area.
participant in the American Petroleum Institute. He has also presented information, by invitation, to the Okla-
John Guarco is a Combustion Specialist at the John Zink homa Senate Select Committee on Waste Incineration. Mr.
Company, LLC, Shelton, CT. He has worked with burners for Melton has been a Registered Professional Engineer for more
utility and industrial boiler applications for seven years and than 20 years.
has a Masters of Science in Mechanical Engineering. He has Robert E. Schwartz, P.E., is a Vice President at John Zink
authored eight publications and holds three patents. Company, LLC, Tulsa, OK. He has worked in the fields of
Robert Hayes is a Test Engineer at the John Zink Com- combustion, heat transfer, and fluid flow for 40 years and has
pany, LLC, Tulsa, OK. He has worked in the fields of com- an M.S. in mechanical engineering. Mr. Schwartz has been
bustion, heat transfer, and experimentation for three years. granted 50 U.S. patents, has had a number of articles pub-
He has an M.S. in mechanical engineering from Brigham lished and has spoken at many national society meetings. He
Young University. is a registered professional engineer, a member of ASME and
Jim Heinlein is a Senior Controls Engineer at John Zink AIChE, and an associate member of Sigma Xi. He was elected
Company, LLC, Tulsa, OK, where he has been employed for to The University of Tulsa Engineering Hall of Fame in 1991
10 years. He has also worked in Nuclear Engineering, Com- and was a 1988 recipient of the University of Missouri Honor
puter Systems Design, and Low Observables Engineering. He Award for Distinguished Service to Engineering. He is a
is a member of ISA, IEEE, and Tau Beta Pi. He is also member of the Mechanical Engineering Industrial Advisory
qualified as a Naval Surface Warfare Expert. Board, The University of Tulsa. He has been a John Zink
Michael Henneke, Ph.D., is a CFD Engineer at John Zink Burner School instructor for 30 years and served as Director
Company, LLC, Tulsa, OK. His academic background is in of the School for 10 years.
the area of reacting flow modeling and radiative transport. He Prem C. Singh, Ph.D., is a Test Engineer at the John Zink
holds a Ph.D. in mechanical engineering from The University Company, LLC, Tulsa, OK. He has worked in the fields of
of Texas at Austin. He has published three refereed journal combustion, energy engineering, and transport phenomena
papers, as well as many non-refereed articles and has given for over 20 years and has a Ph.D. in chemical engineering.
a number of presentations on computational fluid dynamic He has authored more than 50 publications and is a contrib-
modeling of industrial combustion systems. utor to a book on coal technology published by Delft Uni-
Jaiwant D. Jayakaran (Jay Karan) is Director, Burner versity of Technology. He has worked as a reviewer for the
Technology at the John Zink Company, LLC, Tulsa, OK. He Canadian Journal of Chemical Engineering and Industrial
has worked in the fields of combustion, petrochemicals, and and Engineering Chemistry Research, is a member of AIChE,
power, with responsibilities in R&D, plant operations, and ASME, Sigma XI, and the New York Academy of Sciences,
engineering. Jay has an M.S. in mechanical engineering. He and has been cited in Marquis’s Who’s Who in Science and
has authored several technical articles and papers over the Engineering. He has taught courses in chemical and mechan-
years, and has several patents pending. ical engineering to undergraduate and graduate classes during
Jeff Lewallen is an Account Manager at John Zink Com- his long tenure as a faculty /visiting faculty member.
pany, LLC, Tulsa, OK. He has worked in the field of com- Joseph D. Smith, Ph.D., is Director of Flare Technology
bustion for eight years. He graduated from the University of and Computational Fluid Dynamics at the John Zink Com-
Tulsa in 1992 and holds a B.S. in mechanical engineering. pany, LLC, Tulsa, OK. He has worked in the field of CFD for
Michael Lorra, Ph.D., has been a CFD engineer for the nearly twenty years and has a Ph.D. in chemical engineering.
John Zink Company, LLC, Tulsa, OK, since 1999. Previous He has authored 27 peer-reviewed publications, 16 invited
to that, he worked at Gaswaerme Institut, Essen, Germany, lectures, 19 conference papers, two patents, and has organized
e.V for eight years, where he also finished his Ph.D. He gained and directed three special symposia. As a member of the
experience in NOx reduction techniques, especially in reburn- American Institute of Chemical Engineering, he has served as
ing technology. He developed his own software code for the National Chair of the Student Chapters Committee and as
computation of turbulent reacting flow problems using lami- Local Chair of the Mid-Michigan AIChE section. Research
nar flamelet libraries. topics include reaction engineering and turbulent reactive flow
Paul Melton is a Senior Principal Engineer for the Thermal simulation. He has taught undergraduate and graduate courses
Oxidation Systems Group at the John Zink Company, LLC, in chemical engineering at the University of Michigan and at
Tulsa, OK. He received a BSME from Oklahoma State Uni- the University of Illinois/Urbana–Champaign.
versity and has worked in the field of combustion for more Stephen L. Somers is a Senior Process Engineer at the
than 25 years. His specialty is combustion of all types of John Zink Company, LLC, Tulsa, OK. He has 32 years of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xxxvii Thursday, April 7, 2005 4:00 PM

Contributors xxxvii

experience in combustion and process design with 15 years and has a B.S. in mechanical engineering. He has authored
in sales and design of duct burners for supplementary firing. three publications and is a licensed professional mechanical
He has an M.S. in chemical engineering from the University engineer in California.
of Oklahoma.
Jeff White is the Senior Flare Design Consultant at the
Lev Tsirulnikov, Ph.D., is a Senior Research Engineer for
John Zink Company, LLC, Tulsa, OK. He has worked in the
the John Zink Company, LLC, Shelton, CT. He has developed
field of Flare System Design at John Zink Company for 19
low-emission combustion technologies, burners, and other
years. He has an M.S. in mechanical engineering from The
equipment for gas/oil-fired utility and industrial boilers. He
University of Texas at Austin. He has published two articles,
has a Ph.D. in mechanical engineering. He holds 47 patents
one on flare radiation methods and the other on flow mea-
and has published more than 100 technical papers, including
surement by ASME nozzles.
four books in the combustion/boiler field.
Richard T. Waibel, Ph.D., is a Senior Principal Engineer Roger H. Witte is Director of Marketing and Sales-End
in the Burner Process Engineering Group at the John Zink Users at the John Zink Company, LLC, Tulsa, OK. He has
Company, LLC, Tulsa, OK. He works in the field of burner worked at John Zink for 28 years in the development and
design and development and has a doctorate in fuel science application of Zink equipment and at Conoco for seven years
from The Pennsylvania State University. He has authored over in their refining and chemical operations. He is also Director
70 technical papers, publications, and presentations. Dr. of the John Zink Burner School which teaches the art of
Waibel has been the Chairman of the American Flame combustion. Roger has a B.S. in refining and chemical engi-
Research Committee since 1995. neering from the Colorado School of Mines and has 35 years
Timothy Webster is the Sales Manager for New Technol- of experience in the fields of combustion and operations of
ogies at the John Zink Company, LLC, Shelton, CT. He has combustion equipment. Roger is a member of the Tulsa Engi-
worked in the field of industrial combustion for seven years neering Society.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Table of Contents
Prologue.................................................................................................................................................xxxvii

Chapter 1: Introduction...............................................................................................................................3
Charles E. Baukal, Jr.

Chapter 2: Fundamentals .......................................................................................................................... 33

Joseph Colannino and Charles E. Baukal, Jr.

Chapter 3: Heat Transfer........................................................................................................................... 69

Prem Singh, Michael Henneke, Jaiwant D. Jayakaran, Robert Hayes, and Charles E. Baukal, Jr.

Chapter 4: Fundamentals of Fluid Dynamics........................................................................................ 117

Lawrence D. Berg, Wes Bussman, and Michael Henneke

Chapter 5: Fuels ....................................................................................................................................... 157

Terry Dark, John Ackland, and Jeff White

Chapter 6: Pollutant Emissions............................................................................................................... 189

Charles E. Baukal, Jr. and Joseph Colannino

Chapter 7: Noise ....................................................................................................................................... 223

Wes Bussman and Jaiwant D. Jayakaran

Chapter 8: Mathematical Modeling of Combustion Systems............................................................... 251

Lawrence D. Berg, Wes Bussman, Michael Henneke, and I-Ping Chung

Chapter 9: CFD Based Combustion Modeling ...................................................................................... 287

Michael Henneke, Joseph D. Smith, Michael Lorra, and Jaiwant D. Jayakaran

Chapter 10: Combustion Safety .............................................................................................................. 327

Terry Dark and Charles E. Baukal, Jr.

Chapter 11: Burner Design...................................................................................................................... 351

Richard T. Waibel and Michael Claxton

Chapter 12: Combustion Controls .......................................................................................................... 372

Joe Gifford and Jim Heinlein

xxxix --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

xl The John Zink Combustion Handbook

Chapter 13: Experimental Design for Combustion Equipment ...........................................................401

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Joseph Colannino

Chapter 14: Burner Testing .....................................................................................................................431

Jeffrey Lewallen, Robert Hayes, Prem Singh, and Richard T. Waibel

Chapter 15: Installation and Maintenance.............................................................................................449

Roger H. Witte and Eugene A. Barrington

Chapter 16: Burner/Heater Operations..................................................................................................469

Roger H. Witte and Eugene A. Barrington

Chapter 17 Troubleshooting.....................................................................................................................501
Roger H. Witte and Eugene A. Barrington

Chapter 18: Duct Burners........................................................................................................................523

Peter F. Barry and Stephen L. Somers

Chapter 19: Boiler Burners......................................................................................................................547

Lev Tsirulnikov, John Guarco and Timothy Webster

Chapter 20 Flares......................................................................................................................................589
Robert Schwartz, Jeff White, and Wes Bussman

Chapter 21: Thermal Oxidizers...............................................................................................................637

Paul Melton and Karl Graham

Appendices
Appendix A: Physical Properties of Materials ........................................................................................................695
Appendix B: Properties of Gases and Liquids ........................................................................................................715
Appendix C: Common Conversions .....................................................................................................................725

Index...........................................................................................................................................................729

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Prologue
Fred Koch and John Zink
Pioneers in the Petroleum Industry
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Fred Koch John Zink

The early decades of the twentieth century saw the birth and safely and efficiently. Early burner designs made even natural
growth of the petroleum industry in Oklahoma. Drilling der- gas difficult to burn as traditional practice and safety concerns
ricks sprouted like wildflowers throughout the state, making led to the use of large amounts of excess air and flames that
it among the top oil producers in the nation and Tulsa the nearly filled the fire box. Such poor burning qualities hurt
“Oil Capital of the World” by the 1920s. plant profitability.
Refining operations accompanied oil production. Many of Among firms engaged in natural gas gathering and sales
the early refineries were so small that today they would be in the northeastern part of the state was Oklahoma Natural
called pilot plants. They were often merely topping processes, Gas Company (ONG). It was there that John Steele Zink,
skimming off natural gasoline and other light fuel products after completing his studies at the University of Oklahoma in
and sending the remainder to larger refineries with more 1917, went to work as a chemist. Zink’s chemistry and engi-
complex processing facilities. neering education enabled him to advance to the position of
Along with oil, enough natural gas was found to make its manager of industrial sales. But while the wasteful use of
gathering and sale a viable business as well. Refineries fre- natural gas due to inefficient burners increased those sales, it
quently purchased this natural gas to fuel their boilers and troubled Zink, and awakened his talents first as an innovator
process heaters. At the same time, these refineries vented and inventor, and then as an entrepreneur.
propane, butane, and other light gaseous hydrocarbons into Seeing the problems with existing burners, Zink responded
the atmosphere because their burners could not burn them by creating one that needed less excess air and produced a

xliii
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xliv Thursday, April 7, 2005 4:00 PM

xliv The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
compact, well-defined flame shape. A superior burner for that Growth of the company required Zink to relocate his family
era, it was technically a premix burner with partial primary and business to larger facilities on the outskirts of Tulsa. In
air and partial draft-induced secondary air. The use of two 1935, he moved into a set of farm buildings on Peoria Avenue,
airflows led to its trade name, BI-MIX®. The BI-MIX® burner a few miles to the south of the city downtown, a location Zink
is shown in a drawing from one of Zink’s earliest patents. thought would allow for plenty of future expansion.
ONG showed no interest in selling his improved burners As time passed, Zink’s company became engaged in mak-
to its customers, so in 1929 Zink resigned and founded Mid- ing numerous other products, sparked by its founder’s beliefs
Continent Gas Appliance Co., which he later renamed the in customer service and solving customer problems. After
John Zink Company. World War II, Zink was the largest sole proprietorship west
of the Mississippi. Zink’s reputation for innovation attracted
Zink’s BI-MIX® burner was the first of many advances in customers who wanted new burners and, eventually, whole
technology made by his company, which to date has seen new families of products. For example, customers began ask-
almost 300 U.S. patents awarded to nearly 80 of its employing for reliable pilots and pilot igniters, when atmospheric
ees. He carried out early manufacturing of the burner in the venting of waste gases and emergency discharges was
garage of his Tulsa-area home and sold it from the back of replaced by combustion in flares in the late 1940s. This in
his automobile as he traveled the Oklahoma oil fields, gen- turn was followed by requests for flare burners and finally
erating the money he needed to buy the components required complete flare systems, marking the start of the flare equip-
to fabricate the new burners. ment industry. Similar customer requests for help in dealing
The novel burners attracted customers by reducing their with gas and liquid waste streams and hydrocarbon vapor led
fuel costs, producing a more compact flame for more efficient the Zink Company to become a major supplier of gas and
heater operation, burning a wide range of gases, and generally liquid waste incinerators and also of hydrocarbon vapor
being safer to use. Word of mouth among operators helped recovery and other vapor control products.
spread their use throughout not only Oklahoma but, by the Mr. Zink’s great interest in product development and inno-
late 1930s, to foreign refineries as well. vation led to the construction of the company’s first furnace
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xlv Thursday, April 7, 2005 4:00 PM

Prologue xlv

for testing burners. This furnace was specially designed to and the former Soviet Union. One of the first of these pro-
simulate the heat absorption that takes place in a process cessing units was installed in a refinery in Duncan, Oklahoma,
heater. Zink had the furnace built in the middle of the in 1928, one year before Zink started his own company.
employee parking lot, a seemingly odd placement. He had While the two men were not personally acquainted, Koch
good reason for this because he wanted his engineers to pass and Zink’s companies knew each other well in those early
the test furnace every day as they came and went from work years. Winkler–Koch Engineering was an early customer for
as a reminder of the importance of product development to Zink burners. The burners were also used in the Wood River
the Company’s success. refinery in Hartford, Illinois. Winkler–Koch constructed this
Zink went beyond encouraging innovation and motivating refinery in 1940 with Fred Koch as a significant part owner
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

his own employees. During the late 1940s, Zink and his and the head of refining operations. Winkler–Koch Engineer-
technical team leader, Robert Reed (who together with Zink ing, and later Koch Engineering, continued to buy Zink burn-
developed the first smokeless flare) sensed a need for an ers for many years.
industry-wide meeting to discuss technologies and experi- Fred Koch and two of his sons, Charles and David, were
ences associated with process heating. In 1950, they hosted even more successful in growing their family business than
the first of four annual Process Heating Seminars in Tulsa. were Zink and his family. When the Zink family sold the John
Interest in the seminars was high, with the attendance level Zink Company to Sunbeam Corporation2 in 1972, the com-
reaching 300. Attendees of the first Process Heating Seminar pany’s annual revenues were US $15 million. By that time Koch
asked Zink and Reed to conduct training sessions for their Industries, Inc., the parent of Koch Engineering, had revenues
operators and engineers. These training sessions, which com- of almost US $1 billion. Since then Koch has continued to
bined lectures and practical hands-on burner operation in grow, its revenues in the year 1999 were over US $30 billion.
Zink’s small research and development center, were the start When the John Zink Company was offered for sale in 1989,
of the John Zink Burner School®. The year 2000 marks the its long association with Koch made Koch Industries a very
fiftieth anniversary of the original seminar and the fiftieth interested bidder. Acting through its Chemical Technology
year in which the Burner School has been offered. Over the Group, Koch Industries quickly formed an acquisition team,
years, other schools were added to provide customer training headed by David Koch, which succeeded in purchasing the
in the technology and operation of hydrocarbon vapor recov- John Zink Company.
ery systems, vapor combustors, and flares. Koch’s management philosophy and focus on innovation
Included among the 150 industry leaders attending the first and customer service sparked a new era of revitalization and
seminar was Harry Litwin, former President and part owner expansion for the John Zink Company. Koch recognized that
of Koch Engineering Co., now part of Koch Industries of the Peoria Avenue research, manufacturing, and office facil-
Wichita, Kansas. Litwin was a panelist at the closing session. ities were outdated. The growth of Tulsa after World War II
Koch Engineering was established in 1943 to provide engi- had made Zink’s facilities an industrial island in the middle
neering services to the oil refining industry. In the early 1950s of a residential area. The seven test furnaces on Peoria Avenue
it developed an improved design for distillation trays and at the time of the acquisition, in particular, were cramped,
because of their commercial success the company chose to with such inadequate infrastructure and obsolete instrumen-
exit the engineering business. Litwin left Koch at that time tation they could not handle the sophisticated research and
and set up his own firm, the Litwin Engineering Co., which development required for modern burners.
grew into a very sizeable business. A fast-track design and construction effort by Koch
During the same period that John Zink founded his busi- resulted in a new office and manufacturing complex in the
ness, another talented young engineer and industry innovator, northeast sector of Tulsa and was completed at the end of
Fred C. Koch, was establishing his reputation as an expert in 1991. In addition, a spacious R&D facility adjacent to the
oil processing. The predecessor to Koch Engineering Co. was new office and manufacturing building replaced the Peoria
the Winkler–Koch Engineering Co., jointly owned by Fred test facility.
Koch with Lewis Winkler, which designed processing units The initial multimillion dollar investment in R&D facilities
for oil refineries. Fred Koch had developed a unique and very included an office building housing the R&D staff and support
successful thermal cracking process which was sold to many personnel, a burner prototype fabrication shop, and an indoor
independent refineries throughout the United States, Europe, laboratory building. Additional features included steam boil-

2Sunbeam Corporation was primarily known as an appliance maker. Less well known was Sunbeam’s group of industrial specialty companies such
as John Zink Company.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-FM-Frame Page xlvi Thursday, April 7, 2005 4:00 PM

xlvi The John Zink Combustion Handbook

ers, fuel storage and handling, data gathering centers, and a customer’s fuel composition. Multiple cameras will provide
measurement instrumentation and data logging for perfor- video images along with the electronic monitoring and

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
mance parameters from fuel flow to flue gas analysis. recording of a wide range of flare test data, including noise
Koch has repeatedly expanded the R&D facility. When the emissions. The facility will be able to test all varieties of flare
new facility began testing activities in 1992, nine furnaces systems with very large sustained gas flow rates at or near
and a multipurpose flare testing area were in service. Today, those levels which customers will encounter in the field.
there are 14 outdoor test furnaces and two indoor research Indeed, flow capacity will match or exceed the smokeless rate
furnaces. Control systems are frequently updated to keep of gas flow for virtually all customers’ industrial plants, giv-
them state of the art. ing the new flare facility a capability unmatched in the world.
Zink is now able to monitor and control burner tests from These world-class test facilities are staffed with engineers
an elevated Control Center that has a broad view of the entire and technicians who combine theoretical training with prac-
test facility. The Control Center includes complete automa- tical experience. They use the latest design and analytical
tion of burner testing with live data on control panels and tools, such as Computational Fluid Dynamics, physical mod-
flame shape viewing on color video monitors. Fuel mixtures eling, and a Phase Doppler Particle Analyzer. The team can
and other test parameters can be varied remotely from the act quickly to deliver innovative products that work success-
control panels inside the Control Center. Up to four separate fully, based on designs which can be exactly verified before
tests in four different furnaces can be conducted and moni- the equipment is installed in the field.
tored simultaneously. Koch’s investment in facilities and highly trained technical
A new flare testing facility is under construction at the time staff carries on the tradition John Zink began more than 70
of this writing to dramatically expand and improve Zink’s years ago: providing our customers today, as he did in his
capabilities. This project represents the company’s largest time, with solutions to their combustion needs through better
single R&D investment since the original construction of the products, applications, information, and service.
R&D facility in 1991. The new facilities will accommodate
the firing of a wide variety of fuel blends (propane, propylene, Robert E. Schwartz
butane, ethylene, natural gas, hydrogen, and diluents such as October, 2000
nitrogen and carbon dioxide) to reproduce or closely simulate Tulsa, Oklahoma
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 2 Monday, March 14, 2005 10:52 AM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 1
Introduction
Charles E. Baukal, Jr.

TABLE OF CONTENTS

1.1 Process Industries........................................................................................................................................ 2

1.1.1 Hydrocarbon and Petrochemical Industries .................................................................................. 2
1.1.2 Power Generation Industry ........................................................................................................... 3
1.1.3 Thermal Oxidation ........................................................................................................................ 5
1.2 Literature Review........................................................................................................................................ 5
1.2.1 Combustion ................................................................................................................................... 5
1.2.2 The Process Industries .................................................................................................................. 6
1.2.3 Combustion in the Process Industries ........................................................................................... 7
1.3 Fired Heaters ............................................................................................................................................... 7
1.3.1 Reformers...................................................................................................................................... 7
1.3.2 Process Heaters ............................................................................................................................. 9

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
1.4 Burners ...................................................................................................................................................... 12
1.4.1 Competing Priorities ................................................................................................................... 13
1.4.2 Design Factors ............................................................................................................................ 14
1.4.3 General Burner Types ................................................................................................................. 16
References .................................................................................................................................................................. 22

3
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 4 Monday, March 14, 2005 10:52 AM

4 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 1.1 Typical petroleum refinery.

1.1 PROCESS INDUSTRIES processes that are used in a typical plant (see Figure 1.1). This
Process industries encompass the production of a wide range differs significantly from most other industrial combustion
of products like fuels (e.g., oil and natural gas), glass, metals systems that normally fire a single purchased fuel such as
(e.g., steel and aluminum), minerals (e.g., refractories, natural gas or fuel oil. Another important challenge is that
bricks, and ceramics), and power, to name a few. The treat- many of the burners commonly used in the hydrocarbon and
ment and disposal of waste materials is another example of a petrochemical industries are natural draft, where the buoyant
process industry. In this book, only a few of these are consid- combustion exhaust products create a draft that induces the
ered and briefly discussed. The main focus of the book is on combustion air to enter the burners. This is different from
the hydrocarbon, petrochemical, power generation, and ther- nearly all other industrial combustion processes, which utilize
mal oxidation industries. a combustion air blower to supply the air used for combustion
in the burner. Natural draft burners are not as easy to control
as forced draft burners, and are subject to things like the wind,
1.1.1 Hydrocarbon and Petrochemical which can disturb the conditions in a process heater.
Industries According to the U.S. Dept. of Energy, petroleum refining
The hydrocarbon and petrochemical industries present unique is the most energy-intensive manufacturing industry in the U.S.,
challenges to the combustion engineer, compared to other accounting for about 7% of total U.S. energy consumption in
industrial combustion processes. One of the more important 1994.1 Table 1.1 shows the major processes in petroleum refin-
challenges in these industries is the wide variety of fuels, ing, most of which require combustion in one form or another.
which are usually off-gases from the petroleum refining Figure 1.2 shows the process flow through a typical refinery.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 5 Monday, March 14, 2005 10:52 AM

Introduction 5

The U.S. Dept. of Energy Office of Industrial Technologies TABLE 1.1 Major Petroleum Refining Processes
has prepared a Technology Roadmap for industrial combustion.2 Category Major Process
For process heating systems, some key performance targets for Topping (separation of crude oil) Atmospheric distillation
the year 2020 have been identified for burners and for the overall Vacuum distillation
system. For the burners, the targets include reducing criteria Solvent deasphalting
Thermal and catalytic cracking Delayed coking
pollutant emissions by 90%, reducing CO2 emissions to levels
Fluid coking/flexicoking
agreed upon by the international community, reducing specific Visbreaking
fuel consumption by 20 to 50%, and maximizing the ability to Catalytic cracking
use multiple fuels. For the heating system, the targets include Catalytic hydrocracking
Combination/rearrangement of
reducing the total cost of combustion in manufacturing, enhanc- hydrocarbon Alkylation
ing system integration, reducing product loss rate by 50%, max- Catalytic reforming
imizing system robustness, and zero accidents. The following Polymerization
Isomerization
were identified as top-priority R&D needs in process heating: Ethers manufacture
a burner capable of adjusting operating parameters in real time, Treating Catalytic hydrotreating/hydroprocessing
advanced burner stabilization methods, robust design tools, and Sweetening/sulfur removal
Gas treatment
economical methods to premix fuel and air. The following were
Specialty product manufacture Lube oil
also identified as top-priority R&D needs in process heating: Grease
new furnace designs, advanced sensors, cost-effective heat Asphalt
recovery processes, and new methods to generate heat without Source: From the U.S. Dept. of Energy.1
environmental impact. Both the burners and the process heaters
are considered in a number of chapters within this book.
Flares (see Figure 1.3) are used to dispose of unwanted gases industries. Duct burners (see a typical flame in Figure 1.4) are
or liquids. Usually, liquids are separated from the gas and burners that are inserted into large ducts (see Figure 1.5) to
burned in the liquid state or vaporized and burned as a gas. The boost the temperature of the gases flowing through the ducts.
unwanted material is generally composed of hydrocarbon These burners are frequently used in co-generation projects,
gases, but may include hydrogen, carbon monoxide, hydrogen electrical utility peaking stations, repowering programs, and in
sulfide, certain other combustible gases, or some amount of an industrial mechanical driver systems employing gas turbines
inert gas such as nitrogen or carbon dioxide. There are several with site requirements for steam. They are also used in fluid-
conditions that may require flaring. The largest flaring events ized bed combustors and chemical process plants. The effi-
occur during emergency pressure relieving conditions associ- ciency of a duct burner to supply additional heat approaches
ated with a sudden unavoidable failure, such as loss of electrical 95%, which is much higher than, for example, a backup boiler
power, loss of cooling water, fire, or the like. Flaring also occurs system in generating more steam. Duct burners are often easily
when gases are vented in order to maintain control of a process retrofitted into existing ductwork. Several important factors in
or during start-up or shutdown of a plant. Yet another cause of duct burner applications include low pollutant emissions, safe
flaring is the disposal of unwanted gases. Examples include operation, uniform heat distribution from the duct burners to
unmarketable natural gas that is co-produced with oil and off- the gases flowing through the duct, getting uniform gas distri-
specification gases produced from a process. Whatever the rea- bution through the duct burners, and having adequate turn-
son, flares must reliably combust gases whenever they are called down to meet fluctuating demands. Duct burners typically use
upon. One of the challenges for flares is initiating and main- gaseous fuels, but occasionally fire on oil.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

taining ignition. Initiation of burning requires a pilot that can

Boiler burners (see Figure 1.6) are used to combust fuels,
withstand high winds, rain, and inert surroundings. Another
challenge is an extremely wide turndown ratio because of the commonly natural gas or fuel oil, in the production of steam,
variety of venting conditions. Environmental challenges include which is often used to produce electrical energy for power
minimizing smoke, radiant head load and noise, and maximiz- generation. These burners produce radiation and convection
ing combustion efficiency. Flares are covered in Chapter 20. used to heat water flowing through the boiler. The water is
vaporized into steam. Sometimes, the steam is used in the
plant in the case of smaller industrial boilers. Larger utility
1.1.2 Power Generation Industry boilers produce steam to drive turbines for electrical energy
This book contains chapters on duct burners (Chapter 18) and production. While boiler burners have been around for many
boiler burners (Chapter 19) used in the power generation years, there have been many design changes in recent years
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 6 Monday, March 14, 2005 10:52 AM

6 The John Zink Combustion Handbook

FIGURE 1.2 Typical refinery flow diagram. (From the U.S. Dept. of Energy.1)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Introduction 7

FIGURE 1.4 Duct burner flame.

FIGURE 1.3 Offshore oil rig flare.

due to current emphasis on minimizing pollutant emissions.

These burners are discussed in detail in Chapter 19.

1.1.3 Thermal Oxidation

Thermal oxidizers (see Figure 1.7) are used to treat unwanted
by-product materials that may be solids, liquids, or gases.
The composition of the by-products varies widely and may
range from minute quantities of a contaminant up to 100%.
These by-products come from a variety of industrial pro-
cesses and often have some heating value, which aids in their
thermal treatment.
There are often many options to choose from to eliminate FIGURE 1.5 Duct burner in large duct.
the by-product materials. While the most preferable is
recycling where the by-products are reused in the process, this
is not always an option in some processes. Land-filling may
be an option for some of the solid waste materials. However,
it is often preferable to completely destroy the waste in an
environmentally safe way. Many other methods are possible,
but thermal treatment is often the most economical and effec-
tive. The waste products must be treated in a way that any
emissions from the treatment process must be below regula-
tory limits. Thermal oxidation is discussed in Chapter 21.

1.2 LITERATURE REVIEW

Numerous books are available on the subjects of both combus-
tion and the process industries considered here. However, few
books have been written on the combination of the two. This
section briefly surveys some of the relevant literature on the
subjects of combustion, the process industries, and the combi-
nation of combustion in those industries. Most of these com-
bustion books were written at a highly technical level for use in
upper-level undergraduate or graduate-level courses. The books
typically have broad coverage, with less emphasis on practical
applications due to the nature of their target audience. FIGURE 1.6 Front of a boiler burner.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

8 The John Zink Combustion Handbook

FIGURE 1.7 Thermal oxidizer.

1.2.1 Combustion is much data in the book on flame lengths, flame shapes,
Many good textbooks are available on the fundamentals of velocity profiles, species concentrations, and liquid and solid
combustion, but have little if anything on the hydrocarbon fuel combustion, with a limited amount of information on
process and petrochemical industries.3–8 A recent book by heat transfer. Perthuis’ (1983) book has significant discus-
Turns (1996), designed for undergraduate and graduate com- sions of flame chemistry, and some discussion of heat trans-
bustion courses, contains more discussions of practical com- fer from flames. 13 Keating’s (1993) book on applied
bustion equipment than most similar books.9 Khavkin (1996) combustion is aimed at engines and has no treatment of
has written a book that combines theory and practice on gas industrial combustion processes.14 A recent book by Borman
turbines and industrial combustion chambers.10 Of relevance and Ragland (1998) attempts to bridge the gap between the
here, the Khavkin book has a discussion of tube furnaces theoretical and practical books on combustion.15 However,
used in hydrogen production.
the book has little discussion regarding the types of industrial
There have also been many books written on the more
applications considered here. Even handbooks on combus-
practical aspects of combustion. Griswold’s (1946) book has
tion applications have little if anything on industrial com-
a substantial treatment of the theory of combustion, but is
also very practically oriented and includes chapters on gas bustion systems.16–20 The Furnace Operations book by
burners, oil burners, stokers and pulverized-coal burners, Robert Reed is the only one that has any significant coverage
heat transfer (although brief), furnace refractories, tube heat- of combustion in the hydrocarbon and petrochemical indus-
ers, process furnaces, and kilns.11 Stambuleanu’s (1976) book tries. However, this book was last updated in 1981 and is more
on industrial combustion has information on actual furnaces of an introductory book with few equations, graphs, figures,
and on aerospace applications, particularly rockets.12 There
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
pictures, charts, and references.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 9 Monday, March 14, 2005 10:52 AM

Introduction 9

1.2.2 The Process Industries array of tubes located inside a furnace or heater. The tubes
Anderson (1984) has written a general introductory book on are heated by direct-fired burners that often use fuels that are
the petroleum industry, tracing its development from the by-products from processes in the plant and that vary widely
beginning up to some projections for the future of oil.21 There in composition.
is no specific discussion of combustion in petroleum refining. Using tubes to contain the load is somewhat unique com-
Leffler (1985) has written an introductory book on the major pared to the other types of industrial combustion applications.
processes in petroleum refining, including catalytic cracking, It was found that heating the fluids in tubes has many advan-
hydrocracking, and ethylene production, among many tages over heating them in the shell of a furnace.25 Advantages
others.22 The book is written from an overall process perspec- include better suitability for continuous operation, better
tive and has no discussion of the heaters in a plant. Gary and controllability, higher heating rates, more flexibility, less
Handwerk (1994) have written a good overview of petroleum chance of fire, and more compact equipment.
refining.23 The book discusses many of the processes One of the problems encountered in refinery-fired heaters
involved in petroleum refining operations, including coking, is an imbalance in the heat flux in the individual heater
catalytic cracking, and catalytic reforming, among others. passes.26 This imbalance can cause high coke formation rates
However, it does not specifically discuss the combustion pro- and high tube metal temperatures, which reduce a unit’s
cesses involved in heating the refinery fluids. capacity and can cause premature failures (see Chapter 10).
Meyers (1997) has edited a recently updated handbook on Coke formation on the inside of the heater tubes reduces the
petroleum refining processes.24 The book is divided into 14 heat transfer through the tubes, which leads to the reduced
parts, each on a different type of overall process, including capacity. One cause of coking is flame impingement directly
catalytic cracking and reforming, gasification and hydrogen on a tube, which causes localized heating and increases coke
production, hydrocracking, and visbreaking and coking, formation there (see Chapter 17). This flame impingement can
among others. Each part is further divided into the individual be caused by operating without all of the burners in service,
subtypes and variations of the given overall process. Compa- insufficient primary or secondary air to the burner, operating
nies such as Exxon, Dow-Kellogg, UOP, Stone and Webster, the heater at excessive firing rates, fouled burner tips, eroded
and Foster-Wheeler have written about the processes they burner tip orifices, or insufficient draft. The problem of flame
developed, which they license to other companies. Many impingement shows the importance of proper design27 to
aspects of the processes are discussed, including flow dia- ensure even heat flux distribution inside the fired heater.
grams, chemistry, thermodynamics, economics, and environ- Recently, the major emphasis has been on increasing the
mental considerations, but there is very little discussion of capacity of existing heaters rather than installing new heaters.
the combustion systems. The limitations of overfiring a heater include:

• high tube metal temperatures

1.2.3 Combustion in the Process Industries • flame impingement causing high coke formation rates
The standard book on the subject of combustion in the hydro- • positive pressure at the arch of the heater
carbon and petrochemical industries that has been used for • exceeding the capacity of induced-draft and forced-draft
decades is Furnace Operations by Robert Reed, formerly the fans
chief technical officer of John Zink.17 This book has been • exceeding the capacity of the process fluid feed pump
used in the John Zink Burner School for generations and
gives a good introduction to many of the subjects important Garg (1988) noted the importance of good heater specifi-
in burner and heater operation. However, it is somewhat out- cations to ensure suitable performance for a given process.28
dated, especially with regard to pollution regulations and new Some of the basic process conditions needed for the specifi-
trends in burner designs. The present book is designed to be a cations include heater type (cabin, vertical cylindrical, etc.),
greatly expanded version of that book, with many more equa- number of fluid passes, the tube coil size and material, fluid
tions, figures, tables, references, and much wider coverage. data (types, compositions, properties, and flow rates), heat
duty required, fuel data (composition, pressure, and temper-
ature), heat flux loading (heat flux split between the radiant
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

1.3 FIRED HEATERS and convection sections), burner data (number, type, arrange-
Fired or tubestill heaters are used in the petrochemical and ment, etc.), draft requirement, required instrumentation, as
hydrocarbon industries to heat fluids in tubes for further well as a number of other details such as the number of
processing. In this type of process, fluids flow through an peepholes, access doors, and platforms.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 10 Monday, March 14, 2005 10:52 AM

10 The John Zink Combustion Handbook

FIGURE 1.9 Down-fired burner commonly used in top-

fired reformers.

The reformer is a direct-fired combustor containing numer-

ous tubes, filled with catalyst, inside the combustor.30 The
reformer is heated with burners, firing either vertically
downward or upward, with the exhaust on the opposite end,
depending on the specific design of the unit. The raw feed
material flows through the catalyst in the tubes which, under
the proper conditions, converts that material to the desired
end-product. The burners provide the heat needed for the
highly endothermic chemical reactions. The fluid being
reformulated typically flows through a reformer combustor
containing many tubes (see Figure 1.8). The side-fired
reformer has multiple burners on the side of the furnace with
a single row of tubes centrally located. The heat is transferred
primarily by radiation from the hot refractory walls to the
tubes. Top-fired reformers have multiple rows of tubes in the
FIGURE 1.8 Side- (a) and top-fired (b) reformers firebox. In that design, the heat is transferred primarily from
(elevation view).30
radiation from the flame to the tubes. Figure 1.9 shows a

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
down-fired burner commonly used in top-fired reformers. In
a design sometimes referred to as terrace firing, burners may
1.3.1 Reformers be located in the side wall but be firing up the wall at a slight
As the name indicates, reformers are used to reformulate a angle (see Figure 1.10). Foster Wheeler uses terrace wall
material into another product. For example, a hydrogen reformers in the production of hydrogen by steam reformation
reformer takes natural gas and reformulates it into hydrogen of natural gas or light refinery gas.31
in a catalytic chemical process that involves a significant
amount of heat. A sample set of reactions is given below for The reformer tubes are a critical element in the overall
converting propane to hydrogen29: design of the reformer. Because they operate at pressures up
to 350 psig (24 barg), they are typically made from a high-
C3H8 → C2H4 + CH4 temperature and -pressure nickel alloy like inconel to ensure
C2H4 + 2H2O → 2CO + 4H2 that they can withstand the operating conditions inside the
CH4 + H2O → CO + 3H2 reformer. Failure of the tubes can be very expensive because
CO + H2O → CO2 + H2 of the downtime of the unit, lost product, damaged catalyst,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 11 Monday, March 14, 2005 10:52 AM

Introduction 11

and, possibly, damaged reformer. New reactor technologies

are being developed to improve the process for converting
natural gas to precursor synthesis gas (syngas).32

1.3.2 Process Heaters

Process heaters are sometimes referred to as process furnaces
or direct-fired heaters. They are heat transfer units designed
to heat petroleum products, chemicals, and other liquids and
gases flowing through tubes. Typical petroleum fluids include
gasoline, naphtha, kerosene, distillate oil, lube oil, gas oil,
and light ends.33 The heating is done to raise the temperature
of the fluid for further processing downstream or to promote
chemical reactions in the tubes, often in the presence of a
catalyst. Kern noted that refinery heaters can carry liquids at
temperatures as high as 1500°F (810°C) and pressures up to
1600 psig (110 barg). The primary modes of heat transfer in

Bur
process heaters are radiation and convection. The initial part
of the fluid heating is done in the convection section of the

ner
furnace, while the latter heating is done in the radiant section
(see Figure 1.11). Each section has a bank of tubes in it
where the fluids flow through, as shown in Figure 1.12.34 FIGURE 1.10 Elevation view of a terrace firing furnace.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Early heater designs had only a single bank of tubes that
failed prematurely because designers did not understand the
importance of radiation on the process.25 The tubes closest to
the burners would overheat. Overheating caused the hydro-
carbons to form coke inside the tube. The coke further aggra- Stack
vated the problem by reducing the thermal conductivity
Stack
through the coke layer inside the tube. The reduced thermal
Damper
conductivity prevented the process fluids from absorbing
Convection
adequate heat to cool the tubes, resulting in overheating and
failure of the tubes. One of the key challenges for the heater
Section
designer is to get even heat distribution inside the combustor
to prevent coking inside the tubes. Bell and Lowy (1967)
estimated that approximately 70% of the energy is trans-
Radiant
ferred to the fluids in the radiant section of a typical heater
and the balance to the convection section.35 The tubes in the Section
convection section often have fins to improve convective heat
transfer efficiency. These fins are designed to withstand tem-
peratures up to about 1200°F (650°C). If delayed combustion
occurs in the convection section, the fins can be exposed to
temperatures up to 2000°F (1100°C), which can damage the FIGURE 1.11 Schematic of a process heater.
fins.34
Kern noted that process heaters are typically designed
around the burners.33 There may be anywhere from 1 to over forced-draft systems, burners with air preheat typically have
100 burners in a typical process heater, depending on the higher heat releases than burners without air preheat. Accord-
design and process requirements. In the refinery industry, the ing to one survey, 89.6% of the burners in oil refineries are
average number of burners in a heater varies by the heater natural draft, 8.0% are forced draft with no air preheat, and
type, as shown in Table 1.2.36 On average, mechanical draft 2.4% are forced draft with air preheat.37 The mean size of all
burners have higher firing rates than natural draft burners. For process heaters is 72 × 106 Btu/hr (21 MW), which are mostly
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 12 Monday, March 14, 2005 10:52 AM

12 The John Zink Combustion Handbook

FIGURE 1.13 Fixed heater size distribution.36

reformers. The low and medium firebox temperature heaters

represent about 20% of the chemical industry requirements
and are similar to those in the petroleum refining industry.38
The high firebox temperature heaters represent the remaining
80% of the chemical industry heater requirements and are
unique to the chemical industry.
Berman (1979) discussed the different burner designs used
FIGURE 1.12 Typical process heater.34 in fired heaters.39 Burners can be located in the floor, firing
vertically upward. In vertical cylindrical (VC) furnaces, these
TABLE 1.2 Average Burner Configuration by Heater Type burners are located in a circle in the floor of the furnace. The
Ave. Design VC furnace itself serves as part of the exhaust stack to help
Total Heat Ave. Firing Rate create draft to increase the chimney effect.40 In cabin heaters,
Ave. No of Release per Burner which are rectangular, there are one or more rows of burners
Heater Type Burners (106 Btu hr–1) (106 Btu hr–1)
located in the floor. Burners can be at a low level, firing parallel
Natural draft 24 69.4 2.89 to the floor. In that configuration, they may be firing from two
Mechanical draft, 20 103.6 5.18
opposite sides toward a partial wall in the middle of the furnace
no air preheat
Mechanical draft, 14 135.4 9.67 that acts as a radiator to distribute the heat (see Figures 1.14
with air preheat and 1.15). Burners can be located on the wall, firing radially
From U.S. EPA.36 along the wall (see Figure 1.16), and are referred to as radiant
wall burners. There are also combinations of the above in
certain heater designs. For example, in ethylene production
heaters, both floor-mounted vertically fired burners (see Figure
natural draft. The mean size of forced-draft heaters is
1.17) and radiant wall burners are used in the same heater.
110 × 106 Btu/hr (32 MW). Figure 1.13 shows the distribution
for the overall firing rate for fired heaters. Table 1.3 shows Typical examples of process heaters are shown in Figures
the variety of processes in a refinery that use fired heaters. 1.18 and 1.19. A cabin heater is shown in Figure 1.20, burners
firing in a crude unit are shown in Figure 1.21, and typical
Table 1.4 shows the major applications for fired heaters in burner arrangements are shown in Figure 1.22. Berman (1979)
the chemical industry. These can be broadly classified into noted the following categories of process heaters: column
two categories: (1) low- and medium-firebox temperature reboilers; fractionating-column feed preheaters, reactor-feed
applications such as feed preheaters, reboilers, and steam preheaters; including reformers, and heat supplied to heat trans-
superheaters, and (2) high firebox temperature applications fer media (e.g., a circulating fluid or molten salt), heat supplied
such as olefins, pyrolysis furnaces, and steam-hydrocarbon to viscous fluids; and fired reactors, including steam reformers
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Introduction 13

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
TABLE 1.3 Major Refinery Processes Requiring a Fired Heater
Feedstock
Process heat requirements temperature
outlet of
Process Process Description Heaters used KJ/liter 103 Btu/bbl heater, °F

Distillation
Atmospheric Separates light hydrocarbons from crude in a Preheater, reboiler 590 89 700
distillation column under atmospheric conditions
Vacuum Separates heavy gas oils from atmospheric Preheater, reboiler 418 63 750–830
distillation bottoms under vacuum

Thermal Processes
Thermal cracking Thermal decomposition of large molecules into lighter, Fired reactor 4650 700 850–1000
more valuable products
Coking Cracking reactions allowed to go to completion; Preheater 1520 230 900–975
lighter products and coke produced.
Visbreaking Mild cracking of residuals to improve their viscosity Fired reactor 961 145 850–950
and produce lighter gas oils

Catalytic Cracking
Fluidized catalytic cracking Cracking of heavy petroleum products; a catalyst is used Preheater 663 100 600–885
to aid the reaction
Catalytic hydrocracking Cracking heavy feedstocks to produce lighter products Preheater 1290 195 400–850
in the presence of hydrogen and a catalyst

Hydroprocessing
Hydrodesulfurization Remove contaminating metals, sulfur, and nitrogen Preheater 431 65 a 390–850
from the feedstock; hydrogen is added and reacted
over a catalyst
Hydrotreating Less severe than hydrodesulfurization; removes metals, Preheater 497 75b 600–800
nitrogen, and sulfur from lighter feedstocks;
hydrogen is added and reacted over a catalyst

Hydroconversion
Alkylation Combination of two hydrocarbons to produce a higher Reboiler 2500 377 c 400
molecular weight hydrocarbon; heater used on the fractionator
Catalytic reforming Low-octane napthas are converted to high-octane, Preheater 1790 270 850–1000
aromatic napthas; feedstock is contacted with hydrogen
over a catalyst
a Heavy gas oils and middle distillates.
b Light distillate.
c Btu bbl–1 of total alylate.

Source: From the U.S. EPA.36

and pyrolysis heaters.41 Six types of vertical-cylindrical-fired tubes and are heated to the desired temperature for further
heaters were given: all radiant, helical coil, crossflow with processing. The fluids are preheated in the convection section
convection section, integral convection section, arbor or wicket and heated to the desired process temperature in the radiant
type, and single-row/double-fired. Six basic designs were also section. Radiant heat transfer from the flames to the tubes is
given for horizontal tube-fired heaters: cabin, two-cell box, the most critical aspect of this heater because overheating of
cabin with dividing bridgewall, end-fired box, end-fired box the tubes leads to tube failure and shutdown of the heater.42
with side-mounted convection section, and horizontal-tube/ The tubes can be horizontally or vertically oriented, depend-
single-row/double-fired. ing on the particular heater design.
Many commonly used process heaters typically have a A unique aspect of process heaters is that they are often
radiant section and a convection section. Burners are fired in natural-draft. This means that no combustion air blower is used.
the radiant section to heat the tubes. Fluids flow through the The air is inspirated into the furnace by the suction created by
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 14 Monday, March 14, 2005 10:52 AM

14 The John Zink Combustion Handbook

TABLE 1.4 Major Fired Heater Applications in the Chemical Industry

Firebox 1985 Fired Heater % of Known Chemical
Temperature Energy Requirement Industry Heater
Chemical Process Heater Type (°F) (1012 Btu yr–1) Requirements

Low- and Medium-Temperature Applications

Benzene Reformate extraction Reboiler 700 64.8 9.9
Styrene Ethylbenzine Steam superheater 1500–1600 32.1 4.9
dehydrogenation
Vinyl chloride monomer Ethylene dichloride Cracking furnace N/A 12.6 1.9
cracking
p-Xylene Xylene isomerization Reactor-fired preheater N/A 13.0 2.0
Dimethylterephthalate Reaction of p-xylene and Preheater, hot oil furnace 480–540 11.1 1.7
methanol
Butadiene Butylene dehydrogenation Preheater, reboiler 1100 2.6 0.4
Ethanol (synthetic) Ethylene hydration Preheater 750 1.3 0.2

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Acetone Various Hot oil furnace N/A 0.8 0.1

High-Temperature Applications
Ethylene/propylene Thermal cracking Pyrolysis furnace 1900–2300 337.9 51.8
Ammonia Natural gas reforming Steam hydrocarbon reformer 1500–1600 150.5 23.1
Methanol Hydrocarbon reforming Steam hydrocarbon 1000–2000 25.7 4.0

Total Known Fired Heater Energy Requirement 652.4 100.0

Source: From U.S. EPA. 36

FIGURE 1.14 Sketch (elevation view) of center or target

wall firing configuration. FIGURE 1.15 Horizontal floor-fired burners.

the hot gases rising through the combustion chamber and ating conditions of the plant at any given time. In addition to
exhausting to the atmosphere. Another unique aspect of these hydrocarbons ranging up to C5, the gaseous fuels can also
heaters is the wide range of fuels used, which are often by- contain hydrogen and inerts (like CO2 or N2). The composi-
products of the petroleum refining process. These fuels can tions can range from gases containing high levels of inerts to
contain significant amounts of hydrogen, which has a large fuels containing high levels of H2. The flame characteristics
impact on the burner design. It is also fairly common for for fuels with high levels of inerts are very different than for
multiple fuel compositions to be used, depending on the oper- fuels with high levels of H2 (see Chapters 2 and 5). Add to
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 15 Monday, March 14, 2005 10:52 AM

Introduction 15

that the requirement for turndown conditions, and it becomes

very challenging to design burners that will maintain stability,
low emissions, and the desired heat flux distribution over the
range of possible conditions. Some plants use liquid fuels,
like No. 2 to No. 6 fuel oil, sometimes by themselves and
sometimes in combination with gaseous fuels. So-called com-
bination burners use both a liquid and a gaseous fuel, which
are normally injected separately through each burner.
Shires gave a general heat balance for a process heater:43

Q˙ f = Q˙ g + Q˙1 + Q˙ p (1.1)

where Q̇ f is the heat generated by combusting the fuel, Q̇ g is

the heat going to the load, Q̇ 1 is the heat lost through the
walls, and Q̇ p is the heat carried out by the exhaust products.
This is shown schematically in Figure 1.23.
Talmor (1982) has written a book dealing with the predic-
tion, control, and troubleshooting of hot spots in process
heaters.44 The book gives a method for estimating the mag-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

nitude and location of the maximum heat flux in the combus-

tion zone. It takes into account the firing rate of each burner,
FIGURE 1.16 Wall-fired burner (side view). (Courtesy
of John Zink Company LLC.)
the number of burners, the flame length, the flame emissivity,
the spacing between the burner and the tubes, the spacing
between the burners, and the geometry of the firebox. The
book includes much empirical data specific to a variety of
different process heaters and also gives many detailed exam-
ples that have been worked out.

1.4 BURNERS
The burner is the device that is used to combust the fuel with
an oxidizer (usually air) to convert the chemical energy in the
fuel into thermal energy. A given combustion system may
have a single burner or many burners, depending on the size
and type of the application. For example, in a vertical
cylindrical furnace, one or more burners are located in the
floor of a cylindrically shaped furnace (see Figure 1.24). The
heat from the burner radiates in all directions and is efficiently
absorbed by the tubes. Another type of heater geometry is FIGURE 1.17 Sketch (elevation view) of a horizontally
rectangular (see Figure 1.25). This type of system is generally mounted, vertically fired burner configuration.
more difficult to analyze because of the multiplicity of heat
sources and because of the interactions between the flames
and their associated products of combustion. 1.4.1 Competing Priorities
There are many factors that go into the design of a burner. There have been many changes in the traditional designs that
This section briefly considers some of the important factors have been used in burners, primarily because of the recent
to be taken into account for a particular type of burner, as interest in reducing pollutant emissions. In the past, the
well as how these factors impact things like heat transfer and burner designer was primarily concerned with efficiently
pollutant emissions. combusting the fuel and transferring the energy to a heat
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 16 Monday, March 14, 2005 10:52 AM

16 The John Zink Combustion Handbook

FIGURE 1.18 Examples of process heaters.43

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

load. New and increasingly more stringent environmental ble that staged combustion may produce soot in the flame,
regulations have added the need to consider the pollutant which can increase flame radiation. The actual impact of
emissions produced by the burner. In many cases, reducing staging on the heat transfer from the flame is highly depen-
pollutant emissions and maximizing combustion efficiency dent on the actual burner design.
are at odds with each other. For example, a well-accepted In the past, the challenge for the burner designer was to
technique for reducing NOx emissions is known as staging, maximize the mixing between the fuel and the oxidizer to
where the primary flame zone is deficient in either fuel or ensure complete combustion. If the fuel was difficult to burn,
oxidizer.45 The balance of the fuel or oxidizer can be injected as in the case of low heating value fuels such as waste liquid
into the burner in a secondary flame zone or, in a more fuels or process gases from chemical production, the task
extreme case, can be injected somewhere else in the combus- could be very challenging. Now, the burner designer must
tion chamber. Staging reduces the peak temperatures in the balance the mixing of the fuel and the oxidizer to maximize
primary flame zone and also alters the chemistry in a way combustion efficiency while simultaneously minimizing all
that reduces NOx emissions because fuel-rich or fuel-lean types of pollutant emissions. This is no easy task as, for exam-
zones are less conducive to NOx formation than near-stoichi- ple, NOx and CO emissions often go in opposite directions
ometric zones (see Chapter 6). NOx emissions increase rap- (see Figure 1.26). When CO is low, NOx may be high, and
idly with the exhaust product temperature (see Figure 6.5). vice versa. Modern burners must be environmentally friendly,
Because thermal NOx is exponentially dependent on the gas while simultaneously efficiently transferring heat to the load.
temperature even small reductions in the peak flame tempera-
ture can dramatically reduce NOx emissions. However, lower 1.4.2 Design Factors
flame temperatures often reduce the radiant heat transfer There are many types of burners designs that exist due to the
from the flame because radiation is dependent on the fourth wide variety of fuels, oxidizers, combustion chamber geo-
power of the absolute temperature of the gases. Another metries, environmental regulations, thermal input sizes, and
potential problem with staging is that it may increase CO heat transfer requirements. Additionally, heat transfer
emissions, which is an indication of incomplete combustion requirements include things like flame temperature, flame
and reduced combustion efficiency. However, it is also possi- momentum, and heat distribution. Garg (1989) lists the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 17 Monday, March 14, 2005 10:52 AM

Introduction 17

FIGURE 1.19 Typical heater types.

following burner specifications that are needed to properly position, flame dimensions, ignition type, atomization media
choose a burner for a given application: burner type, heat for liquid fuel firing, noise, NOx emission rate, and whether
release and turndown, air supply (natural draft, forced draft, waste gas firing will be used.46 Some of these design factors
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

or balanced draft), excess air level, fuel composition(s), firing are briefly considered next.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 18 Monday, March 14, 2005 10:52 AM

18 The John Zink Combustion Handbook

FIGURE 1.20 Cabin heater.

1.4.2.1 Fuel The fuel choice has an important influence on the heat
transfer from a flame. In general, solid fuels like coal and
Depending on many factors, certain types of fuels are pre-
liquid fuels like oil produce very luminous flames which
ferred for certain geographic locations due to cost and avail-
contain soot particles that radiate like blackbodies to the heat
ability considerations. Gaseous fuels, particularly natural
load. Gaseous fuels like natural gas often produce nonlumi-
gas, are commonly used in most industrial heating applica-
nous flames because they burn so cleanly and completely
tions in the United States. In Europe, natural gas is also com-
without producing soot particles. A fuel like hydrogen is
monly used along with light fuel oil. In Asia and South
completely nonluminous as there is no carbon available to
America, heavy fuel oils are generally preferred, although the
produce soot.
use of gaseous fuels is on the rise.
In cases where highly radiant flames are required, a lumi-
Fuels also vary depending on the application. For example, nous flame is preferred. In cases where convection heat trans-
in incineration processes, waste fuels are commonly used fer is preferred, a nonluminous flame may be preferred in
either by themselves or with other fuels like natural gas. In order to minimize the possibility of contaminating the heat
the petrochemical industry, fuel gases often consist of a blend load with soot particles from a luminous flame. Where nat-
of several fuels, including gases like hydrogen, methane, pro- ural gas is the preferred fuel and highly radiant flames are
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
pane, butane, propylene, nitrogen, and carbon dioxide. desired, new technologies are being developed to produce
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 19 Monday, March 14, 2005 10:52 AM

Introduction 19

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 1.21 Crude unit burners.

more luminous flames. These include things like pyrolyzing fuels simultaneously. Another application where multiple fuels
the fuel in a partial oxidation process,47 using a plasma to may be used is in waste incineration. One method of disposing
produce soot in the fuel,48 and generally controlling the mix- of waste liquids contaminated with hydrocarbons is to combust
ing of the fuel and oxidizer to produce fuel-rich flame zones them by direct injection through a burner. The waste liquids
that generate soot particles.49 are fed through the burner, which is powered by a traditional
Therefore, the fuel itself has a significant impact on the fuel such as natural gas or oil. The waste liquids often have
heat transfer mechanisms between the flame and the load. In very low heating values and are difficult to combust without
most cases, the fuel choice is dictated by the customer as part auxiliary fuel. This further complicates the burner design,
of the specifications for the system and is not chosen by the wherein the waste liquid must be vaporized and combusted
burner designer. The designer must make the best of whatever concurrently with the normal fuel used in the burner.
fuel has been selected. In most cases, the burner design is
optimized based on the choice of fuel. 1.4.2.2 Oxidizer
In some cases, the burner may have more than one type of The predominant oxidizer used in most industrial heating
fuel. An example is shown in Figure 1.27.50 Dual-fuel burners processes is atmospheric air. This can present challenges in
are designed to operate typically on either gaseous or liquid some applications where highly accurate control is required
fuels. These burners are used, usually for economic reasons, due to the daily variations in the barometric pressure and
where the customer may need to switch between a gaseous fuel humidity of ambient air. The combustion air is sometimes
like natural gas and a liquid fuel like oil. These burners nor- preheated to increase the overall thermal efficiency of a pro-
mally operate on one fuel or the other, and sometimes on both cess. Combustion air is also sometimes blended with some of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 20 Monday, March 14, 2005 10:52 AM

20 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 1.22 Typical burner arrangements (elevation view).

the products of combustion, a process usually referred to as flame temperatures resulting in reduced NOx emissions,
flue gas recirculation (FGR). because NOx emissions are highly temperature dependent.
FGR is used to both increase thermal efficiency and reduce
NOx emissions. Capturing some of the energy in the exhaust 1.4.2.3 Gas Recirculation
gases and using it to preheat the incoming combustion oxi- A common technique used in combustion systems is to
dizer increases thermal efficiency. FGR also reduces peak design the burner to induce furnace gases to be drawn into
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 21 Monday, March 14, 2005 10:52 AM

Introduction 21

FIGURE 1.24 Burner (B) arrangement (plan view) in

the floor of vertical cylindrical furnaces: (a) small-diameter
furnace with a single centered burner and (b) larger diameter
furnace with four burners symmetrically arranged at a radius
from the center.

FIGURE 1.23 Process heater heat balance.43

the burner to dilute the flame, usually referred to as flue or

furnace gas recirculation. Although the furnace gases are hot,
they are still much cooler than the flame itself. This dilution
can accomplish several purposes. One is to minimize NOx
emissions by reducing the peak temperatures in the flame, as
in flue gas recirculation. However, furnace gas recirculation
may be preferred to flue gas recirculation (see Figure 1.28)
because no external high-temperature ductwork or fans are
needed to bring the product gases back into the flame zone.
FIGURE 1.25 Burner (B) arrangement (plan view) in
the floor of rectangular cabin heaters: (a) single row of burners
Another reason to use furnace gas recirculation may be to

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
in a narrower heater, (b) two rows of staggered burners in a
increase the convective heating from the flame because of the slightly wider heater, and (c) two rows of aligned burners in
added gas volume and momentum. An example of furnace an even wider heater.
gas recirculation into the burner is shown in Figure 6.12.

1.4.3 General Burner Types of the premixed type. Premixed burners often produce shorter
There are numerous ways that burners can be classified. Some and more intense flames, compared to diffusion flames. This
of the common ones are discussed in this section, along with a can produce high-temperature regions in the flame, leading to
brief description of implications for heat transfer. nonuniform heating of the load and higher NOx emissions.
In diffusion-mixed flames, the fuel and the oxidizer are sep-
1.4.3.1 Mixing Type arated and unmixed prior to combustion, which begins where
One common method for classifying burners is according to the oxidizer/fuel mixture is within the flammability range. An
how the fuel and the oxidizer are mixed. In premixed burners, example of a diffusion flame is a candle (see Figure 1.31). A
shown in a cartoon in Figure 1.29 and schematically in diffusion-mixed gas burner, shown schematically in Figure
Figure 1.30, the fuel and the oxidizer are completely mixed 1.32, is sometimes referred to as a “raw gas” burner because
before combustion begins. Radiant wall burners are usually the fuel gas exits the burner essentially as raw gas, having no
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 22 Monday, March 14, 2005 10:52 AM

22 The John Zink Combustion Handbook

FIGURE 1.26 Adiabatic equilibrium NO and CO as a function of the equivalence ratio for an air/CH4 flame.

FIGURE 1.27 Typical combination oil and gas burner.

FIGURE 1.28 Schematic of flue gas recirculation.
air mixed with it. Diffusion burners typically have longer
flames than premixed burners, a lower temperature hot spot,
and a more uniform temperature and heat flux distribution.
It is also possible to have partially premixed burners, shown
schematically in Figure 1.33, where some fraction of the fuel
is mixed with the oxidizer. Partial premixing is often done FIGURE 1.29 Cartoon of a premixed burner.
for stability and safety reasons because it helps anchor the
flame, but also reduces the chance for flashback, which is
sometimes a problem in fully premixed burners. This type of Another burner classification based on mixing is known as
burner often has a flame length, temperature, and heat flux staging — staged air and staged fuel. A staged air burner is
distribution that is somewhere between the fully premixed shown in a cartoon in Figure 1.34 and schematically in
and diffusion flames. Figure 1.35. A staged fuel burner is shown in a cartoon in
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Introduction 23

FIGURE 1.30 Typical premixed gas burner. (Courtesy of API.50)

Figure 1.36 and schematically in Figure 1.37. Secondary and plant. Burners often need to be able to fire multiple fuels that
sometimes tertiary injectors in the burner are used to inject a may be produced by the plant, depending on the process con-
portion of the fuel and/or the air into the flame, downstream ditions and on start-up vs. normal operation. These gaseous
of the root of the flame. Staging is often done to reduce NOx fuels often have significant amounts of methane, hydrogen,
emissions and to produce longer flames. These longer flames and higher hydrocarbons (e.g., propane and propylene). They
typically have a lower peak flame temperature and more uni- may also contain inerts such as CO2 and N2. The heating
form heat flux distribution than non-staged flames.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

value can range from 500 to 1500 Btu ft –3 (19 to 56 MJ m–3).

Burners firing oil require some type of liquid atomization,
1.4.3.2 Fuel Type
There are three common fuel classifications for burners used commonly mechanical (pressurizing the liquid high enough
in the process industries and generally listed in order of to force it through an atomizer), air, or steam. Steam is most
increasing complexity as follows: gas, oil, or a combination commonly used because it is economical, readily available,
of gas + oil. Gas burners are either diffusion (raw gas or no and gives a wide turndown ratio and good flame control.
premixing), premixed, or partially premixed. The gas compo- Combination burners can usually fire 100% oil, 100% gas, or
sition can vary widely, as it is often a by-product from the any combination in between.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 24 Monday, March 14, 2005 10:52 AM

24 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 1.31 Painting of a diffusion flame by Georges de LaTour. (Courtesy of the Los Angeles County Museum of Art.)

Copyright CRC Press

Introduction 25

FIGURE 1.32 Cartoon of a diffusion burner.

FIGURE 1.33 Cartoon of a partially premixed burner.

FIGURE 1.34 Cartoon of a staged air burner.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 1.35 Schematic of a typical staged air combination oil and gas buner.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 26 Monday, March 14, 2005 10:52 AM

26 The John Zink Combustion Handbook

FIGURE 1.36 Cartoon of a staged fuel burner.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 1.37 Schematic of a typical staged fuel gas burner.

1.4.3.3 Combustion Air Temperature in the combustor. A schematic is shown in Figure 1.38, and
One common way of classifying the oxidizer is by its temp- an example is shown in Figure 1.39. In this type of burner,
erature. It is common in many industrial applications to the pressure drop and combustor stack height are critical in
recover heat from the exhaust gases by preheating the producing enough suction to induce enough combustion air
incoming combustion air, either with a recuperator or a into the burners. This type of burner is commonly used in the
regenerator. Such a burner is often referred to as a preheated chemical and petrochemical industries in fluid heaters. The
air burner. main consequence of the draft type on heat transfer is that the
natural-draft flames are usually longer than the forced-draft
1.4.3.4 Draft Type flames so that the heat flux from the flame is distributed over
Most industrial burners are known as forced-draft or mechan- a longer distance and the peak temperature in the flame is
ical-draft burners. This means that the oxidizer is supplied to often lower.
the burner under pressure. For example, in a forced-draft air
burner, the air used for combustion is supplied to the burner 1.4.3.5 Location
by a blower. In natural-draft burners, the air used for combus- Process burners are often classified by their location in the
tion is induced into the burner by the negative draft produced furnace or heater. Floor or hearth burners are located in the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 27 Monday, March 14, 2005 10:52 AM

Introduction 27

SECONDARY AIR

PRIMARY AIR

GAS PILOT GAS

FIGURE 1.38 Typical natural-draft gas burner.

bottom of the combustor and fire vertically upward. Roof

burners are located in the ceiling and fire vertically down-
ward. Wall burners can be located in the wall or in the floor,
firing along the wall. Their function is to heat a refractory
wall to radiate heat to process tubes.

1.4.4 Potential Problems

There are many potential problems that could affect the per-
formance of burners and therefore the performance of the
heaters, boilers, and furnaces used to process the materials of
interest to the end user. A few examples will illustrate some of
the potential problems that may be encountered. Figure 1.40
shows flames impinging on the process tubes in a cabin
heater. Flame impingement on the tubes can cause premature
coking and significantly reduce the operational run time.
Figure 1.41 shows flames pulled toward the wall of a heater.
This may be caused by burner design problems or by gas flow
currents in the furnace. While flames leaning away from the
tubes may reduce coking, they also reduce performance
because the heater is designed for vertical flames. Less heat is FIGURE 1.39 Natural-draft burner.
transferred to the tubes when the flames lean away from the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

28 The John Zink Combustion Handbook

tubes. This reduces the throughput of the entire process.

Figure 1.42 shows an oil burner that needs some type of ser-
vice or adjustment. In all of the above examples, the perfor-
mance of the combustion system is reduced. Chapter 17
discusses some of the common problems encountered and
how to fix them.

1.2 CONCLUSIONS
This book considers all aspects of combustion, with particu-
lar emphasis on applications in the process industries includ-
ing the petrochemical, hydrocarbon, power generation, and

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
thermal oxidation industries. The fundamentals of combus-
tion, heat transfer, and fluid flow are discussed from a more
applied approach. Many other aspects of combustion, such as
fuel composition, pollutant emissions, noise, safety, and con-
FIGURE 1.40 Flames impinging on tubes in a cabin
heater. trol, are also discussed. Topics of specific interest to burners
are also treated including design, testing, installation, mainte-
nance, and troubleshooting. There are also very detailed con-
siderations of process burners, flares, boiler burners, duct
burners, and thermal oxidizers. Many of these topics have
never been adequately covered in other combustion books.
The extensive use of color illustrations further enhances the
usefulness of this book as an essential tool for the combus-
tion engineer.

REFERENCES

1. U.S. Dept. of Energy Office of Industrial Technology,

Petroleum — Industry of the Future: Energy and
Environmental Profile of the U.S. Petroleum Refining
Industry, U.S. DOE, Washington, D.C., December
FIGURE 1.41 Flames pulled toward the wall. 1998.
2. U.S. Dept. of Energy Office of Industrial Technology,
Industrial Combustion Technology Roadmap, U.S.
DOE, Washington, D.C., April 1999.
3. R.A. Strehlow, Fundamentals of Combustion,
International Textbook Co., Scranton, PA, 1968.
4. F.A. Williams, Combustion Theory, Benjamin/ Cummings,
Menlo Park, CA, 1985.
5. J.A. Barnard and J.N. Bradley, Flame and Combustion,
2nd edition, Chapman and Hall, London, 1985.
6. B. Lewis and G. von Elbe, Combustion, Flames and
Explosions of Gases, 3rd edition, Academic Press,
New York, 1987.
7. W. Bartok and A.F. Sarofim, Eds., Fossil Fuel
Combustion, John Wiley & Sons, New York, 1991.
8. I. Glassman, Combustion, 3rd edition, Academic Press,
FIGURE 1.42 Gas flames needing service. New York, 1996.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 29 Monday, March 14, 2005 10:52 AM

Introduction 29

9. S.R. Turns, An Introduction to Combustion, McGraw-Hill, 29. H. Futami, R. Hashimoto, and H. Uchida,
New York, 1996. Development of new catalyst and heat-transfer design
method for steam reformer, J. Fuel Soc. Japan,
10. Y.I. Khavkin, Combustion System Design: A New
68(743), 236-243, 1989.
Approach, PennWell Books, Tulsa, OK, 1996.
30. H. Gunardson, Industrial Gases in Petrochemical
11. J. Griswold, Fuels, Combustion and Furnaces,
Processing, Marcel Dekker, New York, 1998.
McGraw-Hill, New York, 1946.
31. J.D. Fleshman, FW hydrogen production, in Handbook of
12. A. Stambuleanu, Flame Combustion Processes in
Petroleum Refining Processes, 2nd edition, R.A. Myers,
Industry, Abacus Press, Tunbridge Wells, U.K., 1976.
Ed., McGraw-Hill, New York, 1996, chap. 6.2.
13. E. Perthuis, La Combustion Industrielle, Éditions
32. J.S. Plotkin and A.B. Swanson, New technologies key
Technip, Paris, 1983.
to revamping petrochemicals, Oil & Gas J., 97(50),
14. E.L. Keating, Applied Combustion, Marcel Dekker, 108-114, 1999.
New York, 1993.
33. D.Q. Kern, Process Heat Transfer, McGraw-Hill, New
15. G. Borman and K. Ragland, Combustion Engineering, York, 1950.
McGraw-Hill, New York, 1998.
34. N.P. Lieberman, Troubleshooting Process Operations,
16. C.G. Segeler, Ed., Gas Engineers Handbook, Industrial PennWell Books, Tulsa, OK, 1991.
Press, New York, 1965.
35. H.S. Bell and L. Lowy, Equipment, in Petroleum
17. R.D. Reed, Furnace Operations, 3rd edition, Gulf Processing Handbook, W.F. Bland and R.L. Davidson,
Publishing, Houston, TX, 1981. Eds., McGraw-Hill, New York, 1967, chap. 4.
18. R. Pritchard, J.J. Guy, and N.E. Connor, Handbook of 36. E.B. Sanderford, Alternative Control Techniques
Industrial Gas Utilization, Van Nostrand Reinhold, Document — NOx Emissions from Process Heaters, U.S.
New York, 1977. Envir. Protection Agency Report EPA-453/R-93-015,
February, 1993.
19. R.J. Reed, North American Combustion Handbook, 3rd
edition, Vol. I, North American Mfg. Co., Cleveland, 37. L.A. Thrash, Annual Refining Survey, Oil & Gas J.,
OH, 1986. 89(11), 86-105, 1991.
20. IHEA, Combustion Technology Manual, 5th edition, 38. S.A. Shareef, C.L. Anderson, and L.E. Keller, Fired
Industrial Heating Equipment Assoc., Arlington, VA, Heaters: Nitrogen Oxides Emissions and Controls,
1994. U.S. Environmental Protection Agency, Research
Triangle Park, NC, EPA Contract No. 68-02-4286,
21. R.O. Anderson, Fundamentals of the Petroleum
June 1988.
Industry, University of Oklahoma Press, Norman, OK,
1984. 39. H.L. Berman, Fired heaters. II: Construction materials,
mechanical features, performance monitoring, in Process
22. W.L. Leffler, Petroleum Refining for the Nontechnical
Heat Exchange, V. Cavaseno, Ed., McGraw-Hill, New
Person, PennWell Books, Tulsa, OK, 1985.
York, 1979, 293-302.
23. J.H. Gary and G.E. Handwerk, Petroleum Refining:
40. A.J. Johnson and G.H. Auth, Fuels and Combustion
Technology and Economics, 3rd edition, Marcel
Handbook, 1st edition, McGraw-Hill, New York, 1951.
Dekker, New York, 1994.
41. H.L. Berman, Fired heaters. I: Finding the basic
24. R.A. Meyers, Handbook of Petroleum Refining
design for your application, in Process Heat
Processes, 2nd edition, McGraw-Hill, New York, 1997.
Exchange, V. Cavaseno, Ed., McGraw-Hill, New York,
25. W.L. Nelson, Petroleum Refinery Engineering, 2nd 1979, 287-292.
edition, McGraw-Hill, New York, 1941.
42. V. Ganapathy, Applied Heat Transfer, PennWell Books,
26. G.R. Martin, Heat-flux imbalances in fired heaters Tulsa, OK, 1982.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

cause operating problems, Hydrocarbon Processing,

43. G.L. Shires, Furnaces, in The International
77(5), 103-109, 1998.
Encyclopedia of Heat & Mass Transfer, G.F. Hewitt,
27. R. Nogay and A. Prasad, Better design method for fired G.L. Shires, and Y.V. Polezhaev, Eds., CRC Press,
heaters, Hydrocarbon Processing, 64(11), 91-95, 1985. Boca Raton, FL, 1997, 493-497.
28. A. Garg and H. Ghosh, Good heater specifications 44. E. Talmor, Combustion Hot Spot Analysis for Fired
pay off, Chem. Eng., 95(10), 77-80, 1988. Process Heaters, Gulf Publishing, Houston, 1982.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch01-Frame Page 30 Monday, March 14, 2005 10:52 AM

30 The John Zink Combustion Handbook

45. J.L. Reese, G.L. Moilanen, R. Borkowicz, C. Baukal, 48. R. Ruiz and J.C. Hilliard, Luminosity enhancement of
D. Czerniak, and R. Batten, State-of-the-art of NOx natural gas flames, Proc. of 1989 Int. Gas Research
emission control technology, ASME Paper 94-JPGC- Conf., T.L. Cramer, Ed., Government Institutes,
EC-15, Proc. of Int. Joint Power Generation Conf., Rockville, MD, 1990, 1345-1353.
Phoenix, October 3-5, 1994. 49. A.G. Slavejkov, T.M. Gosling, and R.E. Knorr, Low-
46. A. Garg, Better burner specifications, Hydrocarbon NOx Staged Combustion Device for Controlled
Processing, 68(8), 71-72, 1989. Radiative Heating in High Temperature Furnaces, U.S.
47. M.L. Joshi, M.E. Tester, G.C. Neff, and S.K. Panahi, Patent 5,611,682, March 18, 1997.
Flame particle seeding with oxygen enrichment for 50. API Publication 535: Burner for Fired Heaters in
NOx reduction and increased efficiency, Glass, 68(6), General Refinery Services, 1st edition, American
212-213, 1990. Petroleum Institute, Washington, D.C., July 1995.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 2
Fundamentals
Joseph Colannino and Charles E. Baukal, Jr.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
TABLE OF CONTENTS

2.1 Introduction............................................................................................................................................... 34
2.2 Uses for Combustion................................................................................................................................. 34
2.3 Brief Overview of Combustion Equipment and Heat Transfer................................................................. 34
2.4 Net Combustion Chemistry of Hydrocarbons........................................................................................... 34
2.5 Conservation of Mass................................................................................................................................ 35
2.6 The Ideal Gas Law .................................................................................................................................... 35
2.7 Stoichiometric Ratio and Excess Air ........................................................................................................ 38
2.7.1 Heat of Combustion .................................................................................................................... 38
2.7.2 Adiabatic Flame Temperature..................................................................................................... 46
2.8 Substoichiometric Combustion ................................................................................................................. 46
2.9 Equilibrium and Thermodynamics............................................................................................................ 47
2.10 Substoichiometric Combustion Revisited ................................................................................................. 47
2.11 General Discussion ................................................................................................................................... 54
2.11.1 Air Preheat Effects ...................................................................................................................... 55
2.11.2 Fuel Blend Effects....................................................................................................................... 57
2.12 Combustion Kinetics................................................................................................................................. 60
2.12.1 Thermal NOx Formation: A Kinetic Example............................................................................ 60
2.12.2 Reaction Rate .............................................................................................................................. 60
2.12.3 Prompt-NOx Formation .............................................................................................................. 61
2.12.4 The Fuel-Bound NOx Mechanism.............................................................................................. 61
2.13 Flame Properties ....................................................................................................................................... 61
2.13.1 Flame Temperature ..................................................................................................................... 61
2.13.2 Available Heat............................................................................................................................. 64
2.13.3 Minimum Ignition Energy .......................................................................................................... 64
2.13.4 Flammability Limits ................................................................................................................... 64
2.13.5 Flammability Limits for Gas Mixtures ....................................................................................... 66
2.13.6 Flame Speeds .............................................................................................................................. 67
References .................................................................................................................................................................. 67
33
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 34 Monday, March 14, 2005 11:04 AM

34 The John Zink Combustion Handbook

2.1 INTRODUCTION Most large boilers have water in the tubes and fire outside
Combustion is the controlled release of heat from the chemical — called water-tube boilers. Fire-tube boilers put the fire and
reaction between a fuel and an oxidizer. The fuels in the hot gases in tubes surrounded by water. This system is appli-
refining, petrochemical, and power generation industries are cable to smaller, unattended boilers.
almost exclusively hydrocarbons. Hydrocarbons comprise Process heaters are akin to water-tube boilers, but with some
only hydrogen (H) and carbon (C) in their molecular structure. very important differences. First, process heaters contain a pro-
Natural gas and fuel oil are examples of hydrocarbon fuels. cess fluid in the tubes, rather than water. The process fluid is
Other fuels are described later in this chapter and in Chapter 5. usually a hydrocarbon, for example, crude oil. Process heaters
come in two main varieties: vertical cylindrical (VC) and cabin
style (see Chapter 1). VCs comprise a cylindrical flame zone
2.2 USES FOR COMBUSTION surrounded by process tubes. Cabin-style heaters are rectangu-
lar with wall and roof tubes (see Figure 2.1). The radiant section
Combustion is used either directly or indirectly to produce
comprises the space surrounded by tubes having a direct view
virtually every product in common use. To name a few,
of the flame. Most process heaters also have a convective sec-
combustion processes produce and refine fuel, generate
tion comprised of overhead tubes that cannot directly view the
electricity, prepare foods and pharmaceuticals, and transport
flame. Convective tubes receive their heat from the direct con-
goods. Fire has transformed humankind and separated it from
tact of the combustion gases. The transition from the radiant
the beasts, illumined nations, and safeguarded generations. It
to convective sections is known as the bridgewall. Chapters 1
has been used in war and peace, to tear down and build up; it
and 15 contain more detailed discussion of process heaters.
is both feared and respected. It is a most powerful tool and
Reactors such as cracking furnaces and reforming furnaces
worthy of study and understanding.
are more extreme versions of process heaters. Here, the pro-
cess fluid undergoes chemical transformations to a different
substance. For example, in an ethylene cracking furnace, liq-
2.3 BRIEF OVERVIEW OF uid or gas feedstock transforms to ethylene (C2H4), an inter-
COMBUSTION EQUIPMENT AND mediate in the production of polyethylene and other plastics.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

HEAT TRANSFER There are many specialized types of reactors using combus-
In the process industries, combustion powers gas turbines, tion as the heat source.
process heaters, reactors, and boilers. The burner combusts
fuel and generates products of combustion and heat. A
firebox contains the flame envelope. The fire heats water in 2.4 NET COMBUSTION CHEMISTRY
the tubes to boiling. The steam rises to a steam drum that OF HYDROCARBONS
separates the liquid and vapor phases, returning water to the Consider the combustion of methane (CH4) and air. CH4 is
tubes and passing steam. The steam may be further heated in the major component of natural gas. The combustion of CH4
a superheater. Superheaters raise the temperature of the produces carbon dioxide (CO2) and water vapor (H2O).
steam above the boiling point, using either radiant and/or Equation 2.1 summarizes the net reaction.
convective heat transfer mechanisms.
Radiant heat transfer requires a line-of-sight to the flame. CH 4 + 2O 2 → CO 2 + 2H 2 O (2.1)
Only this heat transfer mechanism can operate in a vacuum.
For example, the Earth receives essentially all its heat from Equation (2.1) is the stoichiometric equation. It gives the
the sun through this mechanism. Convection requires the bulk relative proportions of every element (e.g., C, H, and O) in
movement of a hot fluid. In a boiler, hot combustion gases each molecule, and the relative proportions of each molecule
transfer heat to the outer tube wall via convection. Convection (e.g., CH4, O2, CO2, and H2O) in the reaction. A molecule is
occurs naturally by means of buoyancy differences between the smallest collection of chemically bound atoms that define
hot and cool fluids — termed natural convection, or by motive a substance. An atom is the smallest building block of a
devices such as fans or blowers — termed forced convection. molecule having a unique chemical identity. The arrow
Heat transfers from the outer to inner tube wall by conduction shows the direction of the reaction. Species to the left of the
— the predominant heat transfer mode through metals. Inside arrow are the reactants; those to the right are products.
the tube, convection is the predominant mode of heat transfer In a stoichiometric equation, the subscripted numbers define
to the inside fluid. A more complete discussion of heat trans- the proportions of elements in a molecule. Equation (2.1)
fer is given in Chapter 3. shows that the methane molecule comprises four hydrogen
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 35 Monday, March 14, 2005 11:04 AM

Fundamentals 35

atoms for every carbon atom, and that an oxygen molecule

comprises two oxygen atoms. Antecedent numbers (those that
precede the molecular identity and are not subscripted) define
the proportions of molecules in the reaction. The antecedent
numbers in a stoichiometric equation are the stoichiometric
coefficients. If a molecule appears without an antecedent
number, the stoichiometric coefficient is assumed to be “1.”
Thus, Equation (2.1) shows two oxygen molecules reacting
with one methane molecule to produce one molecule of car-
bon dioxide and two molecules of water.
Because stoichiometric equations deal only with propor-
tions, they can be multiplied by any convenient basis to obtain
more relevant or convenient units. One especially relevant
unit is the mole. A mole comprises 6.02 × 1023 molecules and
is abbreviated mol, where 6.02 × 1023 is known as Avogadro’s
number. The derivation of the mole follows a complicated
history but results in some convenient properties. For exam-
ple, 1 mole of hydrogen atoms has a mass of about 1 g
(actually, 1.01 g). The mass of a mole of atoms is the atomic
weight of the atom. Table 2.1 lists atomic weights for some
common elements.1 However, hydrogen does not exist as H
under normal conditions, but rather as the molecule H2. One
easily derives the molecular weight as the sum of atomic
weights of the molecule’s elements. For example, from Table
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

2.1, calculate the molecular weight of H2 as 2 × 1.01 = 2.02.

The molecular weight of CH4 is 12.01 + 4(1.01) = 16.05.

2.5 CONSERVATION OF MASS FIGURE 2.1 Typical cabin-style process heater.

Stoichiometric equations are always balanced equations.
That is, the number and kind of atoms in the reactants always
ton-mole that scale from, rather than refer to, Avogadro’s
equal the number and kind of atoms in the products. This is
because chemical reactions, including combustion, have no number. One obvious basis for a combustion problem is to
ability to alter atomic identities, only molecular ones. The scale from the actual fuel flow.
above discussion demands that combustion reactions In Equation (2.1), the moles of products equal the moles of
conserve mass; that is, the mass of the reactants must equal reactants; but in general, moles are not conserved. For exam-
the mass of the products. ple, 2H2 + O2 → 2H2O. Therefore, 3 moles of reactants yield
The stoichiometric equation can be used to generate any only 2 moles of product. However, 36.04 lb reactants generate
convenient mass basis. According to Equation (2.1), 1 mole 36.04 lb products, confirming that mass is conserved, but not
of CH4 reacts with 2 moles of O2 to produce 1 mole of CO2 moles. Table 2.2 shows molar ratios for combustion reactants
and 2 moles of H2O. Using Table 2.1, it is possible to calculate and products for some common fuels. Table 2.3 shows the
that 16.05 lb methane react with 32.00 lb air to produce 44.01 moles of O2 required to stoichiometrically combust 1 mole of
lb CO2 and 36.04 lb H2O. According to the principle of some common fuels, and the number of moles of CO2 and
conservation of mass, the mass of products equals the mass
H2O produced during that reaction.
of reactants, or 48.05 lb.
Strictly speaking, the units must be in grams to refer to
Avogadro’s number of molecules. However, the use of grams,
2.6 THE IDEAL GAS LAW
pounds, tons, or any convenient unit is justified as long as
proportions are preserved. This has led to the very common The ideal gas law applies for typical combustion reactions
and justified use of oxymorons like pound-mole (lbmol) and and relates the pressure, volume, and number of moles:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 36 Monday, March 14, 2005 11:04 AM

36 The John Zink Combustion Handbook

TABLE 2.1 Alphabetical List of Atomic Weights for Common Elements1

Name Symbol At. no. Atomic Weight Footnotes Name Symbol At. no. Atomic Weight Footnotes

Actinium Ac 89 [227] Mendelevium Md 101 [258]

Aluminum Al 13 26.981538(2) Mercury Hg 80 200.59(2)
Americium Am 95 [243] Molybdenum Mo 42 95.94(1) g
Antimony Sb 51 121.760(1) g Neodymium Nd 60 144.24(3) g
Argon Ar 18 39.948(1) g r Neon Ne 10 20.1797(6) g m
Arsenic As 33 74.92160(2) Neptunium Np 93 [237]
Astatine At 85 [210] Nickel Ni 28 58.6934(2)
Barium Ba 56 137.32(7) Niobium Nb 41 92.90638(2)
Berkelium Bk 97 [247] Nitrogen N 7 14.00674(7) g r
Beryllium Be 4 9.012182(3) Nobelium No 102 [259]
Bismuth Bi 83 208.98038(2) Osmium Os 76 190.23(3) g
Bohrium Bh 107 [264] Oxygen O 8 15.9994(3) g r
Boron B 5 10.811(7) g m r Palladium Pd 46 106.42(1) g
Bromine Br 35 79.904(1) Phosphorus P 15 30.973761(2)
Cadmium Cd 48 112.411(8) g Platinum Pt 78 195.078(2)
Calcium Ca 20 40.078(4) g Plutonium Pu 94 [244]
Californium Cf 98 [251] Polonium Po 84 [209]
Carbon C 6 12.0107(8) g r Potassium K 19 39.0983(1) g
Cerium Ce 58 140.116(1) g Praseodymium Pr 59 140.90765(2)
Cesium Cs 55 132.90545(2) Promethium Pm 61 [145]
Chlorine Cl 17 35.4527(9) m Protactinium Pa 91 231.03588(2)
Chromium Cr 24 51.9961(6) Radium Ra 88 [226]
Cobalt Co 27 58.933200(9) Radon Rn 86 [222]
Copper Cu 29 63.546(3) r Rhenium Re 75 186.207(1)
Curium Cm 96 [247] Rhodium Rh 45 102.90550(2)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Dubnium Db 105 [262] Rubidium Rb 37 85.4678(3) g

Dysprosium Dy 66 162.59(3) g Ruthenium Ru 44 101.07(2) g
Einsteinium Es 99 [252] Rutherfordium Rf 104 [261]
Erbium Er 68 167.26(3) g Samarium Sm 62 150.36(3) g
Europium Eu 63 151.964(1) g Scandium Sc 21 44.955910(8)
Fermium Fm 100 [257] Seaborgium Sg 106 [266]
Fluorine F 9 18.9984032(5) Selenium Se 34 78.96(3)
Francium Fr 87 [223] Silicon Si 14 28.0855(3) r
Gadolinium Gd 64 157.25(3) g Silver Ag 47 107.8682(2) g
Gallium Ga 31 69.723(1) Sodium Na 11 22.989770(2)
Germanium Ge 32 72.61(2) Strontium Sr 38 87.62(1) g r
Gold Au 79 196.96655(2) Sulfur S 16 32.066(6) g r
Hafnium Hf 72 178.49(2) Tantalum Ta 73 180.9479(1)
Hassium Hs 108 [269] Technetium Tc 43 [98]
Helium He 2 4.002602(2) g r Tellurium Te 52 127.60(3) g
Holmium Ho 67 164.93032(2) Terbium Tb 65 158.92534(2)
Hydrogen H 1 1.00794(7) g m r Thallium Tl 81 204.3833(2)
Indium In 49 114.818(3) Thorium Th 90 232.0381(1) g
Iodine I 53 126.90447(3) Thulium Tm 69 168.93421(2)
Iridium Ir 77 192.217(3) Tin Sn 50 118.710(7) g
Iron Fe 26 55.845(2) Titanium Ti 22 47.867(1)
Krypton Kr 36 83.80(1) g m Tungsten W 74 183.84(1)
Lanthanium La 57 138.9055(2) g Uranium U 92 238.0289(1) g m
Lawrencium Lr 103 [262] Vanadium V 23 50.9415(1)
Lead Pb 82 207.2(1) g r Xenon Xe 54 131.29(2) g m
Lithium Li 3 6.941(2) g m r Ytterbium Yb 70 173.04(3) g
Lutetium Lu 71 174.967(1) g Yttrium Y 39 88.90585(2)
Magnesium Mg 12 24.3050(6) Zinc Zn 30 65.39(2)
Manganese Mn 25 54.938049(9) Zirconium Zr 40 92.224(2) g
Meitnerium Mt 109 [268]

g – geological specimens are known in which the element has an isotopic composition outside the limits for normal material. The difference between the
atomic weight of the element in such specimens and that given in the table may exceed the stated uncertainty.
m – modified isotopic compositions may be found in commercially available material because it has been subjected to an undisclosed or inadvertent isotopic
fractionation. Substantial deviations in atomic weight of the element from that given the table can occur.
r – range in isotopic composition of normal terrestrial material prevents a more precise atomic weight being given; the tabulated atomic weight value should
be applicable to any normal material.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 37 Monday, March 14, 2005 11:04 AM

Fundamentals 37

TABLE 2.1 (continued) Alphabetical List of Atomic Weights for Common Elements
This table of atomic weights is reprinted from the 1995 report of the IUPAC Commission on Atomic Weights and Isotopic Abundances. The Standard Atomic
Weights apply to the elements as they exist naturally on Earth, and the uncertainties take into account the isotopic variation found in most laboratory samples.
Further comments on the variability are given in the footnotes.
The number in parentheses following the atomic weight value gives the uncertainty in the last digit. An entry in brackets indicates that mass number of the
longest-lived isotope of an element that has no stable isotopes and for which a Standard Atomic Weight cannot be defined because of wide variability in isotopic
composition (or complete absence) in nature.
REFERENCE
IUPAC Commission on Atomic Weights and Isotopic Abundances, Atomic Weights of the Elements, 1995, Pure Appl. Chem., 68, 2339, 1996.
Source: Courtesy of CRC Press.1

PV = nRT (2.2) TABLE 2.2 Molar Ratios for Some Combustion Reactions
and Products
where P = Pressure of the gas, psia Reaction Moles Reactants Moles Products
V = Volume of the gas, ft3 H2 + 0.5 O2 → H2O 1.5 1.0
n = Number of moles CO + 0.5 O2 → CO2 1.5 1.0
R = Gas constant = 10.73 psia-ft3/lbmol-°R CH4 + 2 O2 → CO2 + 2 H2O 3.0 3.0
C2H2 + 3 O2 → 2 CO2 + H2O 2.5 3.0
T = Absolute temperature, °R C3H8 + 5 O2 → 3 CO2 + 4 H2O 6.0 7.0
C4H10 + 6.5 O2 → 4 CO2 + 5 H2O 8.5 9.0
Degrees Rankine (°R) are defined as the number of
Note: Most combustion reactions do not conserve moles.
Fahrenheit degrees above absolute zero, the coldest possible
theoretical temperature. Equation (2.2) shows that gas TABLE 2.3 Molecular Weights and Stoichiometric
volume and moles are directly proportional. Coefficients for Common Gaseous Fuels
Another useful form of the ideal gas law is: Molecular O2 CO2 H 2O
Common Name Formula Weight (moles) (moles) (moles)
PM = ρRT (2.3) Hydrogen H2 2.02 0.5 0.0 1.0
Carbon monoxide CO 28.01 0.5 1.0 0.0
where ρ = Density of the gas, lb/ft3 Methane CH4 16.05 2.0 1.0 2.0
M = Molecular weight of the gas, lb/lbmol Ethane C2H6 30.08 5.0 2.0 3.0
Ethene, ethylene C2H4 28.06 4.0 2.0 2.0
Acetylene, ethyne C2H2 26.04 3.0 2.0 1.0
Several units need further explanation. Molecular weights, Propane C3H8 44.11 7.0 3.0 4.0
whether given in g/mol, lb/lbmol, or tons/ton-mole, all have Propene,
identical magnitude. That is, CH4 has a molecular weight of propylene C3H6 42.09 6.0 3.0 3.0
Butane C4H10 58.14 7.0 4.0 5.0
16.05 g/mol or 16.05 lb/lbmol. Because absolute zero is Butene, butylene C4H8 56.12 8.0 4.0 4.0
–459.67°F, to convert °F to °R, add 459.67. For example, Generic
70°F ≈ 530°R. Finally, psia is defined as pounds force per hydrocarbon CxHy 12.01 x + 1.01 y x + y/2 x y/2
square inch, absolute. Normal atmospheric pressure is
14.7 psia. However, pressure gauges read “0” at atmospheric
pressure, denoted 0 psig — pounds force per square inch, Example 2.1
gauge. Therefore, to convert from psig to psia, add local From Equation (2.2),
atmospheric pressure. For example, 35 psig ≈ 50 psia. Note
PV (30 + 14.7)[ psia] ∗ 1000[ft 3 ]
that atmospheric pressure varies with elevation. For example, n= = = 7.86 lbmol
RT  psia ft 3 
 ∗ ( 459.7 + 70)[°R]
the normal atmospheric pressure in Denver (elevation ~5000 10.73
ft) is only 12.3 psia. Thus, a gauge reading of 35 psig in  lbmol °R 
Denver equates only to about 47 psia. One must take into From Equation (2.3) obtain
a c c o u n t e l eva t i o n w h e n p e r f o r m i n g c o m bu s t i o n
calculations. PM (30 + 14.7)[ psia ] ∗ 16.05[lb lbmol]
ρ= =
An example best reinforces these points. A 1000 ft3 vessel RT  psia ft 3 
contains methane at 30 psig at 70°F. How many lbmol of 10.73  ∗ ( 459.7 + 70)[°R ]
 lbmol °R 
methane does the vessel contain? What is the gas density?
= 0.126 lb/ft 3
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

How much does the gas weigh? The solutions follow.

38 The John Zink Combustion Handbook

Finally, multiply the density by the volume to obtain the total wet products: TWP = A F + x 4 (2.9)
weight of gas, m = ρV = 0.126 lb/ft3 ∗ 1000 ft3 = 126 lb.
total dry products: TDP = A F − x 4 (2.10)
2.7 STOICHIOMETRIC RATIO AND In situ analyzers measure the flue gas species in the actual
EXCESS AIR hot wet environment. In contrast, extractive analyzers remove
The stoichiometric coefficient for oxygen identifies the the flue gas, condense the water, and measure the concentra-
theoretical oxygen required for combustion. To find the tion of the flue gas species in the dry gas. Therefore, two sets
theoretical air comprising this amount of oxygen, it is of equations are needed for wet and dry measurements.
necessary to define a mole of air as 0.21 lbmol O2 +
fO2 ,wet = ε(1 + x 4) TWP fO2 ,dry = ε(1 + x 4) TDP (2.11)
0.79 lbmol N2. Accordingly, Equation (2.1) is modified to
account for air [in brackets]:

fCO2 ,wet = 1 TWP fCO2 ,dry = 1 TDP

CH 4 + 2[O 2 + 79 21 N 2 ] →
(2.12)
(2.4)
CO 2 + 2 H 2 O + 2(79 21) N 2 fN2 ,wet = 79 21 (1 + ε ) (1 + x 4) TWP

Equation (2.4) is theoretical in that it presumes that all the fN2 ,dry = 79 21 (1 + ε ) (1 + x 4) TDP (2.13)
oxygen and fuel react and that nitrogen does not. Actually,
trace amounts of nitrogen will react with oxygen to form
nitrogen oxides (NOx). Although important in other contexts, fH2O,wet = x (2 TWP) (2.14)
the amount of reacting nitrogen is too small to consider here.
In industrial practice, perfect mixing cannot be achieved. It is where f is the mole or volume fraction of the subscripted
actually more cost-effective to ensure complete combustion species, 0 < f < 1, and the subscripts wet or dry refer to in situ or
with the addition of excess air. Excess air is that amount extractive measurements, respectively. Because of the strong
beyond theoretical added to ensure complete combustion of relationship between oxygen and excess air, the excess
the fuel. To account for excess air, Equation (2.4) is modified oxygen can be used as a measure of excess air. For this
with ε, the fraction of excess air. purpose, it is usually easier to recast the equations for oxygen
in the following forms:
CH 4 + 2(1 + ε )[O 2 + 79 21 N 2 ] →
(2.5) 0.21ε 0.21ε
fO2 ,wet = fO2 ,dry = (2.15)
CO 2 + 2 H 2 O + 2ε O 2 + 2(1 + ε )(79 21) N 2 K wet + ε K dry + ε

Equation (2.5) shows two important chemical features of

4 + 1.21x 4 + 0.79 x
complete combustion: no carbon monoxide (CO) and some K wet = K dry = (2.16)
4+x 4+x
unreacted oxygen appear in the combustion products. To
account for any hydrocarbon fuel, Equation (2.5) is modified
 fO2 ,wet   fO2 ,dry 
by x, the H/C molar ratio. Equation (2.6) gives a generic ε = K wet  ε = K dry 
  (2.17)
 0.21 − fO2 ,wet   0.21 − fO2 ,dry 

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
equation for hydrocarbons with air.

CH x + (1 + ε )(1 + x 4)[O 2 + 79 21 N 2 ] → The equations are displayed graphically in Figure 2.2a–f for
various fuels on a wet and dry basis.
CO 2 + x 2 H 2 O + ε(1 + x 4) O 2 (2.6)

+(1 + x 4)(1 + ε )(79 21) N 2 2.7.1 Heat of Combustion

In addition to the conservation of mass, energy is also con-
From Equation (2.6), it is possible to derive formulas relating served in a combustion reaction. One measure of the chemi-
volumes of flue gas species to excess air for a given fuel H/C: cal energy of a fuel is the heat of combustion. Table 2.4
gives heats of combustion for some typical fuels on a vol-
oxygen-to-fuel ratio: O F = (1 + x 4)(1 + ε ) (2.7) ume and mass basis.2 Heat of combustion is reported as
either net heating value (lower heating value, LHV) or gross
air-to-fuel ratio: A F = (100 21)(O F ) (2.8) heating value (higher heating value, HHV). To understand
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 39 Monday, March 14, 2005 11:04 AM

Fundamentals 39

FIGURE 2.2 Species concentration vs. excess air for the following fuels: (a) CH4, (b) natural gas, (c) simulated refinery gas (25% H2, 50%
A

CH4, 25% C3H8), (d) C3H8, (e) No. 2 oil, and (f) No. 6 oil.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

40 The John Zink Combustion Handbook

FIGURE 2.2 (continued)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fundamentals 41

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 2.2 (continued)

Copyright CRC Press

42 The John Zink Combustion Handbook

A

FIGURE 2.2 (continued)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fundamentals 43

FIGURE 2.2 (continued)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

44 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 2.2 (continued)

Copyright CRC Press

Fundamentals 45

TABLE 2.4 Combustion Data for Hydrocarbons

Higher Theor. Adiabatic
Heating Value Air/fuel Max Flame Temp Ignition Temp Flammability
(vapor), Ratio, Flame Speed, (in air) (in air) Flash Point Limits (in air)
Hydrocarbon Formula Btu lbm–1 by mass (ft s–1) (°F) (°F) (°F) (% by volume)

Paraffins or Alkanes
Methane CH 4 23875 17.195 1.1 3484 1301 Gas 5.0 15.0
Ethane C 2H 6 22323 15.899 1.3 3540 968–1166 Gas 3.0 12.5
Propane C 3H 8 21669 15.246 1.3 3573 871 Gas 2.1 10.1
n-Butane C4H10 21321 14.984 1.2 3583 761 –76 1.86 8.41
iso-Butane C4H10 21271 14.984 1.2 3583 864 –117 1.80 8.44

n-Pentane C5H12 21095 15.323 1.3 4050 588 < –40 1.40 7.80
iso-Pentane C5H12 21047 15.323 1.2 4055 788 < –60 1.32 9.16
Neopentane C5H12 20978 15.323 1.1 4060 842 Gas 1.38 7.22
n-Hexane C6H14 20966 15.238 1.3 4030 478 –7 1.25 7.0
Neohexane C6H14 20931 15.238 1.2 4055 797 –54 1.19 7.58

n-Heptane C7H16 20854 15.141 1.3 3985 433 25 1.00 6.00

Triptane C7H16 20824 15.151 1.2 4035 849 — 1.08 6.69
n-Octane C8H18 20796 15.093 — — 428 56 0.95 3.20
iso-Octane C8H18 20770 15.093 1.1 — 837 10 0.79 5.94

Olefins or Alkenes
Ethylene C 2H 4 21636 14.807 2.2 4250 914 Gas 2.75 28.6
Propylene C 3H 6 21048 14.807 1.4 4090 856 Gas 2.00 11.1
Butylene C 4H 8 20854 14.807 1.4 4030 829 Gas 1.98 9.65
iso-Butene C 4H 8 20737 14.807 1.2 — 869 Gas 1.8 9.0
n-Pentene C5H10 20720 14.807 1.4 4165 569 — 1.65 7.70

Aromatics
Benzene C 6H 6 18184 13.297 1.3 4110 1044 12 1.35 6.65
Toluene C 7H 8 18501 13.503 1.2 4050 997 40 1.27 6.75
p-Xylene C8H10 18663 13.663 — 4010 867 63 1.00 6.00

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Other Hydrocarbons
Acetylene C 2H 2 21502 13.297 4.6 4770 763–824 Gas 2.50 81
Naphthalene C10H8 17303 12.932 — 4100 959 174 0.90 5.9

Note: Based largely on: “Gas Engineers’ Handbook”, American Gas Association, Inc., Industrial Press, 1967. For heating value in J kg–1, multiply the value
in Btu lbm–1 by 2324. For flame speed in m s–1, multiply the value in ft s–1 by 0.3048.

REFERENCES
American Institute of Physics Handbook, 2nd ed., D.E. Gray, Ed., McGraw-Hill Book Company, 1963.
Chemical Engineers’ Handbook, 4th ed., R.H. Perry, C.H. Chilton, and S.D. Kirkpatrick, Eds., McGraw-Hill Book Company, 1963.
Handbook of Chemistry and Physics, 53rd ed., R.C. Weast, Ed., The Chemical Rubber Company, 1972; gives the heat of combustion of 500 organic compounds.
Handbook of Laboratory Safety, 2nd ed., N.V. Steere, Ed., The Chemical Rubber Company, 1971.
Physical Measurements in Gas Dynamics and Combustion, Princeton University Press, 1954.

the difference, reconsider Equation (2.1). From Equation yields the LHV. A condensing turbine is an example of the
(2.1), when methane burns, it produces two products: CO2 former process, while a typical process heater is an example
and H2O. The CO2 will remain a gas under all conceivable of the latter. Consequently, the process industry usually
industrial combustion conditions. However, H2O can exist uses the LHV. Boiler and turbine calculations usually use
as either a liquid or a vapor, depending on how much heat is the HHV. However, either measure can be used in combus-
extracted from the process. If so much heat is extracted that tion calculations as long as one is consistent. The inconsis-
the H2O condenses, then the combustion yields its HHV. If tent use of LHV and HHV is a major source of error in
water is released from the stack as a vapor, then combustion combustion calculations.

Copyright CRC Press

46 The John Zink Combustion Handbook

2.7.2 Adiabatic Flame Temperature Table B-4 in the Appendix gives the total heat capacity by the
How hot can a flame be? First, there is a difference between mass of each flue gas species. Rearranging Equation (2.17)
heat (Q) and temperature (T). Heat is energy in transit. When for ∆T gives the following:
a body absorbs heat, it stores it as another form of energy, ∆T = ∆H mC p (2.20)
increasing the body’s temperature and expanding it. That is,
the material uses some of the thermal energy to raise the Because there are several species in the flue gas, the
temperature and some of the energy to expand the body contribution of each species must be used for mCp. That is,
against the atmosphere. The same amount of heat absorbed in mCp = mCO2 CpCO2 + mH2O CpH2O + mO2 CpO2 + mN2 CpN2. To
different materials will yield different temperature increases illustrate the calculation, assume that an average heat
and expansions. capacity is ~0.30 lb/MMBTU. Then the adiabatic flame
For example, 100 Btu of heat will raise the temperature of temperature becomes
1 lb water by 100°F and expand the material from 62.4 ft3 to
63.8 ft3. The same 100 Btu of heat absorbed by 1 lb air will Btu  lb°F   1 
∆T = 22, 000 = 3543°F
lb  0.30 Btu   20.7 lb 
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

increase the temperature by 400°F and expand the material

from 13.1 ft3 to 15.6 ft3. The total energy used to raise the For air and fuel and 60°F, the adiabatic fuel temperature
temperature and increase the volume is called enthalpy (H). becomes
Enthalpy relates to temperature by a quantity known as the
isobaric heat capacity, Cp. Table B-4 in the Appendix gives AFT = 3543 + 60 = 3603°F
heat capacities for various gases. For a given mass of fuel, For more about adiabatic flame temperature, see Section 2.13
m, the quantities relate by Equation (2.18): and Table 2.4.
Note that the actual flame temperature will be much cooler
Q = m∆H = mC p ∆T (2.18) than this, because heat will transfer from the flame to the
surroundings via convection and radiation. Also, at high tem-
The symbol ∆ is used to denote a difference between two
peratures, CO2 → CO + 0.5 O2 and H2O → H2 + 0.5 O2,
states. Consider the following example.
reducing the adiabatic flame temperature. Such dissociations
have not been taken into account. Nonetheless, this calcula-
Example 2.2 tion method is useful for calculating the approximate adia-
If 1 lb CH4 combusts in 15% excess air, what is the batic flame temperature.
maximum possible flame temperature.
Solution: From (2.6), the stoichiometric equation is
2.8 SUBSTOICHIOMETRIC
CH 4 + 1.15 ∗ 2(O 2 + 79 21 N 2 ) →
COMBUSTION
CO 2 + 2 H 2 O + 0.30 O 2 + 8.65 N 2 The concept of excess air presumes air in addition to that
required for combustion. However, if one does not provide
Use the basis of 1 lb CH4 and ratio all other components by enough air, combustion may still continue, generating large
the molecular weight of CH4 to obtain the following: quantities of CO and combustibles. This is referred to as sub-
stoichiometric combustion. Process heaters and boilers
IN OUT
should NEVER be operated in this mode. Suddenly adding
CH4: 1.00 lb CO2: 1∗(44.01/16.05) = 2.74 lb air to such a hot mixture could result in explosion. Because
H2O: 2∗(18.02/16.05) = 2.24
O2: 1.15∗2∗(32.00/16.05) = 4.59 O2: 0.30∗(32.00/16.05) = 0.60
substoichiometric combustion may have deadly conse-
N2: 1.15∗2∗79/21∗(28.02/16.05) = 15.11 N2: 8.65∗(28.02/16.05) = 15.11 quences, it is useful to consider the process, observe its fea-
20.7 lb 20.7
tures, and learn to avoid it. The stoichiometric ratio, Φ, is a
fuel:air ratio. It has the following relationship with ε.
The highest possible flame temperature presumes no loss from
the flame whatsoever. This is known as the adiabatic flame Φ = 1 (1 + ε ) (2.21)
temperature. Equation (2.19) gives the total energy from
ε = (1 − Φ) Φ (2.22)
combustion of 1 lb of fuel:
Equation (2.23) shows Equation (2.6) modified for substoichio-
1 lb ∗ 22, 000 Btu/lb = 22, 000 Btu (2.19) metric combustion.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 47 Monday, March 14, 2005 11:04 AM

Fundamentals 47

CH x + a[O 2 + 79 21 N 2 ] → 1 1 2 + β 
1+  = 
x x
a= +  (2.27)
Φ 4  2  1 + β 2(1 + βK ) 
b CO + (1 − b) CO 2 + ( x − c) 2 H 2
One could solve for Φ and substitute into Equation (2.26).
+ c 2 H 2 O + a 79 21 N 2 (2.23)
However, the equation is quadratic and complicated. An
where a, x are specified and b, c are unknown having the easier solution is to solve for both Φ and the desired species
relation 2a = 2 – b + c/2. The reader should note that the using a parametric relation in β. Equation (2.28) gives the
formulation neglects soot. Turns3 has pointed out that using an relation for Φ:
equilibrium calculation with the water gas shift reaction
arrives at a good approximation for substoichiometric species. ( x + 4)(1 + β)(1 + βK )
Φ= (2.28)
This is adequate for investigating the general features of 2(2 + β)(1 + βK ) + x (1 + β)
substoichiometric combustion, which is done in Section 2.10.
Solving for the species as a function of β gives the following:

2.9 EQUILIBRIUM AND x 79  1 β x 

TWP = 1 + +  + + 
THERMODYNAMICS 2 21 1 + β 2(1 + β) 4(1 + βK ) 
Equation (2.24) gives the water gas shift reaction:
CO + H 2 O ↔ CO 2 + H 2 1  βKx  79  1 β x 
(2.24) TDP = 1 +  +  + + 
2  1 + βKx  21 1 + β 2(1 + β) 4(1 + βK ) 
The double-headed arrow indicates that the reaction proceeds
in both directions simultaneously. When the rate of the fO2 ,wet = 0 fO2 ,dry = 0 (2.29)
forward reaction equals that of the reverse, the process is in
dynamic equilibrium. Equilibrium is characterized by the 1  1  1  1 
fCO2 ,wet =   fCO2 ,dry =   (2.30)
following relation: TWP  1 + β  TDP  1 + β 
K = [CO 2 ][H 2 ] [CO][H 2 O] (2.25)
1  β  1  β 
fCO,wet =   fCO,dry =   (2.31)
where the brackets denote wet volume concentrations of the TWP  1 + β  TDP  1 + β 
enclosed species. For substoichiometric combustion, it will
be useful to define the following quantities: α = [H2]/[H2O], 1  x 
fH2O,wet =   (2.32)
β = [CO]/[CO2], then K = α/β. 2 TWP  1 + βK 

2.10 SUBSTOICHIOMETRIC 1  βKx  1  βKx 

fH2 ,wet =   fH2 ,dry =   (2.33)
COMBUSTION REVISITED 2 TWP  1 + βK  2 TDP  1 + βK 
Now that equilibrium and the water gas shift reaction have
79 1 2 + β x 
been defined, one can revisit substoichiometric combustion. fN2 ,wet =  + 
Solving the mass balance for C, H, and oxygen, in turn for α 21 2 TWP  1 + β 2(1 + βK ) 
and β, and using the relation K = α/β, one obtains the (2.34)
following equations: 79 1  2 + β x 
fN2 ,dry =  + 
21 2 TDP  1 + β 2(1 + βK ) 
1 2 + β x  79 
CH x +  +  O + N → Combining the excess air and substoichiometric equations,
2  1 + β 2(1 + βK )  2 21 2 
one can construct a graph of species concentrations vs. Φ, as
shown in Figure 2.3(a–e) for various fuels on a wet and dry
 1   β  1 x 
  CO 2 +   CO +  H O basis. In particular, the substoichiometric portion of the
 1 + β  1 + β 2  1 + βK  2
graphs use K = 0.19, which corresponds to a temperature of
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

~2200°F (1100°C). As Turns 3 has pointed out for propane,

1  βKx 
+   79   1   2 + β + x 
 H2 +       N (2.26) this gives excellent agreement with rigorous equilibrium
2  1 + βK  21 2  1 + β 2(1 + βK )  2
calculations. Note that one can generate considerable CO and
Now, by combining Equations (2.6), (2.10), (2.23), and H2 from substoichiometric combustion. If air is suddenly
(2.26) for the left side of the relation, one knows that a must admitted to such a hot mixture, explosion is likely. The moral
have the following expression. of the story? Do not run a process heater or boiler out of air!
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 48 Monday, March 14, 2005 11:04 AM

48 The John Zink Combustion Handbook

FIGURE 2.3 Species concentration vs. stoichiometric ratio for the following fuels: (a) CH4, (b) natural gas, (c) simulated refinery gas (25% H2,

.

50% CH4, 25% C3H8), (d) C3H8, (e) No. 2 oil, and (f) No. 6 oil.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fundamentals 49

.

FIGURE 2.3 (continued)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

50 The John Zink Combustion Handbook

FIGURE 2.3 (continued)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fundamentals 51

.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 2.3 (continued)

Copyright CRC Press

52 The John Zink Combustion Handbook

Stoichiometeric ratio, Φ
(e) Fuel Oil #2

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 2.3 (continued)

Copyright CRC Press

Fundamentals 53

.
(f)
FIGURE 2.3 (continued)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

54 The John Zink Combustion Handbook

FIGURE 2.4 Adiabatic equilibrium reaction process.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 2.5 Adiabatic equilibrium calculations for the predicted gas composition as a function of the O2:CH4 stoichiometry
for air/CH4 flames, where the air and CH4 are at ambient temperature and pressure.

2.11 GENERAL DISCUSSION proportions to make the mixture flammable. A spark is then
In this section, the concepts discussed so far are applied to initiated to ignite the mixture. The right box represents the
combustion in general. Figure 2.4 shows a cartoon of an adia- process an infinite time later to ensure all the reactions have
batic equilibrium process. The boxes represent perfectly gone to completion (i.e., reached equilibrium). In reality,
insulated enclosures, which do not exist in reality but are use- most combustion reactions are completed in only a fraction
ful for illustrating the concept. The boxes are filled with a of a second. Many species are then present after the reaction
combustible mixture of a fuel and oxidizer, in this case meth- is completed. The exact composition depends on the ratio of
ane and air, respectively. The left box represents the process the fuel to air. For example, if not enough air is present, then
at the time just before a spark is applied to ignite the mixture. CO will be generated. If sufficient air is present, then little or
The only species in the box are CH4 and air (O2 + 3.76N2) in no CO will be present. This is illustrated in Figure 2.5 which
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 55 Monday, March 14, 2005 11:04 AM

Fundamentals 55

FIGURE 2.6 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the major species as
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

a function of the air preheat temperature for air/CH4 flames, where the CH4 is at ambient temperature and pressure.

FIGURE 2.7 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the minor species as
a function of the air preheat temperature for air/CH4 flames, where the CH4 is at ambient temperature and pressure.

shows the predicted species for the adiabatic equilibrium greater than 2.0 are fuel lean, as excess oxygen is present.
combustion of methane and air as a function of the stoichi- This figure shows that the exhaust product composition is
ometry. For methane, the stoichiometric O2:CH4 ratio for the- highly dependent on the ratio of the fuel to the oxidizer.
oretically perfect combustion is 2.0 as shown in Table 2.3. 2.11.1 Air Preheat Effects
Stoichiometries less than 2.0 are fuel rich, as insufficient Figure 2.6 shows the major species for the predicted exhaust
oxygen is present to fully combust the fuel. Stoichiometries gas composition for the stoichiometric combustion of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 56 Monday, March 14, 2005 11:04 AM

56 The John Zink Combustion Handbook

FIGURE 2.8 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the major species as
a function of the fuel preheat temperature for air/CH4 flames, where the air is at ambient temperature and pressure.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 2.9 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the minor species as
a function of the fuel preheat temperature for air/CH4 flames, where the air is at ambient temperature and pressure.

methane with preheated air. There is almost no change up to is due to chemical dissociation. Figure 2.8 shows the
temperatures of about 1000°F (540°C), and only a relatively predicted major species in the exhaust products for the
small change at higher temperatures. Figure 2.7 shows the combustion of preheated methane with ambient air. There is
predicted minor species in the exhaust gas for the same very little change in the species concentration with fuel
reaction of ambient temperature methane with preheated air. preheat. Note that higher fuel preheat temperatures present
This graph shows that there is a dramatic increase in all the safety problems because of the auto-ignition temperature of
minor species as the air preheat temperature increases. This methane, which is approximately 1200°F (650°C) in air.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 57 Monday, March 14, 2005 11:04 AM

Fundamentals 57

FIGURE 2.11 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the minor species as a
function of the fuel blend (H2 + CH4) composition for air/fuel flames, where the air and fuel are at ambient temperature and pressure.

Figure 2.9 shows that the predicted minor species predicted major species for the combustion of air with fuel
concentrations increase with fuel preheat temperature. blends consisting of H2 and CH4. CO2 and N2 decline and
H2O increases as the H2 content in the fuel increases. It is
2.11.2 Fuel Blend Effects important to note that the species concentrations are not lin-
Fuel blends are particularly important in many of the hydro- ear functions of the blend composition, where the change
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
carbon and petrochemical industries. Figure 2.10 shows the occurs more rapidly at higher H2 compositions. Figure 2.11 is
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 58 Monday, March 14, 2005 11:04 AM

58 The John Zink Combustion Handbook

FIGURE 2.12 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the major species as a
function of the fuel blend (N2 + CH4) composition for air/fuel flames, where the air and fuel are at ambient temperature and pressure.

FIGURE 2.13 Adiabatic equilibrium stoichiometric calculations for the predicted gas composition of the minor species as a
function of the fuel blend (N2 + CH4) composition for air/fuel flames, where the air and fuel are at ambient temperature and pressure.

a similar plot of the predicted minor species as functions of concentrations as the N2 content increases. Figure 2.13 shows
the H2/CH4 fuel blend. This graph also shows strong non- the predicted minor species for the combustion of N2/CH4
linearities as the H2 content increases. Figure 2.12 shows the fuel blends. This graph also shows a rapid decline in the spe-
predicted major species for the combustion of air with fuel cies concentration, in this case for the minor species.
blends consisting of an inert (N2) and CH4. At the extreme of Real combustion processes are not adiabatic, as the whole
100% N2, there is no fuel left in the “fuel blend” and no com- intent is to transfer heat from the flame to some type of load.
bustion takes place. There is a rapid change in the species The amount of heat lost from the process determines the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fundamentals 59

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 2.14 Equilibrium calculations for the predicted gas composition of the major species as a function of the
combustion product temperature for air/CH4 flames, where the air and fuel are at ambient temperature and pressure.

FIGURE 2.15 Equilibrium calculations for the predicted gas composition of the minor species as a function of the
combustion product temperature for air/CH4 flames, where the air and fuel are at ambient temperature and pressure.

temperature of the exhaust gases. The higher the heat losses concentration as a function of temperature. Figure 2.15 shows
from the flame, the lower the exhaust gas temperature. the predicted minor species for the combustion of air and
Figure 2.14 shows the predicted major species for the com- methane as a function of the exhaust gas temperature. The
bustion of air and methane as a function of the exhaust gas concentrations are essentially zero up to temperatures of
temperature. The peak temperature is the adiabatic flame tem- about 2000°F (1100°C) and rapidly increase up to the adia-
perature. There is relatively little change in the major species batic flame temperature.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 60 Monday, March 14, 2005 11:04 AM

60 The John Zink Combustion Handbook

2.12 COMBUSTION KINETICS Consider NO formed from the high-temperature reaction

The net results of stoichiometric equations have been considered of N2 and O2, referred to as thermal NOx. There are more
thus far. However, the actual combustion mechanism is quite than 70 steps in the sequence. Often, a single elemental reac-
complex, involving very short-lived species that do not survive tion paces an entire reaction sequence, just as the slowest car
much beyond the flame. For example, the simplest system — on a mountain road paces all the cars following it. This
hydrogen combustion — comprises about 20 elemental slowest step is the rate-limiting step. In the case of thermal
reactions. Elemental reactions denote the actual species NOx, the rate-limiting step is the rupture of the N≡N triple
involved in the reaction. Sometimes, liberties are taken with bond by an oxygen atom:

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
the antecedent numbers in order to balance the net equation.
The equals (=) operator distinguishes the elemental reactions. N 2 + O = NO + N (2.40)
The arrow operator (→) refers to the net results. Some
important elemental reactions for the hydrogen-oxygen system In turn, the nitrogen radical reacts with available oxygen as
are as follows: follows:

N + O 2 = NO + O
2[H 2 + M = 2 H + M]
(2.41)
(2.35)

Adding these two equations together gives the overall

2[H + O 2 = OH + O] (2.36) reaction:

O + O + M = O2 + M (2.37) N 2 + O 2 → 2 NO (2.42)

2[H + OH + M = H 2 O + M] (2.38) 2.12.2 Reaction Rate

Fundamental kinetic principles formulate the rate law. In
general, the rate law for an elemental reaction rR + sS = pP + qQ
Net : 2 H 2 + O 2 → 2 H 2 O (2.39)
is given by the following differential equation:
In Equations (2.35 to 2.38), “M” refers to a third body such
d[ P ]
= k f [ R] [ S] − kr [ P] [Q]
r s p q
as the reactor wall or another nearby molecule that can (2.43)
absorb some of the reaction energy, but participates in no dt
other way. The numbers preceding the brackets [ ] are used
where P and Q are product species, R and S are reactant spe-
merely to balance the net reaction. Equations (2.35 to 2.38)
cies, t is time, kf is the forward reaction rate constant, kr is
show that the combustion of hydrogen is not as simple as
the reverse reaction rate constant, the lowercase letters are
2 H2 colliding with a single O2 to form 2 H2O, as the net
the stoichiometric coefficients of the species (uppercase),
reaction would suggest. About 20 elemental reactions exist
and the brackets [ ] denote wet volume concentrations. For
for the hydrogen-oxygen system.
Equation (2.42), it is possible to neglect the reverse reaction
The situation worsens for hydrocarbons. It takes more than
and write:
100 elemental reactions to describe the combustion of refinery
gases containing hydrogen, methane, and propane. Higher
d[NO]
hydrocarbons and more complex mixtures require even more = k f [N 2 ][O] (2.44)
reactions. Fortunately, the net reaction is enough for most dt
purposes. However, CO and NOx formation cannot be under-
stood using only net reactions. Unfortunately, [O] is not known, nor can it be conveniently
measured. However, if it is presumed that a partial
equilibrium exists between molecular and atomic oxygen,
2.12.1 Thermal NOx Formation: that is, 0.5 O2 → O [see Equation (2.41)], then [O] = KO20.5.
A Kinetic Example The constants can be combined as k = kf ∗ K.
NOx is one pollutant that regulatory agencies are scrutinizing Rate constants follow the temperature relation k = ATne–(b/T),
more and more (see Chapter 6). The “x” in NOx indicates a where A, b, and n are constants and T is the absolute temper-
variable quantity. NOx from boilers and process heaters ature. For the case of thermal NOx, n = 0. Making these
comprises mostly NO and very little NO2. substitutions gives the final differential equation:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 61 Monday, March 14, 2005 11:04 AM

Fundamentals 61

b TABLE 2.5 Adiabatic Flame

[N ] [O ]dt
−
d[NO] = Ae T
2 2 (2.45) Temperatures

This differential equation cannot be integrated because the Air

actual temperature-oxygen-time path in a turbulent diffusion Fuel °F °C

flame is not known. However, Equation (2.45) does show that H2 3807 2097
increasing the temperature, oxygen concentration, or time CH4 3542 1950

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
C2H2 4104 2262
increases NOx formation. Consequently, NOx reduction strat- C2H4 3790 2088
egies usually attempt to reduce one or more of these factors. C2H6 3607 1986
C3H6 4725 2061
C3H8 3610 1988
2.12.3 Prompt-NOx Formation C4H10 3583 1973
Another NOx formation mechanism is prompt NOx. This CO 3826 2108
occurs at the flame front and is responsible for no more than
20 ppm NOx in refinery or natural-gas fueled equipment. The
mechanism can be summarized as:

CH x + N 2 → HCN ↔ CN (not balanced) (2.46)

2.13 FLAME PROPERITIES
The flame temperature is a critical variable in determining the
HCN ↔ CN + O 2 → NO + CO + H (not balanced) (2.47) heat transfer, as will be shown in Chapter 3. This section shows
how the adiabatic flame temperature is affected by the fuel
Both of these reactions are very fast and do not require high composition, the equivalence ratio, and the air and fuel preheat
temperature. It would appear that the only way to reduce temperatures. As previously mentioned, real flame
NOx from the prompt mechanism would be to dilute the temperatures are not as high as the adiabatic flame temperature,
HCN and CN species on the fuel side of the combustion but the trends are comparable and representative of actual
zone, or reduce the available oxygen. conditions.

2.12.4 The Fuel-Bound NOx Mechanism 2.13.1 Flame Temperature

The fuel-bound NOx mechanism is similar to prompt NOx
and proceeds through the same HCN-CN chemistry. Table 2.5 shows the adiabatic flame temperature for common
However, the fuel-bound mechanism differs in the following hydrocarbon fuels combusted with air. Figure 2.16 shows the
ways. adiabatic flame temperature as a function of the equivalence
ratio for three fuels: H2, CH4, and C3H8. The peak temperature
1. The fuel-bound mechanism requires nitrogen as part of
occurs at about stoichiometric conditions (φ = 1.0). In that
the fuel molecule.
case, there is just enough oxidizer to fully combust all the
2. At low fuel-nitrogen concentrations, all of the bound
nitrogen converts to NOx. fuel. Any additional oxidizer absorbs sensible energy from the
3. The fuel-bound mechanism can be responsible for hun- flame and reduces the flame temperature. In most real flames,
dreds of ppm NOx, depending on the amount of nitrogen the peak flame temperature often occurs at slightly fuel lean
bound in the fuel. conditions (φ < 1.0). This is due to imperfect mixing where
slightly more O2 is needed to fully combust all the fuel.
The first steps in the chemistry differ in that the intermediates
Nearly all industrial combustion applications are run at fuel-
are formed directly from pyrolysis of the parent molecule.
lean conditions to ensure that CO emissions are low.
Ambient nitrogen is unimportant.
Therefore, depending on the actual burner design, the flame
CH x N y → HCN ↔ CN (not balanced) (2.48) temperature may be close to its peak, a condition that is often
desirable for maximizing heat transfer. One problem often
The subsequent chemistry (oxidation pathways for HCN and encountered by maximizing the flame temperature is that high
CN) is identical to prompt NOx. flame temperature maximizes NOx emissions. NOx increases
Reducing the available oxygen, reducing the nitrogen content approximately exponentially with gas temperature. This has
in the fuel, or diluting the fuel species with an inert gas reduces led to many design concepts for reducing the peak flame
NOx. temperature in the flame to minimize NOx emissions.4
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 62 Monday, March 14, 2005 11:04 AM

62 The John Zink Combustion Handbook

FIGURE 2.16 Adiabatic flame temperature vs. equivalence ratio for air/H2, air/CH4, and air/C3H8 flames, where the air
and fuel are at ambient temperature and pressure.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 2.17 Adiabatic flame temperature vs. air preheat temperature for stoichiometric air/H2, air/CH4, and air/C3H8
flames, where the fuel is at ambient temperature and pressure.

Figure 2.17 shows how preheating the air in the combustion shows the effect of preheating the fuel on the adiabatic flame
of the three fuels shown dramatically increases the adiabatic temperature. Again, there is a nearly linear rise in the flame
flame temperature. The increase is nearly linear for the air temperature, but the magnitude of the increase is much less
preheat temperature range shown. Air preheating is com- than for air preheating. This is due to the much larger mass
monly done to both increase the overall system efficiency of air compared to the mass of fuel in the combustion process.
(which will be graphically shown later) and to increase the Preheating the air to a given temperature requires much more
flame temperature, especially for higher temperature heating energy than preheating the fuel to that same temperature,
and melting processes like melting metal or glass. Figure 2.18 because of the difference in mass.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 63 Monday, March 14, 2005 11:04 AM

Fundamentals 63

FIGURE 2.18 Adiabatic flame temperature vs. fuel preheat temperature for stoichiometric air/H2, air/CH4 and air/C3H8
flames, where the air is at ambient temperature and pressure.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 2.19 Adiabatic flame temperature vs. fuel blend (CH4/H2 and CH4/N2) composition for stoichiometric air/fuel
flames, where the air and fuel are at ambient temperature and pressure.

Figure 2.19 shows how the flame temperature varies for fuel carbon and petrochemical applications for fluid heating.
blends of H2/CH4 and N2/CH4. The flame temperature increases Because such fuels are by-products of the chemical manufac-
as the H2 content in the blend increases. It is important to note turing process, their use is much less expensive than purchasing
that the increase is not linear; the increase is more rapid at higher H2 from an industrial gas supplier as well as being more cost-
levels of H2. Because of the relatively high cost of H2 compared effective than purchasing other fuels. The graph also shows that
to CH4 and C3H8, it is not used in many industrial applications. the adiabatic flame temperature decreases for N2/CH4 fuel
However, high H2 fuels are often used in many of the hydro- blends as the N2 content increases. Again, the decrease is not
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 64 Monday, March 14, 2005 11:04 AM

64 The John Zink Combustion Handbook

FIGURE 2.20 Adiabatic flame temperature vs. fuel blend (CH4/H2) composition and air preheat temperature for stoichi-
ometric air/fuel flames, where the fuel is at ambient temperature and pressure.

through openings, by air infiltration that will absorb sensible

energy, as well as by other types of energy losses that are
dependent on the burner and heater designs and by the pro-
cess operations. The accounting of the distribution for where
the energy goes in a process is sometimes graphically
depicted using a Sankey diagram. Figure 2.21 presents a very
simplified Sankey diagram showing that only 40% of the
energy goes to the load in that example. The available heat
for that example is 50%, which includes the 40% to the load
FIGURE 2.21 Sample Sankey diagram showing distribu- and the 10% lost to various sources. Figure 2.22 shows the
tion of energy in a combustion system. calculated available heat for three different fuels as a function
of the exhaust or flue gas temperature. As expected, there is a
rapid decrease in available heat as the exhaust gas temperature
linear and rapidly decelerates at higher N2 contents until no
increases. This indicates that more and more energy is being
flame is present for a “fuel” having 100% N2. Figure 2.20 shows
how preheating the combustion air for fuel blends of H2 and carried out of the exhaust instead of being transferred to the
CH4 increases the adiabatic flame temperature. However, the load as the exhaust temperature increases. At the adiabatic
increase is not a dramatic rise from pure CH4 to pure H2. Again, flame temperature for each fuel, there is no available heat as
the change in flame temperature with blend composition is all the energy was carried out in the exhaust. Figure 2.23
nonlinear. shows that the available heat increases with the air preheat
temperature, which simply indicates that energy was recov-
2.13.2 Available Heat ered in the process and was used to preheat the combustion
The available heat in a process is defined as the gross heating air. Figure 2.24 shows that preheating the fuel increases the
value of the fuel, minus the energy carried out of the exhaust efficiency, but to a much lesser extent than air preheating.
stack by the flue gases. This difference is the energy that is The mass of air is much greater than the mass of fuel, so pre-
available to do work. However, some of that energy will be heating the fuel is less effective than preheating the air if the
lost by conduction through the heater walls, by radiation preheat temperature is the same.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fundamentals 65

FIGURE 2.22 Available heat vs. gas temperature for stoichiometric air/H2, air/CH4, and air/C3H8, flames where the air
and fuel are at ambient temperature and pressure.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 2.23 Available heat vs. air preheat temperature for stoichiometric air/H2, air/CH4, and air/C3H8 flames at an
exhaust gas temperature of 2000°F (1100°C), where the fuel is at ambient temperature and pressure.

2.13.3 Minimum Ignition Energy intermediate species such as CH3, H, O, etc. Such species are
Ignition energy graphs usually have the vertical axis as the extremely reactive and recombine to form the final products,
relative energy of the fuel mixture (see Figure 2.25). The CO2 and H2O. Since the net heat release is greater than the
reactants start from an initial state. If the minimum ignition minimum ignition energy, the reaction, once started, will
energy is supplied, the reactant bonds will rupture, producing continue until virtually all of the reactants are consumed. The
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 66 Monday, March 14, 2005 11:04 AM

66 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 2.24 Available heat vs. fuel preheat temperature for stoichiometric air/H2, air/CH4, and air/C3H8 flames at an
exhaust gas temperature of 2000°F (1100°C), where the air is at ambient temperature and pressure.

What the diagram says is that fuel and air comprising a very
high chemical energy may exist in a metastable state, until one
introduces a spark or flame of sufficient energy. Once the sys-
tem reaches the minimum ignition energy, the reaction will be
self-sustaining until the reaction consumes enough of the reac-
tants. At that point, the reaction cannot liberate enough heat to
supply the minimum ignition energy and the flame goes out.

2.13.4 Flammability Limits

Suppose that fuel and air are not provided in stoichiometric
proportions, but have a great excess of fuel or air. Will the flame
continue to propagate if the ignition source is removed? That
FIGURE 2.25 Graphical representation of ignition and depends on whether the fuel/air mixture has enough chemical
heat release. energy to exceed the minimum ignition energy. If not, the flame
will extinguish. This leads to a lower and upper flammability
limit. The lower flammability limit (fuel lean) is where fuel is
horizontal axis shows the progress of the reaction. At the insufficient to achieve the minimum ignition energy. The upper
upper left, the diagram shows that the fuel/air mixture has a flammability limit (fuel rich) is where there is insufficient air.
high potential energy. At the lower right, it is noted that the Table 2.4 gives the flammability limits for many pure gases.
products of combustion have relatively little remaining
chemical energy. Because energy must be conserved, the 2.13.5 Flammability Limits for Gas Mixtures
difference between the upper and lower energy levels must For gas mixtures, one can use Le Chatelier’s rule to estimate
be the amount of heat that the combustion reaction liberates. flammability limits for gas mixtures. Because this is only an
Note, however, that the energy diagram does not slope estimate, one must confirm the flammability limit of the actual
monotonically along the reaction coordinate, but contains a mixture. Such experiments are relatively inexpensive and many
hump. This hump is the minimum ignition energy. third parties exist that can perform this kind of analysis.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch02-Frame Page 67 Monday, March 14, 2005 11:04 AM

Fundamentals 67

Le Chatelier’s rule states that the flammability limit of a this type are used as flame holders because they stabilize the
mixture is equal to the reciprocal of the sum of reciprocal flame front and keep it from moving forward or backward.
flammability limits weighted by their mole fractions. Another concept used in premix burners is quench distance,
the distance needed to remove sufficient heat from the flame
1 (2.49) to put it out. Here, burner slots or orifices have a finite thick-
LFL =
 1   1   1   1  ness that exceeds the quench distance. Because the burner is
K
y1   + y2  LFL  + y3  LFL  + + yn  LFL 
 LFL1   2   3   n cooler than the flame, if the flame does begin to burnback,
the heavy metal will remove sufficient heat and cool the flame
1 (2.50) below its minimum ignition energy. Without this feature, a
UFL =
 1   1   1   1  flame that finds its way into a premix burner could flashback.
K
y1   + y2  UFL  + y3  UFL  + + yn  UFL  With flashback, the combustion occurs in the burner, rather
 UFL1   2  3  n
than at the flame holder. Sustained burnback will destroy the
where LFL is the lower flammability limit, UFL is the upper burner in a short time.
flammability limit, LFLi is the LFL for species i, and UFLi is Diffusion burners supply fuel with no premix chamber. The
the UFL for species i. fuel meets the air outside the fuel nozzle. With diffusion burners,
flashback is not an issue because the fuel alone cannot support
combustion (i.e., the upper flammability limit is exceeded).
2.13.6 Flame Speeds
However, liftoff is still a concern. If the flame lifts off the burner,
The reaction between fuel and air can only occur at a finite it may travel to a place beyond the flammability limits and
speed. That finite speed depends on the speed of the reaction extinguish. Under certain conditions, the flame can repeatedly
(chemical) and the amount of turbulence in the flame (physical). liftoff and reestablish. This behavior is dangerous because the
If the flame has a lot of turbulence, hot pockets of gas recircu- fuel may burn incompletely during one part of the cycle and
late and the mixture burns faster. To first focus on the chemi- reignite later, causing an explosion. The cycle of liftoff and
cal part, suppose a long tube is filled with a flammable burnback can occur many times a second, causing rumble or
mixture. If one end of the tube is ignited, the flame front will vibration. Such rumble can be a sign of dangerous instabilities.
move along the tube at a precise velocity. A flame that has no Modern burners are designed to give high heat release in
turbulence is a laminar flame. Accordingly, the flame speed of short distances. This necessitates fuel velocities that greatly
a laminar flame is known as the laminar flame speed. It is a exceed the laminar flame speed. To stabilize such flames,
function of the kinetics of the combustion reaction. Under various flame holders are used. For example, an ignition ledge
standard conditions, this is a function of the fuel chemistry on a burner is a type of flame holder known as a bluff body.
alone. The laminar flame speeds for various fuels are tabu- Even if the air flows by the ledge at very high speed, the air
lated in Table 2.4. Now suppose that instead of a stationary speed very close to the ledge will be very slow. The flame
fuel mixture with a moving flame front, the fuel is moved. If will then establish very near the ignition ledge and be quite
the fuel is metered exactly at its flame speed, the flame front stable even over a wide range of firing rates. The burner tile
will remain stationary. If the fuel is metered faster than the itself is designed with a sudden expansion into the furnace,
flame speed, the flame front will move forward (called liftoff). which also acts as a flame holder because the gas velocity
If the fuel is metered slower than its flame speed, the flame decreases rapidly just after the expansion.
front will travel backward (called burnback or flashback).
Typical burners operate with fuel flows in excess of the lam-
REFERENCES
inar flame speed. To avoid liftoff, several devices are used.
Consider premix burners first. Fuel flows across an orifice into 1. D.R. Lide, Ed., CRC Handbook of Chemistry and
the throat of a venturi. The venturi is designed to entrain air Physics, 79th edition, CRC Press, Boca Raton, FL, 1998.
near the stoichiometric ratio. Gradual flow passages are used to 2. F. Kreith, Ed., The CRC Handbook of Mechanical
avoid turbulence, and hot gases are recirculated back to the Engineering, CRC Press, Boca Raton, FL, 1998.
burner. The fuel/air mixture is supplied at velocities above the
laminar flame speed. As the fuel jet issues from the burner, the 3. S.R. Turns, An Introduction to Combustion, McGraw-Hill,
velocity slows considerably. The flame front establishes where New York, 1996.
the flame and gas velocities are equal. The sudden expansion 4. J.L. Reese et al., State-of-the-Art of NOx Emission
from the burner avoids liftoff as the velocity rapidly slows. The Control Technology, Proc. Int. Joint Power Generation
high fuel/air velocity avoids burnback. Sudden expansions of Conf., Phoenix, October 3-5, 1994.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 3
Heat Transfer
Prem Singh, Michael Henneke, Jaiwant D. Jayakaran,
Robert Hayes, and Charles E. Baukal, Jr.

TABLE OF CONTENTS

3.1 Introduction............................................................................................................................................... 70
3.2 Conduction ................................................................................................................................................ 71
3.3 Thermal Conductivity ............................................................................................................................... 71
3.3.1 One-dimensional Steady-State Conduction ................................................................................ 73
3.3.2 Transient Conduction: Lumped Capacitance.............................................................................. 77
3.4 Convection ................................................................................................................................................ 82
3.4.1 Newton’s Law of Cooling ........................................................................................................... 82
3.4.2 Laminar Flow Convection........................................................................................................... 83
3.4.3 Turbulent Internal Flow .............................................................................................................. 84
3.4.4 Turbulent External Flow ............................................................................................................. 85
3.5 Radiation ................................................................................................................................................... 87
3.5.1 Blackbody Radiation/Planck Distribution .................................................................................. 88

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
3.5.2 Radiant Exchange Between Black Surfaces ............................................................................... 90
3.5.3 Radiant Exchange Between Gray/Diffuse Surfaces.................................................................... 91
3.5.4 View Factors for Diffuse Surfaces .............................................................................................. 92
3.5.5 Infrared Temperature Measurement............................................................................................ 92
3.5.6 Radiation in Absorbing/Emitting/Scattering Media ................................................................... 93
3.5.7 Mean-Beam-Length Method....................................................................................................... 95
3.5.8 Equation of Radiative Transfer ................................................................................................... 97
3.5.9 Radiation Emitted by a Flame .................................................................................................. 101
3.6 Heat Transfer in Process Furnaces .......................................................................................................... 102
3.6.1 Flame Radiation ........................................................................................................................ 105
3.6.2 Furnace Gas Radiation.............................................................................................................. 105
3.6.3 Refractory Surface Radiation.................................................................................................... 106
3.6.4 Analysis of Radiation Heat Transfer......................................................................................... 106
3.6.5 Heat Transfer Through the Wall of a Furnace........................................................................... 108
69
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 70 Monday, March 14, 2005 11:17 AM

70 The John Zink Combustion Handbook

3.6.6 Heat Transfer in the Process Tube ................................................................................................................109

3.6.7 Furnace Gas Flow Patterns ...........................................................................................................................109
3.6.8 Role of the Burner in Heat Transfer..............................................................................................................110
3.7 Conclusions.................................................................................................................................................................112
References ....................................................................................................................................................................................112
Nomenclature................................................................................................................................................................................114

3.1 INTRODUCTION Desmond give a brief presentation on nonluminous gaseous

Heat transfer is one of the fundamental purposes of combus- radiation, but no discussion of flames or combustion.17
tion in the process industries. The objective of many indus- Ganapathy’s book on applied heat transfer is one of the better
trial combustion applications is to transfer energy, in the form ones concerning heat transfer in industrial combustion and
of heat, to some type of load for thermal processing of that includes a chapter on fired-heater design.18 Blokh’s book is
load.1 An understanding of heat transfer is essential to the also a good reference for heat transfer in industrial combus-
successful design and operation of fired equipment. The tion and is aimed at power boilers.19 It has much information
objective of this chapter is to review helpful concepts of heat on flame radiation from a wide range of fuels, including
transfer, focusing on those topics as applied to combustion, pulverized coal, oils, and gases. A few handbooks on heat
particularly in the process industries. transfer have been written, but these also tend to have little
Numerous excellent books have been written on the subject if anything on industrial combustion systems.20–22
of heat transfer. However, almost none of them have any sig- Heat is a form of energy upon which the majority of all
nificant discussion of combustion. This is not surprising as the refinery processes are based. Heat transfer is that science which
field of heat transfer is very broad, making it very difficult for seeks to understand and predict the energy transfer between
any work to be exhaustive. Many of the heat transfer textbooks masses resulting from differences in temperature.2,3 Heat trans-
have no specific discussion of heat transfer in industrial com- fer is commonly divided into three mechanisms or modes for
bustion but do treat gaseous radiation heat transfer.1–7 classification: conduction, convection, and radiation.
The heat transfer books written specifically about radiation These mechanisms all have importance as applied to com-
often have sections covering heat transfer from luminous and bustion in the process industries. Consider a typical fired
nonluminous flames. Hottel and Sarofim’s book has a good heater, as shown in Figure 3.1, that consists of tubes with
blend of theory and practice regarding radiation.8 It also has flowing fluid (usually some type of oil) to be heated, and a
a chapter specifically on applications in furnaces. Love’s book burner (or group of burners) designed to provide the required
on radiation has short theoretical discussions of radiative heat energy for the desired process, a radiant section, and a con-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

transfer in flames and measuring flame parameters, but no vection section for heating. Major heat transfer processes in
other significant discussions of flames and combustion.9 petrochemical or refinery furnaces include:
Özisik’s book focuses more on interactions between radiation, 1. conduction through the furnace refractory and convection
conduction, and convection, with no specific treatment of from the wall of the furnace to the surrounding air
combustion or flames.10 A short book by Gray and Müller is 2. radiation exchange between the flame, the surrounding
aimed toward more practical applications of radiation.11 Spar- walls, and process tubes
row and Cess have a brief chapter on nonluminous gaseous 3. convection from the hot furnace gases to the process tubes
radiation, in which they discuss the various band models.12 and from process tube walls to the fluid flowing through
the tubes
Some of the older books on heat transfer are more practi-
cally oriented with less emphasis on theory. Kern’s classic See Section 3.6 for a comprehensive discussion on the vari-
book Process Heat Transfer has a chapter devoted specifically ous heat transfer processes taking place in a furnace and how
to heat transfer in furnaces, primarily boilers and petroleum one can calculate various effects.
refinery furnaces.13 Hutchinson gives many graphical solu- The consequences of the performance of these heat transfer
tions of conduction, radiation, and convection heat transfer mechanisms can significantly impact product throughput and
problems, but nothing specifically for flames or combustion.14 quality, furnace efficiency, equipment lifetime, and safety.
Hsu has helpful discussions on nonluminous gaseous radia- Other critical phenomena for consideration include the effect
tion and luminous radiation from flames.15 Welty discusses of heat transfer mechanisms on the fired equipment itself (e.g.,
heat exchangers, but not combustors or flames.16 Karlekar and heat transfer effects on burner fuel tips), or the effect of heat
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 71 Monday, March 14, 2005 11:17 AM

Heat Transfer 71

Stack TABLE 3.1 Thermal Conductivity of Common Materials

Material Btu/h ft F W/m C
Damper
Gases at atmospheric pressure 0.004 to 0.70 0.007 to 1.2
Insulating materials 0.01 to 0.12 0.02 to 0.21
Nonmetallic liquids 0.05 to 0.40 0.09 to 0.70
Nonmetallic solids (brick, stone, concrete) 0.02 to 1.5 0.04 to 2.6
Liquid metals 5.0 to 45 8.6 to 78
Convection Alloys 8.0 to 70 14 to 121
Pure metals 30 to 240 52 to 415
Section

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Heat
Loss
TABLE 3.2 Properties of Various Substances at above 32°F
(0°C) (except for steam as noted below)
Radiant ρ cp k
Section Lb/ft3 Btu/lb °F Btu/h ft °F
Metals
Copper 559 0.09 223
Aluminum 169 0.21 132
Nickel 556 0.12 52
Burner Iron 493 0.11 42
Carbon Steel 487 0.11 25
FIGURE 3.1 A typical fired heater. Alloy Steel 18Cr 8Ni 488 0.11 9.4

Nonmetal Solids
Limestone 105 ~0.2 0.87
transfer on the performance of the fired equipment with Glass Pyrex 170 ~0.2 0.58
respect to NOx emissions, flame stability, and flame shape. Brick K-28 27 ~0.2 0.14
Plaster 140 ~0.2 0.075
Kaowool 8 ~0.2 0.016
3.2 CONDUCTION Gases
Conduction heat transfer refers to the transfer of energy from Hydrogen 0.006 3.3 0.099
the more energetic to the less energetic particles of a sub- Oxygen 0.09 0.22 0.014
stance, resulting from interaction between the particles. Con- Air 0.08 0.24 0.014
Nitrogen 0.08 0.25 0.014
duction is the net transfer of energy by random molecular Steam1 0.04 0.45 0.015
motion — also called diffusion of energy. Conduction in
gases and liquids is by such molecular motion, except that in Liquids

liquids, the molecules are more closely spaced and the Water 62.4 1.0 0.32
Sulfur dioxide (liquid) 89.8 0.33 0.12
molecular interactions are stronger and more frequent. In the
1 Reference temperature for steam is 212°F (100°C). All other temperatures
case of solids, conduction refers to the energy transfer by lat-
are 32°F (0°C).
tice waves induced by atomic motion. When the solid is a
conductor, the translational motion of free electrons transfers
energy. In nonconductors, the transfer of energy takes place thermal conductivity of the material transferring heat. More
only via lattice waves. detailed information on thermal conduction heat transfer is
Heat conduction occurs in both stationary and moving sol- available in books specifically written on that subject.23–28
ids, liquids, and gases. The primary postulate of classical heat
conduction theory is that the rate of heat conduction in a
material is proportional to the temperature gradient. This is 3.3 THERMAL CONDUCTIVITY
consistent with the second law of thermodynamics, indicating Thermal conductivity is a material property that is expressed
that heat flows in the direction of decreasing temperature, or in Btu/(hr-ft-°F) or W/(m-K) and is dependent on the chemi-
from hot bodies to cold bodies. cal composition of the substance. Typical values for some
r materials are shown in Tables 3.1 and 3.2.
q = − k∇T (3.1)
The thermal conductivity of solids is generally higher than
Equation (3.1) states that heat flux is proportional to the tem- liquids, and liquids higher than gases. Among solids, the
perature gradient, and the proportionality constant is called the insulating materials have the lowest conductivities. The
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 72 Monday, March 14, 2005 11:17 AM

72 The John Zink Combustion Handbook

TABLE 3.3 Properties of Selected Gases at 14.696 psi

Temperature ρ cp k ρ cp k
°F lb/ft3 Btu/lb °F Btu/h ft °F lb/ft3 Btu/lb °F Btu/h ft °F

Air CO 2

0 0.0855 0.240 0.0131 0.1320 0.184 0.0076

500 0.0408 0.248 0.0247 0.0630 0.247 0.0198
1000 0.0268 0.263 0.0334 0.0414 0.280 0.0318
1500 0.0200 0.276 0.0410 0.0308 0.298 0.042
2000 0.0159 0.287 0.0508 0.0247 0.309 0.050
2500 0.0132 0.300 0.0630 0.0122 0.311 0.055
3000 0.0113 0.314 0.0751 0.0175 0.322 0.061

O2 N2

0 0.0945 0.219 0.0133 0.0826 0.249 0.0131

500 0.0451 0.235 0.0249 0.0395 0.254 0.0236
1000 0.0297 0.252 0.0344 0.0260 0.269 0.0320
1500 0.0221 0.263 0.0435 0.0193 0.283 0.0401
2000 0.0178 0.0672 0.0156 0.0468
2500 0.0148 0.0792 0.0130 0.0528
3000 0.0127 0.0912 0.0111

0 0.0059 3.421 0.1071

500 0.0028 3.470 0.1610
1000 0.0019 3.515 0.2206
1500 0.0014 3.619 0.2794
2000 0.0011 3.759 0.3444
2500 0.0009 3.920 0.4143
3000 0.0008 4.218 0.4880

Flue Gases
Natural Gas Fuel Oil Coal
k k k

0 — — —
500 0.022 0.022 0.022
1000 0.030 0.029 0.029
1500 0.037 0.036 0.036
2000 0.044 0.043 0.043
2500 0.051 0.049 0.050

thermal conductivities of pure metals typically decrease with decrease with increasing molecular weight. Thus, a light gas
an increase in temperature, while the conductivities of alloys such as hydrogen has a relatively high conductivity.
can either increase or decrease (see Table 3.2). For many When calculating the thermal conductivity of nonhomoge-
heat transfer calculations, it is sufficiently accurate to assume neous materials, one must use the apparent thermal conduc-
a constant thermal conductivity corresponding to the average tivity to account for the porous or layered construction of the
temperature of the material. material. In furnace refractory walls, the thermal conductivity
The thermal conductivities of most nonmetallic liquids can vary from site to site for the same material. This is because
range from 0.05 to 0.15 Btu/hr-ft-°F (0.09 to 0.26 W/m-K), the thermal conductivity of these materials is strongly depen-
and the thermal conductivities of many liquids tend to dent on their apparent bulk density (mass per unit volume).
decrease with increasing temperature. For higher temperature insulations, the apparent thermal con-
The thermal conductivities of gases increase with temperature ductivity of fibrous insulations and insulating firebrick
and are independent of pressure at the conditions at which most decreases as bulk density increases, because the denser mate-
furnace cavities operate. Generally, gas thermal conductivities
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
rial attenuates radiation. However, there is a limit at which
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 73 Monday, March 14, 2005 11:17 AM

Heat Transfer 73

any increase in density increases the thermal conductivity due 3.3.1.1 Plane Wall
to conduction in the solid material. It can be shown that the general heat equation in isotropic
It is known that the specific heats of solids and liquids are media is:
generally independent of pressure. Table 3.2 also shows the
∂  ∂ T  ∂  ∂T  ∂  ∂T  ∂T
specific heats of various metals, alloys, and nonhomogeneous k  + k  + k  + q˙ = ρc p (3.2)
materials at 68°F (20°C). These values can be used at other ∂ x  ∂ x  ∂ y  ∂ y  ∂z  ∂ z  ∂t
temperatures without significant error. where k = thermal conductivity of the media
Gases, on the other hand, demonstrate more temperature q̇ = rate of energy generation within the
dependence with regard to their specific heat. For all practical system
purposes, in furnace analyses one can neglect any pressure ∂T
ρc p = time rate of change of sensible energy of
dependence. Table 3.3 gives the specific heat data for air and ∂t
the system
other gases at different temperatures. In the case of steam and
Equation (3.2) is known as the heat diffusion equation. For
water, the variation of both thermal conductivity and specific
steady-state conduction in a medium with constant thermal con-
heat can be significant over the ranges of temperatures and
ductivity, the above equation, without heat generation, becomes:
pressures encountered in industrial steam systems. Refer to
any standard steam tables for data on water and steam. ∇2T = 0 (3.3)
When using thermal insulators as a heat barrier, one must
where ∇ is the Laplacian operator, defined in cartesian coor-
2
bear in mind that the effectiveness of an insulator depends
dinates as:
significantly on the temperature of its cold face. Thus, it is
not possible to protect a metal object in a furnace by insulating ∇ 2 ≡ ∂ 2 ∂x 2 + ∂ 2 ∂y 2 + ∂ 2 ∂z 2
all around it, unless there is an adequate path for the heat to
escape from the object to a cooler location, such as the atmo- For one-dimensional transfer of heat, Eq. (3.3) becomes:
sphere outside the furnace. Regardless of the thickness of
d 2T
insulation on an object that is in a furnace, if it is not attached =0 (3.4)
dx 2
to a cold sink, the object will eventually attain the furnace
temperature. This heat-up time period is governed by the Equation (3.4) for the plane wall shown in Figure 3.2 can be
specific heat of the material and is merely the time (and heat solved to obtain:
input) required to heat the mass of insulation (and object) up
to the furnace temperature. The quantity of heat required to (
T = Ts,2 − Ts,1 ) Lx + T s ,1 (3.5)
reach furnace temperature is given by:
The energy flux can be evaluated using Fourier’s law of heat
conduction:
Q = mCp ∆T

where m is the mass of the material and Cp is the specific heat

qx =
kA
(
T − Ts,2
L s,1
) (3.6)
of material and ∆T is the temperature difference between Table 3.4 summarizes one-dimensional steady-state solu-
ambient and furnace temperatures. tions to heat equations for different coordinate systems.
Equation (3.6) can also be written as:
3.3.1 One-dimensional Steady-State
Conduction qx =
(T s ,1 − Ts,2 ) (3.7)
In a one-dimensional steady-state conduction situation, the L
temperature change occurs only in one direction. The system kA
is described as steady state when the temperature at every where the quantity L/kA has units of (K/W) and is called the
point remains the same over time. This assumption is usually thermal resistance. Figure 3.2 also includes the equivalent
valid for analysis of a furnace wall under steady-state furnace thermal circuit. One side of the plane wall (x = 0) is being
operation, when the firing rate and temperature gradient heated by the surrounding fluid at T∞,1 and the other side
through the furnace wall can be considered constant for all (x = L) is being cooled by surrounding cold fluid at T∞,2. The
practical purposes. However, during startup and shutdown, thermal circuit in Figure 3.2 includes both convective heat-
the heat input and temperature gradients are changing over ing and cooling, as well as conduction through the material
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

time and must be treated differently. of the wall.

74 The John Zink Combustion Handbook

FIGURE 3.2 Heat transfer through a plane wall: (a) temperature distribution, and (b) equivalent thermal circuit.29

TABLE 3.4 One-dimensional, Steady-State Solutions to the Heat

Equation with No Generation29
Plane Wall Cylindrical Walla Spherical Walla

d 2T 1 d  dT  1 d  2 dT 
Heat equation =0 r =0 r =0
dx 2 r dr  dr  r 2 dr  dr 

x ln(r r2 )  1 − (r1 r ) 
Temperature distribution Ts,1 − ∆T Ts,2 + ∆T Ts,1 − ∆T  
L ln(r1 r2 )  1 − (r1 r2 ) 

∆T k∆T k∆T
Heat flux (q″) k
L r ln(r2 r1 ) r 2
[(1 r ) − (1 r )]
1 2

∆T 2 πLk∆T 4 πk∆T
ln(r2 r1 ) (1 r ) − (1 r )
Heat rate (q) kA
L 1 2

Thermal resistance (Rt. cond)

L ln(r2 r1 ) (1 r ) − (1 r )
1 2
kA 2 πLk 4 πk
a The critical radius of insulation is r = k/h for the cylinder and r = 2k/h for the sphere.
cr cr
From F.P. Incropera and D.P. DeWitt, Fundamentals of Heat and Mass Transfer, 4th edition,
John Wiley & Sons, New York, 1996. With permission.

3.3.1.2 Composite Wall T∞,1 − T∞,2

qx = (3.8)
In industry, different furnace designs are used for different
heat transfer operations. Design economics require that these ∑R t

furnaces often have several walls in series to increase where T∞,1 and T∞,2 are the surrounding temperatures and ΣRt
strength, provide better insulation, or even to enhance is the total thermal resistance of the system. The total thermal
appearance. The one-dimensional steady-state heat transfer resistance is evaluated as:
analysis can also be applied to these cases. Multiple walls in

∑R = h A + k A + k A + k A + h A
series can be considered to be a composite wall, as shown in 1 LA LB LC 1
t (3.9)
Figure 3.3. The heat flux in the x-direction is expressed as: 1 A
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
B C 2
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 75 Monday, March 14, 2005 11:17 AM

Heat Transfer 75

Tx,1

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Ts,1
T2

Ts,4

LA LB LC
Tx,4
Hot fluid kA kB kC
Tx,1, h1
Cold fluid
A B C Tx,2 , h2
x

Tx,1 Ts,1 T2 T3 Ts,4 Tx,4

1 LA LB LC 1
h1A k AA kBA kCA h4 A

FIGURE 3.3 Equivalent thermal circuit for a series composite wall.29

where h1 and h2 are convection heat transfer coefficients on

the two sides of the composite wall, respectively, and kA, kB,
and kC are the thermal conductivities of walls A, B, and C,
respectively. Thus, the heat transfer rate can be expressed as:

T∞,1 − Ts,1 Ts,1 − T2 T2 − T3

qx = = = =K (3.10)
1 LA LB
h1 A kA A kB A

Here A is the cross-sectional area for heat flow. An overall

heat transfer coefficient U can then be defined:

−1
 1   L   L   L   1  
1
U= =   +  A  +  B  +  C  +    (3.11)
Rtot A  h1   k A   k B   kC   h4  

A circuit diagram of the thermal resistance of the composite FIGURE 3.4 Temperature drop due to thermal contact
walls is also shown in Figure 3.3. resistance.

3.3.1.3 Contact Resistance

In composite systems, the temperature drop across the inter- two solid materials A and B, is illustrated in Figure 3.4. If the
face between the walls might be appreciable. This tempera- heat flux for a unit area of interface is qx″ (q x′′ = qcontact
′′ + qgap
′′ ),
ture drop, as caused by the contact resistance, Rt,c , between then the contact resistance can be defined as:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 76 Monday, March 14, 2005 11:17 AM

76 The John Zink Combustion Handbook

Rt′′,c =
TA − TB
(3.12) qr =
(
2 πLk Ts,1 − Ts,2 )
q x′′ (3.17)
ln(r2 r1 )

The thermal contact resistance for different combinations From the above equation, the conduction thermal resistance
of solids is available in standard texts.29 can be given by:

3.3.1.4 Conduction in Radial Coordinate Systems ln(r2 r1 )

Rt ,cond = (3.18)
For steady-state heat conduction with no generation of 2πLk
energy during conduction, the transfer in a hollow cylinder of
radius r is given as: The above concept can be extended to derive the equation for
heat transfer in a system of multiple cylinders (see
Figure 3.5) with radii r1, r2, r3, and r4; corresponding temper-
1 d  dT 
kr =0 (3.13) atures, Ts,1, Ts,2 , Ts,3, and Ts,4; thermal conductivities of kA, kB,
r dr  dr  and kC; and all having length, L:

Energy generation would refer to the exothermic or endo- T∞,1 − T∞,4

qr =
thermic energy release within the conducting medium. This,
1 ln(r2 r1 ) ln(r3 r2 ) ln(r4 r3 ) 1
of course, would not be the case in the wall of a pipe. How- + + + +
2 πr1 LH1 2 πk A L 2 πk B L 2 πkC L 2 πr4 Lh4
ever, if one were considering the transfer of heat through a
gas that is reacting, as is often the case in combustion, then (3.19)
energy generation would have to be factored in.
where T∞,1 and T∞,4 represent surrounding temperatures, and
Fourier’s law of heat conduction for this case can be written h1 and h4 represent convection heat transfer coefficients on
as: surfaces 1 and 4, respectively. Figure 3.5 also shows the ther-
mal circuit of the system.
dT dT In terms of the overall heat transfer coefficient U and total
qr = − kA = − k (2πrL) (3.14) thermal resistance Rtot ,
dr dr

T∞,1 − T∞,4
where 2πrL is the area of the cylinder normal to the direction
of heat flow. The temperature distribution in the cylinder can
qr =
Rtot
(
= UA T∞,1 − T∞,4 ) (3.20)

be determined by integrating Eq. (3.13) twice to give

3.3.1.5 Cartesian Coordinate Systems
Generation of energy during the conduction process causes
T (r ) = C1 ln r + C2 (3.15) the general heat balance equation to become:

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
where C1 and C2 are constants of integration and can be deter- d 2 T q˙
mined by the application of boundary conditions T(r1) = Ts,1 + =0 (3.21)
dx 2 k
and T(r2) = Ts,2.
Thus, the general equation for temperature distribution in where q̇ is the rate of heat generation per unit volume. The
a cylinder is: boundary conditions can be formulated as T(–L) = Ts,1 and
T(L) = Ts,2 , which, when used with the solution of Eq. (3.21),
gives the temperature distribution in the plane as:
Ts,1 − Ts,2 r
T (r ) = ln  + Ts,2 (3.16)
ln(r1 r2 )  r2  ˙ 2
qL  x 2  Ts,2 − Ts,1 x Ts,1 − Ts,2
T ( x) = 1 − 2  + + (3.22)
2k  L  2 L 2
The differentiation of Eq. (3.16) with respect to r and
substitution in Eq. (3.14) yields the expression for heat trans- When both surfaces are maintained at the same temperature,
fer rate: Ts, the above equation simplifies to:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 77 Monday, March 14, 2005 11:17 AM

Heat Transfer 77

FIGURE 3.5 Temperature distribution for a composite cylindrical wall.29

˙ 2
qL  x2  ˙ 02
qr  r2 
T ( x) =  1 −  + Ts (3.23) T (r ) = 1 − r 2  + Ts (3.25)
2k  L2  4k  0 

3.3.1.6 Cylindrical Coordinate Systems Applying Fourier’s law and differentiating the above equa-
The steady-state conduction equation with heat generation in tion with respect to r determines the heat transfer rate at any
the cylindrical coordinate systems can be represented by: radius in the cylinder.

1 d  dT  q˙ 3.3.2 Transient Conduction: Lumped

r + =0 (3.24)
r dr  dr  k Capacitance
Unsteady-state conduction involves storage of heat. For
Here, again, the heat generated within the system per unit example, in heating a furnace during startup, heat must be
volume is given by q̇. The solution to Eq. (3.24) with the supplied to bring the walls to the operating temperature and
boundary conditions T(r0) = Ts, and (dT/dr)r=0 = 0 becomes: to overcome the steady-state losses of normal operation. In a
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

78 The John Zink Combustion Handbook

typical continuous furnace operation, the heat stored in the  1 

walls and the metal of the tubes is insignificant compared to τt =   (ρVc) = Rt Ct (3.28)
the total heat input. In heaters that are heated and cooled peri-  hAs 
odically, such as in batch process work, the heat stored in the
walls can be a significant cost. where Rt is the resistance to convection heat transfer and Ct is
Unsteady-state conduction occurs in heating or cooling the lumped thermal capacitance of the solid. The physical
processes where the temperatures change with time. Exam- significance of Eq. (3.28) is the fact that any increase in the
ples include operating regenerative heaters, raising boiler value of Ct or Rt will cause the system to respond more
pressure, and turndown conditions on process furnaces. By slowly for a given change in the temperature.
introducing time as an additional variable, conduction analy-
ses become more complicated. 3.3.2.1 Applicability of Lumped Capacitance Method
Unsteady-state conduction problems can be easily solved Considering a plane wall with temperatures Ts,1 and Ts,2
if the temperature gradient within a solid material can be (Ts,1 > Ts,2) at sides 1 and 2, and a surrounding temperature of
ignored. The conditions under which temperature gradients T∞, the surface energy balance will give:
can be ignored are considered below. A common example
would be dropping a hot metal sphere into a cold-water bath.
The high thermal conductivity of the metal (compared with
kA
( )
T − Ts,2 = hA Ts,2 − T∞
L s,1
( ) (3.29)
the convection coefficient as discussed below) usually allows
analyzing the time-temperature history of the metal’s temper- where A is the surface area and k is the thermal conductivity
ature without regard to temperature variations within the of the solid. Equation (3.29) can be rearranged to give:
sphere. The simplified approach, based on neglecting the tem-
perature gradient in the metal, is called a lumped capacitance L
method. (T s ,1 ) = kA = hL ≡ Bi
− Ts,2
(T −T )
(3.30)
In a furnace setting, it is usually possible to neglect the 1 k
s ,2 ∞
temperature gradient in metal pipe walls because the thermal hA
conductivity is high compared to the rest of the heat transfer
path. However, insulating refractory cannot be treated as The dimensionless quantity (hL/k) is called the Biot number
lumped capacitance. Obviously, due to its simplicity, a (Bi). It is a measure of the temperature drop in a solid as
lumped capacitance approximation is the first resort in the compared to the temperature drop between the surface and
analysis of a transient problem. However, it should be noted the fluid. It can also be interpreted as the ratio of resistance
that its simplicity also makes it the least accurate approach. due to conduction and resistance due to convection. When the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

If a hot solid initially (t < 0) at a temperature Ti is cooled conduction resistance is negligible as compared to the con-
and attains any temperature T(t) at any time (t > 0), a general vection resistance (i.e., Bi << 1), the lumped capacitance
heat balance equation can be written as: assumption is valid. In the case of uneven surfaces or compli-
cated shapes, the estimation of length, L, is difficult and,
therefore, a characteristic length, Lc, is usually taken, which
− hAs (T − T∞ ) = ρVc 
dT
(3.26) is the ratio of volume to surface area.
 dt 
The exponent in Eq. (3.27) can now be rewritten as:
where h is the convection heat transfer coefficient, As is the
surface area of the solid, ρ is the density of the solid, c is hAs t hLc k t hL αt
= = c 2 = Bi ⋅ Fo (3.31)
the specific heat of the solid, and T∞ is the temperature of ρVc 2
k pc Lc k Lc
the surrounding medium. Solving the above differential
equation gives: where α is the thermal diffusivity and Fo is the Fourier num-
αt
ber ≡ 2 . Thus, the temperature distribution can be
θ T − T∞   hA   Lc
= = exp − s  t  (3.27) expressed as:
θ i Ti − T∞   ρVc  

θ T − T∞
Here, the quantity (ρVc/hAs) is called the thermal time con- = = exp( − Bi ⋅ Fo) (3.32)
stant, expressed as τ t. It can also be written as: θ i Ti − T∞
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 79 Monday, March 14, 2005 11:17 AM

Heat Transfer 79

If the condition of the solid is simultaneously affected by con- This differential equation can be solved with the help of the
vection, radiation, applied surface heat flux, and internal following initial and boundary conditions:
energy generation, the situation could be complicated and dif-
ficult to solve. Thus, the more general form of Eq. (3.26) is: IC: T ( x, 0) = Ti (3.37)

(
− hAs (T − T∞ ) + εAs σ T 4 − Tsur
4
)+ Eg + qs = ρVc
dT
dt
(3.33)  ∂T 
BC 1:  
 ∂x  x =0
=0 (3.38)

where Eg is the heat generated within the system, qs is the

 ∂T 
heat supplied to the system, ε is the emissivity of the solid, BC 2: - k   = h[T ( L, t ) − T∞ ] (3.39)
 ∂x  x = L
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

and σ is the Stefan-Boltzmann constant.

The above equation is a nonlinear, first-order, nonhomoge-
Using the following nondimensional quantities:
neous differential equation that cannot be integrated to obtain
an exact solution. However, when there is no imposed heat
θ T − T∞
flux and negligible convection compared to radiation, Eq θ* = = (3.40)
(3.33) can be simplified and solved to finally give: θ i Ti − T∞

where x* = x/L and t* = (α t/L2). The final functional depen-

ρVc  1 1 dence of temperature can be shown as:
t=  − 3 (3.34)
3εAs σ  T 3
Ti 
(
θ* = f x *, Bi, Fo ) (3.41)
Here it is assumed that the surrounding temperature (Tsur) is
zero. Schneider 31 presented an exact solution for the case when
On the other hand, if radiation is negligible compared to a plane wall is subjected to cooling from both sides. Starting
convection, and the convection coefficient, h, is constant with from the nondimensional form of Eq. (3.36), it was shown
respect to time, the differential equation can be solved to give: that:

T − T∞
Ti − T∞
= exp( − at ) +
ba
Ti − T∞
[1 − exp(− at )] (3.35) θ* = ∑ C exp(−ζ Fo) cos(ζ x )
n
2
n n
*
(3.42)

αt
hAs where Fo =
where a= L2
ρVc 4 sin ζ n
Cn =
q s + Eg 2ζ n + sin 2ζ n
b=
ρVc
The eigenvalues of ζn are positive roots of the transcendental
The second term on the right-hand side of Eq. (3.35) is the equation:
outcome of the applied flux and the heat generated within
the system. ζ n tan ζ n = Bi (3.43)

3.3.2.2 General Solution for the Lumped which are given in mathematical tables.
Capacitance Method
The one-dimensional, unsteady-state heat balance equation, 3.3.2.3 Radial Coordinate Systems
Schneider 30 has given exact solutions for an infinite cylinder
without any internal generation and with no energy input
and sphere. For an infinite cylinder, the dimensionless temp-
from the outside, is written as:
erature distribution is:

∂ 2 T 1 ∂T
=
∂x 2 α ∂t
(3.36) θ* = ∑ C exp(−ζ Fo)J (ζ r )
n
2
n o n
*
(3.44)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 80 Monday, March 14, 2005 11:17 AM

80 The John Zink Combustion Handbook

FIGURE 3.6 Transient conduction through a solid.

αt The eigenvalues of ζn are positive roots of the transcendental

where Fo = equation:
r02
1 − ζ n cot ζ n = Bi (3.49)
2 J1ζ n
Cn = ⋅ (3.45) given by Schneider.30
ζ n J 20ζ n + J12ζ n
3.3.2.4 Transient Conduction in Semi-infinite Solids
The eigenvalues of ζn are positive roots of the transcendental The unsteady-state, one-dimensional conduction equation
equation: given by Eq. (3.36) and the initial condition defined by
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Eq. (3.37) can also be applied to the case of semi-infinite

ζ n J1 (ζ n ) solids. Carslaw and Jaeger31 and Schneider30 have pre-
= Bi (3.46)
J 0 (ζ n )
sented exact solutions for the following three different
surface conditions:
Case (1): Constant Surface Temperature (Ts):
J1 and J0 are Bessel functions of the first kind and their values
are tabulated in standard mathematical tables. T ( x, t ) − Ts  x 
= erf  12 (3.50)
For the spherical case, the dimensionless temperature dis- Ti − Ts  ( 4αt ) 
tribution was given by Schneider30 as follows:
Case (2): Constant Surface Heat Flux (qs′′ = qo′′):
(
sin ζ n r *
)
θ* = ∑ C exp(−ζ Fo)
n
2
n
ζnr *
(3.47)   αt  1 2 
  π   −x2 
T ( x, t ) − Ti = 2 q0′′  exp 
αt  k   4αt 
where Fo = 
r02  

4 sin(ζ n ) − cos(ζ n )  x 
− q0′′  erfc
x
Cn = (3.48)  h 12
(3.51)
2ζ n − sin 2ζ n  ( 4αt ) 
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 81 Monday, March 14, 2005 11:17 AM

Heat Transfer 81

102
k = Thermal Conductivity Btu/h ft2 F/in.

Heavy Duty FB
90% Alumina
Castable
101
1st Quality FB Dense
Castable
Insul Castable
(b) 3000 IFB K-26
Insul Castable
IFB K-30 2500

600
stable 1
1 Insul Ca
K-20 Ceramic Fiber Board
IFB
s
las
e rg
Fib
Med Temp Block

10-1
103 104
Temperature, F
FIGURE 3.7 Thermal conductivity of (a) some commonly used steels and alloys and (b) some refractory materials.

Case (3): Surface Convection Condition: where the complimentary error function (erfc m) has been
defined as (1 – erf m).
 ∂T  h(T∞ − T (0, t ))(T ( x, t ) − Ti ) As a typical example, case (2) has been shown graphically
−k  =
 ∂x  x =0 T∞ − Ti in Figure 3.6. In the above relationships, the thermal conduc-
tivity of the material has been assumed to be constant for
 x    hx h 2 αt   most practical purposes if the temperature change is not
= erfc 1 2  − exp + 2  (3.52)
 ( 4αt )    k k  appreciable. Figure 3.7(a) shows thermal conductivities of
some commonly used steels and alloys as a function of tem-
  ( ) 
12
h 2 αt perature. Thermal conductivities of various refractory mate-
erfc x + 
  ( 4αt )1 2  rials are shown in Figure 3.7(b). Figure 3.8 shows temperature
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

 k 
 variation in a slab, as a function of thickness, under three
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 82 Monday, March 14, 2005 11:17 AM

82 The John Zink Combustion Handbook

is the most commonly employed mechanism in the process

industries. Schematically, hot and cold fluids, separated by a

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
solid boundary, are pumped through the heat-transfer equip-
ment, the rate of heat transfer being a function of the physical
properties of the fluids, the flow rates, and the geometry of
the system. Flow is generally turbulent and the flow duct
varies in complexity from circular tubes to baffled and finned
tubes. Theoretical analyses of forced convection heat transfer
have been limited to relatively simple geometries and laminar
flow. Usually, for complicated geometries, only empirical rela-
tionships are available. However, computational fluid dynam-
ics CFD (see Chapter 9), the science of computer modeling
flows and heat transfer, has advanced enough to provide good
information based on the available semi-empirical CFD mod-
els. In forced convection, heat transfer coefficients are
strongly influenced by the mechanics of flow occurring during
forced-convection heat transfer. Intensity of turbulence,
entrance conditions, and wall conditions are some of the
factors that must be considered for greater accuracy.
FIGURE 3.8 Temperature-thickness relationships cor-
responding to different thermal conductivities. Mixed convection refers to those situations when both nat-
ural and forced convection are at work. A good example
TABLE 3.5 Typical Convective Heat Transfer Coefficients would be the convective heat transfer process taking place on
Condition Btu/h ft2 F W/m2C the outside surface of a furnace wall when there is some wind
blowing. In the absence of wind, the wall would be cooled
Air, free convection 1 to 5 6 to 30
Air, forced convection 5 to 50 30 to 300 purely by natural convection; but with wind, both mechanisms
Steam, forced convection 300 to 800 1800 to 4800 are present simultaneously.
Oil, forced convection 5 to 300 30 to 1800
Water, forced convection 50 to 2000 300 to 1200 One example of the importance of forced convection in the
Water, boiling 500 to 20,000 3000 to 120,000 process industries is the convection section in many process
heaters. This is the downstream section of the heater that is
heated by the combustion exhaust gases exiting the radiant or
different conditions: (a) thermal conductivity is constant, primary heating section. Not all heaters have a convection
(b) thermal conductivity is increasing with temperature, and section, but Garg estimates that heater efficiency can be
(c) thermal conductivity is decreasing with temperature. increased from 55–65% to 80% or more with the addition of
a convection section.32 A number of books are available that
deal specifically with convection heat transfer.33–41
3.4 CONVECTION
Convection heat transfer takes place in fluids. A combination 3.4.1 Newton’s Law of Cooling
of molecular conduction and macroscopic fluid motion con-
tributes to convective heat transfer. Convection takes place Any convective transfer of heat can be represented by a gen-
adjacent to heated surfaces as a result of fluid motion past the eral heat balance equation called Newton’s law of cooling:
surface. All convection processes fall into three categories:
natural convection, forced convection, and mixed convection.
Natural convection occurs when fluid motion is created as q ′′ = h(Ts − T∞ ) (3.53)
a result of local density differences alone. Theoretical analy-
ses of natural convection require the simultaneous solution of where q″ is the heat flux, h is the convective heat transfer
the coupled equations of motion and energy. coefficient, and (Ts – T∞) is the temperature difference
Forced convection results when mechanical forces from between the hot fluid and the cold fluid/surrounding. Table 3.5
devices such as fans give motion to the fluid. Forced convection gives typical values of convective heat transfer coefficients.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 83 Monday, March 14, 2005 11:17 AM

Heat Transfer 83

FIGURE 3.9 Thermal boundary layer development in a heated circular tube.29

3.4.2 Laminar Flow Convection and other heat exchanger manufacturers use heat transfer tubes
with internal “rifling” to enhance the convective heat transfer.
An important factor that influences convective heat transfer is
the laminar sublayer. It is well-known that for a turbulent flow In most cases, the boundary layer effect is dominant in
of a fluid past a solid, in the immediate neighborhood of the gases. In a system transferring heat from a gas to a liquid, the
surface there exists a relatively quiet zone of fluid called the resistance on the liquid side can usually be neglected because
laminar sublayer. As one approaches the wall from the body of it is so much smaller than the resistance on the gas side.
the fluid, the flow slows down, and this slowed-down region is Consider a fluid at a uniform temperature T(r,0) entering a
tube where the heat transfer takes place from the wall of the

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
known as the boundary layer. The boundary layer itself has
fluid that is turbulent closer to the core of the flow, followed by tube (maintained either at a constant temperature or with a
a transition zone, and finally becomes laminar very close to the constant wall heat flux); a thermal entrance region is formed
wall. The portion of the flow that is essentially laminar is as shown in Figure 3.9. For laminar flow conditions, the thermal
called the laminar sublayer. In the laminar sublayer, the heat is entry length is given by Kays and Crawford 37 as:
transferred by molecular conduction. The resistance of the
laminar layer to heat flow will vary according to its thickness (x fd,t )
Dlam = 0.05 Re D Pr (3.54)
and can range from 95% of the total resistance for some fluids
to only 1% for others (e.g., liquid metals). where xfd,t is the thermal entrance length, D is the diameter of
In highly turbulent flows, the sublayer is thinner and, thus, the tube, ReD is the Reynolds number based on tube diameter,
greater turbulence makes for better heat transfer in general. and Pr is the Prandtl number. It is interesting that the equa-
Similarly, surface roughness and other mechanisms, such as tion for thermal entry length is very similar to the equation
oscillating flow or phase change, will aid in heat transfer by for hydrodynamic entry length, which is:
disturbing the boundary layer. So, from a heat transfer point
of view, it is of benefit in reformer and cracking tubes to have (x fd,h )
Dlam = 0.05 Re D (3.55)
the rough surface finish that results from the spin-cast process
used to make the tubes. Similarly, many boiler manufacturers where xfd,h is the hydrodynamic entry length.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 84 Monday, March 14, 2005 11:17 AM

84 The John Zink Combustion Handbook

3.4.2.1 Fully Developed Velocity and 3.4.3 Turbulent Internal Flow

Temperature Profiles Internal or conduit flow is a flow field in which the fluid com-
For the particular situation when the velocity in a tube can be pletely fills a closed stationary duct. On the other hand, exter-
approximated as uniform and the temperature is given by a nal or immersed flow is where the fluid flows past a
parabolic profile, that is: stationary immersed solid. With internal flow, the heat trans-
fer coefficient is theoretically infinite at the location where
u(r ) = C1 and [
T (r ) − Ts = C2 1 − (r r0 )
2
] (3.56) heat transfer begins. The local heat transfer coefficient rap-
idly decreases and becomes constant, so that after a certain
length, the average coefficient in the conduit is independent
The convection coefficient, h, is given by:
of the length. The local coefficient may follow an irregular
pattern, however, if obstructions or turbulence promoters are
qs′′ present in the duct.
h= (3.57)
Ts − Tm Because the analysis of turbulent flow heat transfer is quite
complex, calculations must rely on empirical correlations.
where Ts and Tm are tube surface temperature and the mean The Chilton-Colburn analogy provides an important correla-
temperature of the fluid in the tube, respectively. Also, the tion for the Nusselt number in turbulent flow heat transfer:
mean temperature is given by:
Nu D f
ST Pr 2 3 = Pr 2 3 = (3.64)
∫
r0
2 Re D Pr 8
Tm = uTrdr (3.58)
um r02 0

where St is Stanton number and f is the friction factor, given

Noting that the temperature profile is parabolic and the veloc- by
ity is uniform throughout, Eq. (3.56) shows that:

f = 0.184( Re D )
−1 5
−2 kC2 (3.65)
 ∂T 
qs′′ = k   = (3.59)
 ∂r  r =r r0
0
Thus,
Therefore,
Nu D = 0.023( Re D ) (Pr )1 3
45
(3.66)
qs′′ 4k
h= = (3.60)
Ts − Tm r0 Dittus and Boelter42 suggested a modification in Eq. (3.66)
by replacing the exponent of the Prandtl number by n, where
hD n is 0.4 for heating and 0.3 for cooling. It is to be noted that
or Nu D = =8 (3.61) Eq. (3.66) or its modification, are good for cases where the
k
temperature difference (Ts – Tm) is moderate. When the tem-
where NuD is the Nusselt number based on tube diameter. perature difference is large, the equation suggested by Sieder
Similarly, in a circular tube characterized by uniform sur- and Tate43 is recommended as follows:
face heat flux and fully developed conditions, the Nusselt
number is given as:
Nu D = 0.027( Re D ) (Pr )1 3 (µ µ s )
45 0.14
(3.67)

Nu D = 4.36 (3.62)
where µs is the viscosity of the fluid determined at surface
temperature, and all the other properties are measured at the
and for laminar, fully developed conditions, with a constant
mean temperature. The Dittus–Boelter and Sieder–Tate
surface temperature, the Nusselt number is:
equations are applicable for cases of both uniform surface
temperature and heat flux conditions. Petukhov 44 has given
Nu D = 3.66
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

(3.63) a correlation that gives more accurate results than the

Heat Transfer 85

Dittus–Boelter or Sieder–Tate equations, but is more complex 3.4.4 Turbulent External Flow
to use. The correlation is: External flow or immersed flow occurs when a fluid flows past
a stationary immersed solid. Similar to internal flow, the local

Nu D =
( f 8) Re D Pr (3.68)
coefficient immersed flow is again infinite at the point where
1.07 + 12.7( f 8) ( Pr 2 3 − 1)
12 heating begins. Subsequently, it decreases and may show vari-
ous irregularities, depending on the configuration of the body.
Usually, in this instance, the local coefficient never becomes
where the friction factor, f, is obtained from the Moody constant as flow proceeds downstream over the body.
diagram.45 When heat transfer occurs during immersed flow, the rate
For the special cases of liquid metals, where the Prandtl depends on the configuration of the body, the position of the

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
number is very small (0.003 ≤ Pr ≤ 0.05), Skupinski et al.46 body, the proximity of other bodies, and the flow rate and
have given a correlation for heat transfer in fully developed turbulence of the stream. The heat transfer coefficient varies
turbulent flow. For constant surface heat flux: over the immersed body, because both the thermal and the
momentum boundary layers change even with simple config-
Nu D = 4.82 + 0.0185 Pe 0.827 (3.69) urations immersed in an infinite flowing fluid. For complicated
D
configurations and assemblages of bodies, such as found on
the shell side of a heat exchanger, little is known about the
for 3.6 × 103 < ReD < 9.05 × 105 and 102 < PeD < 104, where
local heat transfer coefficient; empirical relationships giving
the Peclet number is defined as PeD = ReD·Pr. For constant
average coefficients are all that are usually available. Research
surface temperature:47
conducted on local coefficients in complicated geometries has
not been extensive enough to extrapolate into useful design
Nu D = 5.0 + 0.025 Pe 0.8
D (3.70) relationships.
For turbulent flow with Reynolds numbers up to about 108,
Reed48 has presented extensive literature on different the local friction coefficient is given by:49
correlations for heat transfer in laminar and turbulent flow
conditions. C f,x = 0.0592 Re −x1 5 (3.73)

3.4.3.1 Noncircular Tubes/Sections The velocity boundary layer thickness is given by:
Although the correlations discussed thus far have been pre-
sented for circular tubes, these relationships can be extended δ = 0.37 × Re −x1 5 (3.74)
to noncircular tubes and sections by simply replacing the
tube diameter with the hydraulic diameter, defined as: where x is the distance in the direction of flow. Thus, the local
Nusselt number for external turbulent flow is:
Ac
Dh = 4 (3.71)
P Nu x = St Re x Pr = 0.0296 Re 4x 5 Pr 1 3 (3.75)

where Ac is the flow cross-sectional area and P is the wetted where St is the Stanton number.
perimeter. Calculations of Reynolds number and Nusselt Complications arise when the boundary layer formation on
number are based on hydraulic diameter. In the case of a con- the external flow consists of both laminar and turbulent por-
centric annulus, the hydraulic diameter is given by: tions. Under these circumstances, neither laminar nor turbulent
correlations are satisfactory. A reasonably good correlation for
π mixed boundary layer conditions is:
4  Do2 − Di2
 4 ( )
Dh = = Do − Di (3.72)
πDo + πDi [
Nu L = 0.037 Re 4L 5 − 871 Pr 1 3 ] (3.76)

Table 3.6 summarizes the convection correlations in circular with the conditions that 0.6 < Pr < 60; 5 × 105 < ReL < 108;
tubes.29 Rex,c = 5 × 105; and where ReL is based on total length of the
surface and Rex,c is the critical Reynolds number for transition from laminar to turbulent.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 86 Monday, March 14, 2005 11:17 AM

86 The John Zink Combustion Handbook

TABLE 3.6 Summary of Convection Correlations for Flow in a Circular Tube29

Correlation Conditions

f = 64/ReD Laminar, fully developed

NuD = 4.36 Laminar, fully developed, uniform q s′′, Pr ≥ 0.6

NuD = 3.66 Laminar, fully developed, uniform Ts , Pr ≥ 0.6

0.0668( D L) Re D Pr Laminar, thermal entry length (Pr >> 1 or an unheated starting length), uniform Ts
Nu D = 3.66 +
[
1 + 0.04 ( D L) Re D Pr ]
23

0.14 Laminar, combined entry length

 µ
23
or, Nu D = 1.86 Re D Pr  {[ReDPr/(L/D)]1/3 (µ/µs)0.14} ≥ 2; uniform Ts , 0.48 < Pr < 16,700; 0.0044 < (µ/µs) < 9.75
  µ 
 LD   s

f = 0.316 Re −D1 4 Turbulent, fully developed; ReD ≤ 2 × 104

f = 0.184 Re −D1 5 Turbulent, fully developed; ReD ≤ 2 × 104

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
f = (0.790 ln Re D − 1.64)
−2
Turbulent, fully developed; 3000 ≤ ReD ≤ 5 × 106

Nu D = 0.023 Re 4D 5 Pr n Turbulent, fully developed; 0.6 ≤ Pr ≤ 160; ReD ≥ 10,000, (L/D) ≥ 10; n = 0.4 for Ts > Tm and n = 0.3 for Ts < Tm

0.14
 µ
Nu D = 0.027 Re 4D 5 Pr 1 3   Turbulent, fully developed; 0.7 ≤ Pr ≤ 16,700, ReD ≥ 10,000, (L/D) ≥ 10
 µs 

 ( f 8)(Re − 1000) Pr 
Nu D = 
D
 Turbulent, fully developed; 0.5 ≤ Pr ≤ 2000; 3000 ≤ ReD ≤ 5 × 106, (L/D) ≥ 10
 (
 1 + 12.7( f 8)1 2 Pr 2 3 + 1 ) 


Nu D = 4.82 + 0.0185(Re D Pr ) Liquid metals, turbulent, fully developed, uniform q s′′, 3.6 × 103 < ReD < 9.05 × 105, 102 < PeD < 104
0.827

Nu D = 5.0 + 0.025(Re D Pr )
0.8
Liquid metals, turbulent, fully developed; uniform Ts, PeD > 100

Source: F.P. Incropera and D.P. DeWitt, Fundamentals of Heat and Mass Transfer, 4th edition, John Wiley & Sons, New York, 1996. With permission.

Similarly, the suitable correlation for friction coefficient in When L is very high compared to xc (i.e., the entire surface
mixed boundary cases is given by: is covered by turbulent layer) the correlation for heat transfer
simplifies to:

[
C f,L = 0.074 Re1L 5 − [1742 Re L ] ] (3.77)
Nu L = 0.037 Re 4L 5 Pr 1 3 (3.78)

with the conditions that 5 × 105 < ReL < 108 and Rex,c = 5 × 105. and similarly, the friction coefficient becomes:

C f,L = 0.074 Re −L1 5 (3.79)

Heat Transfer 87

3.4.4.1 Convection Heat Transfer for Cylinders in TABLE 3.7 Constants of Equation (3.80)
Cross Flow for a Circular Cylinder in Cross Flow29
Hilpert 50 has presented a correlation for the average Nusselt ReD C m

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
number for convection heat transfer for cylinders in cross 0.4 – 4 0.989 0.330
flow: 4 – 40 0.911 0.385
40 – 4 × 10 3 0.683 0.466
4 × 103 – 4 × 10 4 0.193 0.618
hD
Nu D = = C Re mD Pr 1 3 (3.80) 4 × 104 – 4 × 10 5 0.027 0.805
k

where the Nusselt and Reynolds numbers are based on the TABLE 3.8 Constants of Equation (3.85) for the Tube Bank
diameter of the cylinder, and constants C and m are as in Cross Flow29
presented by Hilpert 50 and Knudsen and Katz (1958)51 in Configuration ReD,max C m
Table 3.7. Aligned 10 – 10 2 0.80 0.40
In Eq. (3.81), Churchill and Bernstein 52 suggest a more Staggered 10 – 102 0.90 0.40
comprehensive correlation covering a wider range of Rey-
Aligned 102 – 103  Approximate as a single
nolds and Prandtl numbers and suitable for the entire range Staggered 102 – 103

 (isolated) cylinder
of experimental data available:
Aligned 103 – 2 × 105 0.27 0.63
(ST/SL > 0.7)a
58 45
0.62 Re1D2 Pr 1 3   Re D  
Nu D = 0.3 + 1 +    (3.81)
[ ]
Staggered 103 – 2 × 105 0.35 (ST/SL)1/5 0.60
1 + (0.4 Pr )
23 14
  282, 000   (ST/SL > 2)

Staggered 103 – 2 × 105 0.40 0.60

where all the physical properties are determined at film
(ST/SL > 2)
temperature.
Aligned 2 × 105 – 2 × 106 0.021 0.84
3.4.4.2 Convection Heat Transfer in Banks of Tubes Staggered 2 × 105 – 2 × 106 0.022 0.84

Grimison53 suggested a correlation for convection heat trans- a For (ST/SL < 0.7), heat transfer is inefficient and aligned tubes should not
fer in aligned or staggered banks of tubes for ten or more be used.
Source: A. Zhukauskas, Advances in Heat Transfer, Vol. 8, J.P. Hartnett and
rows of tubes: T.F. Irvine, Jr., Eds., Academic Press, New York, 1972. With permission.

Nu D = 1.13 C1 Re mD,max Pr 1 3 (3.82)

for the conditions 2000 < ReD,max < 40,000 and Pr ≥ 0.7 Zhukauskas54 has presented a correlation that is more recent
For the number of rows less than ten, a correction factor and widely used:
must be used.29 In the above equation, the Reynolds number
is based on the maximum fluid velocity occurring within the  Pr 
14

tube bank. The maximum velocity for the aligned arrange- Nu D = C ⋅ Re mD,max Pr 0.36   (3.85)
ment is given by:  PrS 

The equation is valid for the following conditions: number

ST
Vmax = V (3.83) of rows, NL ≥ 20; 0.7 < Pr < 500; 1000 < ReD,max< 2 × 106,
ST − D
where all the properties, except Prs , have been determined at
and the maximum velocity for the staggered arrangement is: the arithmetic mean of the inlet and outlet fluid temperatures,
and constants C and m are listed in Table 3.8.
ST
Vmax = V (3.84)
2( SD − D)
3.5 RADIATION
where ST = center-to-center distance in a transverse plane Thermal radiation heat transfer is the movement of energy by
SD = distance between tube centers in a diagonal plane electromagnetic waves. Quantum theory describes electro-
D = diameter of tubes magnetic energy as photons or quanta. Unlike conduction or
V = freestream velocity convection, radiation does not require any intervening
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 88 Monday, March 14, 2005 11:17 AM

88 The John Zink Combustion Handbook

(Kelvin). The product σTs4 is also known as the blackbody

emissive power Eb.
Radiation plays an important role in many industrial pro-
cesses that require heating, cooling, drying, combustion, and
solar energy. Figure 3.10 shows a schematic of the propagation
of electromagnetic waves. The electrical and magnetic oscil-
lations can be seen to be orthogonal to each other. Figure 3.11
shows the spectrum of electromagnetic radiation and Table 3.9
gives approximate wavelengths, frequencies, and energies for
selected regions of the spectrum. Several books dedicated to
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

general radiation heat transfer are available.8,55–60

This section first treats thermal radiation heat transfer
FIGURE 3.10 Orthogonal oscillations of electric and between surfaces in enclosures. Each surface is assumed to
magnetic waves in the propagation of electromagnetic waves. be isothermal, gray, and diffuse. The assumption that the
surfaces are gray means that they emit and absorb thermal
radiation without regard to the wavelength or frequency of
medium for transfer. Electromagnetic radiation, in the wave- the radiation. Because thermal radiation is a wave, it has all
length range of 0.1 to 100 micrometers, is produced solely by of the properties of a wave. It has a wavelength, a frequency,
the temperature of a body. Energy at the body’s surface is and a wavenumber. It can be reflected, refracted, and dif-
converted into electromagnetic waves that emanate from the fracted. The assumption of diffuse surfaces implies that the
surface and strike another body. Some of the thermal radia- surfaces emit, absorb, and reflect radiation energy without
tion is absorbed by the receiving body and reconverted into regard to the direction (relative to the surface). All real sur-
internal energy, while the remaining energy is reflected from faces are neither gray nor diffuse. Many real surfaces have
or transmitted through the body. The fractions of radiation surface imperfections or oxide layers that may be on the order
reflected, transmitted, and absorbed by a surface are known, of 1 × 10–6 m (1 micron) thick. This surface layer will interact
respectively, as reflectivity ρ, transmissivity τ, and absorptiv- with radiation of wavelengths near this value. Additionally,
ity α. The sum of these fractions equals one. Thermal radia- electromagnetic theory can be used to predict both the direc-
tion is emitted by all surfaces whose temperatures are above tional and wavelength dependence of pure metal surfaces.
absolute zero. Analysis of radiation heat transfer in an industrial setting is
Thermal radiation can pass through some gases, like N2, complicated by these factors. Furthermore, the radiative prop-
without absorption taking place. Thus, these gases do not erties of real surfaces can be strongly dependent on the surface
affect radiative transfer. On the other hand, gases like carbon preparation. For example, a polished metal will have an emis-
dioxide, water vapor, and carbon monoxide affect radiation sivity that may be an order of magnitude below the emissivity
to some extent and are known as participating gases. These of the same surface with an oxide layer. Uncertainties in
participating gases are, of course, common constituents of radiative properties are always a consideration when conduct-
furnace flue gases and, as such, play a significant role in the ing radiation heat transfer analyses. However, when used
transfer and distribution of heat to the heater tubes. appropriately and with care, values of properties obtained
All surfaces emit radiation in amounts determined by the from literature surfaces can be used to make reasonably accu-
temperature and the nature of the surface. The perfect radi- rate heat transfer calculations.
ator, commonly known as a “blackbody,” absorbs all the
radiant energy reaching its surface and emits radiant energy 3.5.1 Blackbody Radiation/Planck
at the maximum theoretical limit according to the Stefan-
Distribution
Boltzmann law:
A blackbody is an ideal body, useful as a reference in discus-
sion of radiant heat transfer theory, that is both an ideal and
Qr = AσTs4 diffuse emitter and absorber of radiant energy. It absorbs all
incident radiation, regardless of wavelength and direction. No
where Qr is the radiant energy, A is the area, σ is the Stefan- real surface can emit more energy than a blackbody at a given
Boltzmann constant 0.1713 × 10–8 Btu/hr ft2-R4 (or 5.669 × temperature and wavelength. Because the emission from a
10–8 W/m2-K4), and Ts is the absolute temperature in Rankine blackbody is diffuse, the intensity of a blackbody is given as:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 89 Monday, March 14, 2005 11:17 AM

Heat Transfer 89

Wavelength (µm)
FIGURE 3.11 Spectrum of electromagnetic radiation.

TABLE 3.9 Spectrum of Electromagnetic Radiation

Wavelength Wavelength Frequency Energy
Region (Angstroms) (centimeters) (Hz) (eV)

Radio >10 9 >10 <3 × 10 9 <10–5

Microwave 10 to 10 6
9 10 to 0.01 3 × 109 to 3 × 1012 –5
10 to 0.01
Infrared 106 to 7000 0.01 to 7 × 10 – 5 3 × 1012 to 4.3 × 1014 0.01 to 2
Visible 7000 to 4000 7 × 10 –5 to 4 × 10 –5 4.3 × 1014 to 7.5 × 1014 2 to 3
Ultraviolet 4000 to 10 4 × 10 –5 to 10 –7 7.5 × 1014 to 3 × 1017 3 to 10 3
X-Rays 10 to 0.1 10 –7 to 10 –9 3 × 1017 to 3 × 1019 103 to 10 5
Gamma Rays <0.1 <10 –9 >3 × 1019 >10 5

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Eb 3.5.1.1 Planck Distribution
Ib = (3.86)
π Spectral distribution of a blackbody emission was first deter-
mined by Planck 61 and is given by:
where Eb is the emissive power of the blackbody.
2 hc02
A relatively small opening to a cavity with a uniform inte- I λ ,b ( λ , T ) =
[ ]
(3.87)
rior surface temperature closely approximates the radiation λ5 exp( hc0 λkT ) − 1
characteristics of a blackbody. Radiation that enters the sur-
face will be partially absorbed and partially reflected by the where h = 6.6256 × 10–34 J·s (Planck’s constant)
first internal surface of incidence. If the opening is small k = 1.3805 × 10–23 J/K (Boltzmann constant)
compared to the cavity dimension, then virtually all of the c0 = 2.998 × 108 m/s (speed of light in vacuum)
energy that enters the cavity will undergo multiple internal T = absolute temperature of the blackbody, in K
reflections and eventually be absorbed. Further, although the
On the assumption that the blackbody is a diffuse emitter,
surfaces within the cavity are not black and they do not emit
its spectral emissive power is given by:
blackbody radiation, their radiosity (radiosity is the total radi-
ation leaving a surface; in this case, it will be the emitted
2hπc02
radiation plus the reflected radiation) will be that of a black- Eλ ,b (λ, T ) = πIλ ,b (λ, T ) =
[ ]
(3.88)
body. Proof of this is given by Siegel and Howell.59 λ5 exp(hc0 λkT ) − 1
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 90 Monday, March 14, 2005 11:17 AM

90 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 3.12 Spectral blackbody emissive power.

Equation (3.88) is known as the Planck distribution. The dotted line in Figure 3.12 shows the locus of points of
Figure 3.12 shows the variation of spectral emissive power as the maximum in the spectral distribution curves.
a function of wavelength for selected temperatures. The figure
indicates that as temperature increases, the blackbody emis- 3.5.1.3 Stefan-Boltzmann Law
sive power at every wavelength increases and the wavelength Integration of the Planck distribution equation shows that the
of peak emission decreases. Radiation from the sun is approx- emissive power of a blackbody is given as:
imated by radiation from a 5800 K blackbody source. The
temperature at which radiant energy emissions from a surface Eb = σT 4 (3.90)
become visible to the human eye is called the Draper point,
occurring at approximately 800 K. where σ is the Stefan-Boltzmann constant and has the numerical
value of 5.670 × 10–8 W/m2·K4. (0.1714 × 10–8 Btu/hr·ft2·R4)
3.5.1.2 Wien’s Displacement Law Equation (3.90) is known as the Stefan-Boltzmann law. The
importance of this law is that the emissive power of a black-
From Figure 3.12 it is clear that the blackbody spectral distri-
body can be directly obtained for any temperature. Also, if the
bution has a maximum and the corresponding λmax depends
emissivity of any real surface is known, its emissive power
on temperature. Differentiating Eq. (3.87) with respect to λ
can be calculated using the blackbody emissive power.
and setting the result equal to zero gives:

3.5.2 Radiant Exchange Between

λ max T = 2897.8 µm ⋅ K (3.89) Black Surfaces
Because in a blackbody there is no reflection, energy leaves
Equation (3.89) is known as Wien’s displacement law. exclusively as emission and is absorbed completely by
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 91 Monday, March 14, 2005 11:17 AM

Heat Transfer 91

another blackbody. Figure 3.13 shows exchange between

two black surfaces of arbitrary shape and size. If qi→j is the ni nj
rate at which radiation leaves surface i and is intercepted by
surface j:
Ji=Ebi
Jj=Ebj Aj,Tj
qi→ j = ( Ai Ji ) Fij (3.91)

where Fij is the shape factor. Because the radiosity of a black

Ai,Ti
surface equals the emissive power,
FIGURE 3.13 Radiation transfer between two surfaces
qi→ j = Ai Fij Ebi (3.92) approximated as blackbodies.

Similarly,
Ebi − Ji
qi = (3.99)
(1 − ε i )( Ai ε i )

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
q j→i = Ai Fji Ebj (3.93)

Thus, the net exchange between the two black surfaces is: Thus, the total rate at which radiation reaches surface i from
all surfaces is:
(
qij = Ai Fij Ti 4 − Tj4 ) (3.94)
Ai Gi = ∑F A J ji j j (3.100)
3.5.3 Radiant Exchange Between
Gray/Diffuse Surfaces Using the reciprocity and summation rule, the net rate of
The main problem in the radiation exchange between non- radiation transfer to surface i becomes:
blackbodies is the surface reflection. Consider an exchange
between surfaces in an enclosure. Assume that they are
isothermal, opaque, and gray, with uniform radiosity and
qi = ∑ F A (J − J ) = ∑ q
ij i i j ij (3.101)

irradiation. The net rate of heat transfer from a surface is

given by: Equation (3.101) is a relationship between the net rate of
radiation transfer from surface i to the sum of radiant exchange
with the other surfaces. The above exchange can also be rep-
qi = Ai ( Ji − Gi ) (3.95) resented as:

where J is the radiosity and G is the irradiation. Also,

Ebi − ji
N
Ji − J j

Ji = Ei + ρi Gi (3.96) (1 − ε i ) ( Ai ε i )
= ∑ (A F )
j =1 i ij
−1 (3.102)

where E is emissive power and ρ is the reflectivity of the sur- A network representation of the above equation is shown in
face. Thus, Figure 3.14.
In situations where the net radiation transfer rate is known
qi = Ai ( Ei − α i Gi ) (3.97) and not the temperature, Eq. (3.101) is used in the alternate
form:
where ρi = (1 – αi) = (1 – εi) for an opaque, diffuse, gray sur-
Ji − J j
face. Therefore, the radiosity is given as:
qi = ∑ (A F ) i ij
−1 (3.103)

Ji = ε i Ebi + (1 − ε i )Gi (3.98)

The solutions of Eq. (3.101) or (3.103) are easily accom-
and plished by matrix inversion or iteration methods.29
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 92 Monday, March 14, 2005 11:17 AM

92 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 3.14 Network representation of radiative exchange between surface i and the remaining surfaces of an enclosure.29

3.5.4 View Factors for Diffuse Surfaces cos θ i cos θ j

∫∫
1
Fij = dAi dAj (3.106)
View factor (also called shape factor or configuration factor) Ai πR 2
is defined as the fraction of the radiation leaving surface i that
is intercepted by surface j. In Figure 3.15, two arbitrary sur- Similarly, view factor Fji can be calculated for the radiation
faces are exchanging radiation. They have areas Ai and Aj , leaving surface j and intercepted by surface i.
temperatures Ti and Tj , and are separated by a distance R. View factors follow reciprocity and summation rules given
They are at angles θi and θj from normals ni and nj. It can be as follows:
shown that the total rate at which radiation leaves surface i
and is intercepted by j is: Ai Fij = Aj Fji (reciprocity rule)

cos θ i cos θ j ∑F =1 (summation rule)

∫∫
ij
qi→ j = Ji dAi dAj (3.104)
πR 2

Tables 3.10 and 3.11 show view factors for two- and three-
Thus, the view factor, which is the fraction of radiation that dimensional geometries, respectively, and Figures 3.16, 3.17,
leaves Ai and is intercepted by Aj , is given as: and 3.18 show the view factors for three very common
configurations.
qi→ j
Fij = (3.105)
Ai Ji 3.5.5 Infrared Temperature Measurement
Planck’s distribution relates the radiation emitted by a
or, assuming that the two surfaces are diffuse emitters and blackbody to its temperature. This relation is used in heat
reflectors and have uniform radiosity, the shape factor is transfer analysis to determine how much energy is emitted
given as: by a surface. A further application of this relation is to mea-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 93 Monday, March 14, 2005 11:17 AM

Heat Transfer 93

dAj

nj
Gj
dAj cos Gj
Aj,Tj
R
ni
Gi
dAi
Ai,Ti ni dMj-i

dAi

FIGURE 3.15 View factor of radiation exchange between faces of area dAi and dAj .29
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

sure the intensity of the emitted radiation and use this mea- 3.5.6 Radiation in Absorbing/Emitting/
surement to determine the surface temperature. In practice,
Scattering Media
a number of complicating factors make it impossible to use
The foregoing discussion on radiation heat transfer was lim-
Planck’s distribution to convert the measured intensity to
ited to surface exchange. Surface exchange is radiation heat
the surface temperature. In reality, there are no perfectly
transfer from one surface to another, assuming that the
black surfaces. Real surfaces are at best gray and not always
medium between the two surfaces is a vacuum or a transpar-
diffuse. In addition, there will be radiation reflected from
ent substance. The notion of surface exchange is actually an
the surface that must be compensated for during measure-
idealization. When radiant energy is incident on a surface, it
ment. While it is possible to make these corrections analyti-
actually penetrates that surface some distance when consid-
cally, vendors of infrared temperature measuring devices
ered at the molecular level. For most metals, this distance is
invariably make extensive use of calibration. Calibration
only several Angstroms (Å, 10–10 m), while for most non-
allows the measurement device to be corrected for spectral
metals, it is several microns (µm, 10–6 m).
selectivity of the detector and for nonlinearities in the
Radiation absorption and emission in gases is due to the
detector’s response.
quantum energy levels of the gas molecules. An in-depth anal-
Figure 3.19 shows infrared temperature measurements made ysis and discussion of the topic is beyond the scope of this
on a burner. By selecting an appropriate wavelength for the section, but some understanding of gas spectra is necessary to
intensity measurement (in this case, the wavelength is 3.9 µm), understand gas radiation. Because air is primarily composed
the effects of CO2 and H2O between the emitting surface and of symmetric diatomic molecules (which typically do not emit
the infrared camera are negated. Surface temperatures can or absorb in the infrared) and inerts (N2, O2, and Ar), air is
then be readily measured “through” a flame. It is very difficult usually considered a transparent medium. Humid air, however,
to make reliable gas temperature measurements by measuring does absorb some radiation. Normally, this absorption is
the infrared emission because the emission from gases neglected as it is usually not significant. Other molecules that
depends on the temperature of the gas volume, the composi- are commonly found in combustion applications, such as CH4
tion of the gas volume, as well as the dimension of the gas and other hydrocarbons, CO2, H2O, CO, etc., do emit and
volume. Because most real applications involve nonisother- absorb in the infrared. Unlike many solid surfaces, however,
mal gas volumes (such as a flame in a furnace), IR measure- their emission and absorption do not smoothly vary with wave-
ments are not feasible. length. Rather, their emission and absorption spectra oscillate
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 94 Monday, March 14, 2005 11:17 AM

94 The John Zink Combustion Handbook

TABLE 3.10 View Factors for Two-dimensional Geometries

Geometry Relation

( ) ( )
12 12
 W +W 2
+ 4 −  Wj + Wi
2
+ 4
 i   
Fij = 
j
Parallel plates with midlines connected by
a perpendicular 2Wi

Wi = wi L , Wj = w j L

Inclined parallel plates of equal width and  α

Fij = 1 − sin
a common edge  2

( ) ( ) 
2 12
1 + wi w j − 1 + w j wi
Perpendicular plates with a common edge
Fij = 
2

wi + w j − w k
Three-sided enclosure Fij =
2 wi

violently with wavelength, but only in narrow “bands” centered is meaningless in a participating media. Instead, one must
around wavelengths particular to the species under consider- consider the intensity of radiation. From the study of surface
ation. exchange, one knows that the intensity of radiation emitted
If the medium between surfaces is not transparent to ther- by a diffuse surface is independent of angle, while emissive
mal radiation, it is called a participating medium. The notion power varies as the cosine of the normal angle. Section 3.5.8
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

of emissive power, so useful in analyzing surface exchange, briefly shows how radiant intensity is absorbed, emitted, and
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 95 Monday, March 14, 2005 11:17 AM

Heat Transfer 95

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
TABLE 3.10 (continued) View Factors for Two-dimensional Geometries
Geometry Relation

1 
[ ] − [C ]
12 2 12
Fij = π + C − ( R + 1) − ( R − 1)
2
2 2
K
2π 

R 1  R 1 
Parallel cylinders of different radii +( R − 1) cos −1  −  − ( R + 1) cos −1  +  
C C   C C 
R = rj ri , S = s ri

R = rj ri , S = s ri

r  −1 s1 s 
Cylinder and parallel rectangle Fij = tan − tan −1 2 
s1 − s2  L L

12
  D 2 
12
 D  s 2 − D2 
Infinite plane and row of cylinders Fij = 1 − 1 −  + tan −1  
  S    S  D2 

Source: Adapted from F.P. Incropera and D.P. DeWitt, Fundamentals of Heat and Mass Transfer, 4th edition, John Wiley & Sons, New York, 1996.

scattered by participating media. Radiant intensity within a diatomic gases of symmetrical composition such as O2, N2,
participating medium is a function of location (typically and H2 are transparent to thermal radiation. Important gases
three independent variables in a three-dimensional problem), that absorb and emit radiation are polyatomic gases such as
direction (two independent angles are required to describe CO2 and H2O and asymmetric molecules such as CO. Deter-
direction in a three-dimensional problem), and wavelength mination of radiant flux from gases is highly complex, but it
or wavenumber (one independent variable) if the problem is can be simplified by using Hottel’s assumption62 that
steady state. The fact that radiant intensity is a function of involves determination of emission from a hemispherical
six independent variables immediately indicates that the anal- mass of gas at temperature Tg to a surface element located at
ysis will be significantly more complicated than, for example, the center of the hemisphere’s base as:
conduction heat transfer, where there are only three indepen-
dent variables in a three-dimensional, steady-state problem. Eg = ε g σTg4 (3.107)
Further, if scattering is to be considered, the equation of
radiative transfer will have an integro-differential form. where Eg is emissive power and εg is the gas emissivity. Fig-
ures 3.21 and 3.22 show the emissivity of water vapor and
3.5.7 Mean-Beam-Length Method carbon dioxide, respectively, as a function of gas tempera-
Gases emit and absorb radiation in discrete energy bands dic- ture. The figures are based on experimental data taken in
tated by the allowed energy states within the molecule. While hemispherical shape of the gas at 1 atm total pressure, in a
the energy emitted by a solid shows a continuous spectrum, mixture with nonradiating gases. For pressures other than
the radiation emitted and absorbed by a gas is restricted to 1 atm, corrections must be incorporated. When carbon diox-
bands. Figure 3.20 shows the emission bands of carbon dioxide and water vapor both appear together with nonradiating
ide and water vapor relative to blackbody radiation at 1090 K gases, the total emissivity of the gas is obtained by:
(1500°F). The emission of radiation for these gases occurs in
the infrared region of the spectrum. The inert gases and ε g = ε w + ε c − ∆ε (3.108)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 96 Monday, March 14, 2005 11:17 AM

96 The John Zink Combustion Handbook

TABLE 3.11 View Factors for Three-dimensional Geometries

Geometry Relation

Aligned parallel rectangles

X=X L Y =Y L

( )( )
12
  1 + X 2 1 + Y 2 
2
( ) X
12
Fij = ln   + X 1+ Y 2 tan −1 K
π XY   1 + X + Y
2 2
 (1 + Y ) 2 12



( ) Y
12
+Y 1 + X 2 tan −1 − X tan −1 X − Y tan −1 Y 
(1 + X )
12


Coaxial parallel disks

Ri = ri L , R j = rj L

1 + R j2
S = 1+
Ri2

1 
( ) 
2 12
Fij =  2
S − S − 4 rj ri 
2  

H = Z X, W =Y X
Perpendicular rectangles with a common edge
1  1 1
( )
12
Fij = W tan −1 + H tan −1 − H 2 + W 2 K
πW  W H

× tan −1
1
+ ln 
(
1  1 + W 1 + H K
2 2
)( )
( H +W
2
)
2 12 4  1+ W + H

2
(
2
)

( )  ( ) 
W2 H2
 W2 1+ W2 + H2  H2 1 + H2 + W2 
× ×  
(
 1 + W W + H
2 2
)(
2
)  (
 1 + H H + W
2 2 2
)( )  


Source: Adapted from F.P. Incropera and D.P. DeWitt, Fundamentals of Heat and Mass Transfer, 4th edition, John Wiley & Sons, New York, 1996.

where the correction factor ∆ε can be obtained from Figure 3.23. Using mean beam length Le instead of L (the radius of
hemisphere), gas emissivity is obtained, which in turn gives
Normal emissivity of various surfaces is tabulated in
radiant heat transfer to a surface due to emission from an
Table 3.12.
adjoining gas:
The mean beam length, Le, can be defined as the radius of
a hemispherical gas mass whose emissivity is equivalent to q = ε g As σTg4 (3.110)
that for the geometry of interest. Table 3.13 gives the mean
beam length of numerous gas geometries and shapes from where As is the surface area. The net radiation exchange rate
Hottel.62 For geometries not covered in Table 3.13, the mean between the surface at temperature Ts and the gas at Tg is then
beam length can be approximated as given by

Le = 3.4 (Volume) (Surface area )

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
(3.109) (
qnet = As σ ε g Tg4 − α g Ts4 ) (3.111)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 97 Monday, March 14, 2005 11:17 AM

Heat Transfer 97

The relationship for determining absorptivity of carbon

dioxide and water vapor from their respective emissivity data
was given by Hottel 62 as follows:

( ) ( )
0.45
α w = C w Tg Ts ε w Ts , pw Le Ts Tg (3.112)

( ) ( )
0.65
α c = C c Tg Ts ε c Ts , pc Le Ts Tg (3.113)

When both carbon dioxide and water vapors are present, the
total gas absorptivity is obtained by:

α g = α w + α c − ∆α (3.114)

where ∆α = ∆ε. FIGURE 3.16 View factor for aligned parallel rectangles.29
1.0
3.5.8 Equation of Radiative Transfer 8
j
rj
6
Consider the propagation of a “pencil” beam of radiant 5 L
0.8 ri
energy through a participating medium. The radiant energy is 4 i
3
absorbed by the medium, decreasing the intensity of the radi-
0.6
ant energy according to: rj / L = 2
Fij

1. 5

 ∂I λ  1.25
= − aλ I λ ,
0.4
 
1.0
(3.115)
 ∂s  absorption 0.8

0.6
0.2
where aλ is the spectral absorption coefficient, and s is a coor- 0.4
dinate along the path. Additionally, the intensity of the radia- 0.3
0.0
tion is increased by emission from the medium. The increase 0.1 0.2 0.4 0.6 0.8 1 2 4 6 8 10
in the radiant intensity is given by: L / ri

 ∂I λ  FIGURE 3.17 View factor for coaxial parallel disks.29

  = a λ I bλ (3.116)
 ∂s  emission

A further effect to be considered when particulate media is

present is scattering. When radiant energy strikes a solid parti-
cle within the medium, the radiation may be reflected or dif-
fracted so that its direction changes. As radiation propagates,
its intensity is decreased by out-scattering and increased by
in-scattering. Attenuation by (out) scattering is described by:

 ∂I λ 
  = − aλ I λ (3.117)
 ∂s  outscatter

The increase in intensity due to in-scatter is given by:

 ∂Iλ  σ
  = sλ
 ∂s  inscatter 4 π ∫ I (s )Φ(s , s)dΩ
4π
λ i
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
i i (3.118) FIGURE 3.18 View factor for perpendicular rectangles
with a common edge.29
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 98 Monday, March 14, 2005 11:17 AM

98 The John Zink Combustion Handbook

FIGURE 3.19 Infrared thermal image of a flame in a furnace. (Courtesy of John Zink Co.)

(a) 1
Relative Blackbody

2400K
Radiance

1200K
600K

0
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10,000
Wave Number, cm -1
Carbon Dioxide
1
(b)
Relative Blackbody

2400K
Radiance

1200K
600K

0
1000 2000 3000 4000 5000 6000 7000 8000 9000 10,000
Wave Number, cm -1

Water Vapor
FIGURE 3.20 Emission bands of (a) CO2 and (b) H2O.16
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Heat Transfer 99

0.7
0.6
0.5
0.4 pw L = 20 ft-atm
10

Gas Emissivity, Aw, of Water Vapor

0.3
5
3
0.2 2
1.5
1.0
0.6
0.1 0.4
0.08 0.3
0.2
0.06
0.15
0.05
0.10
0.04
0.08
0.03 0.06

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
0.05
0.04
0.02 0.03
0.02
0.015
0.01
0.01 0.007
0.008 0.005

500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Temperature, R

FIGURE 3.21 Emissivity of water vapor in a mixture with nonradiating gases at 1-atm total pressure and of hemispherical shape.62

0.3

0.2

pw L = 5.0 ft-atm
Gas Emissivity, Ac, of Carbon Dioxide

3
0.1 2
0.08 1
0.6
0.06 0.4
0.05 0.2
0.04 0.10
0.06
0.03
0.04

0.02 0.02
0.010
0.008
0.006
0.01 0.005
0.008 0.004
0.003
0.002
0.005
0.004 0.001
0.003
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Temperature, R

FIGURE 3.22 Emissivity of carbon dioxide in a mixture with nonradiating gases at 1-atm total pressure and of hemispher-
ical shape.62
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 100 Monday, March 14, 2005 11:17 AM

100 The John Zink Combustion Handbook

0.07
Tg » 125° C Tg » 540° C Tg » 930° C
0.06
Mixture Correction, De

L (pw + pc) = 5.0 ft-atm

0.05 3.0
2.0
L (pw + pc) = 5.0 ft-atm
1.5
0.04 1.0
L (pw + pc) = 5.0 ft-atm
0.03 3.0 3.0 0.75
2.0
2.0 1.5
0.02 1.0 0.5
1.5
1.0 0.75
0.01 0.75 0.5
0.5 0.3
0.3
0.3 0.2 0.2 0.2
0.00
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
FIGURE 3.23 Radiation heat transfer correction factor for mixtures of water vapor and carbon dioxide.63

TABLE 3.12 Normal Emissivities, ε, for Various Surfaces TABLE 3.13 Mean Beam Lengths Le for Various Gas
Material Emissivity, ε Temp. (°F) Description Geometries
Aluminum 0.09 212 Commercial sheet Characteristic
Aluminum oxide 0.63–0.42 530–930 Geometry Length Le
Aluminum paint 0.27–0.67 212 Varying age and Sphere Diameter (D) 0.65 D
Al content (radiation to surface)
Brass 0.22 120–660 Dull plate Infinite circular cylinder Diameter (D) 0.95D
Copper 0.16–0.13 1970–2330 Molten (radiation to curved surface)
Copper 0.023 242 Polished Semi-infinite circular cylinder Diameter (D) 0.65D
Cuprous oxide 0.66–0.54 1470–2012 (radiation to base)
Iron 0.21 392 Polished, cast
Circular cylinder of equal height and diameter Diameter (D) 0.60D
Iron 0.55–0.60 1650–1900 Smooth sheet
(radiation to entire surface)
Iron 0.24 68 Fresh emeried
Infinite parallel planes Spacing between 1.80L
Iron oxide 0.85–0.89 930–2190
(radiation to planes) planes (L)
Steel 0.79 390–1110 Oxidized at 1100F
Cube Side (L) 0.66L
Steel 0.66 70 Rolled Sheet
(radiation to any surface)
Steel 0.28 2910–3270 Molten
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Arbitrary shape of volume V Volume to area 3.6(V/A)

Steel (Cr-Ni) 0.44–0.36 420–914 18–8 rough, after
(radiation to surface of area A) ratio (V/A)
heating
Steel (Cr-Ni) 0.90–0.97 420–980 25–20 oxidized in Source: From C.J. Hoogendoorn, C.M. Ballintijn, and W.R. Dorrestijn,
service J. Inst. Fuel, 43, 511–516, 1970. With permission.
Brick, red 0.93 70 Rough
Brick, fireclay 0.75 1832
Carbon, lampblack 0.945 100–700 0.003 in. or thicker
Water 0.95–0.963 32–212 Summing all of these effects results in the Equation of
Note: 1. SI conversion: T, °C = 5/9 (°F – 32). Transfer for radiation in a participating media:

∂Iλ
where the subscript i in the integrand denotes the incident ∂s
( )
= − aλ + σ sλ Iλ + aλ Ibλ

direction. Physically, this integral can be described as a sum- σ sλ

mation over all possible directions of the radiation entering a +
4π ∫ I (s )Φ(s , s)dΩ
4π
λ i i i
(3.119)
particular location, multiplied by the scattering phase function,
which represents the fraction of radiation traveling in a particu- This equation describes the propagation of radiation through
lar direction si that is scattered into a new direction s. absorbing/emitting/scattering media. It is an integro-differ-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 101 Monday, March 14, 2005 11:17 AM

Heat Transfer 101

FIGURE 3.24 Photographic view of a luminous flame. (Courtesy of John Zink Co.)

ential equation when scattering is considered. Analytical emitting radiant energy. This is not true. Radiant energy
solutions of the equation of transfer are possible only for is emitted only by the gases and solids (particularly car-
very simple geometries and boundary conditions. For more bon) present in the flame. The gaseous combustion prod-
complex geometries and boundary conditions, approximate ucts H2O and CO2 are the gases that emit radiation in
solution techniques such as the spherical harmonics method significant quantities, while any radical species (such as
and the discrete ordinates method can be used. These meth- CO) in the flame are present at such small fractions and
such thin pathlengths that their emission is typically neg-
ods are discussed further in Chapter 9. These approximate
ligible. However, determining the concentration and tem-
solution techniques are more fully discussed in texts such as
perature of H2O and CO2 within the flame is nontrivial
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
those by Modest63 and Siegel and Howell.59
and, in fact, the gases are clearly nonisothermal.
3. The presence of solid carbon particles within the flame
3.5.9 Radiation Emitted by a Flame
(which give flames a yellowish color) can dominate the
Accurate estimation of the heat emitted from a flame is very radiant emission from the flame. Again, predicting the
difficult, for reasons that include: concentration (measured as a volume fraction) and tem-
1. The flame temperature is not known. While one can perature of these carbon particles is very difficult.
readily calculate an adiabatic flame temperature for a
given fuel, the actual flame temperature will be below this Figure 3.24 shows a yellow luminous flame. The yellowish
value because the flame emits radiant heat. color of the flame is due to broadband radiation by carbon
2. The often-used term “flame radiation” suggests that some particles. The flame shown was produced by combusting a
special mechanism is at work within the reaction zone fuel oil atomized by steam. Flames with significant soot
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 102 Monday, March 14, 2005 11:17 AM

102 The John Zink Combustion Handbook

FIGURE 3.25 Photographic view of a nonluminous flame from a John Zink gas burner. (Courtesy of John Zink Co.)

fractions significantly radiate directly from the flame. In 3.6 HEAT TRANSFER IN PROCESS
contrast, nonluminous or slightly luminous flames (as shown FURNACES
in Figure 3.25) emit only a small fraction of the energy A complete treatment of heat transfer in process furnaces is
liberated by the combustion process. beyond the scope of this chapter, but the chapter would not
Figure 3.26 shows a radiant wall burner. In this burner, a be complete without describing, albeit briefly, the phenom-
mixture of fuel and air jets out radially from the burner. Very ena at play in process furnace heat transfer.
near the burner, the wall is dark because the flame “stands off” Process furnaces are a good example of systems that incor-
of the burner exit. Further away from the burner, the refractory porate all the heat transfer mechanisms concurrently at work
surface is a bright yellow color. In this particular application, in gases, liquids, and solids. The challenge is to achieve good
the temperature of this refractory is above 2000°F (1100°C). heat transfer in the radiation and convection mechanisms.
As illustrated in the photograph, the visible radiation from the Conduction plays only a minor role in getting the heat from
hot refractory surface dominates any flame radiation in the the flame to the process fluid, but it is the primary mechanism
visible region, rendering the flame invisible to the human eye. at work in preventing heat loss to the surroundings.
This burner is very common in ethylene pyrolysis furnaces, For the following discussion, refer to Figure 3.1 at the
where the burner is used to heat a refractory wall. The primary beginning of this chapter. The main part of the furnace that
heat transfer mode to the refractory wall is probably convec- contains the burner flames is the radiant section. The process
tive, although gas radiation plays a role. The hot wall then heating tubes are located in the radiant section in various
radiates energy to process tubes that run parallel to the wall arrangements. In low temperature furnaces (gases exit the
at a distance of approximately 3 feet (1 m). radiant section at less than 800°C (1500°F)), the process tubes
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Heat Transfer 103

FIGURE 3.26 Photographic view of a radiant wall burner. (Courtesy of John Zink Co.)

are often located close to the walls of the furnaces, and in spectrum, their emission into the infrared spectrum is quite
high temperature furnaces (gases exit the radiant section at large. In typical cracking furnaces and some boilers where
more than 980°C (1800°F)), the tubes are usually suspended temperatures are high (above about 1800°F or 1000°C), the
in the main furnace space, away from the walls. furnace walls will glow bright orange or even yellow. In
In many furnace designs, a convection section is located these furnaces, it is frequently difficult to see a gas flame
downstream of the radiant section. The combustion gases that visually. This is because the walls are radiating in the visual
leave the radiant section flow through the convection section spectrum. However, even though the naked eye cannot see
and then through the stack to be vented to atmosphere. any gas radiation from the flame, radiation from the flame
Depending on the design, an air preheater (not shown in is still a significant contributor to heat transfer. Figure 3.12
Figure 3.1) may be installed in the path to the stack to further (the blackbody emissive power graph) showed that black-
extract heat from the flue gas. The convection section may body curves do not cross. This means that at any given
be used to preheat the process fluid, generate steam, or heat wavelength, the hotter the radiator is, the higher the black-
another process fluid. Since the radiant section is the high body emissive power. The governing factor in heat transfer
temperature section of the furnace, the final passes of the in the radiation section is temperature because the radiant
heating process are located there. heat transfer coefficient is directly proportional to the fourth
power of the temperature.
Radiation is the dominant heat transfer mechanism in the
radiant section. Both participating media (gaseous) radiation Convection also contributes to heat transfer in the radiant
and surface exchange are significant. It should be noted that section. The furnace gases circulate vigorously inside the
even though gas flames emit only a little in the visible radiant section driven by the in-flow of combustion air, gas
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

104 The John Zink Combustion Handbook

In the average heater, about 80% of the total heat is trans-

ferred in the radiant section and approximately 20% comes
10 Load (kg/hr) from the convection section. With this ratio of heat transfer
in the heater sections, it is obvious that the greatest benefit
Gas 700
will result from improvements in the radiant section.
Oil 764
Detailed analysis of furnace heat transfer is complex.
There exist well-defined methods to calculate the heat trans-
8 ferred in the many varieties of heat exchangers (e.g., parallel,
counterflow, shell-and-tube, and compact), but furnace heat
transfer calculation methods are less well-documented.
Many furnace vendors do have proprietary methods for com-
puting the heat transferred to the process load, but these are
6 largely based on empirical data and are not documented in
the open literature. Furthermore, furnace heat transfer seems
Meters

to be the source of controversy and disagreement among

engineering professionals in the field. Much disagreement
exists as to the relative importance of gas radiation and
4
surface radiation, for example.
The function of the burner is to deliver heat to the process
load as uniformly as possible. Reaching this ideal condition
is impossible. The burner equipment used must possess supe-
2 rior ability to disperse heat to the gaseous furnace atmosphere
if the heater operation is to be satisfactory. It is the relative
ability to disperse heat by a particular burner that may make
it suited to firing in a particular heater. If the burner is applied
to a different furnace design, it may not be as effective.67
0 Hoogendoorn et al. made heat flux measurements in a
0 50 100 150 rectangular, vertical tube furnace with two round burners
firing vertically upward.65 Both oil and gas flames were
103 kcal/m2h
tested. The objective of the study was to determine the valid-
FIGURE 3.27 Vertical heat flux distribution for oil and ity of the assumption of a constant furnace temperature often
gas firing in a vertical tube furnace.66 used to calculate the heat flux to process tubes. A 1 in.
(25 mm) diameter, water-cooled heat flux probe, with and
without air screens, was used to measure radiation and total
expansion due to combustion, and the temperature gradients heat flux, respectively. Forced convection was calculated
through the furnace. These flow patterns are very difficult to from the difference between the total and radiant heat flux
predict a priori, but they are very important in assessing heat measurements. The heat flux was found to be significantly
transfer to the process tubes. In the convection section of the nonuniform in the furnace. Gas flames were found to have
furnace, most of the heat is transferred by convection. The a more uniform heat flux distribution than oil flames, as
shown in Figure 3.27.
first row or two of tubes in the convection section that have
Selçuk et al. studied the effect of flame length on the
a “line of sight” view of the radiant section experience a lot
radiative heat flux distribution in a process fluid heater.66 The
of radiant transfer. Consequently, the first row of tubes is
radiative heat flux distribution information is important in
usually not equipped with fins, to avoid excessive localized
the design of the heater. It helps to prevent premature damage
heat flux. The subsequent rows of tubes in the convection to the process tubes due to improper flame heights and helps
section are often finned to maximize the convective heat optimize the heat transfer rate to the tubes, thereby maxi-
transfer per unit tube length. The governing factor in heat mizing thermal efficiency. A two-flux radiation model was
transfer in the convection section is mass velocity, since the used to predict the radiant heat transfer in the heater. The
convective heat transfer coefficient is strongly dependent on predictions were in good agreement with the experimental
the velocity. data. The results showed that the radiant flux at the tube
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Heat Transfer 105

surface was a strong function of the flame height, as shown 0.10

in Figure 3.28.

3.6.1 Flame Radiation 0.09

The radiant sources in the furnace are the flame, the radiant
furnace surfaces, and the radiant furnace gases (H2O and 0.08
CO2). Gas flames produce some radiation directly from the Lf = 0
flame. How much depends on how luminous the flame is. An
0.07
oil flame can radiate three to four times as much as a gas
flame due to the high quantities of soot formed in the flame
that make it luminiscent. A gas flame also may produce soot 0.06
under certain conditions of mixing and can radiate relatively
more, but not as much as an oil flame. The distance from the 0.05

(qt)/Lad
flame to the heat transfer surface at various locations in the Lf = 0.24L
furnace varies widely. Since radiant transfer varies inversely
as the square of the distance between the radiant and the 0.04
absorptive bodies, flame radiation will not be uniformly
Lf = 048L
delivered to all portions of the furnace. The heightened 0.03
requirement of uniformity of heat flux in ethylene cracking
Lf = 0.72L
furnaces has driven the design of furnaces with multiple
small burners distributed uniformly over the furnace walls. 0.02
Flame radiation does not dominate the radiation process in Lf = L
the furnace. As an illustration, consider a combination burner 0.01
that can provide both a gas and oil flame in the same location
in a furnace. Even though the oil flame is three to four times
0.00
as radiant as the gas flame, the furnace performance and tube
surface temperatures are not significantly changed. Issues
such as these introduce controversy in the various schools of -0.01
thought regarding furnace heat transfer. Experimentally deter- 0.0 0.2 0.4 0.6 0.8 1.0
mining the true radiant behavior in the furnace is very difficult Z/L
and would entail several compromises that would make the
data questionable. FIGURE 3.28 Distribution of dimensionless average
radiant flux density at the tube surfaces for various flame
3.6.2 Furnace Gas Radiation lengths (Lf = flame length, L = heater height, Z = height).67
An analysis of the heat capacities of furnace gases indicates
that gases that radiate in the infrared, carbon dioxide and
Consider a lb-mol of methane as 380 scf. At 910 Btu/scf,
water, only carry about 33% of the total heat released. Con-
the lower heating value of a lb-mol of methane is 345,800
versely, 63% of the heat is contained in other gases, namely,
Btu. If it is presumed that 10% of the heat is radiated directly
oxygen and nitrogen.
by the flame burst, the heat content of the gases is 311,220
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Consider the chemistry of burning methane at 10% excess

Btu. The respective heat contents of the component gases
air:
based on their specific heats are as follows:

CH4 + 2.20 O2 + 8.36 N2 = CO2 + 2H2O + 0.20 O2 + 8.36 N2 CO2 = 37,700

H2O = 67,200 = 104,900 Btu = 33.7%
Gaseous products are: O2 = 4,800
N2 = 201,520 = 206,320 Btu = 63.3%
• 1 mol CO2
• 2 mols H2O The total heat content of the radiating gases, carbon dioxide
• 8.36 mols N2 and water, is 104,900 Btu. The heat content of the nitrogen
• 0.20 mols O2 and oxygen is 206,320 Btu immediately following the radiant
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 106 Monday, March 14, 2005 11:17 AM

106 The John Zink Combustion Handbook

part of the flame. These gases are in a homogeneous mixture At steady state, the refractory reaches and maintains a
in which a portion of the gases is radiant-capable and a portion thermal equilibrium. The amount of energy reaching the
is not. The heat energy of a portion of the gases is being refractory is either re-radiated back or lost to the surroundings
dissipated by radiation to produce a steady decrease in heat through conduction to the outside of the furnace. Analysis of
content within these gases. The other portion that does not the possible radiant heat transfer components that could trans-
radiate then transfers its heat to the radiating gases as their fer energy to the refractory do not account for all the heat
temperature decreases. that is in fact reaching the refractory. The difference, a sig-
However, the emissivity of the radiating gases is quite low. nificant amount, is therefore coming from convection. As the
The quantity of heat radiated is a relatively small portion of hot gases sweep down the walls of the furnace, they heat the
the total heat content of the gases, so a significant amount of walls by a combination of radiation and convection. To
the heat transfer occurs when the gases come into close prox- achieve these rates of convection, the gas velocity in the
imity of the heat transfer tubes. First, the contact of the gases proximity of the walls has to be quite high. Reed stated that
with the surface of the tubes and refractory walls transfers gas velocities could reach 50 ft/sec in the vicinity of the walls.
heat by convection. Second, the close proximity of the gases It is common to think of only the refractory areas that glow
to these surfaces makes the radiation transfer higher since as being radiant. In reality, all the hot surfaces, whether they
distance is minimized. Thus, the heat transfer in this combi- are visibly elevated in temperature or not, radiate. The visibly
nation mode depends on the vigorous furnace hot gas currents. glowing surfaces are, of course, radiating more than the
darker surfaces. Generally speaking, refractory surfaces pos-
3.6.3 Refractory Surface Radiation sess high emissivity and thus readily deliver their heat by
radiation.
The furnace interior refractory walls have a much greater sur-
face area than the surface area of the heat transfer tubes.
Therefore, proportionally greater energy is delivered to the 3.6.4 Analysis of Radiation Heat Transfer
furnace refractory surfaces. Refractory has a very high heat In this and following discussions, flame radiation is approxi-
capacity. The refractory’s ability to store heat exceeds that of mated by treating the flame as a isothermal cylinder of gases.
the gases and tube materials. Thus, initially, a significant These gases can be reasonably assumed to be 17% H2O
portion of the furnace heat up time is due to the heat capacity (by mole) and 8% CO2 (by mole) at 1540°C (2800°F). For
of the refractory. illustration purposes, the mean beam length of the flame is
Refer back to Table 3.3 for the specific heats of some assumed to be 1 m (3.3 ft), however, this is only an approxi-
common materials. One can see that an average refractory mation. Better accuracy requires more information to calcu-
brick has a specific heat of approximately 0.2 Btu/lb-°F, late the mean beam length. The pressure-pathlength for H2O
almost double that of carbon steel. The largest heat storage is then 0.56 atm-ft, and for CO2, the pressure-pathlength is
occurs in the refractory. To illustrate the magnitude, let us 0.264 atm-ft. From Figures 3.21 and 3.22, the emissivity of
consider a furnace that is 30 ft × 30 ft × 40 ft (9 m × 9 m × the water vapor is about 0.1, while the emissivity of the CO2 is
12 m). With six in. (15 cm) of refractory thickness, the furnace about 0.065. The total emissivity (uncorrected) is then 0.165.
now has approximately 200,000 lb (90,000 kg) of refractory. Figure 3.23 indicates a 5% correction to the combined emis-
The quantity of heat stored in the refractory can be estimated sivity, so the corrected flame emissivity is 0.157. At 1540°C
using the formula: (2800°F), the blackbody emissive power is Eb–flame = σT4 =
5.67 × 10–8 (1540 + 273)4 = 610 kW/m2 = 194,000 Btu/ft2.
Q = m × Cp × (T1 – T2) Figure 3.29 is an illustration of a vertical/cylindrical fur-
nace. In this style of furnace, the burners (shown here as a
Assume the refractory heats up from an ambient tempera- single flame) are surrounded by process tubes. The furnace
ture of T2 = 70°F to a steady-state average temperature of shell is just outside the process tubes. The radiative circuit
500°F. The average is the temperature midway in the thick- diagram in the figure shows how radiative heat flows from the
ness of the refractory. The hot surface will be considerably flame to the tubes and refractory walls. To use this diagram,
hotter and the cold surface typically around 200°F. The the emissivity values for the flame, refractory wall, and tube
amount of heat stored in the refractory is: surfaces are required. For illustration purposes, the flame
emissivity determined above (0.157 for a flame temperature
Q = 200,000 lb × 0.2 Btu/lb-°F × (500 – 70)°F of 2800°F [1540°C]), a typical refractory emissivity of 0.65,
Q = 17.2 × 106 Btu and a typical tube surface emissivity (oxidized metal) of 0.85
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Heat Transfer 107

FIGURE 3.29 Radiation heat transfer in a cylindrical furnace.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
will be used. To compute view factors, some dimensions need is not that small). For calculation purposes, an inside surface
to be assumed, such as a furnace diameter of 10 m, a flame refractory temperature of 650°C (1200°F) and a tube surface
diameter of 1 m, and a tube diameter of 20 cm. Then, a temperature of 430°C (800°F) is assumed.
calculation using the formula in Table 3.10 gives the view The circuit diagram shown in Figure 3.29 leads to a system
factor from the flame to a single tube as 0.022. If one assumes of three linear equations. Since the height of the furnace has
that there are 16 tubes in the furnace, then the view factor not been specified, all the results will be per unit height. The
from the flame to the tubes is Fflame-tubes = 16 × 0.022 = 0.352. solution of these equations gives the radiative heat flux from
Assume that the flame radiation that is not incident on the the flame as 284 kW/m (294,000 Btu/hr-ft), the heat flux to
tubes is incident on the refractory, so that Fflame-refractory = 1 – the tubes is 251 kW/m (260,000 Btu/hr-ft), and the heat flux
Fflame-tubes = 0.648. Since there is no view factor catalog entry to the furnace refractory is 33 kW/m (34,000 Btu/hr-ft).
to help in computing the view factor from the tubes to the These results mean, for instance, that the heat flux to the
refractory, this value will be assumed to equal one. This tubes is 260,000 Btu per foot of tube length. If the tubes
neglects the view factor from the tubes to the flame (which were 50 ft long, the total heat flux into the tubes would be
is small) and the fact that the tubes “see each other” (which 13 x 106 Btu/hr.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 108 Monday, March 14, 2005 11:17 AM

108 The John Zink Combustion Handbook

L1 L2 L3
L1 = Furnace Wall
L2 = Second Refractory Layer
L3 = First Refractory Layer
Furnace
Reinforcement

T1
T2

Ambient Radiation
Air

T3
Furnace
T5 T4 Gases
T6

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Furnace
Reinforcement

Qradiation from flame

r1 + L1 + L2 + L3 r1 + L1
ln ln
r1 + L1 + L2 r1
2F k3 2F k1
T6 T5 T4 T3

T2 T1
1 r + L1 + L2 1
ln 1
2F (r1 + L1 + L2 )houtside r1 + L1 2F r1 hinside
2 F k2

FIGURE 3.30 Cross section of furnace wall.

3.6.5 Heat Transfer Through the Wall of the flue gases and flame, respectively. This heat is conducted
through the refractory and eventually convected away by
a Furnace
natural and forced (wind) convection on the outside of the
Figure 3.30 illustrates a typical furnace wall. The outer layer shell. A typical inside heat transfer coefficient is 30 W/m2-K
of the furnace wall is the steel furnace shell. The inner layers (5.3 Btu/hr-ft2-°R). A typical outside heat transfer coeffi-
typically consist of refractory brick and, perhaps, soft refrac- cient is 17 W/m2-K (3 Btu/hr ft2-°R). If the refractory (both
tory blanket. The circuit diagram on the figure indicates how blanket and brick) conductivity is assumed to be 0.5 W/m-K
the heat transfer through the wall can be analyzed. All three and the steel conductivity is 100 W/m-K, then the heat flux
heat transfer mechanisms are indicated. The inside surface is through the wall can be computed. Using the radiative heat
subjected to both convective and radiative heat transfer from flux from the previous analysis as 33 kW/m, assume a
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 109 Monday, March 14, 2005 11:17 AM

Heat Transfer 109

Boundary Layer Boundary Layer

Tube Wall

Furnace
Gases

Process Fluid
Furnace
Gases
Coking

FIGURE 3.31 Cross section of process heat transfer tube.

refractory surface temperature of 650°C (1200°F) and a flue tube exterior surfaces can be readily added to get a more
gas temperature of 675°C (1250°F). Assume that the blanket physically realistic calculation.
thickness (L1) is 5 cm (2 in.),the brick thickness (L2) is 5 cm
(2 in.) and the steel thickness is 1.25 cm (0.5 in.); then the 3.6.7 Furnace Gas Flow Patterns
circuit analysis gives the total heat flux through the furnace As previously noted, in low temperature furnaces the tubes
wall as 56kW/m (58 Btu/hr ft), and the outer skin tempera- are usually located very close to the walls. Again, since the
ture is 92°C (200°F). radiation is inversely proportional to the square of the dis-

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
tance, the radiation from the walls to the tube is significant
3.6.6 Heat Transfer in the Process Tube when the tubes are close to the walls.
Figure 3.31 shows a cross-sectional view of a process fluid In the case of radiant wall furnaces used in ethylene crack-
flowing through a tube. Radiant heat is incident on the outer ing operations, the wall is directly heated by the flame in
surface of the tube, along with convection heat transfer from order to capitalize on the high heat capacity of the refractory.
the furnace gases. This heat is conducted through the wall of The quantity of ethylene produced is maximized when the
the tube. Any coking or scaling on the inside or outside sur- heat is applied evenly to the entire length of the tube. The
face of the tube will add to the heat transfer resistance, ideal way to accomplish this is to heat the wall and allow it
which will subsequently increase the outside surface tem- to radiate to the tubes. The high heat capacity of the refractory
perature. Heat transfer into the process fluid can be analyzed acts as a huge capacitance that helps to smooth out peaks in
using the formulae given earlier in this chapter. The circuit the temperature profile.
analysis shown in the previous two examples can also be In the low temperature furnaces, the heating of the refrac-
applied to this example. Additionally, the effects of extra tory walls is caused by the hot gases sweeping down the
heat transfer resistance due to coke or scale buildup on the walls. The flow of gases between the tubes and the walls is
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 110 Monday, March 14, 2005 11:17 AM

110 The John Zink Combustion Handbook

important, as the example below will illustrate. The flames was put back into service, it had regained its original heat
heat the gases and buoyancy causes them to rise. As the gases transfer capability. No change other than the refractory repair
close to the tubes deliver heat to the tubes and walls, they had been implemented. Because of this, it was surprisingly
cool down, become denser, and flow down toward the bottom evident that the increased space between the tubes and the
of the furnace. This establishes a circulation pattern within wall accounted for increased heat absorbing ability.
the furnace, such that the gases rise up from the flames and
reverse direction higher up in the furnace and flow down the 3.6.7.1 Tube-to-Wall Spacing
wall and tubes to the furnace floor. Upon reaching the floor, Reed conducted experiments at a reduced scale to evaluate the
the gases are reheated to either make another circuit or to influence of tube to wall spacing.68 A test heater was constructed
exit from the furnace en route to the stack. The benefit of using tubes with an outer diameter of 1/4 in. (6 mm). The test
recirculation is optimum when the tubes are on typical two- furnace dimensions were 18 × l8 × 27 in. (46 × 46 × 69 cm).
diameter centers, and decreases as the center-to-center dis- There was a provision to accurately adjust the tube-to-wall
tance is reduced to less than two diameters. spacing and the relationship of the burner to the tubes. Accu-
Gases in such recirculation flow pass over the entire tube rately metered, saturated air was passed through the 3/16-in.
areas as well as the wall behind the tubes, scrubbing the tube (4.8 mm) ID tubes as the source of heat absorption. A ther-
surfaces for heat transfer by convection. Far more importantly, mocouple was used in the air stream at the exit from the
they also scrub the refractory wall behind the tubes to contin- tubes to measure the temperature of the exiting air as well as
ually deliver heat to the wall surface. Whether the tubes are the heat absorbed. The firing rate and excess air, as well as
horizontal or vertical does not seem to make much difference. furnace temperature, were closely controlled to identical
There are several methods to visualize furnace gas flow conditions for all tests.
patterns. Today, CFD is the preferred engineering tool to study Reed reported that for a tube spaced one-half diameter off
furnace flue gas patterns. Chapter 9 provides several examples the wall, the heat transfer to the tube is increased 13% over
of such studies. On the other hand, in the furnace one may the condition where the tube is tangent to the wall. If the tube
use various powdery substances such as baking soda, partic- is spaced one diameter off the wall, the heat transfer to the
ulate carbon, etc. to observe furnace flow patterns. The powder tube is increased approximately 29% over the condition
is usually introduced in the air stream to the burners and is where the tube is tangent to the wall. Further increase of the
seen to glow briefly in the furnace. The glowing particles trace tube-to-wall spacing to as much as three to four diameters
the flow patterns. This is a useful but approximate technique provided no increase in heat transfer. Still greater spacing
because the persistence of the glowing is short, and if it is too actually created a decrease in heat transfer.
short, it may adversely bias the conclusions being drawn.
To verify that the gas flow behind the tubes was the con-
Reed67 mentions an incident to illustrate the importance of
tributor to the enhanced heat transfer at a tube-to-wall dis-
furnace gas flow between the tubes and walls. A heater had
tance of one tube diameter, an additional experiment was
operated satisfactorily for years. As operation progressed, it
conducted. In this experiment, the tubes were spaced one
was noted that the heater was rapidly losing heat transfer
diameter off the wall to produce the 29% increase in heat
ability, despite the fact that there had been no change in
transfer. Strips of mica 0.003 in. (0.08 mm) thick were placed
operation which might account for the decrease. There was
in the space between the tubes and the wall at the centerlines
no change in pressure drop, so the possibility of coke lay-
of the tubes to block the space between the tubes and the wall
down was rejected. The deterioration in performance came
with material that is substantially transparent to infrared. The
about in less than six months and was noticeable on a day-to-
purpose was to avoid blocking radiant transfer while com-
day basis as the heater operated.
pletely blocking the flow path for gases in the space between
The heater had been in service for many years and was due
the tubes and the wall behind them. With the mica strips, the
for repairs including replacing the refractory side walls which
heat transfer to the tubes was exactly the same as was
were sagging inwards. The walls, supported independently of
observed with the tubes tangent to the wall. In other words,
the steel that supported the tubes, were gradually moving
the 29% performance gain was lost due to the blockage.
toward the tubes. The space between the walls and the back
sides of the tubes was reduced to such a degree that most of
the side walls were actually resting against the tubes. 3.6.8 Role of the Burner in Heat Transfer
The heater was shut down and the side walls were repaired Simple release of an adequate amount of heat to the furnace
to make the space between the tubes and the wall one full atmosphere is not the only objective for the burner. Proper
tube diameter, which in this case was four in. When the heater
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
choice of burners is critical to the performance of the heater.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 111 Monday, March 14, 2005 11:17 AM

Heat Transfer 111

There are no hard and fast rules to govern the choice of burning excessive heat to any local heat transfer area. Local
ers. The design of the furnace and the burner must be matched overheating and flame impingement must be avoided at all
carefully to achieve good overall performance. There is no costs. Flame impingement does not occur solely due to
single burner design that can be universally applied. burner performance. The burner has only limited control over
The function of the burner equipment is to deliver heat to the characteristic flow patterns of a furnace. It is possible to
the gas content of the furnace as uniformly as possible. modify a burner to eliminate flame impingement by changing
Reaching this ideal condition would require an infinitely the fuel jet configurations, but there is only a narrow window
large number of small burners. Ethylene cracking and hydro- of opportunity here, because radical modifications will
gen reforming furnaces most closely approximate this ideal require compromises in other areas of burner performance
arrangement by using many small burners. Over the last such as capacity or emissions.
100 years, the quest for better heat transfer has resulted in a Flame length is of utmost importance in burner design,
myriad of furnace designs. It follows that many burner designs although flame length and heat dispersion are not necessarily
were developed to fit the various furnace designs. In the past, in a fixed relationship. Providing short flames exclusively for
when emissions were not regulated, the primary requirement all applications is not the answer either, because some appli-
for the burner was effective heat transfer. Typically, the old cations require long flames to reach further into large furnaces
burner designs rapidly mixed the fuel and air, resulting in or to deliver heat to locations further away from the burner.
short flames. Emissions regulations have now driven the For example, in the typical floor-fired steam reformer, it is
design of burners for the last two or three decades. The pri- necessary to drive the hot gases from the furnace floor to the
mary requirement is now to meet the emissions regulations top of the furnace to distribute heat to the tube areas where
without compromising furnace performance. This conflicting maximum heat density is demanded. Nowadays, most steam
challenge has been met with considerable engineering inge- reforming furnaces are down-fired for this reason. The con-
nuity over the years. Low-NOx burners designed in the last trary is true in side-wall fired steam reformers.
two decades tend to have longer flames because the strategy In radiant wall firing, the burners are located in areas where
for NOx reduction was to delay mixing and thereby reduce maximum heat transfer is demanded. The flame is expected
peak flame temperatures. to remain close to the wall and not penetrate forward into the
Over the years, furnace manufacturers as well as burner furnace at all. This is because the furnaces are narrow and
manufacturers have researched heat flux profiles in various the burners are located quite close to the tubes which may be
burner–furnace combinations. Again, CFD is a great help in either vertically or horizontally suspended at the center of the
studying heat flux profiles, but even with today’s sophisticated furnace. The “terrace wall” furnace design for the same appli-
modeling capabilities and advanced instrumentation, exact cation is a prime example of the differentiation between heat
measurements are not possible. Exact measurements are dif- dispersion and flame length. The burners are mounted in
ficult to obtain because apart from the obvious problem of terraces on the side wall in much the same way as a floor-
working in a high temperature zone, the geometry of the mounted burner, and the flame is fired vertically up the wall.
furnace and the re-radiation from various furnace surfaces In this design, the flame is considerably longer than the small
make the analysis complicated. With wall fired burners, either wall burners, yet the service performed is identical.
floor mounted or wall mounted, it is somewhat easier to In a typical process heater, the demand for very precise
predict heat flux patterns. However, in furnaces that have free control of heat density per linear foot of tube is not as great
standing flames, the heat flux patterns tend to be specific to as in a steam-reformer furnace. It is possible to use a smaller
that burner–furnace combination. number of larger burners to obtain satisfactory firing condi-
If a desired heat flux pattern is identified, it is possible to tions and heat dispersion, but the burners must be suited to
engineer the flame shape to attempt to meet the requirement. the service. Sometimes, burners capable of reasonably short
Previous experience can help define the burner design flames have not had satisfactory heat dispersion characteris-
required. On the other hand, the outcome can only be esti- tics and must be replaced to reduce tube damage.
mated if there is no previous experience with that particular Some designs have stand-alone flame burners that are
furnace–burner combination. In such cases, some final testing mounted on the side walls of the furnace with the flame fired
and adjustment is usually required. Consequently, burners straight into the furnace space. This is not the same as the
have been developed over the years with flames of every ethylene cracking or reforming radiant wall arrangements.
reasonably conceivable shape. Those are lower temperature furnaces where the design deci-
Burner flames must be shaped and directed to allow the sion has been made to mount a stand-alone flame burner on
required heat diffusion to the furnace gases without deliver- the side wall instead of the floor to reduce the initial cost of
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

112 The John Zink Combustion Handbook

the furnace. Side wall mounting costs less since the furnace REFERENCES
does not have to be elevated to install burners below it, and,
often, less burners are required. However, floor firing has 1. B. Gebhart, Heat Transfer, 2nd edition, McGraw-Hill,
some advantages over side wall firing. With floor-mounted New York, 1971.
burners the heater can typically be fired 25% harder. This is 2. F. Kreith and M.S. Bohn, Principles of Heat Transfer,
because floor mounting makes better use of the combustion Harper & Row, New York, 1986.
volume and provides more uniform heat distribution. 3. J.P. Holman, Heat Transfer, 7th edition, McGraw-Hill,
Either way, there will be greater service from the heater New York, 1990.
when a relatively large number of small burners are used 4. A. Bejan, Heat Transfer, John Wiley & Sons, New
rather than a small number of large burners. If there is a York, 1993.
relatively small number of large burners, there is a greater
5. F.P. Incropera and D.P. Dewitt, Introduction to Heat
mass of gas issuing from each burner and a greater concen-
Transfer, 3rd edition, John Wiley & Sons, New York,
tration of heat before the burner. This larger mass of gases
1996.
and quantity of heat must then be dispersed evenly to the
furnace atmosphere for good performance. It is far easier to 6. A.F. Mills, Heat Transfer, 2nd edition, Prentice-Hall,
disperse smaller amounts of gas and heat as issued by several Englewood Cliffs, NJ, 1998.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

smaller burners. 7. W.X. Janna, Engineering Heat Transfer, 2nd edition,

To conclude, in general, the heat transfer role of the burner CRC Press, Boca Raton, FL, 2000.
in a furnace is to provide the required amount of heat with 8. H.C. Hottel and A.F. Sarofim, Radiative Transfer,
appropriate flame dimensions without localized hot spots or McGraw-Hill, New York, 1967.
flame impingement. The sizing and selection of the burner 9. T.J. Love, Radiative Heat Transfer, Merrill Publishing,
must help make the temperature of the bulk of the furnace Columbus, OH, 1968.
gases as uniform as possible in as short a distance as possible
10. M. Özisik, Radiative Transfer and Interactions with
from the burner.
Conduction and Convection, John Wiley & Sons, New
York, 1973.
11. W.A. Gray and R. Müller, Engineering Calculations in
3.7 CONCLUSIONS Radiative Heat Transfer, Pergamon, Oxford, U.K., 1974.
This chapter presented the fundamentals of heat transfer. 12. E.M. Sparrow and R.D. Cess, Radiation Heat Transfer,
Basic relations for conduction, convection, and radiation heat augmented edition, Hemisphere, Washington, D.C., 1978.
transfer were provided and discussed. This chapter, by neces-
13. J.B. Dwyer, Furnace calculations, in Process Heat
sity, is a very brief and dense presentation of the subject of
Transfer, D.Q. Kern, Ed., McGraw-Hill, New York, 1950.
heat transfer. The interested reader is strongly encouraged to
consult with heat transfer texts (see references) for more 14. F.W. Hutchinson, Industrial Heat Transfer, Industrial
detailed information or explanation. The subject of heat trans- Press, New York, 1952.
fer is a vast and interesting one. The focus of this chapter is on 15. S.T. Hsu, Engineering Heat Transfer, D. Van Nostrand
heat transfer in combustion systems, but there are a multitude Co., Princeton, NJ, 1963.
of other heat transfer applications to which these basic princi- 16. J.R. Welty, Engineering Heat Transfer, John Wiley &
ples can be applied. The authors hope that this brief introduc- Sons, New York, 1974.
tion has given the reader an interest in pursuing the subject
17. B.V. Karlekar and R.M. Desmond, Engineering Heat
further.
Transfer, West Pub. Co., St. Paul, MN, 1977.
The difficulties that may arise when trying to apply heat
18. V. Ganapathy, Applied Heat Transfer, PennWell Books,
transfer relations to furnace heat transfer problems were con-
Tulsa, OK, 1982.
sidered here. Some of the approximations that can be made
to complete such an analysis were considered. Again, the 19. A.G. Blokh, Heat Transfer in Steam Boiler Furnaces,
reader interested in more information should consult the ref- Hemisphere, Washington, D.C., 1988.
erences given at the end of this chapter. In particular, the 20. W.M. Rohsenow, J.P. Hartnett, and E.N. Ganic, Hand-
recent book by Baukal 64 provides a thorough overview of book of Heat Transfer Applications, McGraw-Hill,
combustion heat transfer and cites many references. New York, 1985.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 113 Monday, March 14, 2005 11:17 AM

Heat Transfer 113

21. N.P. Cheremisinoff, Ed., Handbook of Heat and Mass 40. S. Kakac and Y. Yener, Convective Heat Transfer, 2nd
Transfer, four volumes, Gulf Publishing, Houston, TX, edition, CRC Press, Boca Raton, FL, 1995.
1986. 41. P.H. Oosthuizen, An Introduction to Convective Heat
22. F. Kreith, Ed., The CRC Handbook of Thermal Engi- Transfer, McGraw-Hill, New York, 1999.
neering, CRC Press, Boca Raton, FL, 2000. 42. F.W. Dittus and L.M.K. Boelter, University of California,
23. V.S. Arpaci, Conduction Heat Transfer, Addison-Wes- Berkeley, Publications on Engineering, 2, 443, 1930.
ley, Reading, MA, 1966. 43. E.N. Sieder and G.E. Tate, Ind. Eng. Chem., 28, 1429,
24. M.N. Özisik, Boundary Value Problems of Heat Con- 1936.
duction, Dover, New York, 1968. 44. B.S. Petukhov, Advances in Heat Transfer, Vol. 6,
T.F. Irvine and J.P. Hartnett, Eds., Academic Press,
25. U. Grigull and H. Sandner, Heat Conduction, Hemi-
New York, 1970.
sphere, Washington, D.C., 1984.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
45. L.F. Moody, Trans. ASME, 66, 671, 1944.
26. G.E. Myers, Analytical Methods in Conduction Heat
Transfer, Genium Publishing, Schenectady, NY, 1987. 46. E.S. Skupinski, J. Tortel, and L. Vautrey, Int. J. Heat
Mass Transfer, 8, 937, 1965.
27. B. Gebhart, Heat Transfer and Mass Diffusion,
McGraw-Hill, New York, 1993. 47. R.A. Seban and T.T. Shimazaki, Trans. ASME, 73, 803,
1951.
28. D. Poulikakos, Conduction Heat Transfer, Prentice-Hall,
48. C.B. Reed, Handbook of Single-Phase Convective Heat
Englewood Cliffs, NJ, 1994.
Transfer, S. Kakac, R.K. Shah, and W. Aung, Eds.,
29. F.P. Incropera and D.P. DeWitt, Fundamentals of Heat Wiley Interscience, New York, 1987, chap. 8.
and Mass Transfer, 4th ed., John Wiley & Sons, New
49. H. Schlichting, Boundary Layer Theory, 6th edition,
York, 1996.
McGraw-Hill, New York, 1968.
30. P.J. Schneider, Conduction Heat Transfer, Addison- 50. R. Hilpert, Forsch. Geb. Ingenieurwes., 4, 215, 1933.
Wesley, Reading, MA, 1955.
51. J.D. Knudsen and D.L. Katz, Fluid Dynamics and Heat
31. H.S. Carslaw and J.C. Jaeger, Conduction of Heat in Transfer, McGraw-Hill, New York, 1958.
Solids, 2nd edition, Oxford University Press, London,
52. S.W. Churchill and M. Bernstein, J. Heat Transfer, 99,
1959.
300, 1977.
32. A. Garg, How to boost the performance of fired heaters,
53. E.D. Grimison, Trans. ASME, 59, 583, 1937.
Chem. Eng., 96(11), 239-244, 1989.
54. A. Zhukauskas, Heat transfer from tubes in cross flow,
33. V.S. Arpaci, Convection Heat Transfer, Prentice-Hall, in Advances in Heat Transfer, Vol. 8, J.P. Hartnett and
Englewood Cliffs, NJ, 1984. T.F. Irvine, Jr., Eds., Academic Press, New York, 1972.
34. C.S. Fang, Convective Heat Transfer, Gulf Publishing, 55. W.A. Gray and R. Müller, Engineering Calculations in
Houston, TX, 1985. Radiative Heat Transfer, Pergamon, Oxford, U.K., 1974.
35. S. Kakac, R.K. Shah, and W. Aung, Eds., Handbook of 56. J.A. Wiebelt, Engineering Radiation Heat Transfer,
Single-Phase Convective Heat Transfer, John Wiley & Holt, Rinehart and Winston, New York, 1966.
Sons, New York, 1987.
57. E.M. Sparrow and R.D. Cess, Radiation Heat Transfer,
36. L.C. Burmeister, Convective Heat Transfer, 2nd edition, Augmented Edition, Hemisphere, Washington, D.C., 1978.
John Wiley & Sons, New York, 1993. 58. D.K. Edwards, Radiation Heat Transfer Notes, Hemi-
37. W.M. Kays and M.E. Crawford, Convective Heat and sphere, Washington, D.C., 1981.
Mass Transfer, 3rd edition, McGraw-Hill, New York, 59. R. Siegel and J.R. Howell, Thermal Radiation Heat
1993. Transfer, 2nd edition, McGraw-Hill, New York, 1981.
38. A. Bejan, Convection Heat Transfer, 2nd edition, John 60. M.Q. Brewster, Thermal Radiative Transfer and Prop-
Wiley & Sons, New York, 1994 erties, John Wiley & Sons, New York, 1992.
39. M. Kaviany, Principles of Convective Heat Transfer, 61. M. Planck, The Theory of Heat Radiation, Dover, New
Springer-Verlag, New York, 1994. York, 1959.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch03-Frame Page 114 Monday, March 14, 2005 11:17 AM

114 The John Zink Combustion Handbook

62. H.C. Hottel, Radiant heat transmission, in Heat Trans- k Thermal conductivity, w/m·K
mission, 3rd ed., W.H. McAdams, Ed., McGraw-Hill, Boltzmann’s constant
New York, 1954. L Characteristic length, m
Nu Nusselt number
63. M.F. Modest, Radiative Heat Transfer, McGraw-Hill,
P Perimeter, m
New York, 1993.
Pe Peclet number (Re·Pr)
64. C.E. Baukal, Heat Transfer in Industrial Combustion, Pr Prandtl number
CRC Press, Boca Raton, FL, 2000. p Pressure, N/m2
65. C.J. Hoogendoorn, C.M. Ballintijn, and W.R. Dorresteijn, Q Energy transfer, J
Heat-flux studies in vertical tube furnaces, J. Inst. Fuel, q Heat transfer rate, W
43, 511-516, 1970. q̇ Rate of energy generation per unit volume, W/m3
q″ Heat flux, W/m2
66. N. Selçuk, R.G. Siddall, and J.M. Beér, Prediction of R Cylinder radius, m
the effect of flame length on temperature and radiative Re Reynolds number
heat flux distributions in a process fluid heater, J. Inst. Rt Thermal resistance, K/W
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Fuel, 43, 89-96, 1975. Rt,c Thermal contact resistance, K/W

67. R.D. Reed, Furnace Operations, 3rd edition, Gulf Pub- ro Cylinder or sphere radium, m
lishing, Houston, TX, 1981. r, ϕ, z Cylindrical coordinates
r, θ, Φ Spherical coordinates
St Stanton number
SD, ST Diagonal and transverse pitch of a Tube bank, m
NOMENCLATURE T Temperature, K
Symbol Description t Time, s
U Overall heat transfer coefficient, W/m2·K
A Area, m2 V Volume, m3
Ae Cross sectional area, m2 Fluid velocity, m/s
Bi Biot number x, y, z Rectangular coordinates, m
Ct Thermal capacitance, J/K
xc Critical location for transition to turbulence, m
c Speed of light, m/s
xfd, h Hydrodynamic entry length, m
cf Friction coefficient
xfd, t Thermal entry length, m
cp Specific heat at constant pressure, J/kg·K
D Diameter, m Greek Letters
Dh Hydraulic diameter, m
E Thermal (internal) energy, J; emissive power, W/m2 α Thermal diffusivity, m2/s
Eg Rate of energy generation, W δ Hydrodynamic boundary layer thickness, m
F Fraction of blackbody radiation in a wavelength band; δt Thermal boundary layer thickness, m
view factor ε Emissivity
Fo Fourier number θ Temperature difference, K
f Friction factor λ Wavelength, µm
G Irradiation, W/m2 µ Viscosity, kg/s·m
g Gravitational acceleration, m/s2 ρ Mass density, kg/m3
h Convection heat transfer coefficient, w/m2·K; Reflectivity
Planck’s constant σ Stefan–Boltzmann’s constant
J Radiosity, W/m2 ω Solid angle, sr

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 4
Fundamentals of Fluid Dynamics
Lawrence D. Berg, Wes Bussman, and Michael Henneke

TABLE OF CONTENTS

4.1 Introduction............................................................................................................................................. 118

4.2 Fluid Properties ....................................................................................................................................... 118
4.2.1 Density ...................................................................................................................................... 118
4.2.2 Viscosity.................................................................................................................................... 119
4.2.3 Specific Heat ............................................................................................................................. 125
4.2.4 Equations of State ..................................................................................................................... 126
4.3 Fundamental Concepts............................................................................................................................ 128
4.3.1 Hydrostatics .............................................................................................................................. 128
4.3.2 Bernoulli Equation .................................................................................................................... 130
4.3.3 Control Volumes........................................................................................................................ 132
4.3.4 Differential Formulation (Navier-Stokes Equations)................................................................ 134
4.4 Different Types of Flow .......................................................................................................................... 138
4.4.1 Turbulent and Laminar Flow..................................................................................................... 138
4.4.2 Compressible Flow ................................................................................................................... 144
4.5 Pressure Drop Fundamentals .................................................................................................................. 148
4.5.1 Basic Pressure Concepts ........................................................................................................... 148
4.5.2 Roughness ................................................................................................................................. 149
4.5.3 Loss Coefficient ........................................................................................................................ 151
4.5.4 Discharge Coefficient................................................................................................................ 152
References ................................................................................................................................................................ 153

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
117
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 118 Wednesday, April 6, 2005 3:41 PM

118 The John Zink Combustion Handbook

FIGURE 4.1 Variation in measured density with length scale.

4.1 INTRODUCTION defined at every point within a continuum. Because density

The study of fluid dynamics likely dates back to ancient can vary from point to point within a flowing fluid, the ques-
times when early man hunted to keep himself alive. It is easy tion of how density is defined is not trivial. To measure the
to imagine that he quickly discovered he could throw a density of a fluid at a given point, a sample volume of the
streamlined spear much further than a blunt rock. Perhaps he fluid is taken at the given point and the mass of the sample
also noticed that his horse ran a little faster if he ducked his determined. For example, if a sample of one cubic centimeter
head into the wake of the horse’s neck. Through the course of (1 cm3) of the fluid is taken and the mass of the sample is
scientific development, observations such as these, and determined to be x grams, then the density of the fluid at that
countless others, have become a rigorous science. location would be x grams per cubic centimeter (x g/cm3).
The subject of this chapter is the science of fluid dynamics. The problem with the previous example is the possibility
There are many good textbooks on the subject; Panton,1 that the measured value of density might depend on the cho-
White,2 Fox and McDonald,3 and Vennard and Street 4 are sen volume. If the chosen volume is too large, then density
examples. The subject is so broad that it has a number of variations due to other fluid mechanical variables, such as
subfields, for example, turbulence, acoustics, and aerodynam- temperature and pressure, could cause the density to vary
ics. The purpose of this chapter is to give the reader a funda- within the chosen volume.
mental understanding of some of the fluid mechanical concepts Figure 4.1 illustrates the problem. The x-axis is the length
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

that are important in combustion systems, especially those in scale of the volume for measurement. As the length scale gets
the petrochemical industry. long, the measured density may increase or decrease because
of variations within the sample. If the length scale gets too
short, then the sample may contain only several molecules of
4.2 FLUID PROPERTIES the fluid, and the density will be inaccurate because of sam-
Several thermophysical properties are commonly used in fluid pling. The proper definition of density is then:
mechanics analysis. This section presents a brief description of
these properties as they relate to combustion systems. m
ρ ≡ lim (4.1)
L→ 0 V
4.2.1 Density where m is the measured mass of the sample and V is the vol-
Density is defined as the mass per unit volume of a fluid. ume sampled. The limit in Eq. (4.1) is understood to approach
Density is a point function, which means that it must be zero, but to remain much larger than molecular dimensions.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 119 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 119

4.2.2 Viscosity
4.2.2.1 Definition
“Son, you are slower than molasses in January!” is an epithet
that many a parent routinely applies to a child for his slow
movement. While one hopes that age diminishes the truth of
this description, time has not diminished the truth of the
statement — molasses does pour very slowly when it is cold.
FIGURE 4.2 Velocity profile of a fluid flowing along a
This is true for a significant number of other liquids as well: solid surface.
oils, fluids with glycerin, polymeric, and others. At the same
time, many other liquids seem to pour equally well at any
temperature. Water and alcohol are good examples of this. is the viscosity. Conceptually, viscosity then is the difficulty
These qualitative observations of the ease with which various that two fluid “layers” experience as they flow past each
liquids can be poured provide insight into the liquid property other. This is also expressed mathematically as:
of viscosity. Indeed, at a conceptual level, the “pour-ability”
of a liquid can be utilized to understand viscosity. For exam- dV
ple, viscosity describes the rate at which a liquid pours out of τ=µ (4.3)
dY
a container.
In combustion processes, pouring rates are not of direct or
interest. However, due to requirements for maintaining cer-
τ
tain heat releases, specific volumetric flows are required. In µ= (4.4)
dV
this case, viscosity could be descriptive of the power require-
dY
ments to supply a specific volumetric flow through a pipe.
From this description, it follows that as a property, viscosity where τ = Shear stress
is only useful in describing liquids or gases in motion. There- µ = Viscosity
fore, viscosity is the difficulty that various fluids have main- dV
taining motion. = Velocity gradient
dY
While the above discussion is useful in understanding vis-
cosity from a conceptual point of view, in order to be useful 4.2.2.2 Units
in engineering combustion solutions, a more exact definition 4.2.2.2.1 Absolute Viscosity
that allows for quantitative analysis is required. For a fluid The units of stress are (force/area) and the units for the gradi-
flowing in a pipe or along any solid surface, the velocity at ent are (velocity/length). In English units, this is represented
the surface is zero. As distance from the surface increases, the as follows:
velocity also increases, as shown schematically in Figure 4.2.
Assume that each of the arrows in the figure is a discrete  lbf   lbf 
“packet” of liquid. Because they are moving at different  ft 2   ft 2  (lbf )(sec) (lbm)32.17
= = or (4.5)
velocities, as time passes they will move relative to each other.  ft   1  ft 2 (ft )(sec)
This relative motion gives rise to a shear stress between the  sec   sec 
different layers. If there were no velocity difference between ft
the layers, then no shear stress would exist. As the velocity
In a similar manner, metric units for viscosity would, in general,
difference increases, the shear stress increases. The implica-
be:
tion is that shear stress is proportional to the velocity gradient,
which is expressed mathematically as:
(N) ⋅ (sec) kg
or (4.6)
m2 (m) ⋅ (sec)
dV
τ~ (4.2)
dY where
sec = Seconds
where τ is the shear stress, and the derivative is the velocity N = Newtons
difference between different “layers” in the flow. The con- m = Meters
stant of proportionality between shear stress and the velocity
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
kg = Kilograms
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 120 Wednesday, April 6, 2005 3:41 PM

120 The John Zink Combustion Handbook

The metric unit in either form, if multiplied by 10–1, is called 4.2.2.3.1 Temperature Dependence and
a poise. If the unit is multiplied by 10, it is called a centi- Multicomponent Liquids
poise. Most tabular data for viscosity are either in terms of As a property, viscosity arises from molecular interactions at

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
poise, or the “lbm” formulation of English units. Occasion- the atomic level. These interactions tend to be highly influ-
ally, viscosity is tabulated in “lbf” form, or slugs utilized for enced by the fluid temperature and type. For liquids in which
the mass term. Because of the confusing nature of lbm vs. there are strong secondary intermolecular bonds, an increase
lbf, the reader is cautioned to carefully review the units uti- in temperature would tend to weaken these bonds and viscos-
lized in tabular data prior to utilization. All information pro- ity would tend to decrease with higher temperature. For
vided in the appendix is listed in centipoise. gases, intermolecular forces tend to be very weak and viscos-
ity is due to an exchange of momentum between shear layers.
4.2.2.2.2 Kinematic Viscosity In this case, as temperature increases, molecules will migrate
In addition to pure, absolute, or dynamic viscosity, viscosity to other areas of the flow at faster rates. The increased migra-
information is often tabulated as viscosity divided by density. tion results in increased momentum transfer or greater
This is accomplished for convenience, as absolute viscosity viscosity. The graph in Figure 4.3 clearly shows these trends
often appears divided by density in real flow calculations. for several common liquids and gases.
This fluid property is termed the “kinematic” viscosity, and is In addition to a temperature dependence, the hydrocarbon and
expressed mathematically as: petrochemical industries have the additional challenge of deter-
mining the viscosity for multiple constituent liquids. Oils, fuel
µ gases, and natural gases are rarely of a single molecular type, but
ν= (4.7) rather a mixture with properties within certain boundaries. In this
ρ
environment, empirical rules have been developed to provide
reasonable viscosity estimates for given bulk properties and tem-
where
perature. Figure 4.4 provides a graphical temperature dependence
ν = Kinematic viscosity (m2/sec) for hydrocarbon gases based on molecular weight. The following
µ = Absolute or dynamic viscosity equation6 provides a method for calculating the viscosity of gas
ρ = Fluid density mixtures if the individual components are known:

Meters squared per second are the dimensions for kine-

matic viscosity in metric units. When the m2/sec value is
∑ (Viscosity )( N ) i i Mi
Viscosity Mixture = i
(4.8)
divided by 10–4, this property is termed “stokes.” Most tabular
data is available in terms of centistokes — the m2/sec value
∑ (N ) M i
i i

divided by 10–6. Kinematic viscosity is most useful when

analyzing systems involving liquids in which the density is where
unlikely to change significantly with ambient conditions and N = Mole fraction of component i
temperature. In the case of gases, compression, atmospheric M = Molecular weight of component i
pressure change due to elevation, and fluid temperature
changes will affect the density, and thus the kinematic vis- Liquid mixtures tend to be less predictable. Crude oils and
cosity. For these reasons, absolute viscosity should be utilized heavy fuel oils, in particular, are mostly dependent on cut
for any gas analysis. temperatures and origination of the crude. A typical chart
(Figure 4.5) of viscosity vs. temperature for various oils is
4.2.2.3 Other Units provided to illustrate the point. Normally, it is recommended
In addition to the units discussed above, there are a number that information provided by the supplier be utilized to deter-
of historical units employed in the hydrocarbon and petro- mine viscosity for mixtures of liquids; however, Lederer 7 has
chemical industries. These units are a consequence of various given a correlation to determine the viscosity of a mixture of
historic viscosity measurement techniques, especially for crudes and solvents, as follows:
crude oil and various cuts. Reed’s table 5 compares most of
the various units and provides a method to convert from each ln µ m = χ A ln µ A + χ B ln µ B (4.9)
of the units to a modern standard. This table, along with con-
versions, is provided in Table 4.1. and
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 121 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 121

TABLE 4.1 Viscosity Conversion Table

Seconds Seconds Seconds
Viscosity Saybolt Saybolt Seconds Redwood Degrees Degrees
Centipoises Universal Fruol Redwood Admirality Engler Barbey
1.00 31 29 1.00 6200
2.56 35 32.1 1.16 2420
4.30 40 36.2 5.10 1.31 1440
5.90 45 40.3 5.52 1.46 1050
7.40 50 44.3 5.83 1.58 838
8.83 55 48.5 6.35 1.73 702
10.20 60 52.3 6.77 1.88 618
11.53 65 56.7 7.17 2.03 538
12.83 70 12.95 60.9 7.60 2.17 483
14.10 75 13.33 65.0 8.00 2.31 440
15.35 80 13.70 69.2 8.44 2.45 404
16.58 85 14.10 73.3 8.86 2.59 374
17.80 90 14.44 77.6 9.30 2.73 348
19.00 95 14.85 81.5 9.70 2.88 326
20.20 100 15.24 85.6 10.12 3.02 307
31.80 150 19.3 128 14.48 4.48 195
43.10 200 23.5 170 18.90 5.92 144
54.30 250 28.0 212 23.45 7.35 114
65.40 300 32.5 254 28.0 8.79 95
76.50 350 35.1 296 32.5 10.25 81
87.60 400 41.9 338 37.1 11.70 70.8
98.60 450 46.8 381 41.7 13.15 62.9
110 500 51.6 423 46.2 14.60 56.4
121 550 56.6 465 50.8 16.05 51.3
132 600 61.4 508 55.4 17.50 47.0
143 650 66.2 550 60.1 19.00 43.4
154 700 71.1 592 64.6 20.45 40.3
165 750 76.0 635 69.2 21.90 37.6
176 800 81.0 677 73.8 23.35 35.2
187 850 86.0 719 78.4 24.80 33.2
198 900 91.0 762 83.0 26.30 31.3
209 950 95.8 804 87.6 27.70 29.7
220 1000 100.7 846 92.2 29.20 28.2
330 1500 150 1270 138.2 43.80 18.7
440 2000 200 1690 184.2 58.40 14.1
550 2500 250 2120 230 73.00 11.3
660 3000 300 2540 276 87.60 9.4
770 3500 350 2960 322 100.20 8.05
880 4000 400 3380 368 117.00 7.05
990 4500 450 3810 414 131.50 6.26
1100 5000 500 4230 461 146.00 5.64
1210 5500 550 4650 507 160.50 5.13
1320 6000 600 5080 553 175.00 4.70
1430 6500 650 5500 559 190.00 4.34
1540 7000 700 5920 645 204.50 4.03
1650 7500 750 6350 691 219.00 3.76
1760 8000 800 6770 737 233.50 3.52
1870 8500 850 7190 783 248.00 3.32
1980 9000 900 7620 829 263.00 3.13
2090 9500 950 8040 875 277.00 2.97
2200 10000 1000 8460 921 292.00 2.82
Note: The viscosity is often expressed in terms other than centipoise. Formulas for the various viscosimeters are as follows:

Absolute viscosity in cp = 0.261 × T1 – 188/T1 (T1 = Redwood seconds)

Absolute viscosity in cp = 2.396 × T2 – 40.3/T2 (T2 = Redwood Admiralty seconds)
Absolute viscosity in cp = 0.22 × T3 – 180/T3 (T3 = Saybolt universal seconds)
Absolute viscosity in cp = 2.2 × T4 – 203/T4 (T4 = Saybolt Furol seconds)
Absolute viscosity in cp = 0.147 × T5 – 374/T5 (T5 = Degree Engler × 51.3)

Source: R.D. Reed, Furnace Operations, 3rd ed., Gulf Publishing, Houston, TX, 1981.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

122 The John Zink Combustion Handbook

FIGURE 4.3 Absolute viscosity vs. temperature for various fluids.

where ρm, ρA, and ρB, are the densities of the mixture, com-
χ A = (αVA ) (αVA + VB ) (4.10)
ponent A, and component B, respectively. The value of µm
obtained by the correlation compares very well with the
χB = 1 − χA (4.11) experimental data, especially in the high viscosity region.

where α is an empirical constant having a value between 0 4.2.2.4 Kinetic Theory of Gases
and 1.0; A is the more viscous component and B is the less vis-
Gases can be idealized as very small ping-pong balls bounc-
cous component. VA and VB represent the volume composi-
ing around inside an enclosure. By systematically account-
tion for a given mixture. The above equation has been
ing for the distance that a ball must travel prior to hitting
reported to give satisfactory results for other combinations
another ball (mean free path, or λ), the average molecular
and mixtures of petroleum fractions. Shu 8 developed the fol-
speed (c), and the number of balls per unit volume (n), vari-
lowing correlation to obtain the empirical constant in the
ous properties of a fluid can be derived. These properties
above equation:
include viscosity, thermal conductivity, diffusion, and oth-
ers. Typically, these properties are called transport proper-

α=
[
17.04 ( ∆ρ m )
0.5237
(ρA )
3.2745
(ρB )
1.6316
] (4.12)
ties because they involve the movement of some quantity
(momentum, heat, specie, etc.) throughout the fluid. For vis-
ln(µ A µ B )
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

cosity, these theories can be extended to a mixture of gases.

Fundamentals of Fluid Dynamics 123

FIGURE 4.4 Temperature vs. viscosity for various hydrocarbons. (From J.B. Maxwell, Data Book on Hydrocarbons,
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
D. Van Nostrand Company, Princeton, NJ, 1950.)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 124 Wednesday, April 6, 2005 3:41 PM

124 The John Zink Combustion Handbook

FIGURE 4.5 Viscosity of mid-continent oils. (From J.B. Maxwell, Data Book on Hydrocarbons, D. Van Nostrand Company,
Princeton, NJ, 1950.)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fundamentals of Fluid Dynamics 125

Typical Data from CHEMKIN Database

N2 121286N 2 G 0300.00 5000.00 1000.00 1

0.02926640E+02 0.14879768E–02 –0.05684760E–05 0.10097038E–09 –0.06753351E–13 2
–0.09227977E+04 0.05980528E+02 0.03298677E+02 0.14082404E–02 –0.03963222E–04 3
0.05641515E–07 –0.02444854E–10 –0.10208999E+04 0.03950372E+02 4

H2O 20387H 2O 1 G 0300.00 5000.00 1000.00 1

0.02672145E+02 0.03056293E–01 –0.08730260E–05 0.12009964E–09 –0.06391618E–13 2
–0.02989921E+06 0.06862817E+02 0.03386842E+02 0.03474982E–01 –0.06354696E–04 3
0.06968581E–07 –0.02506588E–10 –0.03020811E+06 0.02590232E+02 4

CO2 121286C 1O 2 G 0300.00 5000.00 1000.00 1

0.04453623E+02 0.03140168E–01 –0.12784105E–05 0.02393996E–08 –0.16690333E–13 2
–0.04896696E+06 –0.09553959E+01 0.02275724E+02 0.09922072E–01 –0.10409113E–04 3
0.06866686E–07 –0.02117280E–10 –0.04837314E+06 0.10188488E+02 4

For most purposes, the following formula (Bird et al.17) can where u is the internal energy and T is the gas temperature;
be applied to a mixture of gases: for a constant pressure process, the specific heat is defined as
the change in enthalpy with temperature at constant volume:

µ mix = ∑ (x µ ) (x φ )
i i j ij (4.13)
C p = (∂h ∂T ) p (4.16)

where
where h is the enthalpy of the gas and T is the gas temperature.
Values for the constants Cv and Cp are provided for a variety
∑ [(1 8) ](1 + M )
−1 2
φ ij
0.5
i Mj of substances in the Appendix.

( ) (M )
14 2
1 + µ µ 12
Mi  (4.14) 4.2.3.1 Polynomial Expressions for
 i j j
 Combustion Gases
For a calorically perfect gas, both Cv and Cp are constant with
where i and j refer to two species. temperature. As noted, for a liquid, Cv and Cp have the same
value and, in general, are fairly constant. For many applica-
4.2.3 Specific Heat tions, assuming a constant Cv and Cp is adequate. However, if
Heat capacity is defined as the heat input required to achieve large temperature variations are anticipated, or if the fluid is
a given temperature change. For gases, heat capacity is a highly non-ideal, then temperature-dependent Cv and Cp val-
function of the processes that the gas is undergoing. This ues are desirable.
occurs because as gases expand or compress they do pres- In the literature, many different expressions have been
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

sure-volume work on the surroundings. In general, two pro- developed to allow for variation of specific heat with temper-
cesses are considered: constant pressure and constant ature. The NASA polynomials are recommended for any
volume, resulting in two different heat capacities for gases: application. They are utilized in the CHEMKIN database and
Cv and Cp , which are defined below. In general, this is not the have a known standard form. There is an existing FORTRAN
case for liquids and solids, which have a single heat capacity program to convert thermodynamic information into NASA
(or, for solids and liquids, Cv = Cp ). polynomials, all of which are generally available.
To enhance the usefulness of this approach, a description
The specific heat for a constant volume process is defined
of the coefficients (from Kee10) is provided. In addition, the
as the change in internal energy with temperature at constant
coefficients for three of the most common gases (N2, CO2,
volume. Therefore, the specific heat for a constant volume
and H2O) are provided.
process is given by:
The NASA polynomial form was developed as part of the
original NASA equilibrium program. As a result, the data
Cv = (∂u ∂T )v (4.15) supplied in the database is always in the form of four-line,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 126 Wednesday, April 6, 2005 3:41 PM

126 The John Zink Combustion Handbook

representing the olden-day’s four punch cards. The first line 4.2.4 Equations of State
contains (1) the molecule, (2) the date of coding the informa-
To accomplish many of the calculations required in the com-
tion, (3) the number of different atoms, (4) the phase of the
bustion field, a thorough knowledge of how density varies with
material, and then (5) the temperature range (in degrees
pressure and temperature is required. In this section, some of
Kelvin) of applicability of the polynomials. Most commonly,
the more common state equations are reviewed, along with
there are two ranges for the polynomials: a high-temperature
comments on possible limitations. An exhaustive review of
range and a low-temperature range. If this is the case, the
every type of equation of state is beyond the scope of this
common temperature for high and low is provided. The com-
chapter. The interested reader is referred to Smith et al.,10
mon temperature is usually 1000 K (1300°F).
Modell et al.,11 or Van Wylen and Sonntag15 for additional
The second line contains coefficients a1 to a5 from the information on different equations of state, and additional
equations below. These coefficients are for the high- information regarding those that have been summarized below.
temperature range. The third line contains coefficients a6 and
a7, for the high-temperature range (see below), and a1, a2, and 4.2.4.1 Ideal Gas

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
a3 for the low-temperature range. The fourth line contains
The ideal gas law has the following form:
coefficients a4 to a7 for the low-temperature range. Equations
for using the polynomials are:
PV = nRT (4.17)

Cp where
= a1 + a2 T + a3T 2 + a4 T 3 + a5T 4
R P = Gas pressure
0
h a a a a a V = Volume under consideration
= a1 + 2 T + 3 T 2 + 4 T 3 + 5 T 4 + 6
RTUnits 2 3 4 5 T n = Number of moles
T = Temperature of gas (absolute, either Rankine or
s0 a a a
= a1 ln(T ) + a2 T + 3 T 2 + 4 T 3 + 5 T 4 + a7 Kelvin)
R 2 3 4
R = Universal gas constant
where = 8.314 kJ/(kmol-K)
= 1545 ft-lbf/(lbmole-°R)
Cp = Specific heat at constant pressure
= 1.986 Btu/(lbmole-°R)
R = Gas constant
h0 = Enthalpy In addition to the above form, if both sides of the equation
s 0 = Entropy are divided by the volume (V) term, and moles are converted
T = Temperature, in degrees K to density by multiplying by molecular weight (MW), the
equation becomes:
TUnits = Temperature in preferred units (see below)

P = ρ
It should be noted that the left-hand side of the above equa- R 
T (4.18)
tions is dimensionless. This means that the units of specific  MW 
heat, enthalpy, or entropy obtained are dependent on the units
of the gas constant and temperature. Temperature must be The term (R/MW) is called the specific gas constant and, by
absolute temperature, either Rankine or Kelvin. Enthalpies inspection, is simply the universal gas constant divided by the
and entropies are measured from a standard state, listed in molecular weight of the specific gas.
Kee.9 The above expression would be most useful when The ideal gas law can be derived from the kinetic theory
calculating changes in enthalpy or entropy. Finally, the of gases.32 The two main assumptions that arise from the
expressions do not account for disassociation. Actual heat derivation are: (1) there exist no forces between molecules,
stored at elevated temperatures (above ~ 2000°F or 1100°C) and (2) all collisions are elastic. The fact that many gases
must be evaluated by accounting for disassociation — either frequently behave in this manner make this equation of state
by including radical species (as in the NASA equilibrium extremely useful. Unfortunately, gases do not always meet
code) or by modification of the polynomial expressions to the two stated criteria, which gives rise to alternative equa-
account for the phenomena. tions of state discussed in the following sections.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 127 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 127

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 4.6 Compressibility factor, Z, as a function of reduced pressure and reduced temperature for different gases.
(From F. Kreith, The CRC Handbook of Mechanical Engineering, CRC Press, Boca Raton, FL, 1998.)

4.2.4.2 Compressiblity All gases behave differently at a given temperature and

The ideal gas equation is very simple to use; however, it can- pressure. However, their behavior has been shown to be sim-
not be applied to many of the cases in refinery operations. ilar when the temperature is divided, or normalized, by the
The deviation of the gases from ideal behavior can accurately critical temperature (Tc) and the pressure is normalized by the
be accounted for by the introduction of a correction factor, critical pressure (Pc). A critical property, in this case pressure
called the compressibility factor, Z, as defined by: or temperature, is the value of the stated property at the top
of the vapor dome. This normalization of temperature and
Z = pv RT (4.19) pressure is done by defining reduced pressure and temperature
as follows:
or
PR = P Pcr and TR = T Tcr (4.22)
pv = ZRT (4.20)
where Pcr and Tcr are the critical pressure and temperature,
The compressibility factor can also be defined as: respectively.
A plot of compressibility factor, Z, as a function of reduced
Z = vactual videal (4.21) pressure and reduced temperature for different gases is shown
in Figure 4.6, which is a plot known as the generalized com-
Obviously, Z = 1.0 for the ideal gas; and the farther the value pressibility chart. It can be observed from this chart that at
of Z deviates from 1.0, the more the gas deviates from ideal very low pressures (PR << 1), all gases approximate ideal gas
behavior. behavior. Also, at high temperatures (TR > 2), ideal gas behav-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 128 Wednesday, April 6, 2005 3:41 PM

128 The John Zink Combustion Handbook

ior can be assumed regardless of pressure, except when difference between points A and B within the pipe, as air
PR >> 1. Finally, the deviation of a gas from ideal gas behav- passes through the restriction. Equations (4.26) and (4.27)
ior is greatest near the critical point. relate the pressure at point a to the values of the other param-
This method results in reasonably good results. Unfortu- eters on each side of the manometer:
nately, the functional form of the compressibility chart usually
Pa = h1 γ H2O + (h3 − h1 )γ air + PA
restricts this method to reading values from the chart.
(4.26)
4.2.4.3 Redlich-Kwong Equation
The development of cubic equations of state started from the Pa = h2 γ H2O + (h3 − h2 )γ air + PB (4.27)
Redlich-Kwong equation in the following form:

where γ is the specific weight defined as γ = ρg. Subtracting

RT a
P= − (4.23) Eq. (4.27) from Eq. (4.26) gives
v − b v(v + b)T 0.5

where constants, a and b are determined from critical temper- (

PA − PB = (h2 − h1 ) γ H2O − γ air ) (4.28)
atures and pressures as shown:
Because the specific weight of air is small as compared to the
R 2 Tc2.5 RT specific weight of water, Equation (4.28) can be written as
a = 0.42748 b = 0.0866 c (4.24)
Pc Pc follows:

PA − PB = (h2 − h1 )γ H2O
In addition, mixing rules exist for the two constants for multi-
(4.29)
ple constituent gases. From Modell and Reid,11 they are as
follows:
This equation states that the difference in pressure between
n n points A and B is equal to the difference in the column height
( am ) ∑y a ∑yb
12
= 12
i i bm = i i (4.25) (h2 – h1) of water times the specific weight of water. For
i =1 i =1 example, suppose that the difference in the column height of
water is 1.0 foot (γH2O = 62.4 lb/ft3), the pressure difference
Except near the critical point, the Redlich-Kwong equation between points A and B is
will provide values that are reasonable within experimental
uncertainties, making it useful for mixtures encountered in

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
lb lb lb
the hydrocarbon and petrochemical industries. The constants PA − PB = 1.0 ft × 62.4 = 62.4 2 = 0.433 2 (4.30)
ft 3 ft in.
a and b can be determined from known critical temperatures,
pressures, and mixture mole fractions.
An inclined-tube manometer can be used to improve the
accuracy of the pressure reading as compared to the vertical
U-tube manometer. An inclined manometer consists of a tube
4.3 FUNDAMENTAL CONCEPTS
oriented at a slope, as shown in the illustration in Figure 4.8.
The pressure difference between points A and B can be writ-
4.3.1 Hydrostatics ten as Eq. (4.31):
4.3.1.1 Manometry
A manometer is an instrument that utilizes the displacement
of a fluid column to evaluate pressure. Manometers can have PA − PB = ( L sin φ) γ H2O ( ) (4.31)
different shapes and can be oriented at various angles,
depending on the application. Two of the most common types where the height of the column of air is neglected. Solving
of manometers used in the flare and burner industry are the Eq. (4.31) for L gives:
U-tube and inclined-tube manometers.
A U-tube manometer consists of a tube shaped in the form
PA − PB
of a U, as illustrated in Figure 4.7. In this illustration, water, L= (4.32)
in the U-tube manometer, is used to measure the pressure γ H2O sin φ
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 129 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 129

Notice that as φ becomes small, the length L of the water

column becomes larger for a given pressure difference between
points A and B. Therefore, for relatively small angles of incli-
nation, the differential pressure reading along the inclined
tube can be made large, even for small pressure differences.

4.3.1.2 Buoyancy
When an object is placed in a fluid, it tends to float if the den-
sity of the object is less than the density of the fluid. The
resultant force acting on the body is called the buoyant force.
The buoyant force is equal to the weight of the fluid dis-

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
placed by the object and is directed vertically upward. This
phenomenon is referred to as Archimedes’ principle, in honor
of the Greek mathematician who first conceived of the idea.
The buoyant force can be written mathematically as follows:
FIGURE 4.7 U-tube manometer.
FB = γ fluid V (4.33)

where FB is the buoyant force, γfluid is the specific weight of the

fluid, and V is the volume of the object. As an example, sup-
pose a helium balloon with a volume of 62.8 ft3 (1.78 m3) is
firmly attached to the ground as illustrated in Figure 4.9.
Assume that the specific weights of the ambient air and helium
are 0.0765 lb/ft3 (1.22 kg/m3) and 0.0106 lb/ft3 (0.170 kg/m3),
respectively. Using Eq. (4.33), the buoyant force acting on the
balloon will be 0.0765 × 62.8 = 4.8 lb (2.2 kg). The force
acting in the upward direction at the ground, however, is not
equal to 4.8 lb (2.2 kg) because the weight of the helium in the
balloon creates a downward force (neglect the weight of the
balloon skin). The force acting in the upward direction at the
ground can be calculated by subtracting the weight of the
helium from the buoyant force. For example, the weight of the
helium is 0.0106 lb/ft3 × 62.8 ft3 = 0.666 lb. Therefore, the
force acting in the upward direction at the ground will be 4.8 lb
FIGURE 4.8 Inclined manometer.
– 0.666 lb = 4.138 lb (1.9 kg). The force per unit area, or pres-
sure force, at the location where the balloon is attached to the
ground is 4.183 lb/(π × 22/4) ft2 = 1.33 lb/ft2 (6.5 kg/m2). In
summary, the force per unit area at the location where the
balloon is attached to the ground can be written mathemati-
cally as:

Pground = ( γ air − γ helium ) Lballoon (4.34)

where γair and γhelium are the specific weight of the atmos-
pheric air and helium, respectively, and Lballoon is the length of
the balloon. Notice that the pressure force created at the
ground is a function of the specific weight difference and the
length of the balloon — not the diameter. Also notice that the
pressure force at the ground increases linearly with the length FIGURE 4.9 Helium balloon attached to the ground.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 130 Wednesday, April 6, 2005 3:41 PM

130 The John Zink Combustion Handbook

of the balloon. This example will be helpful when discussing  

draft (Section 4.3.1.3).  
520
γ hot = γ flue   (4.36)
flue gas  460 + Thot (°F ) 
4.3.1.3 Draft gas ST  
 
flue
gas
The term “draft” is commonly used to describe the pressure
inside fired equipment. Draft is the pressure difference where the subscript flue gas ST represents the specific weight
between the atmosphere and the interior of the fired equip- of the flue gas at the standard temperature of 60°F (16°C),
ment at a particular elevation. Since both the atmospheric and the subscript hot flue gas represents the hot flue gas at the
pressure and the pressure inside the equipment vary with ele- actual temperature.
vation as discussed previously, it is important to make these
pressure measurements at the same elevation. For example, it
is common to measure stack draft by connecting one leg of
an inclined-tube manometer to a furnace and leaving the
other leg open to the atmosphere. If the absolute pressure
inside the furnace is less than the atmospheric pressure, then
the liquid in the manometer will move. The difference in FIGURE 4.10 A small packet of fluid flows from point
pressure is called the furnace draft. Typically, draft levels are A to point B along an arbitrary path.
measured in inches of water column pressure, generally writ-
ten as “w.c.” When recording draft, it is not necessary to place
a minus sign in front of the numeric value. Customarily, it is 4.3.2 Bernoulli Equation
understood that draft pressures are always negative values.
An equation that has proved useful in the calculation of flows
Draft is established in a furnace because the hot flue gases for the combustion industry is the Bernoulli equation. In Fig-
in the vertical stack have a lower specific weight than the ure 4.10, a small packet of fluid flows from point A to point B
outside air. The difference in specific weight of the flue gas along an arbitrary path. Neglecting friction (i.e., there is negli-
and air creates a buoyant force, causing the flues gases to float gible shear stress), the conservation of energy analysis yields
vertically upward. The buoyant force is equal to the product the following:
of the difference in specific weights and the height of the
(Change in pressure energy) + (Change in velocity energy)
stack, and can be written mathematically as:
+ (Change in potential energy) = 0
or
 
Pdraft =  γ atmospheric − γ hot  Hstack (4.35)
 air flue 
 1 
 gas  ∆P + ∆  ρV 2  + ∆(ρgh) = 0 (4.37)
 2 gc 
where Pdraft is the draft pressure, Hstack is the stack height, and where
γatmospheric air and γhot flue gas represent the specific weights of the P = Pressure at fluid packet
ambient atmospheric air and the hot flue gas, respectively. ρ = Density of fluid packet
Notice that this equation is very similar to the equation used V = Velocity of fluid packet
to calculate the pressure at the ground in the helium balloon g = Gravitational acceleration
example discussed in the previous section. h = Height of fluid packet
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

As just shown, the draft created by a column of lighter- gc = Gravitational constant

than-air gas is function of the specific weight difference
between the atmospheric air and flue gas, as well as the height 4.3.2.1 Total Pressure vs. Static Pressure
of the stack. The specific weight of the flue gas is a function Often, the potential energy term can be neglected in the
of both the temperature of the gas in the stack and the com- above equation, which results in just the pressure term and
position. An increase in exhaust flue gas temperature the velocity term. The combination of these two terms is
decreases the specific weight, causing an increase in the draft. often referred to as the total pressure. Pressure drop analysis
The specific weight of a hot flue gas can be calculated as in piping systems, discussed in Section 4.5, exclusively exam-
follows: ines the total pressure loss of a system in order to prevent
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 131 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 131

FIGURE 4.11 Pressure relief vessel venting to a flare.

errors from occurring. For example, it is known that there is The following is a simple example to illustrate the differ-
an energy loss in a piping system if the pipe size increases, ences between these pressure concepts. Suppose there is a
either suddenly or gradually. Despite the energy loss, the pressure relief vessel that is venting to a flare system as
static pressure tends to increase if the pipe size is increased. illustrated in Figure 4.11. Assume that the pressure relief
According to the Bernoulli equation, as velocity is decreased, vessel is maintained at 1 psig during the flaring event. As the
the pressure is expected to increase. This being the case, pip- gas exits the vessel, it enters the flare header. If a pressure
ing pressure drop analysis would be a meaningless term if gage were placed on the pressure vessel and on the flare
only static pressure was included. Any analysis must include header somewhere downstream of the vessel, these gages
the total energy or total pressure of the system. Using total would not read the same pressure. At the vessel, the pressure
energy and total pressure, the following terms are defined: would obviously read 1 psig (7 kPag); however, the pressure
gage on the flare header would read less than 1 psig (< 7 kPag).
Static pressure: the pressure term in the Bernoulli equation. The concepts of static, velocity, and total pressure are used
This is normally understood as a pressure and is measured to explain why the pressures are different.
directly by a pressure gage. The total pressure is defined as the static pressure plus the
Dynamic pressure: the velocity term (1/2 ρV 2) from the Bernoulli velocity pressure:
equation.
Total pressure: the sum of static and dynamic pressures. This PT = PS + PV (4.38)
term is utilized for most pressure loss analyses in a piping
system. When liquids are discharged to the atmosphere, where PT is the total pressure, PS is the static pressure, and PV
this term would also include potential energy changes.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
is the velocity pressure. The static pressure is a measure of the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 132 Wednesday, April 6, 2005 3:41 PM

132 The John Zink Combustion Handbook

pressure where the gas velocity is zero. Therefore, the static 4.3.2.1.1 “K” Factors
pressure of the vessel is obviously 1 psig (7 kPag). Because As implied in the previous section, the Bernoulli equation is
the velocity of the gas inside the vessel is zero, the velocity logically utilized to analyze total pressure losses (or energy
pressure must also be zero. From Eq. (4.38), the total pressure losses) for piping systems. To accomplish the analysis, the
of the vessel must be equal to the static pressure of the vessel. equation is modified as follows:
The pressure gage in the flare header reads the pressure at the
wall of the pipe where the gas velocity is zero; therefore, this
pressure gage must be reading the static pressure in the flare
header. Because the gas is moving in the pipe, the velocity  1   1 
∆P + ∆  ρV 2  = Total pressure loss = K  ρV 2 
pressure must be greater than zero. The velocity pressure in  2 gc   2 gc 
the flare header can be analyzed in two ways.
= K (Dynamic pressure) = KPv ( 4.40)
First, assume that the pressure energy losses from the vessel
to the pressure gage on the flare header are zero. If no energy
losses are assumed, then the total pressure on the flare header The two assumptions utilized to derive the Bernoulli equa-
must be equal to the total pressure at the pressure vessel. It tion were: (1) flow along a stream line, and (2) no friction or
is also true that the pressure gage on the flare header must be shear stresses. Piping systems satisfy the first assumption, but
reading the static pressure at that location. Therefore, from do not satisfy the second. Shear stresses arise when a fluid
Eq. (4.38), the velocity pressure in the flare header must be flows over or past a solid object, as discussed in Section 4.2.2
equal to the total pressure minus the static pressure. For of this book. Experience has shown (see Vennard and Street4)
example, if the pressure gage on the flare header reads 0.8 psig that the magnitude of the force will be proportional to the
(6 kPag) and the total pressure is 1.0 psig (7 kPag), then the dynamic pressure term (1/2 ρV 2). The constant of proportion-
velocity pressure equals 0.2 psig (1 kPag). ality commonly utilized is “K”. Values of “K” for different
piping configurations and different fittings have been exten-
In reality, there will be pressure energy losses in the flare
sively studied and reported in the Crane Piping manual31 and
header due to friction between gas molecules and pipe wall.
in Idelchik.12 In the literature, these constants are often

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Therefore, the total pressure in the header will not equal the
referred to as “K” factors.
total pressure in the pressure vessel. One can measure the
total pressure in the flare header by inserting a pitot tube into
the header, as illustrated in the insert of Figure 4.11. The pitot 4.3.3 Control Volumes
tube will measure the static pressure plus the impact pressure, A common analysis methodology is the utilization of control
or velocity pressure, of the gas. This is a measure of the total volumes — also termed the integral method. Conceptually, the
pressure inside the flare header. The velocity pressure in the integral method entails enclosing the region of interest with
flare header can now be determined by subtracting the total control surfaces, then writing the appropriate conservation
pressure reading from the static pressure reading. For exam- equations for the enclosure. The method is discussed exten-
ple, if the total pressure reads 0.9 psig (6 kPag) in the header sively in Fox and McDonald.3 The interested reader is directed
and the static pressure reads 0.7 psig (5 kPag) in the header, to that reference for additional information.
then the velocity pressure in the header must be equal to
0.2 psig (1 kPag). In general, there are two types of physical quantities of
interest: extensive and intensive. For a given volume, exten-
As just mentioned, the velocity pressure is the pressure sive properties describe the entire volume. They include:
created due to the impact of the gas molecules. The velocity
pressure in a gas stream can be calculated as follows:
Total mass = M
r
ρV 2 Momentum = MV
PV = (4.39)
2 gc r
Angular momentum = Ω

where ρ is the density of the flowing gas, V is the velocity, Enthalpy = H

and gc is the gravitational constant. Equation (4.39) is used Energy = E
extensively when determining the pressure drop through flare
and burner systems. Entropy = S
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 133 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 133

Intensive properties are properties per unit mass. As a prac- 4.3.3.3 Energy Conservation
tical matter for systems of interest, an intensive property is
determined by dividing the extensive property by the mass,
d (E)
or density times the volume, of the control volume. For mass, = δQ + δW (4.43)
dt
energy, and momentum conservation, the intensive variables
associated with the appropriate extensive variables are:
The time rate of change of energy for the piece of mass is
equal to the sum of the heat transferred (Q) and work (W).
Mass conservation : M ⇒1 Conceptually, the above conservation laws are reasonably
r easy to understand, but very difficult to apply. To accomplish
Momentum conservation : ⇒V
an analysis, it would be necessary to track and maintain
Energy conservation : E ⇒ e (enthalpy for information about all the possible pieces of mass that happen
flow applications) to flow through a control volume. The fundamental precept
of utilizing control volumes for analysis is that it is possible
If a control volume is drawn around an area of interest, at to determine the time rate of change for a given extensive
any given instant in time the volume is filled with a certain property by determining: (1) the time rate of change in the
amount of mass, the control mass. At this instant in time, the control volume of the corresponding intensive property, and

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
control mass has the exact same shape as the control volume. (2) the flux in and out of the control volume of the same
The following conservation laws are valid for the control intensive property. If an arbitrary extensive property is
mass, which is coincident with the volume for the instant in denoted as N, and its corresponding intensive property is
time as indicated by: denoted as n, then, according to Fox and McDonald,3 the time
rate of change for any extensive property is given by:

dM
=0  dN  r r
∫ ∫
(4.41) d
dt = nρdV + nρV ⋅ dA (4.44)
 dt  PieceofMass dt
CV CS

4.3.3.1 Mass Conservation

An example of how this is applied is provided, but first some
Because mass is neither created nor destroyed, the total control
explanations of the meaning of the terms are in order:
mass will remain constant. This may seem confusing because
there is a mass flux through the volume and the density of the
volume may change with time. Remember that the mass of  dN  (4.45)
interest is the mass that happens to inhabit the control volume  dt  PieceofMass
at a particular instant in time. During the next instant, while
the mass just examined may not be in the volume, or even Equation (4.45), as discussed, is the total rate of change of any
have the same shape as the volume, it still has the same mass. arbitrary extensive property of the piece of mass coincident
This makes the conservation equation simplistic. For a given with the control volume.
piece of material (or mass), the total mass does not change
with time. Without further discussion of this point, the other
∫ nρdV
conservation equations are reproduced for the given piece of d
(4.46)
material (or mass): dt
CV

4.3.3.2 Momentum Conservation Equation (4.46) is the time rate of change of the arbitrary
extensive property N within the control volume. This is
r expressed as a product of the associated intensive property n,
( )= r
∑F
d MV the density, and the volume (dV). The CV on the integral
i (4.42) represents the control volume.
dt i

r r
The time rate of change of momentum for the piece of mass
is equal to the sum of the forces applied to the mass.
∫ nρV ⋅ dA
CS
(4.47)

Copyright CRC Press

134 The John Zink Combustion Handbook

Equation (4.47) is the net flux of the extensive property N of fluid phenomena is the differential formulation. Numerous
through the boundary of the control volume. The product of excellent references exist (e.g., Panton,1 Potter and Foss,13 and
density, velocity, and area is the mass flow rate. Thus, the Kuo14) for the reader interested in greater detail on this subject.
equation really represents the mass flow in and out of the This treatment concentrates on the conceptual development of
control volume times the associated intensive property. the formulation and the final form. It is the intent of this refer-
The boundary is called the control surface, or CS. The ence to provide insight into the meaning of the equations, how
vector dot product of the velocity and the area signifies that simplifications have been applied to them, and the limitations
only the velocity component normal or perpendicular to the of the simplifications.
surface can be used for calculation purposes. An example of On a conceptual level, the equations of interest are partial
how this is used practically can be seen from the conservation differential equations that describe:
of mass as given by:
1. conservation of mass
2. conservation of momentum
dM
=0 (4.48) 3. conservation of energy
dt 4. conservation of species

The associated intensive property for mass is simply one. Similar to density, the main assumption regarding the differen-
Inserting the above information into the general control vol- tial derivation of these equations is that the volume reduces to a
ume statement yields: point, but a point much larger than the molecular length scale.
Each of the conservation principles are discussed in turn.
r r Because it is fairly simple, a reasonably complete deriva-
∫ ∫
d
0= ρdV + ρV ⋅ dA (4.49) tion of conservation of mass is provided. Insight gained from
dt
CV CS the conservation of mass derivation will be generalized as
appropriate for the other conservation equations. Details will
Notice the first term in Eq. (4.49). Practically speaking, this be referenced, as appropriate.
term takes into account the change in density of the control
volume. For steady-state and incompressible analysis, this 4.3.4.1 Conservation of Mass
term becomes zero. The second term is just a statement of Figure 4.12 is an idealization of a small “differential” control
mass in and mass out. Thus, for steady-state and incompress- volume. For simplicity, Cartesian coordinates (X, Y, Z) are
ible calculations, the equation reduces to: utilized; however, the results could be generalized to any
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

other system. The little “cube” has dimensions of dx by dy by

(Mass in)Control Volume = (Mass out )Control Volume dz, and thus the area of any face is the same (i.e., dx · dy), and
the volume is dx · dy · dz. Reference will be made to these
For constant velocity across any opening, the equation observations in all of the following sections.
becomes: If this small control volume is placed at a fixed location in
a fluid flow field, then fluid will flow into and out of the
(ρAV )1 = (ρAV )2 (4.50) volume continuously. In other words, conservation of mass
states that the amount of mass going into the volume will
equal the amount of mass leaving the volume plus any mass
Equation (4.50) is the standard form for calculating mass bal-
stored in the volume. The mass flow across any face into the
ances. Additional examples of the method are found in Fox
volume is going to be the area of the face, times the velocity,
and McDonald3 and Potter and Foss.13 Application of the
times the density, or in equation form:
method to eductors is provided in Chapter 8 of this book. In
addition to fixed control volumes, it is possible to construct
Mass flow through a face
control volumes such that all or part of the surfaces move.
= (Velocity component)·(Face area)·(Density)
Required modifications are outlined in both Panton1 and Pot-
ter and Foss.14 From Section 4.2.1, it is observed that the differential vol-
ume is chosen so that the density is uniform throughout, but
4.3.4 Differential Formulation is not so small that molecular spacing is sparse. Choosing the
All of the previous fluid flow analysis required various simpli- X-direction for analysis, as shown in Figure 4.13, results in
fications and assumptions. The most fundamental formulation the following:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 135 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 135

(Mass flow) X = U (dy ⋅ dz)(ρ) (4.51)

where
U = Velocity in the X-direction
ρ = Density

For conservation of mass analysis, however, the point of

interest is ensuring that the mass entering, leaving, and being
stored are the same. The real question that the equation must FIGURE 4.12 An idealization of a small “differential”
answer is: “What is the difference between mass entering and control volume.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

exiting in the X-direction?” or:

(Mass entering) X − (Mass leaving) X

= Uρ(dy dz ) − (Uρ + d (Uρ)(dy dz ))

= − d (Uρ)( dy dz ) (4.52)

In Eq. (4.52), dU is simply the change in X-direction veloc-

ity that occurred over the distance dx, so the change in mass
flow rate over distance dx is d(Uρ). FIGURE 4.13 Mass flow into and out of volume in the
X-direction.
The mass of a cube is the volume times the density. In
Eq. (4.52), the rate at which mass is entering or leaving the
cube is represented. This being the case, the conservation of
mass for this system will be a rate equation, that is, how much dρ
has the mass changed with time. This is expressed as: (dx dy dz) = − d (ρU )(dy dz) − d (ρV )(dx dz)
dt
− d (ρW )(dx dy) (4.54)
Mass stored/time

= {(Mass at time 1) − (Mass at time 2)} time Dividing Eq. (4.54) by (dx dy dz) and collecting terms on one
side of the equation:
= {ρ1 (dx dy dz ) − ρ2 (dx dy dz )} dt

dρ ∂(ρ) ∂(ρU ) ∂(ρV ) ∂(ρW )

= (dx dy dz ) (4.53) + + + =0 (4.55)
dt ∂t ∂x ∂y ∂z

where where
dρ = Change in density with time U = X velocity component
dt = Differential change in time V = Y velocity component
dx dy dz = Volume of cube W = Z velocity component

The above expression is very reasonable. The only way The above expression has the following physical implications:
that mass can be stored in a fixed-volume container of any dρ
sort is for the density to change. Repeating the mass flow 1. The time-dependent term represents the change of
analysis for each of the other two directions, collecting the mass inside the volume. dt
expressions, and setting them equal to the expression for the 2. The other terms represent the difference between what
change in mass, results in the following general expression leaves and enters the volume.
for conservation of mass for fluid flow: 3. This is summarized as:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 136 Wednesday, April 6, 2005 3:41 PM

136 The John Zink Combustion Handbook

∂(ρ) ∂(ρU ) ∂(ρV ) ∂(ρW ) the momentum leaving the volume, and (2) the momentum
+ + + =0 (4.56) stored. It can be shown that this is the same type of convec-
{ ∂t ∂x ∂y
14444244443∂ z
Internal Mass difference between
tion and storage term derived for conservation of mass, so
changes entering and leaving volume the substantial derivative can be used to describe a change
in momentum. For momentum, velocity times density is
For incompressible flow, this reduces to: utilized; thus:

∂(ρ) ∂(ρ) ∂(ρ) ∂(ρ)

+U +V +W =0 d ( Momentum) D(ρU )
∂t ∂x ∂y ∂z = (4.60)
dt Dt

As it turns out, the general form of Eq. (4.56) can be used Forces that can be applied fall into two categories: (1) forces
for any physical quantity of interest: momentum, energy, chem- that are applied to the entire volume equally, and (2) forces
ical species, etc. In general, this form of equation is said to that are applied to the surface of the volume. The first type of
represent the time variant and convection terms of the differ- force is sometimes called a body force. Gravity is normally
ential equations. In fact, this form appears so often that it is the only body force encountered, and will be the only one
given the special name of “substantial derivative” and repre- considered herein. Others, however, are possible, such as
sented in shorthand as follows: electromagnetic fields and acceleration. The second type of
force, surface forces, can act either in a direction normal to
D( ) ∂( ) U∂( ) V∂( ) W∂( ) the surface (pressure), or in a direction parallel to the surface
= + + + (4.57)
Dt ∂t ∂x ∂y ∂z (shear stress). The analytical form of each of these forces is
summarized below.
where different physical parameters can be substituted into Gravitational Body Force: The gravitational body force
the blank sets of parentheses. This would allow the conserva- is illustrated in Figure 4.14 and is symbolized by Fx. If there
tion of mass for incompressible flow to be written as: is no body force in the X-direction, then this term is ignored.
Common practice is to consider the Y-direction as up and
D(ρ) down, so the gravitational body force term usually would only
=0 (4.58) apply to the Y equation.
Dt
Normal or Pressure Forces: Normal or pressure forces
Use of the substantial derivative will be routinely referenced are illustrated in Figure 4.15 and take the following form:
in the following sections.
Px
4.3.4.2 Conservation of Momentum − (4.61)
dx
Newton’s second law of motion stated in words is: “the time
rate of change of momentum is equal to the sum of all
A net force will arise only if there is a difference in the
applied forces.” In equation form, this is:
direction of interest. The negative sign occurs because higher
pressures in the direction of interest will induce a flow in the
d ( Momentum)
dt
= ∑ Applied forces (4.59) negative direction.
Parallel or Shear Stress: Parallel or shear stress is illus-
trated in Figure 4.16. For a Newtonian fluid, shear stress or
Instead of change of momentum, many physics textbooks refer
force will be proportional to the velocity gradient (see Section
to mass times acceleration. The two statements are equivalent
4.4.1.1). The force due to shear stress along the Y-direction
for volumes with constant density; but for fluid considerations,
on the X-plane is:
changes in momentum are more appropriate. Momentum is a
vector quantity with three components, so any equation
derived for one direction can be generalized into three similar dV
τ=µ (4.62)
equations, one for each component (x, y, z directions). For sim- dy
plicity, only the X-direction is considered here.
The rate of momentum change is represented as: (1) the However, what is of interest is the change in force (the
difference between the momentum entering the volume and force in this case being shear stress) along any one direction.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fundamentals of Fluid Dynamics 137

This change in force formulation gives rise to terms that look

like:

d 2V
µ (4.63)
dy 2

Collecting all of the above terms results in the following. FIGURE 4.14 Gravitational body force.

X-direction momentum equation:

D(ρU ) −∂P  ∂ 2U ∂ 2 V ∂ 2 W 
= Fx + + µ 2 + 2 + 2  (4.64)
Dt
123 Body ∂x  ∂x ∂y ∂z 
force Pressure Shear forces
Momentum force
changes
14444442444444
(stresses)
3
Summation of forces on volume

In a similar manner, equations for the Y and Z momentum

can be derived. The analytical form of the above equation
may at first be quite intimidating. However, it is nothing more FIGURE 4.15 Normal or pressure forces.
than a logical extension of Newton’s second law of motion
from high school physics (F = ma), which has been applied
to fluid flows. Strictly speaking, the above equation is only
valid for incompressible flow. The interested reader is directed
to previously listed references for additional details and an

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
extension of the momentum equation to include compress-
ibility effects.

4.3.4.3 Conservation of Energy

The conservation of energy equation derivation is similar to
mass and momentum. First, start with a general conservation
principle, in this case the first law of thermodynamics, as dis- FIGURE 4.16 Effect of shear stress on X-direction face.
cussed in Van Wylen:15

(Heat transferred) + (Energy in) + (Energy generated) 5. expansion (work)

6. wall friction (work)
= ( Energy stored) + ( Energy out ) + ( Work ) (4.65)
7. viscous dissipation (work)
or, moving terms around: In general, for combustion in the petrochemical industries,
the common terms utilized are convective and radiative heat
(Energy stored) + (Energy out ) − (Energy in) transfer, chemical heat release, and wall friction. Inclusion of
multiple chemical species complicates the formulation, as
= (Heat transferred) + ( Energy generated) − ( Work ) (4.66) molecular diffusion of the various species provides an energy
transport mechanism. Furthermore, the energy in a flowing
The term on the left will again be the substantial derivative of fluid can be separated to consider thermal and kinetic energy
energy into and out of the volume. Depending on the pro- separately. In low Mach-number flows (i.e., incompressible
cesses involved, terms on the right will vary, but may include flows), conservation of kinetic energy is automatically satisfied
all or none of the following: whenever the momentum equation is satisfied (the conservation
1. convective heat transfer (heat transfer) of kinetic energy equation can be derived by manipulating the
2. conductive heat transfer (heat transfer) momentum equation). The following (thermal) energy equation
3. radiative heat transfer (heat transfer) contains the terms that would normally be important in a com-
4. chemical heat release (energy generated) busting flow (numbers correspond to terms described above):
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 138 Wednesday, April 6, 2005 3:41 PM

138 The John Zink Combustion Handbook

D(Chemical species)
=0 (4.68)
Dt

If reactions are occurring, then the change will be equal to

the formation (or destruction) rate of the compound. Reaction
rate equations are normally of the Arrhenius type (see Kuo14).
This results in the conservation of species equation being:

DYi
ρ = ω i Wi − Vi ⋅ ∇Yi (4.69)
Dt

4.4 DIFFERENT TYPES OF FLOW

4.4.1 Turbulent and Laminar Flow

For centuries, scientists have been fascinated with flows in
nature. Visual observations of fluids in motion have resulted
in two broad classifications of flows: laminar and turbulent.
Laminar flows have no fluctuations and tend to flow for
long distances with very little mixing occurring in the flow.
Figures 4.17 and 4.18 illustrate two common examples of
laminar flow. In Figure 4.17, smoke from burning incense
slowly rises in still air for a distance of 6 to 12 in. (15 to
30 cm) before becoming unsteady. In Figure 4.18, water
exits a water faucet at a very low velocity and falls in a
straight line until it strikes an object or is disrupted by air
FIGURE 4.17 Smoke from incense. (From Visualiza- currents. In both cases, the regions of smooth movement are
tion Society of Japan, The Fantasy of Flow, IOS Press,
examples of laminar flows.
Amsterdam, 1993, 96.)
Turbulent flows have fluctuations and eddies associated
with them that increase mixing rates substantially. The
legendary artist and scientist Leonardo da Vinci’s view on
r
∑
DT turbulence is reproduced in Figure 4.19. As shown, the cir-
ρCP =∇
12
4 ⋅ k∇
43T− C pi Yi Vi C p ⋅ ∇T
1424D3
t culation, induced vortices and fluctuations allow, in this case,
Molecular
1i 44424443
conduction
Energy
change Thermal an introduced stream to mix rapidly with a still liquid. Most

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
diffusion
(4.67) flows in nature will tend to become turbulent as flow rate or
r
∑
DP local velocity is increased.
− ω i hoi + βT + Φ { +∇ ⋅ qr
D t 1 2 3
3 123
Viscous
14
i 24
Pressure dissipation Thermal
radiation 4.4.1.1 Reynold’s Number
Chemical work
reaction
During the 1880s, Osborn Reynolds quantified the previous
qualitative observations of laminar and turbulent flows. His
The reader should consult Bird et al.16 for a much more com- observations were first published in 1883.17 A schematic of
plete discussion of energy conservation. his experimental apparatus is shown in Figure 4.20. In this
device, water flows from a tank through a bell-mouthed inlet
4.3.4.4 Conservation of Species into a glass pipe. Also, at the entrance to the glass pipe is the
If there are no chemical reactions, then the implication is that outlet of a small tube, which leads to a reservoir of dye.
the mass fraction of any particular species can only change Reynolds discovered that at low water velocities, the stream
due to mass flow into and out of the volume. Under this con- of dye issuing from the thin tube did not mix with the water.
straint, conservation of species becomes exactly like the con- Instead, the dye became a distinct flow, parallel to the pipe
servation of mass equation, namely: centerline. As the valve was opened and the water velocity
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 139 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 139

increased, it was observed that at some greater velocity, the

dye would rapidly mix with the water, causing the entire flow
to be colored with the dye. Reynolds deduced from these
experiments that two flow regimes existed. In the first (lam-
inar flow), the fluid streams flow past each other in parallel
layers, or laminae, but do not mix with each other. The second
regime, at higher velocities, is where the two streams mix
rapidly. This mixing is caused by vortices, illustrated in
Figures 4.21 and 4.22, that arise in the flow. Due to the pres-
ence of the vortices, this flow has been called turbulent. Figure
4.23 illustrates laminar flow.
Reynolds found that below a certain velocity, the flow
always became laminar. Figure 4.23 shows laminar flow can
even be achieved by flow over a cube. As the velocity was
increased, turbulent flow could always be achieved, but the
precise velocity depended on how still the water in the tank
was prior to the experiment. In addition, Reynolds was able
to generalize his results into a nondimensional parameter,
which today is known as the Reynolds number, Re. It is
defined as follows:

Vdρ
Re = (4.70)
µ
where
V = Velocity in pipe
d = Pipe diameter
ρ = Density of fluid
FIGURE 4.18 Water exiting a faucet at low velocity.
µ = Fluid viscosity (From Visualization Society of Japan, The Fantasy of Flow,
As a practical matter, pipe flows having a Reynolds number IOS Press, Amsterdam, 1993, 96.)
less than 2300 are laminar, and flows having a Reynolds num-
ber greater than 4000 are turbulent. In addition to determining
type of flow, Reynolds numbers have been proven to also scale Area
( 4)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

the intensity of turbulence in a flow. That is, higher Reynolds Hydraulic diameter = (4.71)
Perimeter
numbers result in greater vortex generation and faster mixing
rates. As a result, Reynolds number calculations are very com- It should be noted that the above definition of hydraulic
mon in the petrochemical industry. They are utilized to scale diameter reduces to the definition of normal diameter for cir-
orifice coefficients,19 friction factors (Vernard and Street,4 and cular pipes. The above equation provides reasonable accu-
Section 4.5 of this book), heat transfer rates (Lienhard,19 and racy for calculations involving turbulent flow, but large errors
Chapter 3 of this book), and mass transfer rates (Bird16). will occur for laminar flow calculations. See Vernard and
Street4 for additional details.
4.4.1.2 Hydraulic Diameter
In addition to circular pipe flows, Reynolds numbers can be 4.4.1.3 Reynolds Averaging
utilized, as described above, for geometries other than round. Turbulent flows are characterized by a highly fluctuating
Square, triangular, and other enclosed flows scale with Rey- instantaneous velocity. The flow fields generated are three-
nolds numbers. For geometric variations with no obvious dimensional in nature and dependent on an unknown time
“diameter,” a concept termed the hydraulic diameter is use- function. To help understand the processes, Reynolds 34 sug-
ful. Essentially, it is the ratio of the flow area to the flow gested utilizing a velocity composed of two components: (1) a
perimeter, multiplied by four. In analytical form, this ratio is time averaged velocity, and (2) a fluctuating velocity. This is a
expressed as: logical substitution, because, for 99% of applications, only the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 140 Wednesday, April 6, 2005 3:41 PM

140 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 4.19 Leonardo da Vinci’s view of turbulence.

FIGURE 4.20 Osborn Reynolds’ experimental apparatus used to study the transition from laminar to turbulent flow.

mean (or average) velocity is of interest, and the time average averaged over time to obtain an expression for average
of the fluctuating velocity would be zero. This is expressed as: velocity. It should be noted that while the time average of
fluctuating velocity is zero, the time average of the product
Velocity = V + V ′ (4.72) of two fluctuating velocities is not zero. Potter and Foss13
where present a good discussion of this observation.
Starting from Eq. (4.64) (the momentum equation), if body
V = Average velocity (the bar represents “time
forces are neglected and density is considered constant, the
averaged”)
“x” only equation can be rewritten as:
V′ = Fluctuating velocity (where V′ = 0)

The two velocity terms can be substituted into the differen- D(ρU ) ∂τ ix
= (4.73)
tial momentum equation (4.64), and the resulting equation Dt ∂x i
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 141 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- 141

FIGURE 4.21 Water from faucet showing transition. FIGURE 4.22 Wake area showing mixing vortices.
(From Visualization Society of Japan, The Fantasy of Flow, (From Visualization Society of Japan, The Fantasy of Flow,
IOS Press, Amsterdam, 1993, 96-97.) IOS Press, Amsterdam, 1993, 3.)

FIGURE 4.23 Photograph showing the laminar flow of smoke over a rectangular obstruction. (From M. Van Dyke, An
Album of Fluid Motion, The Parabolic Press, Stanford, CA, 1982, 10.)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 142 Wednesday, April 6, 2005 3:41 PM

142 The John Zink Combustion Handbook

where
 U ′2 U ′V ′ U ′W ′ 
 
 ∂U ∂V ∂W  ρ U ′V ′ V ′2 V ′W ′  (4.79)
τ ix = − Px + µ + +  (4.74) U ′W ′ V ′W ′ W ′2 
 ∂x ∂y ∂z   
The τ part of the momentum equation contains no cross
Collectively, these terms are called “Reynolds stresses.” The
velocity terms. However, close examination of the substan-
term “stress” is applied to them because they are associated
tial derivative (see Section 4.3.4 of this chapter) reveals that
with the stress term in the momentum equation, although
the left-hand side of Eq. (4.73) does contain the products of
they arise from the convective side of the momentum equa-
different velocities. Substitution of the decomposed velocity
tion. Research over the last 100 years has not yet provided a
into the momentum equation, taking the time average
completely satisfactory model for these terms. Computa-
(to develop a time-averaged momentum equation), and sim-
tional fluid dynamics (CFD, see Chapter 9 of this book)
plifying (see Potter and Foss13 for details), results in the fol-
makes extensive use of various assumption and modeling
lowing x momentum equation:
approximations. Occasionally, the ability to accurately model
a flow field still depends on the modeler’s understanding of
D(ρU ) ∂τ ix  ∂U ′ 2 ∂U ′V ′ ∂U ′W ′ 
the limitations of the chosen Reynolds stress model.
= − ρ + +  (4.75)
Dt ∂x i  ∂x ∂y ∂z 
4.4.1.4 Jets
where When a fluid emerges from a nozzle, it will interact with the
surrounding fluids. This type of system, termed a “free jet,”
 ∂U ∂V ∂W 
τ ix = − Px + µ + + commonly occurs in combustion systems (Figure 4.24).
∂z 
(4.76)
 ∂x ∂y High-pressure fuel from a nozzle, steam spargers, and liquid
fuel sprays are all examples of free jets. Figure 4.25 illus-
Interestingly, the cross fluctuating velocity terms have the
trates the interaction of a free jet with the surrounding fluid.
same derivative operator as the shear stress, so it is common
As the jet travels downstream from the nozzle, its diameter
to simply redefine the shear stress with those terms added to
will increase as it captures ambient fluid into its stream. This
it. In that case, the momentum equation for time-averaged
phenomenon has been extensively studied, both experimen-
velocity will have the exact same form as before, namely:
tally and theoretically. For theoretical treatment, the inter-
ested reader is referred to Schlichting,20 Hinze,21 and
D(ρU ) ∂τ ix Tennekes.22 The historical treatment is to assume that fluctu-
= (4.77)
Dt ∂x i ating velocities have the same magnitude in each of the three
directions (this is called isotropic turbulence), which reduces
But the stress term for the x equation is now modified by the
the Reynolds stresses from six terms to one term. The single
fluctuating velocity terms. It now has the following form:
stress term is then modeled as a function of the local velocity
gradient and the nozzle diameter. This model has a single fit-
 ∂U ∂V ∂W 
τ ix = − Px + µ + + ted parameter, which is deduced from experimental data.
 ∂x ∂y ∂z  Experimental characterization of nozzles has resulted in

[ ]
practically the same results as those obtained theoretically,
− ρ U ′ 2 + U ′V ′ + U ′W ′ (4.78) except near-nozzle effects are quantified. Nevertheless, both
treatments result in the following two equations describing
It should be noted that although the fluctuating terms are jet velocities (from Beer and Chigier23):
historically “lumped” with the stress and pressure terms
(forces, from the derivation), they actually arise from the
U0 x
convective side of the momentum equation. If the same pro- = 0.16 − 1.5 (4.80)
cedure is followed for the y and z momentum equations, a Um d0
total of nine fluctuating velocities arise from the derivation.
Noting, however, that velocity order can be interchanged, where
there are only six different terms. Collecting them in tensor U0 = Initial velocity at nozzle (assuming plug flow)
form (terms from the three different equations are listed Um = Maximum (or centerline) velocity of the jet
together), they are as follows: downstream
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fundamentals of Fluid Dynamics 143

x = Distance from nozzle exit

d0 = Diameter of nozzle

Equation (4.80) describes the velocity decay that occurs as a

nozzle interacts with its surroundings. The above equations
are dimensionless; the only requirement for accuracy is that
consistent units for velocity and length be utilized. As the jet
progresses downstream, it expands radially. As it expands, it
develops a radial velocity profile. This profile can be
described by a Gaussian 24 as follows:

 r 
= exp − K u   
U0
(4.81)
Um   x

where
U = Actual velocity
Um = Maximum or centerline velocity at the particular
x location
r = Radial distance from centerline
Ku = Gauss constant, which has a value of about 8724

4.4.1.5 Entrainment
As a free jet interacts with surrounding fluids, it will “pick up”
or entrain the ambient fluid and carry it downstream. It is this
additional fluid that will cause the jet to expand and the
velocity to decrease. Figures 4.25 illustrates the phenomenon
of entrainment. Equations for concentration decay are similar
to the velocity equations, and are given by:
FIGURE 4.24 Free jet structure. (From M. Van Dyke,

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
An Album of Fluid Motion, The Parabolic Press, Stanford,
C0 x CA, 1982, 99.)
= 0.22 − 1.5 (4.82)
Cm d0

 r 
= exp − K u   
C
(4.83)
Cm   x

where
C = Actual concentration
C0 = Initial concentration at nozzle (assuming flat
profile)
Cm = Maximum (or centerline) concentration of the
jet downstream
x = Distance from nozzle exit
d0 = Diameter of nozzle
r = Radial distance from centerline
Ku = Gauss constant, which has a value of about 55.524

One unexpected result of the above jet laws is that concen- FIGURE 4.25 Free jet entrainment. (From M. Van
tration is independent of velocity. Intuition would lead one to Dyke, An Album of Fluid Motion, The Parabolic Press, Stan-
believe that the faster a jet exits an orifice, the faster it mixes ford, CA, 1982, 97.)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 144 Wednesday, April 6, 2005 3:41 PM

144 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
transfer. In a typical reacting flow, the density of the reactants
can be 500% of the products’ density, yet the flow may still be
analyzed as an incompressible flow. In this section, the Mach
number is introduced and the Mach number will be used alone
to determine whether or not a flow is compressible.
Compressible flow is a vast subject. Texts such as Saad24
and Anderson25 cover the subject in far more detail than space
allows here. The purpose of this section is to describe enough
of the subject that the reader understands the key assumptions
behind the use of compressible flow relations in computing
orifice flows. The choked orifice is an important flow mea-
surement device because of its relative simplicity. In addi-
tion, many gas-fired burners and flares fire fuel through
choked orifices. A further purpose is to give the reader some
appreciation for the interesting flow phenomena, especially
shock waves that occur in compressible flows with little or
FIGURE 4.26 Flow around the air intake of a jet engine
in supersonic flow. (From Visualization Society of Japan, The no analog in incompressible flows.
Fantasy of Flow, IOS Press, Amsterdam, 1993, 23.)
4.4.2.1 Basic Thermodynamics Relations
This section is limited to ideal gases, that is, gases that obey
with the surrounding fluid. While it is true that it is entraining the relation P = ρRT, where P is the absolute pressure, T is
more mass, it is not mixing any faster — downstream con- the absolute temperature, and R is the gas constant (the
centrations are not affected by velocity. universal gas constant divided by the gas molecular weight)
From these equations, the mass entrainment rate can also (see Section 4.2.4.1). For an ideal gas, one can readily show
be deduced. The following equation 23 summarizes mass that the enthalpy, h, and internal energy, e, are functions of
entrainment: temperature alone.15 One can also show that the specific heats

me x  ρa 
= 0.32  ρ  −1 (4.84) ∂h  ∂s 
m0 d0  0 Cp = = T  (4.85)
∂T P
 ∂T  P

where
me = Mass flow entrained by jet and
m0 = Initial mass flow from nozzle
x = Distance from nozzle exit
∂e  ∂s 
d0 = Diameter of nozzle Cv = = T  (4.86)
∂T  ∂T  v
ρa = Density of ambient fluid v

ρ0 = Density of initial jet fluid

where v is the specific volume, are also functions of tempera-
4.4.2 Compressible Flow ture only. The subscript notation indicates that differentiation
All gases and liquids are compressible to some extent. The is performed holding the specified variable constant. Since
subject of this section, however, is the flow of fluids in which h = e + P/ρ = e + RT, dh = de + RdT. For an ideal gas,
the density varies significantly due to pressure gradients dh = CpdT and de = CvdT. Combining these relations and
within the flow. For engineering purposes, it is frequently dividing by dT reveals that Cp = Cv + R, a useful relationship
assumed that a flow is incompressible as long as the density that will be exploited later.
change is less than about 5%. Note here that density changes The Maxwell relation of classical thermodynamics is:
due to compression (or pressure forces) are being discussed.
The density of a flowing fluid can also change because of tem-
perature changes, such as in a reacting flow or a flow with heat dh = Tds + dP ρ (4.87)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 145 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 145

It applies to any “pure simple compressible substance.” This

means any substance composed of a single chemical species
whose only mode of doing work is PdV work. Air is also
considered a pure simple compressible substance because it
is a nonreacting mixture of molecules that acts like a single
species.26 The Maxwell relation is important in the present
context because it allows one to derive ideal gas relations for a
constant entropy fluid. If entropy is constant, then Maxwell’s
relation reduces to dh = dP/ρ or CpdT = dP/ρ. Manipulation of
this relation using the ideal gas law leads to the following rela-
tions for an ideal gas undergoing an isentropic change from
state 1 to state 2: FIGURE 4.27 Shock waves from supersonic fighter.
(From Visualization Society of Japan, The Fantasy of Flow,
IOS Press, Amsterdam, 1993, 59.)
γ γ ( γ −1)
P2  ρ2  T 
=  = 2 (4.88)
P1  ρ1   T1 
dynamic problems, where length scales are considerably
larger (typically tens of feet of modern supersonic aircraft).
where γ is the ratio of specific heats, Cp/Cv . For this reason, compressible flows are almost always
analyzed as inviscid flows. The term “inviscid” does not
4.4.2.2 Compressible Flow Concepts imply that the fluid itself has no viscosity, as air and other
The Mach number is the ratio of a fluid’s velocity (measured gases are viscous. Rather, the term implies that viscous
relative to some obstacle or geometric feature), divided by the effects are unimportant. Although the flow fields are mostly
speed at which sound waves propagate through the fluid. A inviscid, it is known that in the region near solid surfaces
sound wave is a very weak pressure disturbance in a fluid. (called the boundary layer), viscous forces are very impor-
This small pressure wave in the fluid is accompanied by a tant. Several clever analysis techniques have been developed
density disturbance and a velocity disturbance. Analysis of the to asymptotically patch the inviscid far-field solution to a
sound wave is straightforward and is given by Anderson.25 boundary-layer solution. The interested reader should see
The result is that the sound speed for an ideal gas is: Panton1 for details.

4.4.2.3 Quasi-one-dimensional Isentropic Flow

c= (∂P ∂ρ) s = γRu T M
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

A quasi-one-dimensional flow is a flow that is “nearly” one-

dimensional. A truly one-dimensional flow must have a con-
For example, air at 77°F (25°C) has a sound speed of 1135 ft s–1 stant flow area. Here, the flow area is allowed to vary slowly
(346 m s–1 or 774 miles per hour). with the x-coordinate. The y and z variations are neglected
Consider a fluid flowing around an obstacle, such as a cir- because the area varies gradually. Flow variables P, ρ, T, and
cular cylinder placed in the flow. If the fluid is flowing at a u are treated as functions of x only. The isentropic assump-
very low speed compared to the sound speed, then the tion implies that the flow is adiabatic. The following sum-
upstream fluid is able to “sense” (via pressure waves) the mary is from Saad,24 Howell,26 and Ward-Smith.27
presence of the circular cylinder and adjust the flow well ahead
of the cylinder. If the speed of the fluid is greater than its sound 4.4.2.3.1 Basic Relations
speed, no such adjustment can be made. Figure 4.26 shows
The x-direction component of the Navier-Stokes equation can
shock waves at the sharply pointed air-intake of a jet engine,
be used alone to describe a truly one-dimensional flow.
and Figure 4.27 shows shock patterns from a supersonic jet.
Because the flow in this section is a quasi-one-dimensional
Compressible flows are typically very high Reynolds number
flow, relations must be developed that correctly account for
flows. For example, because air has a sound speed of 1135 ft/s
the fact that the flow area is changing. Anderson25 has shown
(346 m/s), air flowing at sonic conditions through a 1/8-in.
that momentum and energy conservation can be written as:
(3.2 mm) orifice would have a jet Reynolds number of about
66,000. This is well into the turbulent regime. The subject
of compressible flows has traditionally focused on aero- dP + ρudu = 0 (4.89)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 146 Wednesday, April 6, 2005 3:41 PM

146 The John Zink Combustion Handbook

and downstream location. If one considers a calorically perfect

gas (specific heats independent of temperature), then equating
dh + udu = 0 (4.90) the stagnation enthalpy to the enthalpy at any location

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
(denoted by subscript ‘1’) within the duct gives:
Maxwell’s relation, Tds = dh + dP/ρ, shows that these two
relations are equivalent in an isentropic flow. The differential ho = h1 + u12 2 (4.92)
energy equation shown above can be integrated once to show
that the total enthalpy, ho = h + u2/2, is a constant (ho is the because the stagnation velocity is zero. And because the specific
stagnation enthalpy, defined below). heat is constant, one notes that ho − h1 = Cp (To − T1 ) = u12 2 .

4.4.2.3.2 Flow Through Converging-Diverging Nozzle M12 Cp (To − T1 )

Dividing by the sound speed squared gives . =
Conservation of mass in a variable-area duct requires that γRT1 2
d(ρuA) = 0. Expanding this differential and manipulating From thermodynamics, one knows that Cp – Cv = R. Because
shows that dρ/ρ + du/u + dA/A = 0. One can write dρ as γ = Cp/Cv , one can readily determine that Cp/(γR) = 1/(γ – 1).
dP(dρ/dP). Recall that the definition of the sound speed is Now the temperature and Mach number can be related by
c2 = (dP/dρ)s. Using this relation and applying the differential
momentum equation given above allows one to write energy To γ −1
conservation as –u du/c2 + du/u + dA/A = 0. The Mach num- = 1 + M12 (4.93)
T1 2
ber is defined as M ≡ u/c, and further manipulation allows
mass conservation to be written as: Using the thermodynamic relations for ideal gases under-
going isentropic processes, one can determine additional

( )
du dA relations that relate the pressure at any location (not sub-
M2 − 1 = . (4.91)
u A scripted now) to the Mach number:

γ −1 2 ( )
This is the area-velocity relation. This equation is significant γ γ −1
Po 
= 1+ M (4.94)
in the study of compressible flow. Notice, for example, that P  2 
when the Mach number is less than 1, area and velocity
changes have opposite signs. This is the intuitive result that if and
the mass flow is constant, decreasing the available flow area
γ −1 2 ( )
1 γ −1
must increase velocity. However, if the Mach number is ρo 
= 1+ M (4.95)
greater than one, the relation gives a counterintuitive result: ρ  2 
that to accelerate a flow, one must increase the available flow
area. For this reason, a nozzle that is designed to accelerate a At the throat, the Mach number is 1; thus, the above rela-
flow from rest (or low speed) to supersonic velocities must be tions [Eqs. (4.93), (4.94), and (4.95)] become, respectively:
a converging-diverging nozzle. Its area must decrease to
accelerate the flow to a Mach number of 1, and then the area To γ −1
= 1+ (4.96)
must begin increasing to accelerate the flow to supersonic T* 2
(M > 1) velocities. The minimum area of the duct is called
γ − 1 ( )
the throat of the duct; at this location, the flow is sonic. γ γ −1
Po 
= 1+ (4.97)
The analysis of flows is simplified by defining stagnation P *  2 
conditions (where the flow velocity is zero) and throat con-
ditions (where the flow is sonic). Following Anderson,26 the and
subscript o is used for stagnation conditions and the super-
γ − 1 ( )
1 γ −1
script * for throat conditions. Because the temperature varies ρo 
= 1 + (4.98)
along the length of the nozzle, the sound speed c varies with x. ρ*  2 
As long as the flow is isentropic, the stagnation properties
and throat properties do not change. These relations allow the computation of fluid properties
Because the flow is isentropic, the thermodynamic relations (u, ρ, P, and T) at any location within the duct. A common
shown above for an isentropic process between two states application of these equations is to ensure that a supersonic
can be applied between the stagnation conditions and any nozzle is properly expanded such that the fluid pressure at the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 147 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 147

FIGURE 4.28 Choked flow test rig.

exit plane is the same as the surrounding pressure. A nozzle through a small orifice where the pressure downstream of the
is termed “overexpanded” if the nozzle expands so that the orifice is the back pressure, Pb, there are four cases to con-
pressure at the exit plane is lower than the surroundings. In sider, as shown in Figure 4.29.
this condition, either a normal shock wave will develop
1. If the back pressure and the stagnation pressure are equal,
within the expanding portion of the nozzle, or a series of there is no flow through the orifice (i.e., the valve is closed).
oblique shock waves will appear downstream of the exit 2. If the back pressure is lower than the stagnation pressure,
plane. A nozzle is “underexpanded” if the pressure at the exit but above the choking pressure, then there will be flow
plane is higher than the surrounding pressure. Prandtl-Meyer through the orifice at a Mach number less than 1. The
expansion waves (discussed in the compressible flow texts mass flow rate continues to increase as the back pressure
cited in the introduction to this section and shown in is lowered.
Figure 4.26) are two-dimensional flow structures that will 3. If the back pressure is lowered to a value termed the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

appear in this situation. choking pressure (Pc), then the velocity at the orifice will
become sonic. Because the velocity is sonic, further
The mass flow at any location is ṁ = ρuA . The mass flow decreases in the back pressure cannot be “communicated”
rate is independent of the x-location, and thus one can write upstream, and the mass flow rate through the orifice can
no longer change. The mass flow at this pressure is the
ρuA = ρ*u * A* . The throat velocity is u * = c * = γRT * .
maximum (in reality, the mass flow will continue to
Manipulation of these expressions using the above derived increase somewhat because the assumption of quasi-one-
expressions for ρo/ρ*, ρo /ρ, To/T*, and To/T yields the area dimensional flow is an idealization; see Ward-Smith27
Mach number relation: for details).
4. As the back pressure is lowered below the choking pressure,
( γ +1) ( γ −1) the mass flow rate does not continue to increase in the ideal
 A  = 1  2 1 + γ − 1 M 2  
2
analysis. In reality, a visually striking series of Prandtl-
M 2  γ + 1   
(4.99)
A 
*
2 Meyer expansion waves, seen also in Figure 4.27, appear
just downstream of the exit plane, producing a diamond
4.4.2.3.3 Choked Orifice Flows pattern of high- and low-pressure zones. The density gra-
dients in the flow cause the fluid’s index of refraction to
This section examines the flow of an ideal gas through an
vary so that these diamond patterns are visible. It should
orifice. Compressible orifice flows have numerous applica-
be noted that as a practical matter, valves are not normally
tions in petrochemical combustion. Ward-Smith 27 provides a placed downstream of a choked orifice. Also, while it is
thorough overview of the use of choked orifices as flowmeters. true that for a given upstream pressure, mass flow cannot
Consider a fluid in a large reservoir at (stagnation) pressure be increased, if the upstream pressure is increased, the
P0 , as shown in Figure 4.28. If this fluid is allowed to pass mass flow will increase due to the increase in density.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 148 Wednesday, April 6, 2005 3:41 PM

148 The John Zink Combustion Handbook

FIGURE 4.29 Types of flow of a fluid passing through a small orifice.

Given the stagnation pressure and the back pressure, the exit become significant. The analysis presented here is not appro-
Mach number can be computed using the isentropic relation priate for these high l/d orifices as viscous effects have been
for Po/P. However, if this computation reveals that the exit neglected. Viscous effects can readily be included in the anal-
Mach number is greater than 1, it is realized that a simple orifice ysis. Compressible flow with viscous effects in constant area
geometry will not accelerate the flow to supersonic velocities ducts is called Fanno flow and is discussed at length in standard
and the flow will be choked. The mass flow rate can be com- compressible flow texts such as Saad 24 and Anderson.25
puted as m˙ = ρ*u * Aorifice . For example, consider air (γ = 1.4)
supplied with stagnation pressure 30 psig (44.7 psia or 2 barg)
and stagnation temperature 70°F (21°C) exhausted through a
0.25-in. (6.4-mm) orifice. Equation (4.87) gives the exit Mach 4.5 PRESSURE DROP
number (assuming atmospheric pressure is 14.7 psia) as 1.36, FUNDAMENTALS
indicating that the flow is choked, which means that the Mach
number at the throat is 1. The throat conditions can then be
computed using Eqs. (4.86) to (4.88) utilizing a Mach number
4.5.1 Basic Pressure Concepts
of 1. The throat pressure is 23.6 psia (1.6 barg). The throat 4.5.1.1 Definition of Pressure
temperature is –18°F (–28°C), and the throat density is
Pressure is created by the collision of gas molecules on a sur-
0.144 lbm/ft3. The sound speed at this temperature is 1029 ft/s
face and is defined as the force exerted per unit area on that
(314 m/s). The (ideal) mass flow through this orifice is then
surface. The pressure of the air around us, called atmospheric
0.051 lbm/sec (=ρ*u*A). If the flow is not choked (the exit Mach
number is less than 1), then the ideal mass flow can be com- pressure, is due to air molecules colliding with our bodies,
puted using the fluid properties at the exit Mach number. objects, and the Earth’s surface. Lower elevations at the
In computing real orifice flows, it is common to use a Earth’s surface will tend to have a higher atmospheric pres-
sure because the height, and therefore the weight, of the col-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

discharge coefficient (Cd). The discharge coefficient is defined

as Cd ≡ m˙ actual m˙ ideal . Ward-Smith27 discusses discharge umn of air above us is greater. A higher column of air above
coefficients for various types of orifices. In general, the dis- the surface of the Earth will compress the air more, causing
charge coefficient for orifices is between 0.4 and 0.9. Fortu- an increase in the number of collisions of air molecules at the
nately, discharge coefficients do not vary significantly with Earth’s surface. This increase in the number of collisions is
Mach number. They are a function of orifice geometry, the mechanism responsible for a higher atmospheric pres-
particularly the l/d of the orifice. For very thin-walled orifices sure. For example, at sea level, the atmospheric pressure is
(i.e., the l/d ratio is less than 1), the discharge coefficient will about 14.7 pounds per square inch (psi) or 1 bar. However, in
be very low. As the l/d ratio is increased to somewhere between Denver, Colorado, which is approximately 1 mile (1.6 km)
4 and 7, the discharge coefficient may increase to 0.85 or more. above sea level, the atmospheric pressure is about 12.2 psi
Further increases in the l/d ratio cause viscous effects to (0.83 bar).
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 149 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 149

4.5.1.2 Units of Pressure TABLE 4.2 Properties of U.S. Standard Atmosphere

Pressure is generally measured in units of Pascals (Pa = N/m2). at Sea Level
However, flare and burner system designers will use a variety Property SI Units English Units
of pressure units because the choice of units varies, depend- Pressure 101.33 kPa (abs) 14.696 lb/in.2 (abs)
ing on the customer. For example, customers in the United Temperature 15.0°C 59.0°F
States will typically use units of psi or inches of water col- Density 1.225 kg/m3 0.07647 lb/ft3

umn, whereas in Europe and Asia the units are usually Pa and
millimeters of water column. The conversion of these pres-
sure units are as follows: 1 lb/in2 = 27.68 in. water column = 1. gas flow in a flare stack and header
6895 Pa = 703.072 mm water column. 2. steam flow in a pipe feeding a steam-assisted flare
3. air from a blower feeding an air-assisted flare
4.5.1.3 Standard Atmospheric Pressure 4. gas flow through a burner manifold
Initially, engineers developed a standard atmospheric pres- 5. oil flow through a pipe feeding an atomizing gun
sure so that the performance of aircraft and missiles could be
evaluated at a standard condition. The idea of a standard Although all of these systems are different, the governing
atmospheric pressure was first introduced in the 1920s.28 In equations used to describe the pressure drop are common.
1976, a revised report was published that defined the U.S. The purpose of this section is to discuss the basic concepts
standard atmosphere that is the currently accepted standard. used to determine the pressure drop of a liquid or gas flowing
This standard is an idealized representation of the mean con- through a straight pipe.
ditions of the Earth’s atmosphere in one year. The roughness of the wall on the inside of a pipe influences
Table 4.2 lists several important properties of air for stan- the pressure drop of a fluid flowing through it if the flow is
dard atmospheric conditions at sea level. turbulent. When a fluid flows turbulent through a straight pipe,
organized structures of fluid near the wall called slow-moving

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
4.5.1.4 Gage and Absolute Pressure streaks, can suddenly move into the central region of the pipe
If the Earth was in a perfect vacuum, there would be no col- by a phenomenon called a “turbulent burst.” Figure 4.30 is a
umn of air above its surface; hence, the atmospheric pressure photograph showing the organized structure of a series of
would be zero. The absolute pressure is measured relative to slow-moving streaks near a wall.29
a perfect vacuum. Therefore, when a pressure measurement When these slow-moving streaks burst from the wall,
is taken at the surface of the Earth, the absolute pressure is momentum interchanges between masses of fluid, extracting
equal to the atmospheric pressure. When writing the units of energy from the overall fluid in the form of heat. A pipe with
pressure, it is customary to designate absolute pressure with a rough wall will experience more turbulent bursts than a
the letter “a” or “abs” after the units. For example: psia, kPa, pipe with a smooth wall operating under the same flowing
psi(abs), or kP(abs). The absolute pressure can never be less conditions. Therefore, a pipe with a rough wall will have a
than zero; however, the gage pressure can. larger pressure drop associated with it than a pipe with a
The gage pressure is always measured relative to the atmo- smooth wall.
spheric pressure. A gage pressure of less than zero can exist. The pressure drop for a fully developed flow of fluid in a
For example, suppose there is a sealed container that holds a pipe can be calculated by relating the velocity pressure of the
vacuum at 10 psia at sea level. The gage pressure, which is fluid to the pipe roughness and geometry as follows:
measured relative to the absolute pressure, would be (10 psia –
14.7 psia) = –4.7 psig. The letter “g” after the pressure units L ρV 2
represents gage pressure. Now suppose the container is pressur- ∆P = f (4.100)
D 2 gc
ized to 20 psia. The gage pressure will then be (20 psia –
14.7 psia) = 5.3 psig. Thus, the gage pressure can either be a
where the term ∆P is the pressure drop, f is the friction fac-
positive or negative number, and is just the difference in pressure
tor, L is the length of pipe, D is the inside pipe diameter,
between the atmospheric pressure and the pressure of interest.
ρ is the density of the fluid in the pipe, and V is the mean
velocity of the fluid in the pipe. Equation (4.100) is called
4.5.2 Roughness the Darcy-Weisbach equation, named after two engineers of
The transport of a liquid or gas in a pipe system is very com- the nineteenth century. Weisbach first proposed the use of
mon in the flare and burner industry. Such applications include: the friction factor term, f, and Darcy conducted numerous
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 150 Wednesday, April 6, 2005 3:41 PM

150 The John Zink Combustion Handbook

FIGURE 4.30 Photograph showing slow-moving streaks near a wall using the hydrogen bubble technique. (From
W. Bussman, A theoretical and experimental investigation of near-wall turbulence in drag reducing flows, Ph.D. thesis, The
University of Tulsa, Tulsa, OK, 1990.)

pressure drop experiments using water flowing through VD

pipes with various wall roughness. Re = (4.102)
ν
The friction factor is defined as the dimensionless fluid
friction loss, in velocity heads, per diameter length of pipe. where ν is the kinematic viscosity of the fluid. Again, for
In 1933, Nikuradse conducted several experiments using arti- fully developed laminar flow in a pipe, the friction factor —
ficially roughened pipes made by attaching grains of sand, and hence pressure drop — are independent of the roughness
of known size, to the inside of a pipe wall.30 Nikuradse’s of the pipe wall.
experiments revealed two important characteristics of flow For fully developed turbulent flow in a pipe, the value of
through rough pipes. First, if the flow is laminar, the pressure the friction factor can be determined using experimental data.
drop through the pipe is independent of the roughness of the Figure 4.31 is a plot, based on experimental data, showing f
pipe wall. Second, for highly turbulent flows, the friction as a function of Re and ε/D. This plot is called the Moody
factor is only dependent on the diameter of the pipe, D, and chart or Moody diagram, in honor of L.F. Moody, who cor-
the height of the sand grains, ε. Nikuradse defined a non-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
related Nikuradse’s original data in terms of the relative
dimensional term, called the relative roughness, as ε/D. roughness with commercial pipe material.
For fully developed laminar flow in a pipe, the friction Notice that, for flow in the laminar regime, the friction factor
factor is independent of the relative roughness and can be is independent of the relative roughness. This straight line, in
simply written as: the laminar flow region, is a plot of Eq. (4.101). Also notice
that, for very high Reynolds numbers, the friction factor no
64 longer becomes a function of the Reynolds number as discov-
f = (4.101) ered by Nikuradse. However, for flows with moderate Reynolds
Re
numbers, the friction factor is a function of both the Reynolds
The term “Re” in Eq. (4.101) is the Reynolds number, number and the relative roughness. Also note that, for a smooth
defined as: pipe (ε = 0), the friction factor is not zero. Therefore, regardless
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 151 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 151

FIGURE 4.31 Moody diagram. (From J.A. Roberson and C.T. Crowe, Engineering Fluid Mechanics, Houghton Mifflin,
Boston, MA, 1980.)

of the smoothness of a pipe wall, there will always be a pressure TABLE 4.3 Equivalent Roughness for New Pipes
drop as the fluid flows through the pipe. There is no such thing Equivalent Roughness, ε
as a perfectly smooth wall. On a microscopic level, a wall will Pipe (ft) (mm)
always have a surface roughness. Riveted steel 0.003–0.03 0.9–9.0
Concrete 0.001–0.01 0.3–3.0
Values of relative roughness are available for commercially Wood stave 0.0006–0.003 0.18–0.9
manufactured pipe. Typical roughness values, for various pipe Cast iron 0.00085 0.26
Galvanized iron 0.0005 0.15
material and surfaces, are provided in Table 4.3. Commercial steel or wrought iron 0.00015 0.045
Drawn tubing 0.000005 0.0015
It should be mentioned that the buildup of corrosion or scale
Plastic, glass 0.0 (smooth) 0.0 (smooth)
on the inside of a pipe can significantly increase the relative
Source: J.A. Roberson and C.T. Crowe, Engineering Fluid Mechanics,
roughness. Very old pipes can also be so badly eroded away Houghton Mifflin, Boston, MA, 1980.
on the inside that the effective diameter of the pipe is altered.
Several researchers have attempted to develop an analytical
expression for the friction factor as a function of the Reynolds 4.5.4 Loss Coefficient
number and relative roughness. One well-known equation is The previous section (Section 4.5.2) presented equations to
the Colebrook formula: calculate the pressure drop for a fully developed flow of a fluid
through a straight pipe having a constant cross-sectional area.
In the flare and burner industry, however, it is very common for
1 ε D 2.51 
= −2.0 log +  (4.103) pipe systems to include inlets, elbows, tees, and other fittings
f  3.7 Re f  that can create additional pressure losses. The methods and
procedures used for determining the pressure drop through
This formula is typically used to generate curves in the Moody fittings, however, are not as convenient as for straight pipe
diagram. The difficulty in using the Colebrook formula flow. Pressure losses through fittings are the result of addi-
[Eq. (4.103)] is that, in order to solve for the friction factor f, tional turbulence and/or flow separation created by sudden
an iterative scheme must be used. This is not too difficult, changes in the fluid momentum. Therefore, pressure losses are
however, if a computer is used. significantly influenced by the geometry of the fitting. This
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

152 The John Zink Combustion Handbook

TABLE 4.4 Loss Coefficients for Various Fittings

Description Sketch Additional Data KL

Contraction D2/D1 φ = 60° φ = 180°

0.0 0.08 0.50

0.20 0.08 0.49
0.40 0.07 0.42
0.60 0.06 0.32
0.80 0.05 0.18
∆P = KL V2/2 gc 0.09 0.04 0.10

Expansion D1/D2 φ = 10° φ = 180°

0.0 1.00
0.20 0.13 0.92
0.40 0.11 0.72
0.60 0.06 0.42
∆P = KL V2/2 gc 0.80 0.03 0.16

90° Smooth Bend

r/d KL

1 0.35
2 0.19
4 0.16
6 0.21
8 0.28
10 0.32
∆P = KL V2/2 gc

Source: J.A. Roberson and C.T. Crowe, Engineering Fluid Mechanics, Houghton Mifflin, Boston, MA, 1980.

section discusses the general procedure for estimating the pres- approaching fluid stream, ρV2/2. The loss coefficient is
sure drop through fittings in piping systems. strongly dependent on the geometry of the fitting and the
A complete theoretical analysis for calculating the flow Reynolds number in the pipe approaching the fitting. The
through fittings has yet to be developed. Thus, the pressure loss coefficient for various fittings is given in Table 4.4. For
drop through fittings is based on equations that rely heavily additional information on loss coefficients through various
on experimental data. The most common method used to fittings, refer to Idelchik12 and Crane.31
determine the pressure loss is to specify the loss coefficient,
KL, defined as follows: 4.5.3 Discharge Coefficient
∆Pgc Flare and burner engineers use equations based on the ideal
KL = (4.104) gas law and assumptions of ideal flow to calculate the flow
1
ρV 2 rate of a fluid through a burner nozzle or flare tip. To com-
2
pensate for the results of these ideal equations and assump-
Notice that the loss coefficient is dimensionless and is tions, a constant is introduced to account for the complexity
defined as the ratio of the pressure drop through a fitting to of the flow that makes it non-ideal. This constant is called the
the approaching velocity pressure of the fluid stream. Solving discharge coefficient.
Eq. (4.104) for ∆P relates the pressure drop through a fitting: The discharge coefficient is defined as the ratio of the actual
mass flow rate of a fluid through a nozzle to the ideal mass
ρV 2
∆P = K L (4.105) flow rate and is written as:
2 gc

If the loss coefficient is equal to 1.0, then the pressure loss m˙ actual
Cd =
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

(4.106)
through that fitting will equal the velocity pressure of the m˙ ideal
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 153 Wednesday, April 6, 2005 3:41 PM

Fundamentals of Fluid Dynamics 153

The ideal mass flow rate is defined as the mass flow rate cal-
culated using the ideal gas law and assumptions of ideal flow
L
— no pressure losses due to the internals of the nozzle or tip.

a
The value of the discharge coefficient for a burner nozzle or
flare tip must be determined experimentally. Typically, the
discharge coefficient varies from about 0.60 to 1.0, with 1.0
being ideal in most burner and flare applications. Factors that
D d V
can affect the discharge coefficient include:
1. length-to-diameter ratio of the port
2. the Reynolds number of the fluid in the port
3. beta ratio Length-to-diameter ratio = L / d
4. port angle Reynolds number = V x d / n
5. manufacturing tolerances Beta ratio = d / D
See Figure 4.32 for a description of these variables. Port angle = a

FIGURE 4.32 Factors that can affect the discharge

coefficient.
REFERENCES

1. R.L. Panton, Incompressible Flow, John Wiley and

12. I.E. Idelchik, Handbook of Hydraulic Resistance,
Sons, New York, 1984.
Hemisphere Publishing, Washington, D.C., 1986.
2. F.M. White, Viscous Fluid Flow, McGraw-Hill, New
York, 1991. 13. M.C. Potter and J.F. Foss, Fluid Mechanics, Great
Lakes Press, Okemos, MI, 1975.
3. R.W. Fox and A.T. McDonald, Introduction to Fluid
Mechanics, 2nd ed., John Wiley & Sons, New York, 14. K.K. Kuo, Principles of Combustion, John Wiley &
1978. Sons, New York, 1986.

4. J.K. Vennard and R.L. Street, Elementary Fluid 15. G. Van Wylen and R. Sonntag, Fundamentals of Classi-
Mechanics, 5th ed., John Wiley & Sons, New York, cal Thermodynamics, 2nd ed., John Wiley & Sons,
1975. New York, 1973.
5. R.D. Reed, Furnace Operations, 3rd ed., Gulf Publish- 16. R.B. Bird, W.E. Stewart, and E.N. Lightfoot, Transport
ing, Houston, 1981. Phenomena, John Wiley & Sons, New York, 1960.
6. J.B. Maxwell, Data Book on Hydrocarbons, D. Van 17. O. Reynolds, An experimental investigation of the cir-
Nostrand Company, Princeton NJ, from the Standard cumstances which determine whether the motion of
Oil Development Company, 1950. water shall be direct or sinuous and of the law of resis-
tance in parallel channels, Phil. Trans. Roy. Soc.,
7. E.L. Lederer, Proceedings of the World Petrochemical
174(III), 935, 1883.
Congress, London, 1933, 526.
18. Mark’s Standard handbook for Mechanical Engineers,
8. W.R. Shu, A viscosity correlation for mixtures of heavy
9th ed., McGraw-Hill, New York, 1987.
oil, bitumen, and petroleum fractions, Soc. Petr. Engr.
J., June 1984, 272. 19. J.H. Lienhard, A Heat Transfer Textbook, Prentice-Hall,
Englewood Cliffs, NJ, 1987.
9. R.J. Kee, F.M. Rupley, and J.A. Miller, The CHEMKIN
Thermodynamic Data Base, Sandia Report SAND87- 20. H. Schlichting, Boundary Layer Theory, McGraw-Hill,
8215B, March 1990. New York, 1979.
10. J.M. Smith, H.C. Van Ness, and M.M. Abbott, Introduc- 21. J.O. Hinze, Turbulence, Classic Textbook Re-issue
tion to Chemical Engineering Thermodynamics, 5th ed., series, McGraw-Hill, New York, 1987.
McGraw-Hill, New York, 1996. 22. H. Tennekes, A First Course in Turbulence, MIT Press,
11. M. Modell and R. Reid, Fundamentals of Classical Cambridge, MA, 1973.
Thermodynamics, 2nd ed., John Wiley & Sons, New 23. J.M. Beer and N.A. Chigier, Combustion Aerodynam-
York, 1973. --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- ics, Krieger Publishing, Malabar, FL, 1983.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch04-Frame Page 154 Wednesday, April 6, 2005 3:41 PM

154 The John Zink Combustion Handbook

24. M.A. Saad, Compressible Fluid Flow, Prentice-Hall, 31. Crane Engineering Division, Flow of Fluids through
Englewood Cliffs, NJ, 1985. Valves, Fitting, and Pipe, Crane Co., New York, 1969.
25. J.D. Anderson, Modern Compressible Flow with His- 32. B. Humiston, General Chemistry — Principles and
torical Perspective, McGraw-Hill, New York, 1982. Structure, John Wiley & Sons, New York, 1975,
169–177.
26. J.R. Howell and R.O. Buckius, Fundamentals of Engi-
33. O. Reynolds, On the dynamical theory of imcompress-
neering Thermodynamics, 2nd ed., McGraw-Hill, New
ible viscous fluids and the determination of the crite-
York, 1992.
rion, Phil. Trans. Roy. Soc., 186, A123–164, 1895; and
27. A.J. Ward-Smith, Critical flowmetering: the character- Sci Papers I, 355.
istics of cylindrical nozzles with sharp upstream edges, 34. Visualization Society of Japan, The Fantasy of Flow,
Int. J. Heat and Fluid Flow, 1, 123–132, 1979. IOS Press, Amsterdam, 1993.
28. R.R. Munson, D.F. Young, and T.H. Okiishi, Funda- 35. M. Van Dyke, An Album of Fluid Motion, The Para-
mentals of Fluid Mechanics, John Wiley & Sons, New bolic Press, Stamford, CA, 1982.
York, 1990, 52. 36. F. Kreith, The CRC Handbook of Mechanical Engi-
29. W. Bussman, A Theoretical and Experimental Investi- neering, CRC Press, Boca Raton, FL, 1998.
gation of Near-Wall Turbulence in Drag Reducing 37. J.A. Roberson and C.T. Crowe, Engineering Fluid
Flows, Ph.D. thesis, The University of Tulsa, Tulsa, Mechanics, Houghton, Mifflin, Boston, MA, 1980.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
OK, 1990.
30. J. Nikuradse, Stromungsgesetze in rauhen Rohren,
VDI-Forschungsheft, 1933, 361. Translation available
in N.A.C.A. Tech. Memorandum, 1292.

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 5
Fuels
Terry Dark, John Ackland, and Jeff White

TABLE OF CONTENTS
5.1 Gaseous Fuels ......................................................................................................................................... 158
5.1.1 Introduction............................................................................................................................... 158
5.1.2 Natural Gas ............................................................................................................................... 158
5.1.3 Liquified Petroleum Gas (LPG) ................................................................................................ 159
5.1.4 Refinery Gases .......................................................................................................................... 159
5.1.5 Combustible Waste Gas Streams .............................................................................................. 160
5.1.6 Physical Properties of Gaseous Fuels ....................................................................................... 165
5.1.7 Photographs of Gaseous Fuel Flames ....................................................................................... 165
5.2 Liquid Fuels ............................................................................................................................................ 165
5.2.1 Introduction and History ........................................................................................................... 165
5.2.2 Oil Recovery ............................................................................................................................. 173
5.2.3 Production, Refining, and Chemistry........................................................................................ 175
5.2.4 Oils............................................................................................................................................ 178
5.2.5 Liquid Naphtha ......................................................................................................................... 179
5.2.6 Physical Properties of Liquid Fuels .......................................................................................... 179
5.3 Gas Property Calculations....................................................................................................................... 183
5.3.1 Molecular Weight...................................................................................................................... 183
5.3.2 Lower and Higher Heating Values ............................................................................................ 184
5.3.3 Specific Heat Capacity.............................................................................................................. 184
5.3.4 Flammability Limits ................................................................................................................. 184
5.3.5 Viscosity.................................................................................................................................... 185
5.3.6 Derived Quantities .................................................................................................................... 185
5.4 Typical Flared Gas Compositions ........................................................................................................... 185
5.4.1 Oil Field/Production Plant Gases.............................................................................................. 186
5.4.2 Refinery Gases .......................................................................................................................... 186
5.4.3 Ethylene/Polyethylene Gases.................................................................................................... 186
5.4.4 Other Special Cases .................................................................................................................. 186
References ................................................................................................................................................................ 187
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
157
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 158 Monday, March 14, 2005 12:29 PM

158 The John Zink Combustion Handbook

5.1 GASEOUS FUELS TABLE 5.1 Example Pipeline Quality Natural Gas
Minimum Maximum
5.1.1 Introduction Major and Minor Components, vol%
The term “gaseous fuel” refers to any combustible fuel that Methane 75% —
exists in the gaseous state under normal temperatures and Ethane — 10.0%
Propane — 5.0%
pressures. Gaseous fuels are typically composed of a wide Butane — 2.00%
range of chemical compounds. Low boiling point hydrocar- Pentane and heavier — 5.00%
bons (both paraffins and olefins), hydrogen, carbon monox- Nitrogen and other inerts — 3–4%
Carbon dioxide — 3–4%
ide, and inert gases (nitrogen and carbon dioxide) are among
Trace Components —
the many chemical constituents of common gaseous fuels. Hydrogen sulfide — 0.25–1.0 grains/100 scf
The purpose of this section is to introduce many of the com- Mercaptan sulfur — 0.25–1.0 grains/100 scf
mon fuel gas mixtures used as fuel in the hydrocarbon and Total sulfur — 5–20 grains/100 scf
Water vapor — 7.0 lb/mmcf
petrochemical industries. Commonly occurring waste gas Oxygen — 0.2–1.0 ppmv
mixtures in flare systems are also described. Other characteristics
Heating value, Btu/scf-gross saturated 950 1150
5.1.2 Natural Gas Liquids: Free of liquid water and hydrocarbons at delivery temperature and
pressure.
Natural gas is a gaseous fossil fuel that is formed naturally Solids: Free of particulates in amounts deleterious to transmission and utilization
beneath the Earth and is typically found with or near crude oil equipment.
reservoirs. Proven natural gas reserves in the United States in Adapted from Gas Processors and Suppliers Association, GPSA Engineering
1993 totaled approximately 4.58 × 1012 m3 (1.62 × 1014 ft3).1 Data Book, Vol. 1, 10th ed., Tulsa, OK, 1987. With permission.
In 1989, the U.S. Department of Energy estimated total natu-
ral gas consumption in the United States at 19.384 quadril-
lion Btu, 23.8% of the total U.S. energy consumption.2 increased corrosion rates, formation of solid hydrate com-
Natural gas consists of a fluctuating range of low boiling pounds that can restrict or interrupt gas flow, and freezing of
point hydrocarbons. Methane is the primary chemical com- valves and regulators during cold weather conditions.5 Tech-
ponent, and can be present in amounts ranging from 70 to niques for the dehydration of natural gases include:
99.6% by volume. Ethane can be present in amounts ranging 1. Absorption with liquid desiccants: glycols (typically tri-
from 2 to 16% by volume. Carbon dioxide, nitrogen, hydro- ethylene glycol) are used to absorb water vapor via coun-
gen, oxygen, propane, butane, and heavier hydrocarbons are tercurrent-flow, packed-bed absorption columns.6
also typically present in the fuel analysis.3 The exact analysis 2. Adsorption with solid desiccants: water vapor is adsorbed
usually varies somewhat depending on the source of the gas onto a bed of inorganic porous solid material (silica gel,
and on any heating value adjustments or supplementation. alumina, molecular sieves, etc.).6,7
Natural gas quality specifications have historically been 3. Dehydration with calcium chloride: solid anhydrous
negotiated in individual contracts between the natural gas calcium chloride (CaCl2) absorbs water from the wet
producer and the purchaser or pipeline company. Specifica- natural gas and forms various calcium chloride hydrates
tion parameters often include upper and lower limits for heat- (CaCl2 · xH2O). These hydrates are removed from the
ing value, chemical composition, contaminants, water natural gas stream as a calcium chloride brine solution.6
content, and hydrocarbon dew point. Table 5.1 outlines gen- 4. Refrigeration: a refrigeration coil is used to cool and con-
eral specifications for pipeline-quality natural gas, as pro- dense water vapor from the wet natural gas stream. Sep-
aration of the liquid phase is accomplished via a two-
vided by the Gas Processors Suppliers Association.4 Typical
phase, vapor/liquid separation drum.5
commercial natural gas compositions, listed by production
region, are contained in Table 5.2.3 Hydrogen sulfide must be removed from the raw natural gas
In addition to the primary combustible and inert chemical stream due to air pollution considerations and corrosion haz-
components discussed above, raw natural gas can also contain ards. The hydrogen sulfide content of commercial natural gas
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

undesirable amounts of water, hydrogen sulfide, and/or rarely exceeds 1.0 grains per 100 std. ft3 (0.023 g/m3). A major-
carbon dioxide. Before the raw natural gas can be deposited ity of pipeline companies responding to a 1994 poll limited
into a pipeline transmission network, these undesirable com- hydrogen sulfide concentrations to less than 0.3 g per 100 std.
ponents must be removed. ft3 (0.007 g/m3).1 In addition, carbon dioxide is often removed
Failure to remove the water vapor from raw natural gas from the raw gas because the inert component weakens the
prior to introduction to the pipeline network will result in overall heating value of the gas stream.5 There are numerous
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 159 Monday, March 14, 2005 12:29 PM

Fuels 159

TABLE 5.2 Commercial Natural Gas Components and Typical Ranges of Composition
Sample Gas Compositions by Production Region (vol%)
Fuel Gas
Component Tulsa, OK Alaska Algeria Netherlands Kuwait Libya North Sea Alabama Ohio Missouri Pennsylvania

CH4 93% 100% 87% 81% 87% 70% 94% 90% 94% 84% 83%
C2H6 3% — 9% 3% 9% 15% 3% 5% 3% 7% 16%
C3H8 1% — 3% <1% 2% 10% 1% — <1% — —
C4H10 <1% — 1% <1% 1% 4% <1% — <1% — —
C5 & higher — — — — — — — — — — —
CO2 1% — — 1% 2% — <1% — 1% 1% —
N2 2% 1% <1% 14% 1% 1% 2% 5% 1% 8% 1%
O2 — — — — — — — — <1% — —
H2 — — — — — — — — <1% — —

Total 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100%

Adapted from Reed, R.J., North American Combustion Handbook, Vol. 1, North American Mfg. Co., Cleveland, OH, 1986.

commercial processes (chemical reaction, absorption, and may contain varying small amounts of olefinic hydrocarbons
adsorption) for the removal of acidic components (H2S and CO2) such as propylene and butylene.1
from raw natural gas streams. The Gas Processors Suppliers Most of the LPG used in the United States consists prima-
Association (GPSA) Engineering Data Book discusses many rily of propane.8 Due to their relatively high boiling point,
of these hydrocarbon treatment processes in detail.6 Hydrogen LPG mixtures containing high concentrations of normal
sulfide removed from the raw gas is generally converted to butane (boiling point = 31°F or –1°C at atmospheric pressure)
elemental sulfur via the Claus process.6 or isobutane (boiling point = 11°F or –12°C at atmospheric
After the necessary purification processes have been com- pressure) are preferred for use in warm climates. Conversely,
pleted, the commercial-grade natural gas is compressed to LPG mixtures containing high concentrations of propane
approximately 1000 psig (6.9 MPag) and is introduced to a (boiling point = –44°F or –49°C) are typically preferred for
natural gas pipeline distribution network.5 The gas is recom- use in cold climates.5
pressed along the path to the consumer as necessary. Oper- LPG is often used in the hydrocarbon or petrochemical
ating pressure at an individual natural gas burner located at industry as a fuel gas supplement or as a standby/start-up
a process furnace inside a petrochemical or hydrocarbon pro- fuel. However, due to its value as both a common petrochem-
cessing facility is reduced to an operating pressure range that ical feedstock and a marketable commodity, LPG is not typ-
typically varies between 5 and 30 psig, depending on the ically preferred as a primary processing fuel.9
furnace’s heating requirements and the individual burner’s
design specifications. 5.1.4 Refinery Gases
Although commercial natural gas and LPG are often used as
5.1.3 Liquified Petroleum Gas (LPG) fuels in processing plants, internally generated refinery fuel
Liquefied petroleum gas (LPG) is the general term used to gases serve as the primary fuel component for most refineries,
describe a hydrocarbon that is stored as a liquid under mod- petrochemical plants, and hydrocarbon facilities. It is not
erate pressure but is a gas under normal atmospheric condi- usual for a process unit to produce its own fuel supply. Often,
tions. LPG is vaporized for use as a fuel. The primary fuel gas streams from various processing units are delivered to
chemical components of LPG are propane, propylene, nor- a common mixing point within the plant, before the new gas
mal butane, isobutane, and butylene.1 The Gas Processors mixture is returned to the processing units as refinery gas.
Suppliers Association (GPSA) Engineering Data Book con- Refinery fuel gases contain an extremely wide variety of
tains industry standard product specifications for commercial chemical constituents, including paraffins, olefins, diolefins,
propane (predominantly propane and/or propylene), com- aromatics, mercaptans, organic sulfides, ammonia, hydrogen
mercial butane (predominantly butane and/or butylene), and sulfide, carbon monoxide, carbon dioxide, etc. Because plants
commercial butane-propane mixtures.4 LPG produced via the must operate in a manner best suited to maximize profit, the
separation of heavier hydrocarbons from natural gas is individual fuel gas streams originating at each process unit
mainly paraffinic, containing primarily propane, normal will vary in composition and quantity, depending on numer-
butane, and isobutane. LPG derived from oil-refinery gas ous economic and technical factors.10 Table 5.3 contains
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

160 The John Zink Combustion Handbook

TABLE 5.3 Composition of a Typical Refinery Gas

Refinery Fuel Gas Source (Dry Gas)
Fuel Gas Cracked Coking Reforming FCC Combined Refinery Combined Refinery
Component Gas Gas Gas Gas Gas – Sample 1 Gas – Sample 2

CH4 65% 40% 28% 32% 36% 53%

C2H4 3% 3% 7% 7% 5% 2%
C2H6 16% 21% 28% 9% 18% 19%
C3H6 2% 1% 3% 15% 8% 6%
C3H8 7% 24% 22% 25% 20% 14%
C4H8 1% — — — — —
C4H10 3% 7% 7% 0% 2% 1%
C5 & Higher 1% — — — — —
H2 3% 4% 5% 6% 3% 3%
CO — — — — — —
CO2 — — — — — —
N2 — — — 7% 8% 3%
H2O — — — — — —
O2 — — — — — —
H2S — — — — — —

Total 100% 100% 100% 100% 100% 100%

Adapted from Nelson, 1949. 11

typical chemical compositions of fuel gas streams originating streams, and lubrication oil contamination of the fuel gas
from various process units within a petroleum refinery.11 stream. Potential solutions for the problems associated with
these liquid hydrocarbons include liquid extraction of the
It is very important that the refinery fuel gas leaving the
heavier chemical components (C5 and heavier) and filtration/
common mixing point is a homogenous mixture of the fuel gas
coalescence of liquid components from the gas stream. In
streams supplied. If the individual fuel gas supply streams vary
addition, increasing the velocity of the flowing gas through
significantly in calorific value, and if the supply streams are
burner components (tips, risers, etc.) has been proven to cool
not combined in a homogeneous manner, the calorific value of
the hardware and inhibit the cracking reactions that eventually
the nonhomogeneous refinery fuel gas mixture will also vary
lead to plugging and coking.
widely and often instantaneously. Unless the gas burners and
Wet fuel gas can introduce problems in cooler climates asso-
control systems at each processing furnace have been designed
ciated with the condensation and subsequent freezing of water
to accommodate instantaneous changes in fuel gas calorific
vapor inside the fuel gas system. If the water vapor reaches the
value, the process will likely be impossible to control. All of
dew point in a cold atmospheric environment, there is danger
the combustion performance parameters — including burner
of frost stoppage, freezing, or bursting of lines — a considerable
stability, emissions control, heat transfer efficiency, and heat
fire safety hazard that merits serious thought. Options to combat
flux — will suffer as a result of the nonhomogeneous fuel
water present in the fuel gas system include dehydration systems
mixture.10 Static mixers are often used in various segments of
(as discussed in Section 5.2.1) and steam/electric tracing of
industry to ensure a well-mixed, homogeneous fuel gas mix-
refinery fuel gas lines.10
ture. However, static mixers are often impractical in the petro-
chemical and hydrocarbon processing industries, typically due 5.1.5 Combustible Waste Gas Streams
to pressure drop limitations of the refinery fuel gas system.
The quantity and variety of combustible waste gas streams in
Another problem often associated with the combustion of the hydrocarbon and petrochemical industries are virtually
refinery fuel gases is the presence of liquid hydrocarbons in unlimited. Many of these waste gas streams are relatively
the refinery fuel gas stream, which can accelerate the coking high in inert concentration, with large amounts of nitrogen
and plugging rates of downstream gas burner components. and carbon dioxide often present. As a result, these waste
Sources of unwanted liquid hydrocarbons in refinery fuel gas fuels are often low in heat content, with lower heating values
streams include condensation of heavier fuel gas components in the range of 400 to 800 Btu/scf (0.42 to 0.84 MJ/scm). For
(C5 and higher) due to natural cooling of the fuel gas stream, these reasons, waste fuels are not usually compressed into the
liquid entrainment into absorber or fractionator overhead gas
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- main refinery fuel gas system. Two of the most widely used
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 161 Monday, March 14, 2005 12:29 PM

Fuels 161

FIGURE 5.1 Simplified process flow diagram for hydrogen reforming/pressure swing adsorption. (Adapted from Meyers, 199712).

combustible waste gas fuels, Pressure Swing Adsorption TABLE 5.4 Typical Composition of Steam Reforming/PSA
(PSA) tail gas and Flexicoking gas are discussed in detail in Tail Gas
the sections below. PSA Tail Gas Composition
Fuel Gas Component (vol%)
5.1.5.1 Pressure Swing Adsorption (PSA) Tail Gas CH4 17%
Pressure swing adsorption (PSA) tail gas is a low-pressure, H2O <1%
low-Btu fuel gas produced as a by-product of a PSA process, H2 28%
CO2 44%
a key purification component in the steam reforming hydro- CO 10%
gen production process. Table 5.4 contains the approximate N2 <1%
composition of a typical PSA tail gas fuel stream.
Total 100%
PSA is a cyclic process that uses beds of solid adsorbent
to remove impurities such as carbon dioxide, carbon mon-
oxide, methane, and nitrogen from the hydrogen production
stream. A simplified process flow diagram of a typical steam
reforming hydrogen production unit using PSA is shown in 3. Shift conversion: the water-gas shift reaction is employed
Figure 5.1.12 to convert the carbon monoxide produced in the reforming
step into additional hydrogen and carbon dioxide:
The steam reforming process is conducted in four stages:8,9,12
1. Feedstock preparation: feedstock (light hydrocarbons CO + H 2 O → H 2 + CO 2 (5.2)
such as methane, propane, butane, and light liquid naph-
tha) at approximately 450 psig (31 bar) is preheated and The shift conversion step is exothermic and is conducted
purified to remove reformer catalyst poisons such as halo- at approximately 650°F (343°C) in the presence of a
gens and sulfur-containing compounds. chromium/iron oxide catalyst.
2. Reforming: the purified feedstock is reacted with steam
to form carbon monoxide and hydrogen: 4. Hydrogen purification/PSA: following the shift conver-
sion step, the hydrogen production stream enters the PSA
C n H m + nH 2 O 1500 ° F & Ni Catylst
 →(n + m 2)H 2 + nCO (5.1) portion of the process. Adsorbent beds remove the impu-
rities (carbon dioxide, carbon monoxide, methane, and
The reaction is endothermic and occurs within the process nitrogen) and a small portion of the product. Typical
tubes of a reformer furnace in the presence of nickel hydrogen recovery is 80% or greater, with product purity
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

catalyst at approximately 1500°F (815°C). of approximately 99.9 vol%.

162 The John Zink Combustion Handbook

FIGURE 5.2 Simplified process flow diagram for Flexicoking.13

TABLE 5.5 Typical Composition of Flexicoking Waste Gas stability problems associated with firing the low-pressure,
Flexicoking Waste Gas Composition high-inert-concentration (carbon dioxide and nitrogen) PSA
(by volume) tail gas alone, the PSA tail gas is typically supplemented by
Fuel Gas Component Sample 1 Sample 2 a light refinery fuel gas. The PSA and refinery fuel gases are
CH4 1.0% 0.8% fired in a dual-fuel burner specifically designed for the steam
H2 20.0% 21.0% reforming/PSA process. In this arrangement, the PSA and
CO2 10.0% 10.5%
refinery fuel gases enter the combustion zone through sepa-
CO 20.0% 18.6%
N2 45.0% 45.6% rate fuel connections and burner nozzles. The dual-fuel burn-
H2O 4.0% 3.5% ers are capable of firing the two fuel mixtures separately or
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

H2S 150 ppm 0 simultaneously, with PSA gas never providing more than 85
COS 120 ppm 120 ppm
vol% of the total reformer fuel.
Total 100% 100%

Adapted from Meyers, R.A., Handbook of Petroleum Refining Processes,

5.1.5.2 Flexicoking Waste Gas
2nd ed., McGraw-Hill, New York, 1997. Flexicoking waste gas is a low-pressure, low-Btu fuel gas pro-
duced by petroleum refiners as a by-product of the Exxon
Flexicoking process. Flexicoking is a continuous fluidized-
The PSA unit must be frequently regenerated via depres- bed thermal cracking process used in the conversion of
surization of the adsorbent beds. When depressurization heavy hydrocarbon feedstocks (typically heavy gas oils
occurs, PSA tail gas (sometimes referred to as PSA waste gas) from atmospheric and vacuum distillation) to various
is produced at a pressure of about 5 psig (0.35 kg/cm2) or gaseous and liquid hydrocarbon products. Table 5.5 contains
less. The PSA tail gas consists of the impurities removed the approximate composition of two sample Flexicoking
by the adsorbent beds, as well as the hydrogen that is not waste gas fuel streams.13
recovered in the product stream. The tail gas serves as the A simplified process flow diagram of the Flexicoking
primary fuel for the reformer furnace burners. Due to flame process is shown in Figure 5.2.13 In the Flexicoking process,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
Copyright CRC Press
TABLE 5.6 Volumetric Analysis of Typical Gaseous Fuel Mixtures --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Fuels
Natural Gas LPG Refinery Gases (Dry) Waste Gases
Fuel Gas Cracked Coking Reforming FCC Refinery Gas Refinery Gas PSA Flexicoking
Component Tulsa Alaska Netherlands Algeria Propane Butane Gas Gas Gas Gas Sample 1 Sample 2 Gas Gas

CH4 93.4% 99% 81% 87% — — 65% 40% 28% 32% 36% 53% 17% 1%

Provided by IHS Markit under license with CRC Press

C2H4 — — — — — — 3% 3% 7% 7% 5% 2% — —
C2H6 2.7% — 3% 9% — — 16% 21% 28% 9% 18% 19% — —
C3H6 — — — — — — 2% 1% 3% 15% 8% 6% — —

No reproduction or networking permitted without license from IHS

C3H8 0.6% — 0.4% 2.7% 100% — 7% 24% 22% 25% 20% 14% — —
C4H8 — — — — — 100% 1% — — — — — — —
C4H10 0.2% — 0.1% 1.1% — — 3% 7% 7% 0% 2% 1% — —
C5 & Higher — — — — — — 1% — — — — — — —
H2 — — — — — — 3% 4% 5% 6% 3% 3% 28% 21%
2337-ch05-Frame Page 163 Monday, March 14, 2005 12:29 PM

CO — — — — — — — — — — — — 10% 20%
CO2 0.7% — 0.9% — — — — — — — — — 44% 10%
N2 2.4% 1% 14% 0% — — — — — 7% 8% 3% <1% 45%
H2O — — — — — — — — — — — — <1% 3%
O2 — — — — — — — — — — — — — —
H 2S — — — — — — — — — — — — — —

Total 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 100%

Data compiled from a variety of sources.

TABLE 5.7 Physical Constants of Typical Gaseous Fuel Mixtures

Natural Gas LPG Refinery Gases (Dry) Waste Gases
Fuel Gas Cracked Coking Reforming FCC Refinery Gas Refinery Gas PSA Flexicoking
Component Tulsa Alaska Netherlands Algeria Propane Butane Gas Gas Gas Gas Sample 1 Sample 2 Gas Gas

Molecular weight 17.16 16.1 18.51 18.49 44.1 58.12 22.76 28.62 30.21 29.18 28.02 24.61 25.68 23.73

Not for Resale, 01/15/2018 21:50:21 MST

Lower heating value (LHV), Btu/SCF 913 905 799 1025 2316 3010 1247 1542 1622 1459 1389 1297 263 131
Higher heating value (HHV), Btu/SCF 1012 1005 886 1133 2517 3262 1369 1686 1769 1587 1515 1421 294 142
Specific gravity 0.59 0.56 0.64 0.64 1.53 1.1 0.79 0.99 1.05 1.01 0.97 0.85 0.89 0.82
(14.696 psia/60°F, Air = 1.0)
Wobbe number, HHV/(SG1/2) 1318 1343 1108 1416 2035 3110 1540 1694 1726 1579 1538 1541 312 157
Isentropic coefficient (Cp/Cv) 1.30 1.31 1.31 1.28 1.13 1.10 1.24 1.19 1.19 1.20 1.21 1.23 1.33 1.38
Stoichiometric air required, SCF/MMBtu 10554 10567 10554 10525 10369 10371 10402 10379 10322 10234 10311 10375 9667 8265
Stoichiometric air required, lbm/MMBtu 805 806 805 803 791 791 794 792 787 781 787 792 738 630
Air required for 15% excess air, SCF/MMBtu 12138 12152 12138 12104 11925 11926 11962 11936 11870 11769 11858 11931 11117 9505

Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher

Air required for 15% excess air, lbm/MMBtu 923 924 923 920 907 907 910 908 903 895 902 907 845 723
Volume of dry combustion products, SCF/MMBtu 10983 10956 11141 10953 10962 10996 10890 10909 10871 10847 10911 10904 11722 13517
Weight of dry combustion products, lbm/MMBtu 865 862 876 863 870 874 861 864 862 860 864 862 985 1103
Volume of wet combustion products, SCF/MMBtu 13257 13258 13415 13163 12788 12757 12935 12862 12771 12689 12821 12902 14198 15585
Weight of wet combustion products, lbm/MMBtu 973 971 984 968 957 958 958 957 952 948 864 957 1102 1201
Adiabatic flame temperature, °F 3306 3308 3284 3317 3351 3351 3342 3348 3359 3371 3353 3345 3001 2856

Note: All values calculated using 60°F fuel gas and 60°F, 50% relative humidity combustion air.
163
Copyright CRC Press
TABLE 5.8 Physical Constants of Typical Gaseous Fuel Mixture Components
Heating Value Unit Volume per Unit Volume of Combustible Unit Mass per Unit Mass of Combustible Flammability
Limits
Gas Density (vol% in air
Specific Heat Latent Heat of Theoretical
Ideal Gas, 14.696 psia, 60°F Btu/scf Btu/lbm Required for Combustion Flue Gas Products Required for Combustion Flue Gas Products mixture)
Boiling Vapor Capacity, Cp Vaporization Air
Point Pressure 60°F & 14.696 psia & Specific Gas Specific Required
Fuel Gas Chemical Molecular 14.696 psia 100°F 14.696 psia Boiling Point Gravity Density Volume LHV HHV LHV HHV (lbm/10,000

Provided by IHS Markit under license with CRC Press

No. Component Formula Weight (°F) (psia) (Btu/lbm /°F) (Btu/lbm) (Air = 1) (lbm/ft3) (ft3/lbm) (Net) (Gross) (Net) (Gross) O2 N2 Air CO2 H2O N2 SO2 O2 N2 Air CO2 H2O N2 SO2 Btu) Lower Upper No.

Paraffin (alkane) Series (CnH2n+2)

No reproduction or networking permitted without license from IHS

1 Methane CH4 16.04 –258.69 — 0.5266 219.22 0.554 0.042 23.651 912 1,013 21,495 23,845 2.0 7.547 9.547 1.0 2.0 7.547 — 3.989 13.246 17.235 2.743 2.246 13.246 — 7.219 5.0 15.0 1
2 Ethane C2H6 30.07 –127.48 — 0.4097 210.41 1.038 0.079 12.618 1,639 1,792 20,418 22,323 3.5 13.206 16.706 2.0 3.0 13.206 — 3.724 12.367 16.092 2.927 1.797 12.367 — 7.209 2.9 13.0 2
3 Propane C3H8 44.10 –43.67 190 0.3881 183.05 1.522 0.116 8.604 2,385 2,592 19,937 21,669 5.0 18.866 23.866 3.0 4.0 18.866 — 3.628 12.047 15.676 2.994 1.624 12.047 — 7.234 2.1 9.5 3
4 n-Butane C4H10 58.12 31.10 51.6 0.3867 165.65 2.007 0.153 6.528 3,113 3,373 19,679 21,321 6.5 24.526 31.026 4.0 5.0 24.526 — 3.578 11.882 15.460 3.029 1.550 11.882 — 7.251 1.8 8.4 4
5 Isobutane C4H10 58.12 10.90 72.2 0.3872 157.53 2.007 0.153 6.528 3,105 3,365 19,629 21,271 6.5 24.526 31.026 4.0 5.0 24.526 — 3.578 11.882 15.460 3.029 1.550 11.882 — 7.268 1.8 8.4 5
6 n-Pentane C5H12 72.15 96.92 15.57 0.3883 153.59 2.491 0.190 5.259 3,714 4,017 19,507 21,095 8.0 30.186 38.186 5.0 6.0 30.186 — 3.548 11.781 15.329 3.050 1.498 11.781 — 7.267 1.4 8.3 6
7 isopentane C5H12 72.15 82.12 20.44 0.3827 147.13 2.491 0.190 5.259 3,705 4,017 19,459 21,047 8.0 30.186 38.186 5.0 6.0 30.186 — 3.548 11.781 15.329 3.050 1.498 11.781 — 7.283 1.4 8.3 7
2337-ch05-Frame Page 164 Monday, March 14, 2005 12:29 PM

8 Neopentane C5H12 72.15 49.10 35.9 0.3666 135.58 2.491 0.190 5.259 3,692 3,994 19,390 20,978 8.0 30.186 38.183 5.0 6.0 30.186 — 3.548 11.781 15.329 3.050 1.498 11.781 — 7.307 1.4 8.3 8
9 n-Hexane C6H14 86.18 155.72 4.956 0.3664 143.95 2.975 0.227 4.403 4,415 4,767 19,415 20,966 9.5 35.846 45.346 6.0 7.0 35.846 — 3.527 11.713 15.240 3.064 1.463 11.713 — 7.269 1.2 7.7 9

Napthene (cycloalkane) Series (CnH2n)

10 Cyclopentane C5H10 70.13 120.60 9.917 0.2712 137.35 2.420 0.180 5.556 3,512 3,764 19,005 20,368 7.5 27.939 35.180 5.0 5.0 28.939 — 3.850 11.155 14.793 3.146 1.283 11.155 — 7.262 — — 10
164 The John Zink Combusion Handbook

11 Cyclohexane C6H12 84.16 177.40 3.267 0.2901 153.25 2.910 0.220 5.545 4,180 4,482 18,849 20,211 9.0 33.528 42.970 6.0 6.0 33.528 — 4.620 13.386 17.750 3.146 1.283 11.155 — 7.848 1.3 8.4 11

Olefin Series (CnH2n)

12 Ethene C2H4 28.05 –154.62 — 0.3622 207.57 0.969 0.074 13.525 1,512 1,613 20,275 21,636 3.0 11.320 14.320 2.0 2.0 11.320 — 3.422 11.362 14.784 3.138 1.284 11.362 — 6.833 2.7 34.0 12
(Ethylene)
13 Propene C3H6 42.08 –53.90 226.4 0.3541 188.18 1.453 0.111 9.017 2,185 2,336 19,687 21,048 4.5 16.980 21.480 3.0 3.0 16.980 — 3.422 11.362 14.784 3.138 1.284 11.362 — 7.024 2.0 10.0 13
(Propylene)
14 1-Butene C4H8 56.11 20.75 63.05 0.3548 167.94 1.937 0.148 6.762 2,885 3,086 19,493 20,854 6.0 22.640 28.640 4.0 4.0 22.640 — 3.422 11.362 14.784 3.138 1.284 11.362 — 7.089 1.6 9.3 14
(Butylene)
15 Isobutene C4H8 56.11 19.59 63.4 0.3701 169.48 1.937 0.148 6.762 2,868 3,069 19,376 20,737 6.0 22.640 28.640 4.0 4.0 22.640 — 3.422 11.362 14.784 3.138 1.284 11.362 — 7.129 1.6 — 15
16 1-Pentene C5H10 70.13 85.93 19.115 0.3635 154.46 2.421 0.185 5.410 3,585 3,837 19,359 20,720 7.5 28.300 35.800 5.0 5.0 28.300 — 3.422 11.362 14.784 3.138 1.284 11.362 — 7.135 1.4 8.7 16

Aromatic Series (C0H2n–6)

17 Benzene C6H8 78.11 176.17 3.224 0.2429 169.31 2.697 0.206 4.857 3,595 3,746 17,421 18,184 7.5 28.300 35.800 6.0 3.0 28.300 — 3.072 10.201 13.274 3.380 0.692 10.201 — 7.300 1.38 7.98 17
18 Toluene C7H8 92.14 231.13 1.032 0.2598 154.84 3.181 0.243 4.118 4,296 4,497 17,672 18,501 9.0 33.959 42.959 7.0 4.0 33.959 — 3.125 10.378 13.504 3.343 0.782 10.378 — 7.299 1.28 7.18 18
19 o-Xylene C8H10 106.17 291.97 0.264 0.2914 149.1 3.665 0.280 3.574 4,970 5,222 17,734 18,633 10.5 39.619 50.119 8.0 5.0 39.619 — 3.164 10.508 13.673 3.316 0.848 10.508 — 7.338 1.18 6.48 19
20 m-Xylene C8H10 106.17 282.41 0.326 0.2782 147.2 3.665 0.280 3.574 4,970 5,222 17,734 18,633 10.5 39.619 50.119 8.0 5.0 39.619 — 3.164 10.508 13.673 3.316 0.848 10.508 — 7.338 1.18 6.48 20
21 p-Xylene C8H10 106.17 281.05 0.342 0.2769 144.52 3.665 0.280 3.574 4,970 5,222 17,734 18,633 10.5 39.619 50.119 8.0 5.0 39.619 — 3.164 10.508 13.673 3.316 0.848 10.508 — 7.338 1.18 6.48 21

Not for Resale, 01/15/2018 21:50:21 MST

Additional Fuel Gas Components

22 Acetylene C2H2 26.04 –119 — 0.3966 — 0.899 0.069 14.572 1,448 1,499 20,769 21,502 2.5 9.433 11.933 2.0 1.0 9.433 — 3.072 10.201 13.274 3.380 0.692 10.201 — 7.300 2.5 80 22

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
23 Methyl alcohol CH3OH 32.04 148.1 4.63 0.3231 473 1.106 0.084 11.841 767 868 9,066 10,258 1.5 5.660 7.160 1.0 2.0 5.660 — 4.498 4.974 6.482 1.373 1.124 4.974 — 6.309 6.72 36.5 23
24 Ethyl alcohol C2H5OH 46.07 172.92 2.3 0.3323 367 1.590 0.121 8.236 1,449 1,600 11,918 13,161 3.0 11.320 14.320 2.0 3.0 11.320 — 2.084 6.919 9.003 1.911 1.173 6.919 — 6.841 3.28 18.95 24
25 Ammonia NH3 17.03 –28.2 212 0.5002 587.2 0.588 0.045 22.279 364 441 7,966 9,567 0.75 2.830 3.582 — 1.5 3.330 — 1.409 4.679 6.008 — 1.587 5.502 — 6.298 15.50 27.00 25
26 Hydrogen H2 2.02 –423.0 — 3.4080 193.9 0.070 0.005 188.217 274.6 325.0 51,625 61,095 0.5 1.887 2.387 — 1.0 1.887 — 7.936 26.323 34.290 — 8.937 26.353 — 5.613 4.00 74.20 26
27 Oxygen O2 32.00 –297.4 — 0.2186 91.6 1.105 0.084 11.858 — — — — — — — — — — — — — — — — — — — — — 27
28 Nitrogen N2 29.16 –320.4 — 0.2482 87.8 0.972 0.074 13.472 — — — — — — — — — — — — — — — — — — — — — 28
29 Carbon CO 28.01 –313.6 — 0.2484 92.7 0.967 0.074 13.546 321.9 321.9 4,347 4,347 0.5 1.877 2.387 1.0 — 1.887 — — 1.897 2.468 1.571 — 1.870 — 5.677 12.50 74.20 29
monoxide
30 Carbon CO2 44.01 –109.3 — 0.1991 238.2 1.519 0.116 8.621 — — — — — — — — — — — — — — — — — — — — — 30

Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher

dioxide
31 Hydrogen H2S 34.08 –76.6 394.0 0.2380 235.6 1.177 0.090 11.133 595 646 6,537 7,097 1.5 5.660 7.160 — 1.0 5.660 1.0 1.410 4.682 6.093 — 0.529 4.682 1.880 8.585 4.30 45.50 31
sulfide
32 Sulfur dioxide SO2 64.06 14.0 88 0.1450 166.7 2.212 0.169 5.923 — — — — — — — — — — — — — — — — — — — — — 32
33 Water vapor H2O 18.02 212.0 0.9492 0.4446 970.3 0.622 0.047 21.061 — — — — — — — — — — — — — — — — — — — — — 33
34 Air — 29.97 –317.6 — 0.2400 92 1.000 0.076 13.099 — — — — — — — — — — — — — — — — — — — — — 34
2337-ch05-Frame Page 165 Monday, March 14, 2005 12:29 PM

Fuels 165

hot (500 to 700°F or 260 to 370°C) gas oil is injected into the result in the production of yellow flame. However, both
reactor vessel containing hot, fluidized coke particles. Thermal yellow and blue flame burning can occur at virtually any
cracking reactions inside the reactor vessel produce fresh condition of deficient or surplus combustion air.
petroleum coke that is deposited as a thin film on the surface Yellow flame burning is the direct result of the cracking of
of existing coke particles inside the reactor bed. Cracked vapor a hydrocarbon fuel into its hydrogen and carbon components,
products exit the Flexicoking process through the reactor followed by separate burning of the two constituents. The
vessel overhead stream for additional downstream processing. hydrogen constituents are burned in a rapid process that
Coke from the reactor vessel is continuously injected into the produces a pale lavender-pink flame that is very difficult to
top of a second fluidized vessel, the coke heater, where it is see except against a dark background. When yellow flame
heated and recycled to maintain a reactor bed temperature of burning occurs, the heavier carbon constituents burn in a
950 to 1000°F (510 to 540°C). A portion of the coke fed into relatively slower process that typically results in a lumines-
the top section of the coke heater is injected into the bottom cent yellow flame.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
of a third fluidized vessel, the gasifier. Inside the gasifier, the Blue flame burning is the direct result of progressive oxygen-
coke is reacted with air and steam at approximately 1500 to ation of the fuel in a manner that does not allow uncombined
1800°F (820 to 980°C), producing a low-Btu fuel gas, or carbon to be present in the reaction (flame) envelope. Inadequate
Flexicoking gas, consisting primarily of nitrogen, hydrogen, fuel/air mixing can severely limit this reaction process, produc-
carbon monoxide, and carbon dioxide. The Flexicoking gas ing a greater tendency toward yellow flame. Both yellow and
flows from the top of the gasifier to the bottom of the heater, blue flame burnings are possible with any hydrocarbon fuel, and
where it provides the heat necessary to maintain the reactor both kinds of flame produce equivalent quantities of heat.
bed temperature and helps fluidize the coke heater bed. The The hydrogen-to-carbon weight ratio (H:C) is a good indica-
high-temperature Flexicoking gas leaving the coke heater is tor of a fuel mixture’s relative tendency to produce yellow flame
used for high-pressure steam generation before entrained coke burning, with low H:C ratios corresponding to an increased
fines are removed in a cyclone/venturi scrubber system. movement toward yellow flame burning. Pure hydrogen
Because the low-Btu gas stream leaving the Flexicoking pro- (H:C = ∞) typically burns as a pale lavender-pink flame that is
cess contains substantial concentrations of H2S (~150 ppm by very difficult to see except against a dark background. Pure
volume), the gas must first be sent through a hydrogen sulfide methane (H:C = 0.33) typically burns as a light blue flame. Fuel
removal system before it can be burned as fuel.9,13 mixtures containing propane (H:C = 0.22), butane (H:C = 0.21),
and the olefins (H:C = 0.166) all have a greater tendency to
5.1.6 Physical Properties of Gaseous Fuels exhibit yellow flame burning than pure methane fuels.10
Tables 5.6, 5.7, and 5.8 provide physical and combustion
property data for a large variety of common fuel gas mixtures
and their chemical components. 5.2 LIQUID FUELS

5.1.7 Photographs of Gaseous Fuel Flames 5.2.1 Introduction and History

Figures 5.3 through 5.18 are photographs of a John Zink Liquid fuels are a key component in today’s energy pro-
PSFG staged fuel gas burner, firing a wide variety of fuel gas cesses. An example of an oil flame is shown in Figure 5.34.
mixtures into open air (i.e., not in a furnace). In each photo- During the Industrial Revolution, starting in the mid-18th
graph, the burner is being operated at the same fuel flow rate century, the major energy source used in the world changed
(in terms of energy released per unit time) and under the same from charcoal (wood) to various forms of coal. As technol-
general ambient conditions. Fuel composition is the only ogy developed, the world began moving from the use of coal
parameter that is varied throughout the series. The images are to crude oil (the most abundant liquid fuel used in industry
provided to illustrate the differences in flame appearance today), and its derivatives, to provide the energy and heating
(shape and color) produced by various fuel compositions. requirements needed. The modern era of viable crude oil pro-
Figures 5.19 to 5.33 show a similar series of photos of a duction and use began with commercial wells in the mid-
John Zink VYD raw gas burner firing a wide variety of fuels 1800s. An increasing need for oil products in technology
inside a furnace all at the same firing rate. (such as gasoline for the internal combustion engine and
There is a widely held misconception that yellow-flame automobiles) spurred massive efforts in oil exploration and
burning is solely the direct result of combustion air deficiency. recovery in the early 1900s. Figure 5.35 depicts an oil der-
Inadequate or unsatisfactory fuel/air mixing will certainly rick, circa 1900.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 166 Monday, March 14, 2005 12:29 PM

166 The John Zink Combustion Handbook

FIGURE 5.3 100% TNG flame. FIGURE 5.4 90% TNG/10% N2 flame.

FIGURE 5.5 80% TNG/20% N2 flame. FIGURE 5.6 90% TNG/10% H2 flame.

Major oil deposits found in the United States prompted it provided approximately two thirds of the world’s oil supply
to become a major world oil producer. The successes of at prices near U.S. $1 per barrel (approximately 3000% lower
American oil discovery and production inspired oil compa- than the current per barrel price).14 In 1960, the Organization
nies in other countries to start a worldwide exploration for of Petroleum Exporting Countries (OPEC) was founded by
oil reserves. In the mid-1950s, major U.S. oil companies
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
the governments of major oil-exporting countries for the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 167 Monday, March 14, 2005 12:29 PM

Fuels 167

FIGURE 5.7 75% TNG/25% H2 flame. FIGURE 5.8 50% TNG/50% H2 flame.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 5.9 25% TNG/75% H2 flame. FIGURE 5.10 100% H2 flame.

purpose of stabilizing oil production and prices. As the exploited oil reserves throughout the world. The vast majority
demand for oil increased, the production inevitably increased. of the known oil reserves in the world are located in the Middle
The consumption of crude oil in 1998 was approximately East (approximately two thirds), while the United States ranks
70 million barrels per day.15 Technological developments in eighth on the known reserve list. The United States produces
drilling and exploration techniques have identified and only about 17% of the world’s oil, yet it consumes nearly
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 168 Monday, March 14, 2005 12:29 PM

168 The John Zink Combustion Handbook

FIGURE 5.11 50% TNG/25% H2/25% C3H8 flame. FIGURE 5.12 50% TNG/50% C3H8 flame.

FIGURE 5.13 100% C3H8 flame. FIGURE 5.14 100% C4H10 flame.

30%.16 The Energy Information Administration (Office of States’ remaining oil reserves range from capacities of 25 to
Oil & Gas) estimates that the total U.S. crude oil stocks are 100 billion barrels. Based on our current consumption
310 million barrels, excluding the strategic petroleum patterns, these reserves could supply us with enough oil for
reserve, as of May 19, 2000.14 Estimates of the United only 10 to 30 more years.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fuels 169

FIGURE 5.15 Simulated cracked gas flame. FIGURE 5.16 Simulated coking gas flame.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 5.17 Simulated FCC gas flame. FIGURE 5.18 Simulated reforming gas flame.

Oil exploration, which initially was confined to land, has ficulty with, and the high cost of extracting oil from, these
led to recovery efforts on the bottom of the ocean floor. The complicated mediums keeps conventional crude oil recovery
most abundant forms of oil deposits found in the world today as the leading source of usable raw material for refining pro-
are oil shale, heavy oil deposits, and tar sands. However, dif- cesses. Figure 5.36 shows the capping of a burning oil well.

Copyright CRC Press

170 The John Zink Combustion Handbook

FIGURE 5.19 100% Tulsa natural gas. FIGURE 5.20 100% hydrogen.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 5.21 100% propane. FIGURE 5.22 50% hydrogen/50% propane.

Fuels 171

FIGURE 5.23 50% hydrogen/50% Tulsa natural gas. FIGURE 5.24 50% propane/50% Tulsa natural gas.

FIGURE 5.25 25% hydrogen/75% propane.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 5.26 75% hydrogen/25% propane.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 172 Monday, March 14, 2005 12:29 PM

172 The John Zink Combustion Handbook

FIGURE 5.27 25% hydrogen/75% Tulsa natural gas. FIGURE 5.28 75% hydrogen/25% Tulsa natural gas.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 5.29 25% propane/75% Tulsa natural gas. FIGURE 5.30 75% propane/25% Tulsa natural gas.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 173 Monday, March 14, 2005 12:29 PM

Fuels 173

FIGURE 5.31 25% hydrogen/25% propane/50% Tulsa FIGURE 5.32 25% hydrogen/50% propane/25% Tulsa
natural gas. natural gas.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

5.2.2 Oil Recovery

Crude oil is found in deep, high-pressure reservoirs, encased in
rock, beneath the Earth’s surface. Oil companies use compli-
cated drilling techniques to tap into these pockets and bring the
crude oil to the surface so that it can be collected. Oil drilling is
an expensive process that can be complicated by the location of
the oil in the earth. Therefore, oil companies spend millions of
dollars annually in exploration and cost analysis of potential,
new oil reserves. Incredibly hard rock and deep reserves (some-
times greater than 3000 ft below the surface) necessitate the use
of specially designed drill bits that will stand up to the high
pressures and constant mechanical trauma encountered in drill-
ing. Once an oil reservoir is “hit,” the oil, now having an avenue
to expand, will rush out of the drilling channel that was cleared
by the drilling rig. The oil will be continuously extracted until
the reservoir becomes depleted to the extent that it is no longer
economically viable for a company to spend time and money
to retrieve it. When the oil pressure in the reservoir becomes
too low for natural extraction, pumps can be used to help with
the extraction. Other means of keeping reservoirs “active”
include injecting water, steam, or chemicals into the reservoir
to help make low-pressure, or viscous oil easier to extract.
Once the crude oil has been collected, and temporary stor- FIGURE 5.33 50% hydrogen/25% propane/25% Tulsa
age facilities are nearing their capacity, it must be off-loaded natural gas.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 174 Monday, March 14, 2005 12:29 PM

174 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 5.34 Viewing oil flame through a burner plenum.

FIGURE 5.35 Oil derrick, circa 1900.

Fuels 175

FIGURE 5.36 Capping a burning oil well.

so that further collection is possible. The most common meth- TABLE 5.9 Quantitative Listing of Products Made by
ods of off-loading and transporting crude oil is with pipelines the U.S. Petroleum Industry
(such as the Great Alaskan Pipeline), seafaring oil tankers, Product Classification Number of Individual Products
and barges. These transportation methods deliver the crude Lubricating oils 1156
oil to locations around the world for refining into usable Chemicals, solvents, misc. 300
petroleum products. Greases 271
Asphalts 209
Waxes 113
5.2.3 Production, Refining, and Chemistry White oils 100
The primary concern for a typical refinery is to convert a Rust preventatives 65
Diesel and light fuel oils 27
barrel of crude oil (42 U.S. gallons) into usable products.
Motor gasolines 19
A barrel of crude oil can typically be refined to provide Residual fuel oil 16
11 gallons of gasoline, 5.3 gallons of kerosene, 20.4 gallons Liquified gases 13
of gas-oil and distillates, and 5.3 gallons of heavier Other gasolines 12
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Transformer and cable oils 12

distillates.9 The end products derived from crude oil number Kerosenes 10
in the thousands. Table 5.9 provides a listing of many of these Aviation gasolines 9
products. The processes that produce these different products Jet fuels 5
Carbon blacks 5
are vast and complicated. Figure 5.37 provides a general
Cokes 4
refinery flow diagram. Fuel gas 1
The primary chemical components of crude oil are carbon,
Total 2347
hydrogen, sulfur, oxygen, and nitrogen. The percentages of
these elements found in a crude oil are most frequently used From Gary, 1994. 9

Copyright CRC Press

176 The John Zink Combustion Handbook

FIGURE 5.37 Refinery flow diagram.9

to characterize the oil. Two terms frequently used when refer- group of gases. Any time a fuel is burned in air with a hot
ring to crude oil are “sweet” crude and “sour” crude. Sweet flame, NOx are produced. The greater the flame temperature
crude is oil that contains less than 0.5 wt.% sulfur, while sour of the combustion, the greater the amount of NO that will be
crude contains greater than 0.5 wt.% sulfur. Sulfur content is produced. NO is then oxidized to form NO2 (over a period of
of importance and concern, due to the sulfur oxides that are minutes or hours), which is a major contributor to photochem-
produced during combustion. SO2, for example, is a gas that ical smog. In general, the fate of SO2 and NO are intertwined,
has been shown to contribute significantly to several different as can be seen by the following reaction sequence17:
environmental problems — namely in acid rain formation and
in its ready conversion to sulfuric acid, H2SO4. The nitrogen
SO 2 + OH • → HSO •3
content of crude oil is of special interest to the combustion
industry due to the high levels of nitrogen oxides or NOx (see HSO •3 + O 2 → SO 3 + HOO •
Chapter 6) produced during combustion of these fuels (e.g.,
approximately 0.2 lb per MMBtu NOx or 142 ppm can be SO 3 + H 2 O → H 2 SO 4 ( g)
attributed to “Fuel NOx” for an oil that contains 0.47 wt.%
H 2 SO 4 ( g)   → H 2 SO 4 (aq)
H O
nitrogen). Like SOx, NOx is an environmentally damaging 
2

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fuels 177

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Below is a parallel reaction that takes place between nitro- TABLE 5.10 General Fraction Boiling Points
gen oxide and the hydroperoxy radical, thus producing more Distillation Fraction Temperature Range
of the hydroxyl radical to feed the initial reaction above: Butanes and lighter <90°F
Gasoline 90–220°F
Naphtha 220–315°F
HOO • + NO • → OH • + NO •2 Kerosene 315–450°F
Fuel oils 450–800°F
Residue >800°F
The overall reaction is then:
From Leffler, W.L., Petroleum Refining for the Non-technical Person,
Penn Well Publishing, Tulsa, OK, 1985. With permission.
→ H 2 SO 4 (aq)
H O
SO 2 + NO • + O 2  

2

Crude oil compositions are relatively constant. However, It is worth mentioning the group of compounds called
slight deviations in composition can result in vastly different alkenes (olefins). Alkene compounds do not occur naturally
refining methods. Crude oils also contain inorganic elements in crude oil, but are produced by reaction during the refining
such as vanadium, nickel, and sodium, and usually contain process. Therefore, it should be expected that a refined end
some amount of water and ash (noncombustible material). product will have some percentage of ethylene, propylene, or
The main hydrocarbon constituents of crude oils are alkanes butylene, for example. Alkenes have the general formula of
(paraffins), cycloalkanes (naphthenes), and aromatics. CxH2x and contain a carbon-carbon double bond. Properties
Alkanes (also called paraffins after the Latin parum affinis, of some of the alkenes are contained in Table 5.8.
“little affinity”) are those chemical structures that are based
When a crude oil is refined, the first step is, invariably,
on carbon atoms having only single bonds and that are com-
distillation. The purpose of distillation is to separate lighter
pletely saturated with hydrogen atoms. Some of the alkane
components from heavier ones, based on their respective vol-
hydrocarbons are listed in Table 5.8. The basic chemical
atility. The target of distillation is to separate the crude oil
formula for an alkane is CxH2x+2, where “x” is the number
into different fractions. Each fraction consists of a boiling
of carbon atoms present. Crude oils can contain structures
point range that will yield a mixture of hydrocarbons; see
with up to 70 carbon atoms.9 However, the vast majority of
Table 5.10. Some of these mixtures can then be used as
the compounds contain 40 carbon atoms or less. When the
product (fuels, solvents, etc.) or further refined into gasoline
number of different constitutional isomers (different chemi-
or other desirable mixtures. Catalytic cracking is a typical
cal connectivity and different physical properties, yet iden-
process used to break down and rearrange alkane mixtures
tical chemical formulae) is considered (tetracontane [C40H82]
produced via distillation into smaller, highly branched
has over 62 trillion possible isomers18), it is evident that the
alkanes by heating the mixtures to high temperatures in the
compositional diversity between differing crude oils is
presence of a variety of catalysts. Figure 5.38 shows a fluid
almost limitless.
catalytic cracking process. Due to the reactions that take place
Cycloalkanes (cylcoparaffins or naphthenes) are alkanes in
during catalytic cracking, the product streams are generally
which all or some of the carbon atoms are arranged in a ring.
heavier than the feed streams. Alkanes that are more highly
When a cycloalkane contains only one ring, the general for-
branched are desirable because they have a higher octane
mula is CxH2x. The most stable cycloalkane is cyclohexane,
rating than their unbranched cousins.
while cyclobutane and cyclopropane are the least stable. The
properties of cycloalkanes are very similar to those of alkanes, Of particular interest are the liquid fuels produced during
as shown in Table 5.8. the various refining processes that are used by the hydrocar-
Aromatic compounds are those compounds that contain at bon and petrochemical industries. Refineries frequently burn
least one benzene-like ring. Benzene, discovered in 1825, has these liquid fuels in process heaters so that the heat liberated
a chemical formula of C6H6, and is stable and nonreactive during combustion can be used to drive a more profitable
relative to alkanes and cycloalkanes. Aromatics, such as the process. Light fuel oils are relatively easy to burn and pro-
heterocyclic compounds pyridine and furan, are composed of duce flames similar to gas flames, while heavier oils require
rings that contain elements other than carbon. For example, a more complicated process and produce flames that are quite
the benzene ring contains six carbon atoms, whereas the radiant and more highly dependent on atomization tech-
pyridine ring contains five carbon atoms and one nitrogen niques than the light oils. Oils are fired in burners by them-
atom. Properties of some of the aromatic compounds are selves, or in combination with fuel gas, waste gas, or both.
contained in Table 5.8. Naphtha is frequently fired in combination with a PSA or
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 178 Monday, March 14, 2005 12:29 PM

178 The John Zink Combustion Handbook

Fuel Gas

C3C4 to Treating

To
Flue Gas Debutanized
System Gasoline

Naphtha

Light Cycle Oil

Heavy Cycle Oil

Main-Column Bottoms

Heavy Recycle (Optional)

Air Fresh Feed

FIGURE 5.38 Flow diagram of UOP fluid catalytic cracking complex.27

other waste gas, and requires good vaporization to provide hot surface. No. 2 oil is quite frequently used as fuel for pro-
a quality flame. cess burners because it will readily burn when injected
through a nozzle into a combustion chamber. No. 2 oil is sig-
5.2.4 Oils nificantly easier to burn than residual oil due to the lack of
According to the American Standard Testing Methods atomization and preheating requirements. Atomization is the
(ASTM) D-396, fuel oils are divided into grades, based on the breaking apart of a liquid into tiny, more easily combustible,
types of burners for which they are suitable.20 The grades are droplets using steam, air, fuel gas, or mechanical means.
determined by those values determined to be most significant These light distillate oils will typically distill out between
in figuring performance characteristics. The two classifica- 450 and 800°F (230 and 430°C).
tions that separate these fuel oils are “distillates” and “residu-
als,” where distillates indicate a distillation overhead product 5.2.4.2 Heavy Oils
(lighter oils) and residuals indicate a distillation bottom prod- No. 4 oil is a heavy distillate oil typically blended from, and
uct (heavier oils). Table 5.11 helps in differentiating between thus having characteristics of, both light distillates and residual
these various classifications; and Table 5.12 reveals typical oils. These oils do not readily combust and therefore require
analyses for these oils. some type of atomization, but still fall into a viscosity range
that does not require preheating prior to burning.
5.2.4.1 Light Oils
Grade 1 and 2 oils are light distillate (fuel) oils used primar- 5.2.4.3 Residual Oils
ily in applications that do not require atomization by air or No. 6 oil is a heavy residual oil sometimes referred to as
steam in order to reduce droplet size for proper burning. No. Bunker C oil. This oil requires significant atomization for
1 oil will typically vaporize when it comes into contact with a proper combustion. Due to its high viscosity, No. 6 oil
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fuels 179

TABLE 5.11 Requirements for Fuel Oils (per ASTM D 396)

No. 1 No. 2 No. 4 No. 6
Classification Distillate Distillate Distillate (Heavy) Residual

Density (kg/m3) @ 60°F (15°C), max 850 876 — —

Viscosity @ 104°F (40°C) mm/s2
min 1.3 1.9 >5.5 —
max 2.1 3.4 24 —
Viscosity @ 212°F (100°C) mm/s2
min — — — 15
max — — — 50
Flash point °F (°C), min 100 (38) 100 (38) 131 (55) 140 (60)
Pour point °F (°C), max –0.4 (–18) 21 (–6) 21 (–6) —
Ash, % mass, max — — 0.1 —
Sulfur, % mass, max 0.5 0.5 — —
Water & sediment, % vol., max 0.05 0.05 0.5 2.0
Distillation temperature °F (°C)
10% volume recovered, max 419 (215) — — —
90% volume recovered, min — 540 (282) — —
90% volume recovered, max 550 (288) 640 (338) — —

TABLE 5.12 Typical Analysis of Different Fuel Oils

No. 1 Fuel Oil No. 2 Fuel Oil No. 4 Fuel Oil No. 6 Fuel Oil (sour)

Ash (%) <0.01 <0.01 0.02 0.05

Hydrogen (%) 13.6 13.6 11.7 11.2
Nitrogen (%) 0.003 0.007 0.24 0.37
Sulfur (%) 0.09 0.1 1.35 2.1
Carbon (%) 86.4 86.6 86.5 85.7
Heat of combustion (HHV), Btu/lb 20,187 19,639 19,382 18,343
Specific gravity 60/60°F 0.825 0.84 0.898 0.97
Density (lb/U.S. gal) 6.877 6.96 7.488 8.08

requires heating during handling and further heating prior to 5.2.6 Physical Properties of Liquid Fuels
combustion chamber injection. No. 6 oil is usually preheated When liquid fuels are encountered, there are certain proper-
to 150 to 200°F (66 to 93°C), to decrease its viscosity, before ties that determine into which category they are divided, and
being atomized and injected into the burner. John Zink rec- for what processes they are suitable for.
ommends a maximum viscosity of 200 SSU (Seconds Say-
bolt Universal) for use in its standard oil guns. Figures 5.39 5.2.6.1 Flash Point
and 5.40 show burners firing heavy oil. The flash point of a liquid is the lowest temperature at which
enough vapors are given off to form a mixture that will ignite
5.2.5 Liquid Naphtha when exposed to an ignition source. The standard method for
determining flash point is ASTM D-93. Under certain condi-
Liquid naphtha is similar in its characteristics to kerosene tions, ASTM D-56 can be used for light distillate oils. Some
(Table 5.13). Figure 5.41 shows a typical naphtha distillation flash point values are provided in Table 5.11. The flash point
curve. In general, naphtha will boil out of a mixture between is an important property for indication of volatility and for
220 and 315°F (100 and 157°C). Naphtha is categorized, storage requirements.
based on its volatility, into light, intermediate, and heavy
naphtha. Naphtha is a major constituent of gasoline; how- 5.2.6.2 Pour Point
ever, it generally requires further refining to make suitable The pour point of a liquid is determined by ASTM D-99 and
quality gasoline. Prior to firing naphtha in a burner, care must indicates the lowest temperature at which an oil will flow at a
be taken to vaporize it so that the combustion will be more controlled rate. If the fluid temperature goes below this point,
complete and uniform.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
flow will be inhibited.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 180 Monday, March 14, 2005 12:29 PM

180 The John Zink Combustion Handbook

TABLE 5.13 Naphtha Elemental Analysis

Component Vol. %

n-Heptane 1.610
Methylcyclohexane 2.433
2-methylheptane 5.618
4-methylheptane 1.824
3-methylheptane 4.841
1c,3-dimethylcyclohexane 3.252
1t,4-Dimethylcyclohexane 1.040
1t,2-Dimethylcyclohexane 1.169
n-Octane 16.334
1c,2-Dimethylcyclohexane 1.674
1,1,4-Trimethylcyclohexane 3.500
2,6-Dimethylheptane 2.094
1c,3c,5-Trimethylcyclohexane 2.638
m-xylene 2.426
p-xylene 0.797
2,3-Dimethylheptane 1.475
4-methyloctane 3.417
2-methyloctane 4.491
3-methyloctane 4.576
o-xylene 1.137
n-Nonane 10.120
Other 23.534

Total 100.000

FIGURE 5.39 Burner firing heavy oil (1).

5.2.6.3 Distillation
The distillation of a liquid gives an indication of its volatility,
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

as well as the ease with which it can be vaporized. The test

evaluates the vaporization range of a fuel between its end
point (the point at which 100% of the volume has vaporized)
and the initial boiling point (the point at which the liquid
begins to vaporize). Figure 5.42 shows a typical crude oil dis-
tillation curve.

5.2.6.4 Viscosity
In layman’s terms, the viscosity is a fluid’s resistance to
flow. Technically, the viscosity is the ratio of shear stress to
shear rate of a fluid in motion. Most fluids under consider-
ation in this chapter (gases, fuel oils) are Newtonian fluids
because the ratio given above is constant with respect to
time, at a given temperature and pressure. A very important
factor in the determination of fluid flow is the dimensionless
quantity called the Reynolds number. The Reynolds number
is calculated as:

DVρ DV
Re = or Re = (5.3)
µ ν

where D = pipe diameter, V = fluid velocity, ρ = fluid density,

µ = fluid absolute viscosity, and ν = fluid kinematic viscosity.
FIGURE 5.40 Burner firing heavy oil (2). When the Reynolds number is less than 2100, the flow is
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 181 Monday, March 14, 2005 12:29 PM

Fuels 181
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 5.41 Naphtha distillation curve.

1000
900 Residue
Boiling Te mperature oF

800
700
Gas-Oil
600
500
400 Kerosene
300
Naphtha
200
Gasoline
100
Butanes and lighter
0
0 10 20 30 40 50 60 70 80 90 100
Cumulative Percent Volume

FIGURE 5.42 Crude oil distillation curve.

182 The John Zink Combustion Handbook

typically streamlined and smooth, and called laminar. specific condition. Water is frequently used as a reference sub-
However, when the Reynolds number increases above 2100, stance and, at 60°F, has a specific gravity of 1.0 and a density
internal agitation takes place, and the flow is considered tur- of 1.94 slugs/ft3 (999 kg/m3), where 1 slug = 1 lbf ft/s2. Specific
bulent. As seen in the Eq. (5.3), as the viscosity increases, the gravity for gases requires an additional assumption relating to
flow becomes more laminar, assuming the other properties pressure and temperature. Gas specific gravity is defined rela-
stay constant. Viscosity is divided into two different catego- tive to air as the reference substance and is generally deter-
ries: kinematic viscosity and absolute viscosity. mined at a standard temperature and pressure. Under those
Kinematic viscosity (ν) is dependent on fluid density, and conditions, gas-specific gravity can be calculated as the ratio of
has units of length2 time–1. Typical units for kinematic vis- molecular weights.
cosity are stokes (0.001 m2 s–1), centistokes (stoke/100), °API runs opposite that of specific gravity; therefore, as
Seconds Saybolt Universal (SSU), and Seconds Saybolt Furol °API increases, the density decreases. When a fluid and water
(SSF). Because the density of a fluid is dependent on tem- are compared at 60°F, the °API can be calculated as:
perature, the viscosity of a fluid is likewise dependent on
temperature. As the temperature increases, the viscosity of a 141.5
°API = − 131.5 (5.5)
fluid will decrease (become more fluid, or less viscous), and SG
vice versa.
Absolute viscosity (µ) can be calculated by multiplying the The specific volume (volume per unit mass) is the recip-
kinematic viscosity by the density of the fluid. The most rocal of the density, and is commonly used in thermodynamic
common units for absolute viscosity are the poise (1 Pa sec) calculations.
and the centipoise (cp), which is poise/100. The specific weight of a fluid (γ) is defined as its weight
The viscosity of oil is a very important consideration in per unit volume. The relationship that relates specific weight
proper burner design. As previously mentioned, the more to the density is γ = ρ × g, where ρ is the density, and g is
viscous the fluid, the more preheating required prior to burning. the local acceleration (32.174 ft/s2). The specific weight of
Several useful conversions are listed below: water at 60°F is 62.4 lbm/ft3 (9.80 kN/m3).

1 lbm/ft hr = 0.00413 g/cm s 5.2.6.6 Heat Capacity (Specific Heat)

0.000413 kg/m s The heat capacity, or specific heat, of a fluid is defined as the
1 centipoise = 0.01 poise
amount of heat that is required per unit mass to raise the
0.01 g/cm s
temperature by one degree. Typical units of heat capacity are
0.001 kg/m s
6.72 × 10–4 lbm/ft s
Btu/(lbm-°R) or kJ/(kg-K) in SI units. Heat capacity is
1 stoke = 0.0001 m2/s = 100 centistokes temperature dependent, and is defined in terms of constant
centistokes = (0.266 × SSU) – (195/SSU) for SSU 32 to 100 volume or constant pressure, as can be seen by the following
(0.220 × SSU) – (135/SSU) for SSU > 100 equations:

Density is a fluid’s mass per unit volume, and is important  δh 

due to its effect on other properties, such as viscosity. Addi- Cv   (5.7)
 δT  v
tionally, the density is used to calculate the heat capacity of
an oil. The densities of liquids are frequently given as the where Cp = the heat capacity at constant pressure, Cv = the
°API or the specific gravity (SG). Density and specific grav- heat capacity at constant volume, δh = change in enthalpy,
ity are related in that a liquid with a specific gravity of 1 has a and δT = change in temperature.
density of 1 kg/m3 (0.0624 lb/ft3). The specific gravity of a
To calculate the heat capacity of a petroleum liquid, to
liquid can be calculated by the formula:
within 2 to 4% accuracy, the following equations can be
employed:
SG = ρ ρref (5.4)

where ρ is the density of the substance in question at specific 0.388 + (0.00045 ∗ °F )

C= for units of Btu/(lb m − °R) (5.8)
conditions, and ρref is the density of a reference substance at a SG
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 183 Monday, March 14, 2005 12:29 PM

Fuels 183

Viscosity, SSU

500,000

100,000

20,000
50,000

10,000
5,000
3,000

1,000
2,000

300
600
400

200
100

32
60

400
200

360
180
160

320
140

Temperature, Degrees Fahrenheit

280
Temperature, Degrees Celsius
120

240
40 50 60 70 80 90 100

100 120 140 160 180 200

HT
NO EAV

G
LI
.6

H
.5
NO

.5
NO
30

80
20

60
NO
10

40
.2
NO
0

20
-10
-20

0
-20
-30

40
30

15
200,000

10;000

400

6
5

2
50

10
5,000
2,000
1,000

200
100

0.5
50,000
20,000

Kinematic Viscosity, Centistokes

FIGURE 5.43 Viscosity of fuel oils.

1.685 + (0.039 ∗ °C) 5.3 GAS PROPERTY CALCULATIONS

C= for units of kJ/(kg − K) (5.9)
SG
5.3.1 Molecular Weight
where C = heat capacity, and SG = specific gravity (relative Molecular weight is the mass in grams of 1 gram-mole of a
density), so long as the liquid temperature is between 32 and chemical compound. Avogadro’s Number defines the num-
400°F (0 and 205°C) and the specific gravity is between 0.75 ber of molecules in a gram-mole to be 6.02252 × 1023, a
and 0.96 at 60°F (16°C).1 fundamental constant. To determine the molecular weight of
Further information about gaseous and liquid fuels and a mixture of gases, it is necessary to know the molecular
their properties can be obtained from the references listed at weight of each compound and the composition of the gases
the end of this chapter.21–24
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
in terms of mole or mass fractions. Having assembled this
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 184 Monday, March 14, 2005 12:29 PM

184 The John Zink Combustion Handbook

information, the following formulae are used to calculate in terms of mole or mass fractions. Having assembled this
molecular weight: information, the following formulae are used to calculate
specific heat (cp and cv formulae are analogous, only cp
MW = ∑ MW × y = 1
(5.10)
formulae are shown):

∑
i i
xi
MWi c p ( vol.) = ∑c p ,i ( vol.) × yi (5.13)
where MW = molecular weight of mixture, MWi = molecular
weight of component i, yi = mole fraction of component i,
and xi = mass fraction of component i.
c p ( mass ) = ∑c p ,i ( mass ) × xi (5.14)

where cp (vol) = specific heat of mixture, volume basis; cp,i (vol)

5.3.2 Lower and Higher Heating Values = specific heat of component i, volume basis; cp (mass) =
The lower heating value (LHV) of a gas is the heat released specific heat of mixture, mass basis; and cp,i (mass) = specific
by combustion of a specific quantity of that gas with the heat of component i, mass basis.
products of combustion remaining as vapor. The higher heat-
ing value (HHV) adds to the LHV the latent heat of any 5.3.4 Flammability Limits
steam produced as a combustion product. It represents the Flammability limits define the range of fuel concentrations in
total heat obtained by first burning a fuel and then cooling the air that will sustain a flame without additional air or fuel. The
products to standard temperature. Heating values may be upper flammability limit (UFL) is the maximum fuel concen-
provided on a volume basis, typically Btu/scf, or a mass basis tration that can sustain a flame and the lower flammability
such as Btu/lbm. limit (LFL) is the minimum. These limits are often tabulated
To determine the heating value of a mixture of gases, it is for fuels at some standard temperature, typically 60°F. Flam-
necessary to know the heating value of each compound and mability limits are not constants for a given gas; they are func-
the composition of the gases in terms of mole or mass fractions of the air/fuel mixture temperature. An extensive
tions. Having assembled this information, the following for- discussion of this subject can be found in Coward and Jones.25
mulae are used to calculate heating values: Wierzba and Karim26 present a method for estimating the
flammability limits as a function of mixture temperature by
HVv = ∑ HV v ,i × yi (5.11) calculating adiabatic flame temperature (AFT). First, the AFT
for the standard temperature mixture is determined. Next, the
mixture temperature is set to the desired level and the fuel
HVm = ∑ HV m ,i × xi (5.12) concentration is varied until the calculated AFT for the non-
standard temperature matches the AFT for the standard tem-
where HVv = heating value of mixture, volume basis, HVv,i = perature. They provide an approximating method for
heating value of component i, volume basis; HVm = heating calculating AFT for sub-stoichiometric mixtures.
value of mixture, mass basis; and HVm,i = heating value of Both the Coward and Jones manuscript and the Wierzba
component i, mass basis. and Karim article indicate that a form of Le Chatelier’s rule
can be used calculate LFL and UFL for many combinations
5.3.3 Specific Heat Capacity of fuels and inerts. Both references also mention that this rule
The specific heat capacity of a gas is the energy that must be fails to accurately predict for a few important situations. One
added to a specific amount of the gas to raise its temperature notable example is a mixture of ethylene and carbon dioxide
by one (1) degree. If the gas is maintained at constant pres- that differs substantially from normal calculated LFL and
sure during this heating process, the value is referred to as cp . UFL. Another example is any mixture of chemicals that is
If the gas is maintained at constant volume, the value is prone to react with another at temperatures below the ignition
referred to as cv . Specific heat is not a constant for a given point, such as ethylene and hydrogen. Mixtures involving
gas; it is a function of temperature. Specific heat can be significant amounts of inert compounds (for example, H2O,
defined on a volume basis, typically Btu/lbmole-°F; or on a N2, and CO2) require special treatment either by the AFT
mass basis such as Btu/lb-°F. method described above or by grouping the inerts with fuel
To determine the specific heat of a mixture of gases, it is components in known proportions matching conditions for
necessary to know the specific heat of each compound at which LFL and UFL have been measured. This latter method
the mixture temperature and the composition of the gases is described in detail by Coward and Jones.25
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Fuels 185

With these exceptions in mind, the following mixing rules determined for a stoichiometric fuel/air mixture, although
can be used to calculate LFL and UFL for most common gas other mixtures such as LFL and UFL are sometimes studied
mixtures: for special purposes, as discussed in Section 5.3.4 above.

100 5.3.6.3 Heat Release

LFL = (5.15)
∑ yi
LFLi
Heat release is the product of the flow rate and the heating
value of the fuel using compatible units. This quantity is used
throughout many areas of interest in combustion, including
100 equipment sizing, radiation, and emissions. Unless the pro-
UFL = (5.16)
∑ yi cess involves the recovery of the heat of vaporization of the
UFLi water vapor, the LHV is usually used when calculating heat
release:
5.3.5 Viscosity
Viscosity is discussed in detail in Chapter 4: “Fundamentals HR = w ↔ LHVm = Q ↔ LHVv (5.18)
of Fluid Flow.” A useful mixing rule Eq. (4.8) is also pro-
vided. where HR = heat release (BTU/hr), w = mass flow (lb/hr),
LHVm = lower heating value (BTU/lb), Q = volumetric flow
5.3.6 Derived Quantities (SCFH), and LHVv = lower heating value (BTU/scf).
In addition to the specific properties described above, there
are a number of useful derived parameters that may be of 5.3.6.4 Volume Equivalent of Flow
interest when studying combustion systems. Volume equivalent of flow (Veq) is the volumetric flow of air
at standard temperature and pressure that produces the same
5.3.6.1 Partial Pressure velocity pressure in the same size line. This quantity is often
Partial pressure is the pressure exerted by a single component
used to provide generalized capacity curves for equipment
of a mixture when that component alone occupies the entire
that may need to handle several different gas streams. When
volume at the mixture temperature. Dalton’s law states that
designing equipment for this situation, the stream with the
the total pressure of a mixture is the sum of the partial pres-
highest Veq will often dominate the hydraulic design, unless
sures of the components. While this law has been demon-
the different streams have different allowable pressure drops.
strated to be somewhat in error, especially at high pressures,
Caution should be used in cases where friction is expected to
it is often useful for estimating purposes to determine
be a major factor in the system pressure drop because Veq
whether a more detailed analysis is justified. The basic rela-
does not account for variations in viscosity.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
tionship is:

pi = yi ↔ TP (5.17) MW Tgas 29 Tgas

Veq = Q = 13.1w (5.19)
29 520 MW 520
Partial pressures are of interest when estimating the probabil-
ity of forming condensate in a gas mixture. When the partial where Veq = volume equivalent (SCFH), and Tgas = gas temp-
pressure of a component exceeds the vapor pressure of that erature (°R).
component at the mixture temperature, condensation is
likely.

5.3.6.2 Adiabatic Flame Temperature

5.4 TYPICAL FLARED GAS
The adiabatic flame temperature is the temperature at which COMPOSITIONS
the enthalpy of the products of combustion equals the sum of Gas compositions sent to flares include a large variety of
the enthalpy of the reactants plus the heat released by the individual compounds. The proportions of these compounds
combustion process. Heat loss due to radiation, convection, vary widely from one facility to another and even within a
or conduction is not included; hence the reference to adia- single facility from minute to minute. The following sections
batic. Accounting for dissociation of combustion products is describe in general terms the kinds of gas streams commonly
important. Customarily, the adiabatic flame temperature is encountered in flare systems.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 186 Monday, March 14, 2005 12:29 PM

186 The John Zink Combustion Handbook

5.4.1 Oil Field/Production Plant Gases the ethylene is mixed with heavier hydrocarbons (pentane,
Gases produced in oil fields generally consist of saturated hexane, hexene, etc.) to alter the properties of the polymer.
hydrocarbon gases (paraffins), together with a certain amount Random mixtures of ethylene and other hydrocarbons may
of inerts. Oil field gases range in MW from 19 to 25. Such be sent to the flare from the main process area. In addition,
gases may contain significant amounts of H2S (sour gas wells) reliefs from various special chemical storage areas may
or CO2. In some cases, especially offshore, these associated send relatively pure materials such as hexane or hexene to
gases are burned continuously in the immediate vicinity of the the flare.
oil wells. In other cases, the gas is sent to a production plant
where it is treated in preparation for pipeline use. 5.4.4 Other Special Cases
Production plants convert the raw associated gas into
several, more valuable products. Undesirable components Landfills and digester facilities produce an off-gas that must
such as H2S, CO2, and water vapor are removed in treatment be disposed of to prevent odor problems in the community.
units. Depending on the composition of the feedstock, pro- The gas is generally a mixture of CO2 and CH4. Landfills are
duction plants may include a debutanizer, a depropanizer, rarely above 30 to 40% methane, while digesters may be as
and a deethanizer to separate the large majority of these high as 60 to 70% methane. In some landfills, perimeter
valuable components. The remainder, mostly methane, wells are used to draw air into the edges of the landfill, which
becomes pipeline-quality natural gas after the addition of prevents the spread of anaerobic bacteria and methane. In
odorants such as mercaptans. Within the production plant, these cases, the methane content is even lower and some air
it may become necessary to flare the raw associated gas, the is also sent to the flare.
pipeline product, or the overhead streams from any of the Marine and truck loading facilities burn the vapor displaced
separation units. from the tankers or trucks during the loading operation. In
many cases, the displaced vapor is mostly air with some
5.4.2 Refinery Gases amount of evaporated gasoline or diesel fuel. Depending on
Refineries treat the liquids produced in the oil fields to gen- the ambient temperature, the resulting mixture could be very
erate many essential materials for public consumption as rich in hydrocarbon vapor, or very lean.
well as further chemical processing. As a result of various
Medical equipment, such as bandages or hypodermic nee-
treatment processes, hydrogen and unsaturated hydrocarbon
dles, is often sterilized by contact with ethylene oxide (ETO)
gases (olefins, diolefins, aromatics, etc.) are produced in
vapors. ETO sterilizer flares are designed to receive the ETO
abundance in a refinery. Due to the wide variety of treat-
vapor after the sterilization process is complete. The compo-
ment processes, the composition of flared gases in a refin-
sition coming to these flares generally consists of a mixture
ery is almost entirely unpredictable. Refinery flaring
of ETO and either air or nitrogen. It should be noted that ETO
generally involves hydrogen, paraffins up to decane, olefins
has a flammability range from 3 to 100% and a very low
up to hexene, diolefins up to butadiene, and aromatics up to
ignition temperature.
ethylbenzene, as well as contaminants such as H2S, CO2,
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

and water vapor. Flares are often used as backup equipment for incinerators
during maintenance or malfunctions. In this type of service,
5.4.3 Ethylene/Polyethylene Gases the waste gas is usually enriched with a substantial amount
of clean fuel gas to ensure reliable burning. Steel mills pro-
Ethylene plants use cracking furnaces to convert feedstock
duce off-gases that consist mainly of H2, H2O, CO, CO2, and
into high-quality ethylene. Some plants use ethane as feed-
air. These are generally low LHV mixtures that also require
stock. The gas produced by such plants is often referred to
enrichment and supplemental fuel firing to maintain ignition.
as light cracked gas, and consists of approximately equal
Fertilizer plants and other chemical plants produce ammonia,
portions of hydrogen, ethane, and ethylene with relatively
which may be sent to a flare in an emergency. Waste gases
little else. Other plants use oil as feedstock and produce
that are sent to flares in these facilities may be pure ammonia
heavy cracked gas. Heavy cracked gas is also approxi-
or diluted with nitrogen or water vapor.
mately equal portions of hydrogen, ethane, and ethylene,
but a substantial fraction of the composition consists of The variety of gases and the hazards associated with each
heavy hydrocarbon gases. requires careful review of all aspects of system design to
Polyethylene plants take the ethylene from the ethylene ensure that these fuels are safely handled, whether in a flare,
plant and polymerize it in a variety of ways. In some cases, a furnace, or an incinerator.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch05-Frame Page 187 Monday, March 14, 2005 12:29 PM

Fuels 187

REFERENCES 15. Microsoft Encarta Encyclopedia, Microsoft Corpora-

tion, 1993–1999.
1. Perry, R.H., Green, D.W., and Maloney, J.O., Eds.,
16. Press, F. and Siever, R., Understanding Earth, W.H.
Perry’s Chemical Engineers’ Handbook, 7th ed.,
Freeman, New York, 1994.
McGraw-Hill, New York, 1997, chap. 27.
2. U.S. Department of Energy, State Energy Data Report, 17. Baird, C., Environmental Chemistry, W.H. Freeman,
Consumption Estimates 1960-1989, DOE/EIA- New York, 1995.
0214(89), May 1991.
18. Solomons, T.W., Organic Chemistry, 5th ed., John
3. Reed, R.J., North American Combustion Handbook,
Wiley & Sons, New York, 1992.
Vol. I, North American Mfg. Co., Cleveland, OH, 1986.
4. Gas Processors and Suppliers Association, GPSA Engi- 19. Peyton, K., Fuel Field Manual, McGraw-Hill, New
neering Data Book, Vol. I, 10th ed., Tulsa, OK, 1987. York, 1998.
5. Austin, G.T., Shreve’s Chemical Process Industries, 20. American Standard Testing Methods, ASTM D-396:
5th ed., McGraw-Hill, New York, 1984, chap. 6.
Standard Specification for Fuel Oils, 1998.
6. Gas Processors and Suppliers Association, GPSA Engi-
neering Data Book, Volume II, tenth edition, Tulsa, 21. Dean, J., Lange’s Handbook of Chemistry, 14th ed.,
OK, 1987. McGraw-Hill, New York, 1992.
7. McCabe, W.L., Smith, J.C., and Harriot, P., Unit Oper- 22. Munson, B.R., Young, D.F., and Okiishi, T.H., Funda-
ations of Chemical Engineering, 5th ed., McGraw-Hill, mentals of Fluid Mechanics, 2nd ed., John Wiley &
New York, 1993. Sons, New York, 1994.
8. Leffler, W.L., Petroleum Refining for the Non-Technical
Person, PennWell Publishing, Tulsa, OK, 1985. 23. Heald, C.C., Cameron Hydraulic Data, 18th ed., Ingersoll-
Dresser Pumps, New Jersey, 1994.
9. Gary, J.H. and Handwerk, G.E., Petroleum Refining,
3rd ed., Marcel Dekker, New York, 1994. 24. Van Wylen, G.J., Sonntag, R.E., and Borgnakke, C.,
10. Reed, R.D., Furnace Operations, 3rd ed., Gulf Pub- Fundamentals of Classical Thermodynamics, 4th ed.,
lishing, Houston, 1981. John Wiley & Sons, New York, 1994.
11. Nelson, W.L., Petroleum Refining Engineering, 3rd ed., 25. Coward, H.F. and Jones, G.W., Limits of Flammability
McGraw-Hill, New York, 1949. of Gases and Vapors, U.S. Bureau of Mines, Dept. of
12. Meyers, R.A., Handbook of Petroleum Refining Processes, Interior, Bulletin 503, Pittsburgh, PA, 1952.
2nd ed., McGraw-Hill, New York, 1997, chap. 6.2.
26. Wierzba, I. and Karim, G.A., Prediction of the flamma-
13. Meyers, R.A., Handbook of Petroleum Refining Processes,
bility limits of fuel mixtures, AFRC/JFRC Inter-
2nd ed., McGraw-Hill, New York, 1997, chap. 12.1.
national Symposium, October, Maui, Hawaii, 1998.
14. Web site for Energy Information Administration,
Office of Gas & Oil, www.eia.doe.gov, Crude Oil 27. Meyers, R.A., Handbook of Petroleum Refining Processes,
Watch, May 24, 2000. 2nd ed., McGraw-Hill, New York, 1997, chap. 3.3.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 6
Pollutant Emissions
Charles E. Baukal, Jr. and Joseph Colannino

TABLE OF CONTENTS
6.1 Introduction............................................................................................................................................. 190
6.1.1 Emissions in the Hydrocarbon and Petrochemical Industries .................................................. 190
6.1.2 Conversions............................................................................................................................... 190
6.2 Nitrogen Oxides (NOx)........................................................................................................................... 191
6.2.1 Theory ....................................................................................................................................... 192
6.2.2 Regulations ............................................................................................................................... 196
6.2.3 Measurement Techniques.......................................................................................................... 197
6.2.4 Abatement Strategies ................................................................................................................ 198
6.2.5 Field Results ............................................................................................................................. 204
6.3 Combustibles........................................................................................................................................... 214
6.3.1 CO and Unburned Fuel ............................................................................................................. 214
6.3.2 Volatile Organic Compounds .................................................................................................... 215
6.4 Particulates .............................................................................................................................................. 217
6.4.1 Sources...................................................................................................................................... 217
6.4.2 Treatment Techniques ............................................................................................................... 218
6.5 Carbon Dioxide ....................................................................................................................................... 218
6.6 SOx ......................................................................................................................................................... 219
6.7 Dioxins and Furans ................................................................................................................................. 219
References ................................................................................................................................................................ 219
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

189
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 190 Monday, March 21, 2005 3:29 PM

190 The John Zink Combustion Handbook

TABLE 6.1 Typical Uncontrolled Combustion Emission provided in 18 groups. Some of those groups of relevance
Factors (lb/106 Btu) by Fuel Type here include both refinery gas-fired and fuel oil-fired boilers,
Fuel Type SOx NOx CO Particulates VOCs and heaters fired on natural gas, refinery gas, oil, and a com-
Distillate fuel 0.160 0.140 0.0361 0.010 0.002 bination of natural gas and refinery oil. The U.S. EPA has
Residual fuel 1.700 0.370 0.0334 0.080 0.009 compiled an extensive list of emission factors for a wide
Other oils 1.700 0.370 0.0334 0.080 0.009 range of industrial processes.5 Chapter 1 of U.S. EPA AP-42
Natural gas 0.000 0.140 0.0351 0.003 0.006
Refinery gas 0.000 0.140 0.0340 0.003 0.006
concerns external combustion sources and focuses on the fuel
LPG 0.000 0.208 0.0351 0.007 0.006 type. Sections 1.3, 1.4, and 1.5 of AP-42 focus on fuel oil com-
Propane 0.000 0.208 0.0351 0.003 0.006 bustion, natural gas combustion, and liquefied petroleum gas
Steam coal 2.500 0.950 0.3044 0.720 0.005
combustion, respectively. Chapter 5 of AP-42 focuses on the
Petroleum coke 2.500 0.950 0.3044 0.720 0.005
Electricity 1.450 0.550 0.1760 0.400 0.004 petroleum industry, where the reader is referred to Sections 1.3
and 1.4 for boilers and process heaters using fuel oil and
Source: From Table 1-11 on p.16 of U.S. Dept. of Energy, Energy &
Environmental Profile of the U.S. Petroleum Refining Industry, 1998. natural gas, respectively. Chapter 6 of AP-42 concerns the
organic chemical process industry. Reis (1996) has written a
general book on environmental issues in petroleum engineer-
ing, including drilling and production operations.6
6.1 INTRODUCTION
The purpose of this chapter is to alert the interested reader to
the potential effects on pollutant emissions of the combustion
6.1.2 Conversions
processes in the petrochemical and hydrocarbon industries. It is often necessary to convert pollutant measurements
There continues to be increasing interest in reducing pollut- (e.g., NOx and CO) into a standard basis for both regulatory
ant emissions of all types from all combustion processes. and comparison purposes. One conversion often necessary is
One prognosticator predicts this will continue well into the from the measured O2 level in the exhaust gases to a standard
future.1 These pollutants have deleterious effects on the envi- basis O2 level. The method for converting measurements to a
ronment and there is evidence they may have an impact on standard basis is given by7:
the health of humans and animals. Efforts are underway from
a broad cross-section of organizations to improve existing  20.9 − O 2 
ppm corr = ppm meas  ref
 (6.1)
 20.9 − O 2meas 
techniques and to develop new techniques for minimizing
pollution. While there are other pollutants potentially pro-
duced in the hydrocarbon and petrochemical industries, this where ppmmeas = Measured pollutant concentration in flue
chapter is only concerned with the air pollutants resulting gases (ppmvd)
from combustion processes. ppmcorr = Pollutant concentration corrected to a
There are numerous factors that affect the pollutant emis- reference O2 basis (ppmvd)
sions generated from the combustion of fuels. The U.S. O2meas = Measured O2 concentration in flue gases
Dept. of Energy has classified emission factors by fuel type (vol.%, dry basis)
for petroleum refining, as shown in Table 6.1.2 A U.S. Envi- O2ref = Reference O2 basis (vol.%, dry basis)
ronmental Protection Agency (U.S. EPA) report identified
the following heater design parameters that affect NOx Example 6.1
emissions from process heaters: fuel type, burner type, com- Given: Measured NOx = 20 ppmvd; measured O2 = 2%
bustion air preheat, firebox temperature, and draft type.3 The on a dry basis.
important factors that influence pollution are considered Find: NOx at 3% O2 on a dry basis.
here. A brief general discussion of emissions from heaters Solution: ppmmeas = 20; O2meas = 2; O2ref = 3
in refineries is given in API 560, section F.10.2.40
20.9 − 3 
ppm corr = 20 = 18.9 ppmvd
 20.9 − 2 
6.1.1 Emissions in the Hydrocarbon and
Petrochemical Industries This example shows that corrected NOx values will be lower
The Western States Petroleum Association (WSPA) and the when the basis O2 is higher than the measured O2 because
American Petroleum Institute (API) worked with the Califor- higher O2 levels mean more air dilution and therefore lower
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

nia Air Resources Board (CARB) to develop air toxics emis- NOx concentrations. The reverse is true when the basis O2 is
sion factors for the petroleum industry.4 Source data was lower than the measured O2 level.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 191 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 191

Another correction that may be required is to convert the NO + O 2 NO 2

measured pollutants from a measured furnace temperature to
a different reference temperature. This may be required when
a burner is tested at one furnace temperature and needs to be
modified to find out the equivalent at another furnace tem-
perature. The correction for temperature is:

 T − Tbasis 
ppm corr = ppm meas  ref  (6.2)
 Tmeas − Tbasis  FIGURE 6.1 Cartoon of NO exiting a stack and com-
bining with O2 to form NO2.
where ppmmeas = Measured pollutant concentration in flue
gases (ppmvd)
ppmcorr = Pollutant concentration corrected to a Example 6.3
reference temperature basis (ppmvd) Given: Measured NOx = 20 ppmvd; measured O2 = 2%
Tref = Reference furnace temperature (°F) on a dry basis; measured furnace temperature =
Tmeas = Measured furnace temperature (°F) 1800°F.
Tbasis = Basis furnace temperature (°F) Find: NOx at 3% O2 on a dry basis at a reference
temperature of 2000°F.
Example 6.2 Solution: ppmmeas = 20; O2meas = 2; O2ref = 3; Tmeas = 1800°F;
Given: Measured NOx = 20 ppmvd; measured furnace Tref = 2000°F assume Tbasis = 400°F
20.9 − 3   2000 − 400 
ppm corr = 20
temperature = 1800°F.
= 21.6
Find: NOx at a reference temperature of 2000°F.  20.9 − 2   1800 − 400 
Solution: ppmmeas = 20; Tmeas = 1800°F; Tref = 2000°F
In this case, the increase in NOx due to the temperature cor-
assume Tbasis = 400°F
rection is greater than the reduction in NOx due to the higher
2000 − 400 
ppm corr = 20
O2 reference.
= 22.9 ppmvd
 1800 − 400 

There are two things to notice in the above example. The 6.2 NITROGEN OXIDES (NOx)
first is that the basis temperature was chosen as 400°F, which NOx refers to the oxides of nitrogen. These generally include
is an empirically determined value that applies to many burn- nitrogen monoxide, also known as nitric oxide (NO), and
ers commonly used in the hydrocarbon and petrochemical nitrogen dioxide (NO2). They may also include nitrous oxide
industries. However, this equation should be used with care (N2O) (also known as laughing gas), as well as other less
for more unique burner designs and when there is a very large common combinations of nitrogen and oxygen such as nitro-
difference between the measured and the reference furnace gen tetroxide (N2O4).
temperatures. The second thing to notice is that the NOx In most high-temperature heating applications, the majority
increases when the reference temperature is higher than the of the NOx exiting the exhaust stack is in the form of nitric
measured temperature, and vice versa. As will be shown later, oxide (NO).8 NO is a colorless gas that rapidly combines with
NOx generally increases with the furnace temperature. O2 in the atmosphere to form NO2 (see Figure 6.1). In the
These two corrections can also be combined into a single lower atmosphere, NO reacts with oxygen to form ozone, in
correction when both the measured O2 level and furnace tem- addition to NO2. NO2 is extremely reactive and is a strong
perature are different from the reference O2 level and furnace oxidizing agent. NO2 decomposes on contact with water to
temperature: produce nitrous acid (HNO2) and nitric acid (HNO3), which
are highly corrosive (see Figure 6.2). When NO2 forms in the
 20.9 − O 2   T − T atmosphere and comes in contact with rain, acid rain is pro-

ppm corr = ppm meas  ref

ref basis
 (6.3) duced. Acid rain is destructive to anything it contacts, includ-
 20.9 − O 2meas   Tmeas − Tbasis  ing plants, trees, and man-made structures like buildings,
bridges, etc. In addition to acid rain, another problem with
where the variables are defined above. NO2 is its contribution to smog. When sunlight contacts a
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

192 The John Zink Combustion Handbook

by the high-temperature reaction of nitrogen with oxygen,

H 2O via the well-known Zeldovich mechanism,11 as given by the
simplified reaction:
NO + O 2 NO 2
N 2 + O 2 → NO, NO 2 (6.4)
HNO 2, HNO 3
Thermal NOx increases exponentially with temperature.
Above about 2000°F (1100°C), it is generally the predomi-
nant mechanism in combustion processes, making it impor-
tant in most high-temperature heating applications. This
means that this mechanism becomes more important when
FIGURE 6.2 Cartoon of acid rain. air preheating or oxygen enrichment22 of the combustion air
are used, which normally increases the flame temperature.
Prompt NOx is formed by the relatively fast reaction between
nitrogen, oxygen, and hydrocarbon radicals. It is given by the
SUN overall reaction:

CH 4 + O 2 + N 2 → NO, NO 2 , CO 2 , H 2 O, trace species (6.5)

NOx + O2 + Reactive HCs
In reality, this very complicated process consists of hundreds
of reactions. The hydrocarbon radicals are intermediate spe-
cies formed during the combustion process. Prompt NOx is
Smog generally an important mechanism in lower temperature
combustion processes.
Fuel NOx is formed by the direct oxidation of organo-
nitrogen compounds contained in the fuel. It is given by the
overall reaction:
FIGURE 6.3 Cartoon of photochemical smog formation.
R x N + O 2 → NO, NO 2 , CO 2 , H 2 O, trace species (6.6)

mixture of NO2 and unburned hydrocarbons in the atmos-

In reality, there are many intermediate reactions for this
phere, photochemical smog is produced (see Figure 6.3).
formation mechanism, as indicated in Figure 6.4. Fuel NOx
Many combustion processes are operated at elevated temper- is not a concern for high-quality gaseous fuels like natural
atures and high excess air levels. The combustion products may gas or propane, which normally have no organically bound
have long residence times in the combustion chamber. These nitrogen. However, fuel NOx may be important when oil
conditions produce high thermal efficiencies and product (e.g., residual fuel oil), coal, or waste fuels are used that may
throughput rates. Unfortunately, such conditions also favor the contain significant amounts of organically bound nitrogen.
formation of NOx. NOx emissions are among the primary air Table 6.2 shows typical thermal and fuel NOx emissions for
pollutants because of their contribution to smog formation, acid process heaters.12 The conversion of fuel-bound nitrogen to
rain, and ozone depletion in the upper atmosphere. It is interest- NOx ranges from 15 to 100%.12 The conversion efficiency is
ing to note that only about 5% of typical NOx sources in an generally higher the lower the nitrogen content in the fuel.
industrial region of the United States come from industrial
sources, compared to 44% from highway and off-road vehi-
6.2.1.2 Important Factors Affecting NOx
cles.10
There are many factors that have an impact on NOx formation.
These include the oxidizer and fuel compositions and tempera-
6.2.1 Theory tures, the ratio of the fuel to the oxidizer, the burner and heater
6.2.1.1 Formation Mechanisms designs, the furnace and flue gas temperatures, and the opera-
There are three generally accepted mechanisms for NOx pro- tional parameters of the combustion system. Some of these are
duction: thermal, prompt, and fuel. Thermal NOx is formed--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
considered next.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 193 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 193

FIGURE 6.4 Schematic of fuel NOx formation pathways. (Adapted from W. Bartok and A.F. Sarofim, Eds., Fossil Fuel
Combustion: A Source Book, John Wiley & Sons, New York, 1991.)

Figure 6.5 shows the predicted NO as a function of the TABLE 6.2 Uncontrolled NOx Emission Factors for
flame stoichiometry for air/fuel flames. NO increases at Typical Process Heaters
fuel-lean conditions and decreases at fuel-rich conditions. Uncontrolled Emission Factor, lb/106 Btu
Figure 2.16 shows a plot of the adiabatic equilibrium flame Model Heater Type Thermal NOx Fuel NOx Total NOxa
temperature for air fuel flames as a function of the flame
ND, natural gas-firedb 0.098 N/A 0.098
equivalence ratio. There are several things to notice. The MD, natural gas-firedb 0.197 N/A 0.197
flame temperature for the air/CH4 flame is very dependent on ND, distillate oil-fired 0.140 0.060 0.200
the stoichiometry. Figure 2.16 helps to explain why, for exam- ND, residual oil-fired 0.140 0.280 0.420
ple, NOx is dramatically reduced under fuel-rich conditions. MD, distillate oil-fired 0.260 0.060 0.320
ND, residual oil-fired 0.260 0.280 0.540
One reason is the dramatic reduction in the flame temperature;
ND, pyrolysis, natural 0.104 N/A 0.104
another reason concerns the chemistry. In a reducing atmos- gas-fired
phere, CO is formed preferentially over NO. This is exploited ND, pyrolysis, high-hydrogen 0.140 d N/A 0.140
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

in some of the NOx reduction techniques. An example is fuel gas-firedc

methane reburn.13 The exhaust gases from the combustion Note: N/A = Not applicable
process flow through a reduction zone that is at reducing ND = Natural draft
MD = Mechanical draft
conditions. NOx is reduced back to N2. Any CO that may
a Total NOx = Thermal NOx + Fuel NOx
have formed in the reduction zone and other unburned fuels b Heaters firing refinery fuel gas with up to 50 mole percent hydrogen can
are then combusted downstream of the reduction zone. How- have up to 20% higher NOx emissions than similar heaters firing natural gas.
c
ever, they are combusted at temperatures well below those High-hydrogen fuel gas is fuel gas with 50 mole% or greater hydrogen
content.
found in the main combustion process. These lower temper- d Calculated assuming approximately 50 mole% hydrogen.
atures are not favorable to NOx formation.
Table 2-1 on p. 2-3 of Sanderford, EPA document, 1993.
Figure 6.6 shows the importance of the gas temperature on Source: From E.B. Sanderford, U.S. EPA Report EPA-453/R-93-015,
thermal NOx formation. The NOx rises rapidly at temperatures February 1993.

Copyright CRC Press

194 The John Zink Combustion Handbook

Equivalence Ratio
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5
5000 5000
Fuel Lean Fuel Rich

4000 H2 4000
CH4
C3H8
NO (ppmvw)

NO (ppmvw)
3000 3000

2000 2000

1000 1000

0 0
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5

Equivalence Ratio

FIGURE 6.5 Adiabatic equilibrium NO as a function of equivalence ratio for air/fuel flames.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Gas Temperature ( oC)

1000 1200 1400 1600 1800 2000 2200
2800 2800

2400 2400

2000 H2 2000
NO (ppmvw)

NO (ppmvw)
C3H8
1600 CH4 1600

1200 1200

800 800

400 400

0 0
1600 2000 2400 2800 3200 3600 4000

Gas Temperature ( oF)

FIGURE 6.6 Adiabatic equilibrium NO as a function of gas temperature for stoichiometric air/fuel flames.

above 2000°F (1100°C) for all three fuels shown. This is a Figure 6.7 shows how NOx increases when the combustion
demonstration of the increase in thermal NOx as a function air is preheated. Air preheating is commonly performed to
of temperature. Many combustion modification strategies for increase the overall thermal efficiency of the heating process.
reducing NOx involve reducing the flame temperature However, it can dramatically increase NOx emissions because
because it has such a large impact on NOx. For example, one of the strong temperature dependence of NO formation.
strategy is to inject water into the flame to reduce NOx by Figure 2.17 shows how the adiabatic flame temperature
cooling down the flame to a lower temperature where NOx increases with air preheating. The increase in NO emissions
formation is less favorable. mimics the increase in flame temperature.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 195 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 195

Air Preheat Temp. ( oC)

0 100 200 300 400 500 600 700 800 900 1000
9000 9000

8000 H2 8000
CH4
7000 C3H8 7000

6000 6000
NO (ppmvw)

NO (ppmvw)
5000 5000

4000 4000

3000 3000

2000 2000
0 200 400 600 800 1000 1200 1400 1600 1800 2000

Air Preheat Temp. ( oF)

FIGURE 6.7 Adiabatic equilibrium NO as a function of air preheat temperature for stoichiometric air/fuel flames.

Fuel Preheat Temperature ( oC)

0 50 100 150 200 250 300 350 400 450 500
3400 3400

H2
3200 3200
CH4
3000 C3H8 3000

2800 2800
NO (ppmvw)

NO (ppmvw)

2600 2600

2400 2400

2200 2200

2000 2000

1800 1800
0 100 200 300 400 500 600 700 800 900 1000

Fuel Preheat Temperature ( oF)

FIGURE 6.8 Adiabatic equilibrium NO as a function of fuel preheat temperature for a stoichiometric air/CH4 flame.

Figure 6.8 shows how NOx increases with the fuel preheat increases as the H2 content in the blend increases. This is
temperature. Fuel preheating is another method used to improve similar to the effect on the adiabatic flame temperature as
the overall thermal efficiency of a heating process. Figure 2.18 shown in Figure 2.19. The second thing to note is that the
shows how the adiabatic flame temperature increases due to fuel effect is not linear between pure CH4 and pure H2. NOx
preheating. The increase in NOx emissions follows the same increases more rapidly as the H2 content increases. The third
pattern as the increase in flame temperature. thing to notice is that there is a significant difference between
Figure 6.9 shows how the fuel composition affects NO for the two extremes as the NOx ranges from a little less than
a blend of CH4 and H2. First, it is important to note that NO 2000 ppmvw to a little more than 2600 ppmvw.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

196 The John Zink Combustion Handbook

2700 2700

2600 2600

2500 2500

2400 2400
NO (ppmvw)

NO (ppmvw)
2300 2300

2200 2200

2100 2100

2000 2000

1900 1900
0 10 20 30 40 50 60 70 80 90 100

H2 in Fuel (vol. %)

100 90 80 70 60 50 40 30 20 10 0
CH4 in Fuel (vol. %)

FIGURE 6.9 Adiabatic equilibrium NO as a function of fuel composition (CH4/H2) for a stoichiometric air/fuel flame.
2000 2000

1800 1800

1600 1600

1400 1400

1200 1200
NO (ppmvw)

NO (ppmvw)
1000 1000

800 800

600 600

400 400

200 200

0 0
0 10 20 30 40 50 60 70 80 90 100

N2 in Fuel (vol. %)

100 90 80 70 60 50 40 30 20 10 0
CH 4 in Fuel (vol. %)

FIGURE 6.10 Adiabatic equilibrium NO as a function of fuel composition (CH4/N2) for a stoichiometric air/fuel flame.

Figure 6.10 shows how the fuel composition affects NO 6.2.2 Regulations
for a blend of CH4 and N2. NO (ppmvw) drops off rapidly as Regulations for NOx vary by country and region. The United
the N2 in the fuel blend increases. At 100% N2, the “fuel” States, Japan, and Germany have some of the strictest regula-
produces no NO. The additional quantity of N2 in the fuel tions. Perhaps the most stringent standards in the world are
does not increase NOx because of the increased availability those enforced by the South Coast Air Quality Management
of N2 to make NOx since there is already plenty of N2 avail- District (SCAQMD). SCAQMD governs the greater Los
able from the combustion air. Angeles area and has proposed rules restricting NOx from
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Pollutant Emissions 197

burners to less than 5 ppmvd corrected to 3% O2 for new regulation for burners. For reference, traditional burners gener-
sources. Currently, there are no burners that can meet these ate ~100 ppm NOx when firing gaseous fuels.
emissions without post-combustion controls. Some states have even more local agencies such as
SCAQMD regulating the greater Los Angeles area, or the
6.2.2.1 Units Bay Area Air Quality Management District (BAAQMD) reg-
Baukal and Eleazer (1995)14 have discussed potential sources ulating the greater San Francisco area. Additionally, there
of confusion in the existing NOx regulations. These sources are various voluntary standards recommended by various
of confusion can be classified as either general or specific. institutes. The general trend is toward more stringent regu-
General sources of confusion include, for example, the wide lation. The large number of governing bodies shows the
variety of units that have been used, reporting on either a dry general public support for stricter pollution control at all
or wet sample basis, measuring NO but reporting NO2, and levels of government.
reporting on a volume vs. a mass basis.
Historically, governing bodies have sprung up regionally 6.2.3 Measurement Techniques
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

for the purpose of regulating specific sources. The governing Accurate measurements of pollutants, such as NO and CO,
bodies have generally adopted units related to a traditional from industrial sources are of increasing importance in view
industry metric. This has led to a wide variety of NOx units. of strict air-quality regulations. Based on such measure-
For example, internal combustion (IC) engines are generally ments, companies may have to pay significant fines, stop pro-
regulated on a gram-per-brake-horsepower (g/bhp) basis — a duction, install expensive flue-gas treatment systems, buy
mass-based unit normalized by the output power of the engine. NOx credits in certain non-attainment areas, or change the
Gas turbines, on the other hand, are generally regulated on a production process to a less polluting technology. If compli-
part-per-million (ppm) basis. Because this unit is volume ance is achieved, however, the company may continue their
based, it must be referenced to a standard condition. Gas processes without interruption and, sometimes, sell their
turbines usually operate near 15% excess oxygen, and tradi- NOx credits. Mandel (1997) notes that the equipment cost for
tionally NOx measurement requires removal of water before the gas analysis system is relatively small compared to the
analysis. Thus, gas turbines often use a ppm measurement maintenance and repair costs.15
referenced on a dry volume basis (ppmdv) to 15% oxygen. Numerous studies have been done and recommendations
In contrast, one typically operates industrial boilers and made on the best ways to sample hot gases from high-
process heaters nearer to 3% excess oxygen. Thus, NOx emis- temperature furnaces. For example, U.S. EPA Method 7E16
sions from those units are generally referenced as ppmvd applies to gas samples extracted from an exhaust stack that are
corrected to 3% oxygen. However, these units can also be analyzed with a chemiluminescent analyzer. A typical sampling
regulated on a mass basis normalized by the heat release of system is shown in Figure 6.11. The major components are: a
the burner, for example, pounds per million Btu (lb/MMBtu). heated sampling probe, heated filter, heated sample line, mois-
Large electrical utilities operate their boilers under very tight ture removal system, pump, flow control valve, and then the
oxygen limits. Therefore, some U.S. agencies regulate utility analyzer. The U.S. EPA method states that the sample probe
boilers on a pound-per-megawatt basis (lb/MW). A further may be made of glass, stainless steel, or other equivalent mate-
complication is whether to normalize the unit by gross output rials. The probe should be heated to prevent water in the com-
power (Gross MW), or to subtract parasitic power losses bustion products from condensing inside the probe.
(Net MW). Foreign regulatory agencies use SI units such as The U.S. EPA method is appropriate for a lower tempera-
grams per normal cubic meter (g/Nm3). ture, nonreactive gas sample obtained, for example, from a
utility boiler. However, this method should not be used to
6.2.2.2 Regulations in the Hydrocarbon and obtain samples from higher temperature industrial furnaces
Petrochemical Industries used in glass or metals production. Flue-gas temperatures
The U.S. Environmental Protection Agency (U.S. EPA) regu- from such furnaces, as well as from some incinerators, can
lates emissions in the hydrocarbon and chemical processing be as high as 2400°F (1300°C). This would cause the probe
industries (HPI and CPI, respectively) nationwide. At the state to overheat and affect the measurements because of high-
level, additional agencies are free to adopt more stringent regula- temperature surface reactions inside the probe.
tions. Examples are the California Air Resources Board (CARB) The effects of probe materials, such as metal and fused
and the Texas National Resource Conservation Commission quartz, as well as the probe cooling requirements, have been
(TNRCC). The TNRCC is proposing a very strict sub-10-ppm investigated for sampling gases in combustion systems.17
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 198 Monday, March 21, 2005 3:29 PM

198 The John Zink Combustion Handbook

Stack Wall

Heated Filter

Moisture Sample Gas

Removal Pump Analyzers
System

Heated Sample Line

Heated Probe
Pump

FIGURE 6.11 Sampling system schematic as recommended by the U.S. EPA.

TABLE 6.3 Reduction Efficiencies for NOx Control Several studies have found that both metal and quartz probe
Techniques materials can significantly affect NO measurements in air/fuel
Draft and Total Effective combustion systems, especially under fuel-rich conditions
Fuel Type Control Technique NOx Reduction Percent with high CO concentrations.18,19 However, the NO readings
ND, distillate (ND) LNB 40 were not affected under fuel-lean conditions.
(MD) LNB 43
(ND) ULNB 76
(MD) ULNB 74 6.2.4 Abatement Strategies
SNCRa 60 Before air-quality regulations, the flue gases from combus-
(MD) SCR 75
(MD) LNB + FGR 43
tion processes were vented directly to the atmosphere. As air-
(ND) LNB + SNCR 76 quality laws tightened and the public’s awareness increased,
(MD) LNB + SNCR 77 industry began looking for new strategies to curb NOx emis-
(MD) LNB + SCR 86
sions. The four strategies for reducing NOx are discussed
ND, residual (ND) LNB 27
(MD) LNB 33 next. Table 6.3 shows typical NOx reduction efficiencies as
(ND) ULNB 77 functions of the burner draft type (natural or forced), fuel
(MD)ULNB 73 (distillate or residual oil), and reduction technique.12 The
SNCR 60
(MD) SCR 75 NOx emissions from gas-fired process heating equipment are
(MD) LNB + FGR 28 highly variable (see Table 6.4).29 Therefore, the technique or
(ND) LNB + SNCR 71 techniques chosen to reduce NOx emissions are very site and
(MD) LNB + SNCR 73
(MD) LNB + SCR 83 equipment dependent. This section is not intended to be
MD, distillate (MD) LNB 45 exhaustive, but is comprehensive and includes many of the
(MD) ULNB 74 commonly used techniques for minimizing NOx emissions.
(MD) SNCR 60
(MD) SCR 75
(MD) LNB + FGR 48 6.2.4.1 Pretreatment
(MD) LNB + SNCR 78 The first NOx reduction strategy can be referred to as pre-
(MD) LNB + SCR 92
treatment. Pretreatment is a preventative technique to mini-
MD, residual (MD) LNB 37
(MD) ULNB 73 mize NOx generation. In pretreatment, the incoming feed
(MD) SNCR 60 materials (fuel, oxidizer, and/or the material being heated)
(MD) SCR 75 are treated in such a way as to reduce NOx. Some of these
(MD) LNB + FGR 34
(MD) LNB + SNCR 75 treatments include fuel switching, using additives, fuel treat-
(MD) LNB + SCR 91 ment, and oxidizer switching.
Note: MD = mechanical draft, ND = natural draft, LNB = low-NOx burner,
ULNB = ultra-low-NOx burner, SNCR = selective noncatalytic reduction, 6.2.4.1.1 Fuel Switching
SCR = selective catalytic reduction, FGR = flue gas recirculation. Fuel switching is simply replacing a more polluting fuel with
a Reduction efficiencies for ND or MD SNCR are equal. a less polluting fuel. For example, fuel oils generally contain
Source: From E.B. Sanderford, U.S. EPA Report EPA-453/R-93-015, some organically bound nitrogen that produces fuel NOx.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
February 1993. Natural gas does not normally contain any organically bound
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 199 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 199

TABLE 6.4 NOx Control Technologies in Process Heaters

Controlled Percent
Control Technology Emissions Reduction

Low-NOx burners 0.1–0.3 lb/106 Btu 25–65

Staged air lances N/A 35–51
Fiber burner 10–20 ppm
Ammonia injection N/A 43–70
Urea injection + Low-NOx burner N/A 55–70
Selective catalytic reduction 20–40 ppm 65–90
Selective catalytic reduction + Low-NOx burner 25–40 ppm 70–90

Note: Uncontrolled emissions are in the range of 0.1 to 0.53 lb/106 Btu. N/A = not available.
Source: From J. Bluestein, Gas Research Institute Report GRI-92/0374, Gas Research
Institute, Chicago, IL, 1992.

nitrogen and usually has only low levels of molecular nitro- 6.2.4.1.4 Oxidizer Switching
gen (N2). Partial or complete substitution of natural gas for The fourth type of pretreatment is oxidizer switching, where
fuel oil can significantly reduce NOx emissions by reducing a different oxidizer is used. Air is the most commonly used
the amount of nitrogen in the fuel. Figure 6.9 shows that CH4 oxidizer. It can be shown that substantial NOx reduction can
produces less NOx than H2. Fuels composed entirely of be achieved using pure oxygen, instead of air, for combus-
hydrogen can produce twice as much NOx as fuels with no tion.21 For example, in the extreme case of combusting a fuel
hydrogen.20 Fuel switching may or may not be an option, like CH4 with pure O2, instead of air that contains 79% N2 by
depending on the availability of fuels and on the economics volume, it is possible to completely eliminate NOx as no N2
of switching to a different fuel. is present to produce NOx. For example, if H2 is combusted
with pure O2, the global reaction can be represented by:
6.2.4.1.2 Additives
Another type of pretreatment involves adding a chemical to
the incoming feed materials (raw materials, fuel, or oxidizer) 2H 2 + O 2 → 2 H 2 O (6.7)
to reduce emissions by changing the chemistry of the com-
bustion process. One example would be injecting ammonia By drastically reducing the N2 content in the system, NOx is
into the combustion air stream as a type of in situ de-NOx minimized. However, there are significant challenges to
process, but only under certain conditions (see Section using high-purity oxygen — instead of air — for combus-
6.2.4.4.1). Several factors must be considered in determining tion.22 This technique has not been used widely in the hydro-
the viability of this option. These include economics, the carbon, petrochemical, and power generation industries, but
effects on the process, and the ease of blending chemicals could become more popular in the future as the cost of oxy-
into the process. gen continues to decline as less expensive methods for sepa-
rating oxygen from air are developed.
6.2.4.1.3 Fuel Pretreatment
A third type of pretreatment involves treating the incoming
fuel prior to its use in the combustion process. An example 6.2.4.2 Combustion Modification
would be removing fuel-bound nitrogen from fuel oil or The second strategy for reducing NOx is known as combus-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

removing molecular nitrogen from natural gas, which can tion modification. Combustion modification prevents NOx
reduce NOx in air- or O2-fuel, and O2-fuel combustion, respec- from forming by changing the combustion process. There are
tively. This is normally an expensive process, depending on numerous methods that have been used to modify the com-
how much treatment must be done and how the fuel is treated. bustion process for low NOx. A popular method is a low
For example, it is generally more difficult to remove nitrogen NOx burner design in which specially designed burners gen-
from fuel oil than from natural gas. In Europe, some natural erate less NOx than previous burner technologies. Low NOx
gas supplies have as much as 15% N2 by volume. If only a few burners may incorporate a number of techniques for mini-
percent N2 needs to be removed from that type of natural gas, mizing NOx, including flue-gas recirculation, staging, pulse
this can usually be done relatively easily and inexpensively combustion, and advanced mixing. Common combustion
with adsorption or membrane separation techniques. modification techniques are discussed next.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 200 Monday, March 21, 2005 3:29 PM

200 The John Zink Combustion Handbook

6.2.4.2.1 Air Preheat Reduction lems, especially the high-frequency cycling of the switching
One combustion modification technique is reducing the com- valves, that have not been satisfactorily resolved yet.
bustion air preheat temperature. As shown in Figure 6.8,
reducing the level of air preheat can significantly reduce NOx 6.2.4.2.3 Staging
emissions. Air preheat greatly increases NOx for processes Staged combustion is an effective technique for lowering
that use heat recuperation. However, reduction of air preheat NOx. Staging means that some of the fuel or oxidizer, or
also reduces the overall system efficiency, as shown in both, is added downstream of the main combustion zone. The
Figure 2.23. The loss of efficiency can be somewhat mitigated fuel, oxidizer, or both can be staged into the flame. For exam-
if the heater is equipped with a convection section. This is a ple, there may be primary and secondary fuel inlets where a
fairly easy technique to implement and may be cost-effective portion of the fuel is injected into the main flame zone, and
if the lost efficiency is more than offset by alternative NOx the balance of the fuel is injected downstream of that main
reduction techniques. flame zone. In fuel staging, some of the fuel is directed into
the primary combustion zone, while the balance is directed
6.2.4.2.2 Low Excess Air into secondary and even tertiary zones in some cases (see
As shown in Figure 6.5, excess air increases NOx emissions. Figures 1.36 and 1.37). This makes the primary zone fuel-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

The excess air generally comes from two sources: the combus- lean, which is less conducive to NOx formation when com-
tion air supplied to the burner and air infiltration into the heater. pared to stoichiometric conditions. The excess O2 from the
Excess air produced by either source is detrimental to NOx primary zone is then used to combust the fuel added in the
emissions. Excess air increases NOx formation by providing secondary and tertiary zones. While the overall stoichiometry
additional N2 and O2 that can combine in a high-temperature can be the same as in a conventional burner, the peak flame
reaction zone to form NO. In many cases, NOx can be reduced temperature is much lower in the staged fuel case because the
by simply reducing the excess air through the burners. combustion process is staged over some distance while heat
Air infiltration, sometimes referred to as tramp air, into a is simultaneously being released from the flame. The lower
combustion system affects the excess air in the combustor temperatures in the staged fuel flame help to reduce the NOx
and can affect NOx emissions. The quantity and location of emissions. Thus, fuel staging is effective for two reasons: (1)
the leaks are important. Small leaks far from the burners are the peak flame temperatures are reduced, which reduces
not nearly as deleterious as large leaks near the flames. By NOx; and (2) the fuel-rich chemistry in the primary flame
reducing air infiltration (leakage) into the furnace, NOx can zone also reduces NOx. Waibel et al. (1986) have shown that
be reduced because excess O2 generally increases NOx. fuel staging is one of the most cost-effective methods for
There is also an added benefit in reducing excess air. reducing NOx in process heaters.23
Reducing the excess O2 in a combustion system is also useful In air staging, some of the combustion air is directed into
for maximizing thermal efficiency because any unnecessary the primary combustion zone, while the balance is directed
air absorbs heat that is then carried out of the stack with the into secondary and even tertiary zones in some cases
exhaust products. However, there is a practical limit to how (see Figures 1.34 and 1.35). This makes the primary zone fuel-
low the excess O2 can be. Because the mixing of the fuel and rich, which is less conducive to NOx formation when com-
air in a diffusion flame burner is not perfect, some excess air pared to stoichiometric conditions. The unburned combusti-
is necessary to ensure both complete combustion of the fuel bles from the primary zone are then combusted in secondary
and minimization of CO emissions. The limit on reducing the and tertiary zones. While the overall stoichiometry may be the
excess air is CO emissions. If the excess O2 is reduced too same as in a conventional burner, the peak flame temperature
much, then CO emissions will increase. CO is not only a is much lower in the staged air case because the combustion
pollutant, but also an indication that the fuel is not being fully process is staged over some distance while heat is simulta-
combusted, resulting in lower system efficiencies. neously being released from the flame. The lower temperatures
There are some special techniques that control the O2 in in the staged air flame help reduce the NOx emissions.
the flame to minimize NOx. One example is pulse combus-
tion, which has been shown to reduce NOx because the alter- 6.2.4.2.4 Gas Recirculation
nating very fuel-rich and very fuel-lean combustion zones Furnace gas recirculation is a process that causes the prod-
minimize NOx formation. The overall stoichiometry of the ucts of combustion inside the combustion chamber to be
oxidizer and fuel is maintained by controlling the pulsations. recirculated back into the flame (see Figure 6.12). This is
Pulse combustion is not being used in many industrial com- sometimes referred to as internal flue gas recirculation.
bustion processes at this time due to some operational prob- External flue gas recirculation (see Figure 1.28) is similar.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 201 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 201

FIGURE 6.12 Schematic of furnace gas recirculation.

External flue gas recirculation causes the exhaust gases in the capital intensive than most post-treatment methods. In many
flue to be recirculated back through the burner into the flame cases, there is a limit to how much NOx reduction can be
via ductwork external to the furnace. Although the furnace or achieved using these methods.
flue gases are hot, they are considerably cooler than the flame Another form of water injection is to inject water in the form

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
itself. The cooler furnace or flue gases act as a diluent, reduc- of steam. There are several reasons for this. One is that steam
ing the flame temperature, which in turn reduces NOx (see is much hotter than liquid water and already includes the latent
Figure 6.6). Advanced mixing techniques use carefully heat of vaporization needed to change the liquid water to a
designed burner aerodynamics to control the mixing of the vapor. When liquid water is injected into a combustion process,
fuel and the oxidizer. The goals of this technique are to avoid it can put a large heat load on the process because liquid water
hot spots and make the flame temperature uniform, to can absorb a large amount of energy before becoming a vapor
increase the heat release from the flame, which lowers the due to its high latent heat of vaporization. Steam puts a much
flame temperature, and to control the chemistry in the flame smaller load on the process because it absorbs less energy than
zone to minimize NOx formation. liquid water. Another reason for using steam instead of liquid
External flue gas recirculation requires some type of fan to water is that steam is already in vapor form and mixes readily
circulate the gases external to the furnace and back through with the combustion gases. Liquid water must be injected
the burner. The burner must be designed to handle both the through nozzles to form a fine mist to disperse it uniformly
added volume and different temperature of the recirculated with the combustion gases. Therefore, it is often easier to blend
gases that are often partially or fully blended with the com- steam into the combustion products compared to liquid water.
bustion air. Garg (1992) estimates NOx reductions of up to At equivalent injection rates, water injection is usually more
50% using flue gas recirculation.24 effective at reducing NOx than steam injection.

6.2.4.2.5 Water or Steam Injection 6.2.4.2.6 Reburning

Many of the combustion modification methods attempt to Reburning is a technique similar to fuel staging but uses a
reduce the temperature of the flame to lower NOx emissions. different strategy. An example is methane reburn, where
In many cases, this may result in a reduction of the combus- some methane is injected in the exhaust gases, usually well
tion efficiency.25 For example, if water is injected into the after the primary combustion zone, in which the gases are at
flame to lower NOx, the water absorbs heat from the flame a lower temperature. As previously shown in Figure 6.5, fuel-
and carries most of that energy out with the exhaust gases, rich conditions are not favorable to NOx. As the exhaust
thus preventing the transfer of much of that energy to the gases from the combustion process flow through this fuel-
load. Combustion modification methods are usually less rich reducing zone, NOx is reduced back to N2. Any CO and
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 202 Monday, March 21, 2005 3:29 PM

202 The John Zink Combustion Handbook

TABLE 6.5 NOx Reductions for Different Low-NOx the air needed to fully combust the fuel. Rather than air stag-
Burner Types ing in individual burners, the BOOS technique stages air over
Typical NOx the entire boiler. This technique is relatively inexpensive to
Reductions implement. Ensuring proper heat distribution is important to
Burner Type (%)
prevent overheating the tubes or derating the firing capacity.
Staged-air burner 25–35
Staged-fuel burner 40–50
Low-excess-air burner 20–25 6.2.4.3 Process Modification
Burner with external FGRa 50–60 There are a number of techniques that can be employed to
Burner with internal FGRa 40–50 change the existing process in such a way as to reduce NOx
Air or fuel-gas staging with internal FGRa 55–75
Air or fuel-gas staging with external FGRa 60–80
emissions. These methods are often more radical and expen-
sive, and are not often employed except under somewhat
a FGR = Flue gas recirculation.
unique circumstances. These must be analyzed on a case-by-
Source: A. Garg, Chem. Eng. Prog., 90(1), 46–49, 1994. With permission.
case basis to see if they are viable.

6.2.4.3.1 Reduced Production

other unburned fuels in the exhaust gases are then combusted
If the mass of NOx emitted from a plant is too high, an alter-
downstream of the reduction zone at temperatures well below
native is to reduce the firing rate, which means a correspond-
those found in the main combustion process. These lower
ing reduction in production. The reduction in NOx is
temperature reactions are not favorable to NOx formation, so
proportional to the reduction in firing rate as less fuel is
the net effect is that NOx is reduced.
burned and therefore less NOx is formed. However, this is
There are some challenges with this technique. One is to
generally not a preferred alternative, for obvious reasons, as
get proper injection of the reburning gas and the exhaust
less of the product being made is available to sell. Depending
products. Another is that the reburn zone must be capable of
on the costs to reduce NOx, this may be the most economic
sustaining combustion. It needs to be done in a lower tem-
alternative in some cases.
perature part of the process to minimize the subsequent for-
In boilers, reducing the firing rate reduces the overall
mation of NOx. It may be done, for example, in a previously
temperature inside the boiler, which reduces thermal NOx
uninsulated portion of ductwork. This may require replace-
formation.28 This technique is known as derating and is not
ment with higher temperature materials and the new duct may
desirable if the boiler is capacity limited, but in certain limited
need to be insulated. A third challenge is trying to take advan-
applications it may be a viable alternative.
tage of some of the energy produced during the reburning. A
heat recovery system may need to be added. 6.2.4.3.2 Electrical Heating
One process modification that is sometimes used to minimize
6.2.4.2.7 Low NOx Burners
Garg (1994) discusses the use of various low NOx burners to or eliminate NOx emissions is to replace some or all of the
achieve emissions reductions compared to standard gas fossil-fuel-fired energy with electrical energy. The electrical
burners.26 Table 6.5 shows typical NOx reductions using energy produces no NOx emissions at the point of use and
various low NOx burner techniques. An EPA study found that moves the emissions to the power plant. In general, the
ultra low-NOx burners were the most cost-effective means to resulting NOx emissions at the power plant are often lower
reduce NOx.12 Low NOx burners typically incorporate one or than at an industrial site because of the strict limits imposed
more of the techniques discussed in Section 6.2.4.2. on the power plant and the various methods employed to
minimize NOx, which are often more cost-effective on a unit
6.2.4.2.8 Burner Out-of-Service (BOOS) mass basis because of the economies of scale.
This is a technique primarily used in boilers where the fuel is There are a number of potential problems with this method.
turned off to the upper burners, while maintaining the air flow The first is that the economics are usually very unfavorable
to all.27 The fuel removed from the upper burners is then redi- when replacing fossil fuels with electrical energy. In most
rected to the lower burners, while maintaining the same air hydrocarbon and petrochemical processes, the fuel used in
flow to the lower burners. Therefore, the overall fuel and air the heaters is a by-product that is available at little or no cost.
flow to the boiler remains the same but is redistributed. This On the other hand, electrical energy is often much more
makes the lower burners fuel-rich, which is less conducive to expensive than even purchased fossil fuels like natural gas or
NOx formation due to the lower flame temperatures and fuel-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
oil. Besides the higher operating costs, there would be
rich chemistry. The upper burners (air only) provide the rest of substantial capital costs involved in converting some or all of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 203 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 203

the existing fossil-energy heating to electricity. Besides the energy consumption means less NOx generated. However,
removal of the existing burners, there would be the cost of the this is obviously not an option in most cases, and is only con-
new electrical heaters and often large costs of installing elec- sidered under extreme circumstances. In the above example,
trical substations that would be required for all the additional purer or “sweeter” crudes are much more expensive raw
power. In many parts of the country, large additional sources materials than less pure or more “sour” crudes. Therefore, the
of electricity are not readily available, so a new source of savings in energy may be more than offset by the higher raw
electricity may need to be built at the plant, such as a co- material costs.
generation facility. However, although the electrical costs can
be reduced in that scenario because the transmission losses are 6.2.4.4 Post-treatment
much lower, the NOx emissions are now at different locations The fourth strategy for minimizing NOx is known as post-
at the site and little may then be gained in reducing overall NOx treatment. Post-treatment removes NOx from the exhaust
emissions for the plant. It is likely in the future that regulations gases after the NOx has already been formed in the combus-
will consider the net NOx generated during the production of tion chamber. Two of the most common methods of post-
a product and would include the NOx formed in the generation treatment are selective catalytic reduction (SCR) and selec-
of electricity. This will make replacement of fossil energy with tive noncatalytic reduction (SNCR).29 Wet techniques for
electricity less attractive as most of the power generated in the post-treatment include oxidation-absorption, oxidation-
United States is by fossil-fuel-fired power plants. absorption-reduction, absorption-oxidation, and absorption-
reduction. Dry techniques for post-treatment, in addition to
6.2.4.3.3 Improved Thermal Efficiency SCR and SNCR, include activated carbon beds, electron
By making a heating process more efficient, less fuel needs beam radiation, and reaction with hydrocarbons. One of the
to be burned for a given unit of production. Because the firing advantages of post-treatment methods is that multiple
rate is directly proportional to NOx emissions, less fuel used exhaust streams can be treated simultaneously, thus achiev-
equals less NOx produced. There are many ways to improve ing economies of scale. Most of the post-treatment methods
the efficiency of a process. A few representative examples are relatively simple to retrofit to existing processes.
will be given. One is to repair the refractory and air infiltra- Many of these techniques are fairly sophisticated and are not
tion leaks on an existing heater. This is often relatively trivial to operate and maintain in industrial furnace environ-
inexpensive and saves fuel while reducing NOx. Another is ments. For example, catalytic reduction techniques require a
to add heat recovery to a heating process that does not cur- catalyst that may become plugged or poisoned fairly quickly by
rently have it. The heat recovery could be in several forms. dirty flue gases. Post-treatment methods are often capital inten-
One method is to preheat the incoming combustion air. As sive. They usually require halting production if there is a mal-
previously discussed, this can increase NOx emissions due to function of the treatment equipment. Also, post-treatment does
the higher flame temperatures if it is not done properly. not normally benefit the combustion process in any way. For
Another method is to add a convection section onto a heater example, it does not increase production or energy efficiency.
that does not presently have it. This has other operational It is strictly an add-on cost. A good reference for post-treatment
benefits as well and is often a good choice. A more drastic NOx control for heaters used in refineries is API 536.39
method of increasing the thermal efficiency of a heating pro-
cess is to replace an old, existing heater with a new, more 6.2.4.4.1 Selective Catalytic Reduction (SCR)
modern design. This may make sense if the existing heater is Selective catalytic reduction (SCR) involves injecting an
very old, is high maintenance, and is not easily repairable or NOx-reducing chemical into an exhaust stream in the pres-
upgradable. However, new sources often must meet more ence of a catalyst within a specific temperature window. The
stringent NOx standards than existing sources. chemical is typically ammonia and the temperature window
is approximately 500 to 1100°F (230 to 600°C). The NOx
and NH3 react on the catalyst surface to form N2 and H2O.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

6.2.4.3.4 Product Switching

Another radical process modification that can reduce NOx is The important reactions are:42
to switch the product being produced to one that requires less
6 NO + 4 NH 3 → 5N 2 + 6H 2 O
energy to process. In a process heater, this would involve
replacing the existing process fluid with one that requires less 2NO + 4 NH 3 + 2O 2 → 3N 2 + 6H 2 O
energy to heat. For example, the heavier crude oils require
more energy to process than lighter, purer crudes, so less There are a number of potential problems and challenges
energy would be needed to process the purer crudes. Less with SCR techniques. The catalyst introduces a pressure drop
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 204 Monday, March 21, 2005 3:29 PM

204 The John Zink Combustion Handbook

into the system, which often increases the power require- chemicals slip through the exhaust without reacting, referred
ments for the gas-handling equipment. The catalyst may to as ammonia slip when ammonia exits the stack).12
become plugged or fouled in dirty exhaust streams, which CO can be elevated with large quantities of NH3 because
is especially a challenge when firing liquid fuels like residual the NH3 competes for the OH radical, the main oxidation
oil. The ammonia must be properly injected into the flue route for CO:41
gases to get proper mixing, at the right location to be in the
proper temperature window, and in the proper amount to get CO + OH → CO 2 + H
adequate NOx reduction without allowing ammonia to slip
NH 3 + OH → NH 2 + H 2 O
through unreacted. SCR systems are not very tolerant of
constantly changing conditions, as a stable window of oper- There are also safety concerns with regard to the transport
ation is required for optimum efficiency. Another problem is and storage of the ammonia (NH3) used in SNCR. Other
handling the spent catalyst. Regeneration is often most major challenges of this technology include finding the
attractive but may be more expensive than buying new cat- proper location in the process to inject the chemicals (the
alyst. Disposal of the spent catalyst may be expensive as it chemicals must be injected where the flue gases are within a
may be classified as a hazardous waste, especially if the relatively narrow temperature window for optimum effi-
catalyst contains vanadium, as is commonly the case. A U.S. ciency); injecting the proper amount of chemicals (too much
EPA study found that SCR was the most expensive means will cause some chemicals to slip through unreacted, and too
to reduce NOx.12 little will not get sufficient NOx reductions); and getting
proper mixing of the chemicals with the flue gas products
6.2.4.4.2 Selective Noncatalytic Reduction (SNCR) (there must be both adequate mixing and residence time for
Selective noncatalytic reduction (SNCR) involves injecting the reactions to go to completion). Both physical and com-
NOx-reducing chemicals into the exhaust products from a puter modeling are often used to determine the optimal place,
combustion process within a specific temperature window.39 amount, and method of injection.
No catalyst is involved in the process. The most commonly
used chemicals are ammonia and urea. Other chemicals
6.2.5 Field Results
(e.g., hydrogen, hydrogen peroxide, and methanol) can be
6.2.5.1 Conversions
added to improve the performance and lower the minimum
It is important to be able to convert field measurements to
threshold temperature. The Exxon thermal de-NOx process is
specific units to determine whether the emissions from a
one common SNCR technique using ammonia, and is
specific burner or heater are below their allowable limits. In
employed in a wide variety of industrial applications. A
nearly all cases, NOx is measured on a ppmvd basis. The
typical global reaction for this technique can be written as:
following examples will show how to convert these units to a
specific basis.
2 NO + 4 NH 3 + 2O 2 → 3N 2 + 6H 2 O
Example 6.4
The optimum temperature window, without the addition of Given: Fuel = methane with a gross or higher heating
other chemicals to increase the temperature window, is 1600 value of 1012 Btu/ft3; NO = 20 ppmvd; measured
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

to 2200°F (870 to 1200°C). The Nalco Fuel Tech NOxOUT® O2 = 2% on a dry basis.
is a common SNCR technique employing urea: Find: NOx as NO2 in lb/106 Btu (gross).
Solution: First calculate dry flue gas products.
CO( NH 2 )2 + 2 NO + 1 2O 2 → 2 N 2 + CO 2 + 2 H 2 O Global chemical reaction:

CH4 + x(O2 + 3.76N2) = CO2 + 2H2O + yO2 +

The optimum temperature window, without the addition of 3.76xN2
other chemicals to increase the temperature window, is 1600
to 2000°F (870 to 1100°C). where O2 + 3.76N2 is the composition of air
There are some problems with SNCR. The first is the cost, (79% N2, 21% O2)
which is usually significantly more than non-post-treatment
1. given 2% O2 in dry flue gases:
techniques like low-NOx burners. Although the use of SNCR
decreases NOx, it may increase other undesirable emissions y
= 0.02
such as CO, N2O and NH3 (which can occur if the injected 1 + y + 3.76 x
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 205 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 205

2. O atom balance: 6500

2x = 2 + 2 + 2y = 4 + 2y, or x = 2 + y 6000

NO (ppmvd)
5500
Solving 1 and 2 simultaneously:
5000
CH4 + 2.188(O2 + 3.76N2) = 4500
CO2 + 2H2O + 0.188O2 + 8.23N2
4000
0 10 20 30 40 50 60 70 80 90 100
This shows the moles of products for each mole
of CH4. Note that NO in the flue products has H 2 in CH 4 (vol. %)
been ignored because it is only present in trace
FIGURE 6.13 Adiabatic equilibrium NO as a function
amounts. Assume that all NO is converted to NO2
of the fuel blend composition for H2/CH4 blends combusted
in the atmosphere. with 15% excess air where both the fuel and the air are at
ambient temperature and pressure.
(988 ft3 CH4)(1012 Btu/ft3 CH4) =
1 × 106 Btu (gross)
4400
3
(988 ft CH4)(1 + 0.188 + 8.23) = 4350
4300
NO (ppmvd)

9305 ft3 dry combustion products at STP per 106 Btu

4250
4200
Given 20 ppmvd NO2 =
4150
(20 ft3 NO2/106 ft3 dry products)(9305 ft3 dry 4100
products/106 Btu) = 0.186 ft3 NO2/106 Btu 4050
4000
Density of NO2 = 0.111 lb/ft3 0 10 20 30 40 50 60 70 80 90 100
C3H8 in CH 4 (vol. %)
Mass of NO2 in exhaust products =
(0.186 ft3 NO2/106 Btu)(0.111 lb NO2/ft3 NO2) = FIGURE 6.14 Adiabatic equilibrium NO as a function
0.021 lb NO2/106 Btu (gross) of the fuel blend composition for C3H8/CH4 blends combusted
with 15% excess air where both the fuel and the air are at
ambient temperature and pressure.
6.2.5.2 Fuel Composition Effects
The composition of the fuel supplied to a combustion system
has a significant impact on the NOx emissions. In the petro- 6500
chemical and chemical process industries, there is a very 6000
NO (ppmvd)

wide range of fuel blends used for process heating. These

5500
fuels are often by-products of a refining process. They typi-
cally contain hydrocarbons ranging from C1 to C4, hydro- 5000
gen, and inert gases like N2 and CO2. In a given plant or 4500
refinery, burners used in process heaters may need to be
4000
capable of firing on multiple fuels that are present at different
0 10 20 30 40 50 60 70 80 90 100
times (e.g., start-up, normal operation, upset conditions,
etc.). In many cases, the NOx emissions from the heaters H2 in C3H 8 (vol. %)
may not exceed a given value regardless of what fuel compo-
FIGURE 6.15 Adiabatic equilibrium NO as a function
sition is being fired. Therefore, it is critical that the effects of
of the fuel blend composition for H2/C3H8 blends combusted
the fuel composition on NOx emissions be understood and with 15% excess air where both the fuel and the air are at
quantified to ensure that permitted values are not ambient temperature and pressure.
exceeded. --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

206 The John Zink Combustion Handbook

0 0
1 1

0.2 0.2
0.8 0.8

Fra

Fra
0.4 0.4
2

2
0.6

ctio

ctio
nH

nH
0.6

nT
ctio

ctio
NG

NG
Fra

Fra
0.6 0.6
0.4 0.4

0.8 0.8
0.2 0.2

1 1
0 0

1 0.8 0.6 0.4 0.2 0 1 0.8 0.6 0.4 0.2 0

Fraction C3H8 Fraction C3H8

Adiabatic Flame Relative NOx (ppmv) <= 0.500 <= 0.575
<= 3475.000 <= 3620.714
Temperature (F) <= 3511.429 <= 3657.143 Equilibrium <= 0.650 <= 0.725

<= 3547.857 <= 3693.571 Combustion <= 0.800 <= 0.875

<= 0.950 > 0.950

<= 3584.286 > 3693.571 Model
FIGURE 6.16 Ternary plot of adiabatic equilibrium NO (fraction of the maximum value) as a function of the fuel blend
composition for H2/CH4/C3H8 blends combusted with 15% excess air where both the fuel and the air are at ambient temperature
and pressure.

This section shows the results of an extensive series of tests

to study the effects of fuel composition on NOx emissions
from an industrial-scale burner.43 The data provide additional
insight into effects on NOx over the entire range of fuel
compositions consisting of various fractions of three primary
components: H2, C3H8, and CH4. Figures 6.13 – 15 show how
NOx theoretically varies for two-component fuel mixtures of
CH4–C3H8, CH4–H2 and C3H8–H2, respectively. These figures
show the predicted adiabatic equilibrium NO concentrations
for flames with 15% excess air. Figure 6.16 shows a ternary
diagram of the calculated adiabatic flame temperatures (figure
on the left) over the range of three-component fuel blends
tested and another ternary diagram showing the predicted
adiabatic equilibrium NO (figure on the right) for three-com-
ponent fuel blends containing CH4, C3H8 and H2 combusted
with 15% excess air.
Experiments were conducted using a conventional-type
FIGURE 6.17 Raw gas (VYD) burner (Courtesy of John burner (see Figure 6.17) with a single fuel gas tip and flame-
Zink Co.). holder. The burner was fired vertically upward in a rectangular
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Pollutant Emissions 207

FIGURE 6.18 Test furnace.

furnace (see Figure 6.18). The test furnace was a rectangular The experimental matrix consisted of firing the burner at
heater with internal dimensions of 8 ft (2.4 m) wide, 12 ft a constant heat release (7.5 × 106 Btu/hr or 2.2 MW) and
(3.7 m) long, and 15 ft (4.6 m) tall. The furnace was cooled excess air level (15%) with 15 different fuel blends comprised
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

by a water jacket on all four walls. The interior of the water- of varying amounts of H2, C3H8, and Tulsa Natural Gas
cooled walls was covered with varying layers of refractory (TNG).* For testing and analysis purposes TNG was treated
lining to achieve the desired furnace temperature. The burner as a single fuel component for convenience. TNG, which is
was tested at a nominal heat release rate of 7.5 × 106 Btu/hr comprised of approximately 93% CH4, is a more economical
(2.2 MW). choice than pure CH4 for experimental work, and the analysis
A velocity thermocouple (also known as a suction thermo- is simplified by treating it as a single component. All 15 fuel
couple or suction pyrometer — see Chapter 14) was used to compositions were tested on each of six different fuel gas
measure the furnace and stack gas temperatures. The furnace tips, which differed in port diameter sizes, to enable the
draft was measured with an automatic, temperature-compen- acquisition of additional information regarding effects result-
sated, pressure transducer as well as an inclined manometer ing from differing fuel pressures.
connected to a pressure tap in the furnace floor. Fuel flow Figure 6.19 shows the variation in relative measured NOx
rates were measured using calibrated orifice meters, fully emissions resulting from different concentrations (volume
corrected for temperature and pressure. Emission levels were basis) of H2 in a fuel blend composed with a balance of TNG
measured using state-of-the-art continuous emissions moni-
tors (CEMs) to measure emissions species concentrations of *The nominal composition by volume of TNG is 93.4% CH4, 2.7% C2H6,
NOx, CO, and O2. 0.60% C3H8, 0.20% C4H10, 0.70% CO2, and 2.4% N2.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 208 Monday, March 21, 2005 3:29 PM

208 The John Zink Combustion Handbook

FIGURE 6.19 Measured NOx (percent of the maximum ppmv value) as a function of the fuel blend composition for
H2/TNG blends combusted with 15% excess air where both the fuel and the air were at ambient temperature and pressure.

FIGURE 6.20 Measured NOx (percent of the maximum ppmv value) as a function of the fuel blend composition for
C3H8/TNG blends combusted with 15% excess air where both the fuel and the air were at ambient temperature and pressure.

for each of the six different fuel gas tips tested. The plot, which effect of H2 is significant, with the sharpest increase in NOx
illustrates NOx levels on a concentration basis, clearly shows levels taking place as concentration levels of H2 in the fuel
the correlation between increased H2 content and higher NOx mixture rise from 75% to 100%.
emission levels. The slope of the profile is exponentially The variation in relative measured NOx emissions result-
increasing, qualitatively similar to that predicted by the plotted ing from different concentrations (volume basis) of C3H8 in
theoretical calculations shown previously in Figure 6.13. The a fuel blend composed with a balance of TNG is shown in
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Pollutant Emissions 209

FIGURE 6.21 Measured NOx (percent of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel
blend composition for H2/C3H8 blends combusted with 15% excess air where both the fuel and the air were at ambient temperature
and pressure.

Figure 6.20. The slope of the increase in NOx levels corre- of NOx emissions produced. The lower plot, which shows
sponding to increased concentrations of C3H8 is shown to be the variation in measured NOx levels on a mass per unit heat
relatively constant or slightly declining over the gradient in release basis, illustrates that the overall emissions of NOx on
C3H8 concentration, in contrast with the exponentially a mass basis decrease with increasing fuel hydrogen content
increasing profile of the H2–TNG plot in Figure 6.19. The and continue to decrease or remain relatively flat, even in the
profile showing the effect of C3H8 content is also seen to be high-hydrogen content region which produced a sharp
similar to the corresponding calculated trends shown previ- increase in NOx levels on a volume concentration basis.
ously in Figure 6.14.
Figure 6.21 shows the final two-component fuel blend
6.2.5.3 Fuel Gas Tip Design
results being examined, which describe the variation in rela-
tive measured NOx emissions resulting from different con- Three-component interaction results were also examined by
centrations (volume basis) of H2 in a fuel blend composed considering results from several of the tested fuel gas tip
with a balance of C3H8. The upper plot, which shows mea- designs. Figures 6.22 through 6.24 show contoured ternary
sured relative NOx on a volume concentration basis, illus- plots of variation in relative measured NOx levels corre-
trates that for a given tip geometry and port size, the measured sponding to different fractions of H2, C3H8, and TNG in the
NOx concentrations actually decrease slightly with increasing fuel blend. Plots for three tip designs are shown, the tips dif-
H2 content up to 75% H2 content, then sharply increase with fering only in fuel port area size, which results in different
H2 concentration. fuel pressures for a given heat release on each tip. The results
Due to the decrease in total dry products of combustion are shown for tips in order of increasing port area size, or in
from the burning of H2, expressing NOx in terms of concen- other words, decreasing fuel pressure levels for the design
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
tration (ppmv) does not fully represent the actual mass rate heat release. Two plots are shown for each of three tips, with
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 210 Monday, March 21, 2005 3:29 PM

210 The John Zink Combustion Handbook

FIGURE 6.22 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel
blend composition for TNG/H2/C3H8 blends combusted with 15% excess air where both the fuel and the air were at ambient
temperature and pressure, for gas tip #2.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 6.23 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel
blend composition for TNG/H2/C3H8 blends combusted with 15% excess air where both the fuel and the air were at ambient
temperature and pressure, for gas tip #4.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 211 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 211

FIGURE 6.24 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel
blend composition for TNG/H2/C3H8 blends combusted with 15% excess air where both the fuel and the air were at ambient
temperature and pressure, for gas tip #6.

one illustrating NOx levels on a volume concentration basis composition effects on NOx emissions relatively indepen-
and the other illustrating NOx levels on a mass per unit heat dently from fuel pressure variations. A qualitative compar-
release basis. ison of the plot on the left with the theoretical plots
For each given tip, the highest NOx emissions on a con- previously shown in Figure 6.16 reveals that, on a volume
centration basis occur in the high-hydrogen content region, concentration basis, the change in NOx level as a function
while the highest NOx emissions on a mass per unit heat of fuel composition, for a relatively constant pressure and
release basis occur in the high-propane region. The contoured constant heat release, varies similarly to the trends predicted
gradients illustrate the interaction of the three fuel compo- by the adiabatic flame temperature variation and predicted
nents and how each of the components affects NOx emission relative NOx concentrations from the equilibrium combustion
in different regions of the fuel mixture, such as the steep NOx model over the same regions. This result is expected due to
concentration gradients in the high-hydrogen content regions. the well-established correlation of the dependence of thermal
The effect of C3H8 content can be seen to dominate the NOx NOx formation on flame temperature. The mass basis plot
level gradients on a mass per unit heat release basis with a on the right in Figure 6.25, shows that variation in NOx
relatively constant slope. It is also interesting to note that levels with fuel composition, from a constant fuel pressure
NOx levels overall appear to increase as fuel gas tips change perspective, are less severe than seen in the analysis of a
from having less open fuel port area (higher fuel pressures single fuel gas tip with fixed port sizes, for which fuel
for a given heat release) to having greater open fuel port area pressures may vary greatly to maintain a given heat release
(lower fuel pressure for a given heat release). with fuel composition variation.
Figure 6.25 shows ternary plots of fuel composition From both the two-component and three-component anal-
effects on NOx at a nominal constant fuel pressure of yses, it is evident that fuel pressure has a significant effect
21 psig (145 kPag). This analysis, made possible by testing on NOx emission levels. Figure 6.26 shows a plot of relative
a range of fuel gas tips, enables the examination of fuel
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
NOx levels vs. fuel pressure for each of the 15 different fuels
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 212 Monday, March 21, 2005 3:29 PM

212 The John Zink Combustion Handbook

FIGURE 6.25 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel
blend composition for TNG/H2/C3H8 blends combusted with 15% excess air where both the fuel and the air were at ambient
temperature and pressure, for a constant fuel gas pressure of 21 psig.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 6.26 Measured NOx (fraction of the maximum value in ppmvd) as a function of the fuel pressure for all 15
different TNG/H2/C3H8 blends (A through O) combusted with 15% excess air where both the fuel and the air were at ambient
temperature and pressure.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 213 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 213

FIGURE 6.27 Measured NOx (fraction of the maximum value in both ppmv and lb/MMBtu) as a function of the fuel

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
blend composition, fuel gas pressure, and calculated adiabatic flame temperature for TNG/H2/C3H8 blends combusted with 15%
excess air where both the fuel and the air were at ambient temperature and pressure.

tested. This plot shows a consistent decrease in NOx levels highest NOx concentration levels are found in the region of
correlated with an increase in fuel pressure. This phenomena high adiabatic flame temperatures and low fuel pressures,
is explained by the burner configuration which allows signif- when high concentrations of hydrogen are present. The mass
icant amounts of inert flue gas to be entrained into the flame per unit heat release NOx levels are also at a minimum in the
zone with increasing fuel jet momentum, thus decreasing same region as the concentration-based profiles, however the
thermal NOx formation. maximum NOx levels, when measured on a mass basis, are
not found in the same region, but occur in areas of lowest fuel
6.2.5.4 Summary pressures with a mildly elevated adiabatic flame temperature,
Figure 6.27 shows an overall view of the data collected from which correspond to high C3H8 concentration regions. These
all six tips with each of the 15 different fuel compositions overall trends concur with the previously discussed results
(90 data points in total) from both a NOx volume concentra- and agree with the correlations shown by the three-component
tion basis and mass per unit heat release viewpoint. The plots and two-component interaction analyses.
use fuel pressure and adiabatic flame temperatures as the pri- Adiabatic flame temperature and fuel pressure are both
mary axes to usefully illustrate some overall trends. The plot identified as significant fundamental parameters affecting
of relative NOx concentration levels shows the minimum NOx emission levels when considering the effect of fuel
NOx levels occur in the region with the lowest adiabatic composition on NOx levels. For a conventional burner, with
flame temperature and highest fuel pressures. Inversely, the NOx on a concentration basis, the adiabatic flame temperature
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 214 Monday, March 21, 2005 3:29 PM

214 The John Zink Combustion Handbook

Equivalence Ratio
0.5 0.6 0.7 0.8 0.9 1.0 1.1
10000 10000

9000 9000

8000 CH4 8000

7000 C3H8 7000

6000 6000
CO (ppmvw)

CO (ppmvw)
5000 5000

4000 4000

3000 3000

2000 Fuel Fuel 2000

1000 Lean Rich 1000

0 0
0.5 0.6 0.7 0.8 0.9 1.0 1.1

Equivalence Ratio

FIGURE 6.28 Adiabatic equilibrium CO as a function of equivalence ratio for air/fuel flames.

is dominant, with fuel pressure remaining significant in affect- 6.3.1 CO and Unburned Fuel
ing NOx emission levels. The highest NOx levels on a volume Carbon monoxide (CO) is generally produced in trace quan-
concentration basis occurred at the highest hydrogen content tities in many combustion processes as a product of incom-
fuel compositions at lower fuel pressures. On a mass per heat plete combustion (see Figure 6.14). CO is a flammable gas,

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
release basis, however, the highest relative NOx levels were which is nonirritating, colorless, odorless, tasteless, and nor-
achieved for fuel compositions containing large fractions of mally noncorrosive. CO is highly toxic and acts as a chemi-
C3H8. This appears to result from some combined character- cal asphyxiant by combining with hemoglobin in the blood
istics of a high-propane mixture including: very low fuel that normally transports oxygen inside the body. The affinity
pressure for a given heat release in comparison with the other of carbon monoxide for hemoglobin is approximately 300
fuels; somewhat higher adiabatic flame temperature than CH4; times more than the affinity of oxygen for hemoglobin.30 CO
and a substantially larger amount of total dry products of preferentially combines with hemoglobin to the exclusion of
combustion produced for a given heat release when compared oxygen so that the body becomes starved for oxygen, which
with H2. In summary, these results provide both quantitative can eventually lead to asphyxiation. Therefore CO is a regu-
and qualitative information to improve emission performance lated pollutant with specific emissions guidelines depending
prediction and design of burners with application to a wide on the application and the geographical location.
variation of fuel compositions. CO is generally produced by the incomplete combustion of
a carbon-containing fuel. Normally, a combustion system is
operated slightly fuel lean (excess O2) to ensure complete com-
6.3 COMBUSTIBLES
bustion and to minimize CO emissions. Figure 6.28 shows the
This section has been broken into two types of combustibles. calculated CO as a function of the equivalence ratio (ratio of
The first involves the incomplete combustion of the fuel, 1 is stoichiometric, >1 is fuel rich, and <1 is fuel lean). Because
which usually produces carbon monoxide, and in some lim- these are adiabatic calculations with very high flame tempera-
ited cases, not all of the hydrocarbon fuel is consumed and tures, the dissociation in the flame produces high quantities of
passes through the combustor unreacted. The second type of CO even under fuel lean conditions. This is graphically shown
combustible is volatile organic compounds (VOCs), which in Figure 6.29 where much more CO is produced at higher gas
are generally only important in a limited number of pro- temperatures while all other variables remain the same.
cesses, typically those involving contaminated or otherwise Figures 6.30 and 6.31 show the effects on CO production of
hazardous waste streams. air and fuel preheating, respectively. In both cases, the higher
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 215 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 215

Gas Temperature ( oC)

900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900
9000 9000

8000 8000

7000 CH4 7000

C3H8
6000 6000

5000 5000
CO (ppmvw)

CO (ppmvw)
4000 4000

3000 3000

2000 2000

1000 1000

0 0
1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600

Gas Temperature ( oF)

FIGURE 6.29 Adiabatic equilibrium CO as a function of gas temperature for stoichiometric air/fuel flames.

Air Preheat Temp. ( oC)

0 100 200 300 400 500 600 700 800 900 1000
45000 45000

40000 CH4 40000

C3H8
35000 35000

30000 30000
CO (ppmvw)

CO (ppmvw)

25000 25000

20000 20000

15000 15000

10000 10000

5000 5000
0 200 400 600 800 1000 1200 1400 1600 1800 2000

Air Preheat Temp. ( oF)

FIGURE 6.30 Adiabatic equilibrium CO as a function of air preheat temperature for stoichiometric air/fuel flames.

flame temperatures produced by preheating cause more CO the concentration of carbon available to make CO, which both
formation as the preheat temperature increases. Figure 6.32 reduce CO generation.
shows the effect of fuel composition for H2/CH4 blends. As
expected, higher concentrations of H2 produce less CO, and
at pure H2, no CO is generated. Similarly, Figure 6.33 shows
6.3.2 Volatile Organic Compounds
the effect of fuel composition for CH4/N2 blends. Higher Volatile organic compounds (VOCs) are generally low
concentrations of N2 reduce both the flame temperature and --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
molecular weight aliphatic and aromatic hydrocarbons such
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 216 Monday, March 21, 2005 3:29 PM

216 The John Zink Combustion Handbook

Fuel Preheat Temperature ( oC)

0 50 100 150 200 250 300 350 400 450 500
15000 15000

CH4
14000 14000
C3H8

13000 13000

12000 12000
CO (ppmvw)

CO (ppmvw)
11000 11000

10000 10000

9000 9000

8000 8000
0 100 200 300 400 500 600 700 800 900 1000

Fuel Preheat Temperature ( oF)

FIGURE 6.31 Adiabatic equilibrium CO as a function of fuel preheat temperature for a stoichiometric air/CH4 flame.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 6.32 Adiabatic equilibrium CO as a function of fuel composition (CH4/H2) for a stoichiometric air/fuel flame.

as alcohols, ketones, esters, and aldehydes.31 Typical VOCs combustion process, but they may be contained in the mate-
include benzene, acetone, acetaldehyde, chloroform, toluene, rial that is being heated, such as in the case of a contaminated
methanol, and formaldehyde. These compounds are consid- hazardous waste in a waste incinerator. In that case, the
ered regulated pollutants because they can cause photochemi- objective of the heating process is usually to volatilize the
cal smog and depletion of the ozone layer if they are released VOCs out of the waste and combust them before they can be
into the atmosphere. They are not normally produced in the emitted to the atmosphere.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 217 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 217

9000 9000

8000 8000

7000 7000

6000 6000

5000 5000
CO (ppmvw)

CO (ppmvw)
4000 4000

3000 3000

2000 2000

1000 1000

0 0
0 10 20 30 40 50 60 70 80 90 100
N2 in Fuel (vol. %)

100 90 80 70 60 50 40 30 20 10 0
CH 4 in Fuel (vol. %)

FIGURE 6.33 Adiabatic equilibrium CO as a function of fuel composition (CH4/N2) for a stoichiometric air/fuel flame.

There are two strategies for removing VOCs from the off- 1. dry fine particles being carried out of the process from
gases of a combustion process.33 One is to separate and the raw materials being processed
recover them using techniques like carbon adsorption or 2. particles generated in the combustion process
condensation. The other method involves oxidizing the 3. fuel carryover, where some of the solid fuel passes
VOCs to CO2 and H2O. This process includes techniques through the combustor essentially unreacted
like thermal oxidation (see Chapter 21), catalytic oxidation, 4. other particle carryover
and bio-oxidation. One common way to ensure complete The first and third mechanisms are not usually a major prob-
destruction of VOCs in waste incinerators is to add an after- lem in most hydrocarbon and petrochemical applications.
burner or secondary combustion chamber, which may or
may not have a catalyst, after the main or primary combus- 6.4.1.1 Particle Entrainment
tion chamber.32 The gas flow through the combustor may entrain particles
from the raw materials used in the process. This is often
6.4 PARTICULATES referred to as carryover. An example of this would be in the
There are two common sources of particulates that can be glass-making process where fine dust materials such as sand
carried out of a combustion process with the exhaust gases. are used to make the glass and can be carried out of the glass
One is entrainment and carryover of incoming raw materials, furnace if the gas velocity in the combustion space is high
and the other is the production of particles as a result of the enough. This type of particulate is expensive because not
combustion process. A particular health concern regarding only must the particles be captured by some type of flue-gas
particulate emissions is the hazardous materials that can scrubbing equipment, but some of the raw materials needed
condense on the particle surfaces and be carried into the for the process are also lost.
atmosphere.33 For example, heavy metals vaporized during a
high-temperature combustion process can condense on solid 6.4.1.2 Combustion-Generated Particles
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

particles and be carried out with the exhaust products. The second method by which particles can be emitted from
the combustion system is through the production of particles
6.4.1 Sources in the combustion process. For example, in the combustion of
Principal sources of particulates in most industrial combus- solid fuels, (e.g., coal), ash is normally produced. The air-
tion applications are: borne portion of the ash, usually referred to as fly-ash, can be
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 218 Monday, March 21, 2005 3:29 PM

218 The John Zink Combustion Handbook

carried out of the combustor by the exhaust gases. Heavier oil combustor and are emitted from the exhaust stack. Refractory
flames also tend to generate particulates due to the high car- particles also may be generated from refractory-lined com-
bon contents and increased difficulty in fully oxidizing those bustors. This refractory dust can also be emitted into the
particles prior to exiting the exhaust stack. atmosphere.
Another source of combustion-generated particles is soot
that may be produced in a flame. Under certain conditions,
even gaseous fuels can produce soot. To a certain extent, soot 6.4.2 Treatment Techniques
is desirable because it generally enhances the radiant heat There are a variety of techniques used to control particulate
transfer between the flame and the process. Fuels that have emissions from combustion processes. The specific method
a higher carbon-to-hydrogen mass ratio tend to produce more chosen will depend on many factors, including economics,
soot than fuels with a lower ratio. For example, propane particle size distribution and composition, volumetric flow
(C3H8), which has a C:H mass ratio of about 4.5, is more rate, exhaust stream temperature, and particle moisture con-
likely to produce soot than methane (CH4), which has a C:H tent. The preferred method in most cases is to minimize par-
mass ratio of about 3.0. For clean-burning fuels like natural ticulate formation in the first place by modifying the process.
gas, it is much more difficult to produce sooty flames com- For example, substituting a gaseous fuel or lighter fuel oil for
pared to other fuels (e.g., oil and coal) that have little or no
a heavy fuel oil can significantly reduce particulate emissions
hydrogen and a high concentration of carbon. Flames contain-
resulting from the fuel. Another strategy is to capture the par-
ing more soot are more luminous and tend to release their
ticles for recycling back into the process. One example is a
heat more efficiently than flames containing less soot, which
fluidized bed reactor, in which the majority of the particles
tend to be nonluminous. Soot particles generally consist of
are recirculated back into the process. Because of typically
high-molecular-weight polycyclic hydrocarbons and are
higher costs, the last choice is usually to remove the particles
sometimes referred to as “char.”
from the exhaust stream before they are emitted into the
Ideally, soot would be generated at the beginning of the
atmosphere. This can be done with electrostatic precipitators
flame so that it could radiate heat to the load, and then it
(wet or dry), filters (baghouses), or venturi scrubbers.33
would be destroyed before exiting the flame so that no par-
ticles would be emitted. Soot can be produced by operating
a combustion system in a very fuel-rich mode or by incom-
plete combustion of the fuel due to poor mixing. If the soot
6.5 CARBON DIOXIDE
particles are quenched or “frozen,” they are more difficult to Carbon dioxide (CO2) is a colorless, odorless, inert gas that
incinerate and more likely to be emitted with the exhaust does not support life because it can displace oxygen and act as
products. The quenching could be caused by contact with an asphyxiant. CO2 is found naturally in the atmosphere at con-
much colder gases or possibly by impingement on a cool centrations averaging 0.03%, or 300 ppmv. Concentrations of
surface (e.g., a boiler tube). Soot particles tend to be sticky 3 to 6% can cause headaches, dyspnea, and perspiration. Con-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

and can cling to the exhaust ductwork, clogging the ductwork centrations of 6 to 10% can cause headaches, tremors, visual
and the pollution treatment equipment in the system. If the disturbances, and unconsciousness. Concentrations above 10%
soot is emitted into the atmosphere, it can contribute to smog can cause unconsciousness, eventually leading to death.
in addition to being dirty. The emitted soot particles become
a pollutant because they produce a smoky exhaust that has Carbon dioxide emissions are produced when a fuel contain-
high opacity. Most industrial heating processes have a regu- ing carbon is combusted near or above stoichiometric condi-
lated limit for opacity. tions. Some studies indicate that CO2 is a greenhouse gas that
may contribute to global warming. Many schemes have been
6.4.1.3 Solid Fuel Carryover suggested for “disposing” of CO2, including injection deep into
This mechanism for generating particulates involves some of the ocean and deep-well injection for oil recovery. In some
the solid fuel passing through the combustor essentially unre- European countries, CO2 emissions are considered a pollutant
acted. This is not usually a concern in the hydrocarbon and and as such are regulated. Any technique that improves the
petrochemical industries where solid fuels are rarely used. overall thermal efficiency of a process can significantly reduce
CO2 emissions because less fuel needs to be burned for a given
6.4.1.4 Other Particle Carryover unit of available heat output. Some predict that reductions in
Particles are sometimes generated by scale formation in the CO2 emissions will become increasingly important for the pet-
piping. These scale particles (iron oxide) travel through the rochemical industry.34
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 219 Monday, March 21, 2005 3:29 PM

Pollutant Emissions 219

6.6 SOx In the vast majority of cases, dioxin/furan emissions result

Sulfur oxides, usually referred to as SOx, include SO, S2O, from some contaminant in the load materials being heated in
SnO, SO2, SO3, and SO4, of which SO2 and SO3 are of particu- the combustor. A quick scan of most of the textbooks on
lar importance in combustion processes.35 SO2 is preferably combustion shows that these emissions are essentially ignored
produced at higher temperatures, while SO3 is favored at lower because they are not generally produced in the flame, except
temperatures.36 Because most combustion processes are at in certain limited cases. This is primarily because usually
high temperatures, SO2 is the more predominant form of SOx there are not any halogens in either the fuel or the oxidizer
emitted from systems containing sulfur. Sulfur dioxide (SO2) to produce dioxins or furans. An exception is the case when
is a colorless gas with a pungent odor that is used in a variety waste materials are burned as a fuel by direct injection into
of chemical processes. SO2 can be very corrosive in the pres- a flame. One example is the destruction of waste solvents that
ence of water. It is considered a pollutant because of the chok- may be injected into an incinerator through the burner.
ing effect it can cause on the human respiratory system. It is
also damaging to green plants, which are more sensitive to SO2
than people and animals. When SO2 is released into the atmo- REFERENCES
sphere, it can produce acid rain by combining with water to
produce sulfuric acid (H2SO4). Sulfuric acid is very corrosive 1. J.A. Stanislaw, Petroleum industry faces tectonic shifts
and can cause considerable damage to the environment. changing global energy map, Oil Gas J., 97(50), 8-14,
It is often assumed that any sulfur in a combustor will be 1999.
converted to SO2, which will then be carried out with the 2. U.S. Dept. of Energy Office of Industrial Technology,
exhaust gases.37 The sulfur may come from the fuel or from Petroleum — Industry of the Future: Energy and
the raw materials used in the production process. Fuels like Environmental Profile of the U.S. Petroleum Refining
heavy oil and coal generally contain significant amounts of Industry, U.S. DOE, Washington, D.C., December 1998.
sulfur, while gaseous fuels like natural gas tend to contain
3. S.A. Shareef, C.L. Anderson, and L.E. Keller, Fired
little or no sulfur. The two strategies for minimizing or elim-
Heaters: Nitrogen Oxides Emissions and Controls, U.S.
inating SOx are: (1) removing the sulfur from the incoming
Environmental Protection Agency, Research Triangle
fuel or raw materials, and (2) removing the SOx from the
Park, NC, EPA Contract No. 68-02-4286, June 1988.
exhaust stream using a variety of dry and wet scrubbing
techniques.38 One dry scrubbing technique is limestone injec- 4. D. Hansell and G. England, Air Toxic Emission Factors for
tion. After use, the combined limestone and sulfur can be Combustion Sources Using Petroleum Based Fuels,
used in gypsum board. New membrane separation technology 3 volumes, Energy and Environmental Research Corp.,
is another reduction technique being developed. Irvine, CA, 1998 (available at www.api.org/step/ piep.htm).
5. U.S. EPA, AP-42: Compilation of Air Pollutant Emis-
sion Factors, 5th ed., U.S. Environmental Protection
6.7 DIOXINS AND FURANS Agency, January 1995.
This class of pollutants includes the carbon-hydrogen-oxygen-
halogen compounds and has received considerable attention 6. J.C. Reis, Environmental Control in Petroleum Engineer-
from both the general public and from regulatory agencies ing, Gulf Publishing, Houston, TX, 1996.
because of the potential health hazards associated with them. 7. American National Standards Institute/American Society
Dioxins generally refer to polychlorinated dibenzo-p-dioxin Mechanical Engineering, Performance Test Code PTC
(PCDD) compounds, while furans generally refer to poly- 19.10, Part 10: Flue and Exhaust Gas Analyses, Ameri-
chlorinated dibenzofuran (PCDF) compounds. Some of the can Society of Mechanical Engineers, New York, 1981.
potential health risks include toxicity because of the poison- 8. U.S. Environmental Protection Agency, Nitrogen
ing effect on cell tissues, carcinogenicity because cancerous Oxide Control for Stationary Combustion Sources,
growth may be stimulated, mutagenicity because of possible U.S. EPA Report EPA/625/5-86/020, 1986.
mutations in cell structure or function, and teratogenicity
9. M. Sandell, Putting NOx in a box, Pollution Engineering,
because of the potential changes to fetal tissue.32 The over
30(3), 56-58, 1998.
200 dioxin/furan compounds are regulated in certain indus-
tries, particularly in waste incineration, and also in certain 10. M. Moreton and S. Beal, Controlling NOx emissions,
geographical locations for a wide range of applications — Pollution Engineering Int., Winter, 14-16, 1998.
especially in Europe. 11. Y.B. Zeldovich, Acta Physecochem (USSR), 21, 557, 1946.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

220 The John Zink Combustion Handbook

12. E.B. Sanderford, Alternative Control Techniques 29. J. Bluestein, NOx Controls for Gas-Fired Industrial
Document — NOx Emissions from Process Heaters, Boilers and Combustion Equipment: A Survey of
U.S. EPA Report EPA-453/R-93-015, February 1993. Current Practices, Gas Research Institute (Chicago, IL)
13. U.S. Environmental Protection Agency, Alternative Report GRI-92/0374, 1992.
Control Techniques — NOx Emissions from Utility 30. K. Ahlberg, Ed., AGA Gas Handbook, AGA AB,
Boilers, U.S. EPA Report EPA-453/R-94-023, 1994. Lidingö, Sweden, 1985.
14. C.E. Baukal and P.B. Eleazer, Quantifying NOx for 31. S. Setia, VOC emissions — Hazards and techniques for
Industrial Combustion Processes, J. Air Waste Manage. their control, Chemical Engineering World, XXXI(9),
Assoc., 48, 52-58, 1997. 43-47, 1996.
15. S.B. Mandel, What is the total cost for emissions moni- 32. W.R. Niessen, Combustion and Incineration Processes,
toring?, Hydrocarbon Processing, 76(1), 99-102, 1997. 2nd ed., Marcel Dekker, New York, 1995.
16. U.S. Government, Code of Federal Regulations 40, 33. I. Ray, Particulate emissions: evaluating removal
Part 60, Revised July 1, 1994. methods, Chem. Eng., 104(6), 135-141, 1997.
17. M.C. Drake, Kinetics of Nitric Oxide Formation in Lam- 34. M. Thorning, How climate change policy could shrink
inar and Turbulent Methane Combustion, Gas Research the federal budget surplus and stifle US economic
Institute (Chicago, IL) Report No. GRI-85/0271, 1985. growth, Oil Gas J., 97(50), 22-26, 1999.
18. M.F. Zabielski, L.G. Dodge, M.B. Colket, and 35. E.D. Weil, Sulfur compounds, in Kirk-Othmer Encyclo-
D.J. Seery, The optical and probe measurement of NO: a pedia of Chemical Technology, 3rd ed., Vol. 22, John
comparative study, Eighteenth Symp. (Int.) on Combus- Wiley & Sons, New York, 1983.
tion, The Combustion Institute, Pittsburgh, 1981, 1591. 36. C.T. Bowman, Chemistry of gaseous pollutant forma-
19. A. Berger and G. Rotzoll, Kinetics of NO reduction by tion and destruction, in Fossil Fuel Combustion,
CO on quartz glass surfaces, Fuel, 74, 452, 1995. W. Bartok and A. F. Sarofim, Eds., John Wiley & Sons,
New York, 1991.
20. H.M. Gomaa, L.G. Hackemesser, and D.T. Cindric,
NOx/CO emissions and control in ethylene plants, 37. C.R. Bruner, Handbook of Incineration Systems,
Environmental Progress, 10(4), 267-272, 1991. McGraw-Hill, New York, 1991.
21. C.E. Baukal and A.I. Dalton, Nitrogen oxide measure- 38. S.R. Turns, An Introduction to Combustion, McGraw-Hill,
ments in oxygen enriched air-natural gas combustion New York, 1996.
systems, Proc. 2nd Fossil Fuel Combustion Symp., 39. API Recommended Practice 536: Post-Combustion
ASME PD-Vol. 30, pp.75-79, New Orleans, LA, Janu- NOx Control for Fired Equipment in General Refinery
ary 15, 1990. Services, 1 ed., American Petroleum Institute, Washing-
22. C.E. Baukal, Ed., Oxygen-Enhanced Combustion, CRC ton, D.C., March 1998.
Press, Boca Raton, FL, 1998. 40. API Recommended Practice 560: Fired Heaters for
23. R. Waibel, D. Nickeson, L. Radak, and W. Boyd, Fuel General Refinery Services, 2nd ed., American Petro-
Staging for NOx Control, in Industrial Combustion leum Institute, Washington, D.C., September 1995.
Technologies, M.A. Lukasiewicz, Ed., American Society 41. J. Colannino, Results of a statistical test program to assess
of Metals, Warren, PA, 1986, 345-350. flue-gas recirculation at the Southeast Resource Recovery
Facility (SERRF), Paper 92-22.01, presented at Air &
24. A. Garg, Trimming NOx, Chem. Eng., 99(11), 122-124,
Waste Management Association, 85th Annual Meeting
1992.
and Exhibition, Kansas City, MO, June 21-26, 1992.
25. H.L. Shelton, Find the right low-NOx solution, Environ-
42. M. Takagi, T. Kawai, M. Soma, T. Onishi, and K. Tamaru,
mental Engineering World, Nov.–Dec., 24, 1996.
The Mechanism of the Reaction Between NOx and
26. A. Garg, Specify better low-NOx burners for furnaces, NH3 and V2O5 in the Presence of Oxygen, J. Catal.,
Chem. Eng. Prog., 90(1), 46-49, 1994. 50(3), 441-446, 1977.
27. J. Colannino, Low-cost techniques reduce boiler NOx, 43. R.R Hayes, C.E. Baukal, and D. Wright, Fuel composi-
Chem. Eng., 100(2), 100-106, 1993. tion effects on NOx, presented at the 2000 American
28. J. Colannino, NOx reduction for stationary sources, Flame Research Committee International Symposium,
AIPE Facilities, 23(1), 63-66, 1996. --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Newport Beach, CA, September 2000.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch06-Frame Page 221 Monday, March 21, 2005 3:29 PM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 7
Noise
Wes Bussman and Jaiwant D. Jayakaran

TABLE OF CONTENTS

7.1 Fundamentals of Sound........................................................................................................................... 224

7.1.1 Introduction............................................................................................................................... 224
7.1.2 Basics of Sound ........................................................................................................................ 224
7.1.3 Measurements ........................................................................................................................... 228
7.2 Industrial Noise Pollution ....................................................................................................................... 231
7.2.1 OSHA Requirements ................................................................................................................ 232
7.2.2 International Requirements....................................................................................................... 232
7.2.3 Noise Sources and Environment Interaction............................................................................. 234
7.3 Mechanisms of Industrial Combustion Equipment Noise ...................................................................... 234
7.3.1 Combustion Roar and Combustion Instability Noise ............................................................... 234
7.3.2 Fan Noise .................................................................................................................................. 237
7.3.3 Gas Jet Noise ............................................................................................................................ 238
7.3.4 Valve and Piping Noise............................................................................................................. 239
7.4 Noise Abatement Techniques.................................................................................................................. 239
7.4.1 Flare Noise Abatement Techniques........................................................................................... 239
7.4.2 Burner Noise Abatement Techniques........................................................................................ 242
7.4.3 Valve and Piping Noise Abatement Techniques ....................................................................... 243
7.4.4 Fan Noise Abatement Techniques............................................................................................. 243
7.5 Analysis of Combustion Equipment Noise............................................................................................. 243
7.5.1 Multiple Burner Interaction ...................................................................................................... 243
7.5.2 High-Pressure Flare .................................................................................................................. 244
7.5.3 Atmospheric Attenuation .......................................................................................................... 246
7.6 Glossary .................................................................................................................................................. 246
References ................................................................................................................................................................ 248
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Bibliography ............................................................................................................................................................. 248

223
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 224 Monday, March 21, 2005 2:33 PM

224 The John Zink Combustion Handbook

7.1 FUNDAMENTALS OF SOUND reasonable duration. In extreme cases, the effects of vibration
may be more rapidly manifested, such as in the case of crack-
7.1.1 Introduction ing and falling of hard refractory linings in furnaces.
This chapter is written as a practical guide, as well as a
Silence is golden
reference on noise, for engineers involved in the design, oper-
— Anonymous ation, or maintenance of combustion equipment — be it burn-
ers, furnaces, flares (see Chapter 20), or thermal oxidizers
Noise is a common by-product of our mechanized civiliza- (see Chapter 21). In addition, because this chapter provides
tion and is an insidious danger in industrial environments. comprehensive coverage of the fundamentals of sound, the
Noise pollution is usually a local problem and thus is not creative engineer will also be able to extend his or her knowl-
viewed on the same scale of importance as the more notori- edge to analyze other noise-producing industrial equipment.
ous industrial emissions like NOx, CO, and particulates.
Nonetheless, it is an environmental pollutant that has signifi-
cant impact. 7.1.2 Basics of Sound
Serious concern is merited when a pollutant can result in
either environmental damage or human discomfort. Consid- If a tree falls in the forest and nobody is around to hear
ering the impact on people, noise is most often a source of it, does it still make a sound?
annoyance, but it can also have much more detrimental
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

effects, such as causing actual physical injury. Noise-related Webster’s dictionary defines sound as “that which is heard.”
injuries range from short-term discomfort to permanent hear- Obviously, an engineer will find this definition woefully inad-
ing loss. equate for his or her purposes. The authors resort to the defini-
The sense of hearing is a fragile and vital function of the tion provided in many engineering handbooks: “Sound is the
human body. It resembles vision, more than the other senses, vibration of particles in a gas liquid or solid.”1
because permanent and complete damage can be sustained Sound is propagated through any medium in waves that
quite commonly in an industrial environment. So it follows take the form of pressure peaks (compressions) and troughs
that noise pollution has been recognized as a safety concern (rarefactions), as illustrated in Figure 7.1. The pressure wave
for a long time and has been appropriately regulated. travels through a given medium at the speed of sound for that
Although personnel safety is the most important consider- medium. The auditory system in humans and most animals
ation, noise pollution has several other significant side effects. senses the impingement of these pressure waves on a tissue
Combustion equipment designers are often asked why they membrane and converts them to electrical impulses that are
would want to constrain the combustion process to reduce then sent to the brain and interpreted there.
noise. Typically, these questions come from persons working
Figure 7.2 is a cross-section of the human ear. Sound is
in plants situated in remote areas who often do not realize
collected and funneled into the ear canal by the outer ear. At
that given the age and economic drivers of the petroleum
the end of the ear canal, the sound impinges on the eardrum.
refining and chemical industries, it is now common to find
The bones of the middle ear convey the eardrum’s vibration
plants located in densely populated areas. With industry that
to the inner ear. The inner ear, or cochlea, consists of a fluid-
is situated close to residential areas or busy commercial facil-
filled membrane that has tiny hair cells on the inside. The
ities, high levels of noise become objectionable to people in
hair cells sense the vibration conveyed to the cochlea and
the neighborhood. These emissions eventually lead to gov-
convert this vibration into electrical signals, which are then
ernment regulations to control noise. Within the industrial site
conveyed to the brain.
itself, the immediate issue with noise is one of employee
safety. Furthermore, it is not surprising to find that employee Any given sound can be uniquely identified by two of its
morale and performance improves when noise is reduced, properties: pressure level and frequency. Most naturally occur-
since its presence increases stress level. ring sounds are composites of different pressure levels at var-
Equipment is also affected by noise. In most cases, these ious frequencies. A “pure tone” however, is a sound at only
effects lie in the area of vibration control and are beyond the one frequency. A tuning fork is a good example of a pure tone
scope of this chapter. Suffice it to say, that noise is a form of generator. Such naturally occurring pure tone generators are
vibration and eventually contributes to fatigue, which reduces rare. Even musical instruments create notes that have signifi-
equipment life. The effects of fatigue are frequently accepted cant pressure levels at two or three multiples, or harmonics,
as normal wear and tear if the equipment life cycle spans a of the fundamental frequency of the note.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 225 Monday, March 21, 2005 2:33 PM

Noise 225

FIGURE 7.1 Pressure peaks and troughs.

Nerve that sends

electrical signals
Eardrum to the brain

Cochlea

Ear Canal
Fluid filled
membrane

Hair cells

FIGURE 7.2 Cross-section of the human ear.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
7.1.2.1 Sound Pressure Level and Frequency hearing in humans in the later stages of life typically manifests
Pressure level defines the loudness of the sound, while fre- itself as diminished sensitivity to frequencies from 10 to 20
quency defines the pitch or tone of the sound. Pressure level kHz. Mechanically, this is due to the deterioration of the fine
is the amplitude of the compression, or rarefaction, of the hair cells in the cochlea.
pressure wave. The common unit of pressure level is the deci- It is important to note that the ear is not equally sensitive
bel, abbreviated dB. Frequency is the number of pressure over the entire range from 20 Hz to 20 kHz. This is vital to
waves that pass by an arbitrary point of reference in a given understanding how noise affects humans and how noise con-
unit of time. As such, the typical measure of sound frequency trol is implemented. The human ear is much less sensitive to
is cycles per second (cps), and as with electricity, the com- sound at the extremes of low and high frequencies, as is
monly used unit is the hertz (Hz); 1 Hz = 1 cps. discussed later in the chapter.
The typical range of human hearing extends from 20 Hz The wide range of frequencies in the human hearing range
to 20 kHz. Young children can hear frequencies slightly higher can be conveniently handled by breaking it up into octave
than 20 kHz, but this ability diminishes with age. Loss of bands. Each octave band represents a doubling in frequency.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 226 Monday, March 21, 2005 2:33 PM

226 The John Zink Combustion Handbook

TABLE 7.1 The Ten Octave Bands the relationship between decibel and watts. The “bel,” of
course, stands for Bell Labs.
Full Octave Band Standards In Figure 7.3, the y-axis represents power, in watts. The
Octave Band, Hz Center Frequency, Hz y-axis follows a base-10 scale. The x-axis gives dB values
22–44 31.5 and the line provides the relationship; 120 dB is equal to 1 W.
44–88 63 As an illustration of the log10 relationship, note that 110 dB
88–177 125
is equal to one tenth of a W (0.1 W), and 100 dB is equal to
177–355 250
355–710 500 a hundredth of a watt (0.01 W).
710–1420 1000
1420–2840 2000 7.1.2.3 Sound Power Level
2840–5680 4000
5680–11,360 8000
There is a subtle but important difference between the terms
11,360–22,720 16,000 sound power level and sound pressure level. Sound power
level is used to indicate the total energy-emitting ability of a
1x101 sound source. In other words, sound power is an attribute of
1x100 the source itself, while the sound pressure level (SPL) is used
1x10-1
to indicate the intensity of sound received at any point of
1x10-2
1x10-3 interest, from one or more sources. The illustration in Figure
Power, W

1x10-4 7.4 shows the formula to calculate the sound pressure level
1x10-5 that is expected at a distance r from a spherically radiating
1x10-6
source of power level Lw:
1x10-7
1x10-8
 1 
1x10-9 L p = L w + 10 log10   + 10.5 (7.1)
1x10-10  4π r 2 
1x10-11
1x10-12 where r is in feet.
0 10 20 30 40 50 60 70 80 90 100 110 120 13
In practice, the sound intensity at the location of the listener
dB
is of interest, and this is easily achieved by making measure-
FIGURE 7.3 Relationship of decibels to watts. ments at the point of interest. However, the sound pressure
level can be analytically derived at different locations in com-
plex industrial environments containing multiple sound
Table 7.1 shows the ten octave bands that cover the human sources if the power level of the sources is known. The equa-
hearing range and the center frequencies that can be used to tion can also be used to back-calculate the power level of a
represent the octave band. Each octave band extends over source from a measurement made at a known distance from
seven fundamental musical notes. the source.
The following are useful equations that can be used to
calculate sound pressure and power levels, in dB, from the
7.1.2.2 The Decibel
equivalent pressure and power units:
The unit of pressure level, the decibel, can be difficult to con-
ceptualize and merits some explanation. While it is possible to (
L p (dB) = 20 log10 P 2 × 10 −5 )
quantify the sound pressure level (SPL) in units of either
power or pressure, neither unit is convenient to use because, in (
L w (dB) = 10 log10 W 1 × 10 −12 )
practice, one has to deal every day with sounds that extend
where
over a very large range of power and/or pressure levels. For
Lp = Sound pressure level, in dB
example, the sound power of a whisper is 10–8 watts (W),
Lw = Sound power level, in dB
while the sound power of a jet plane is 105 W. The range of P = Pressure, in N/m2
these two sound sources thus spans 1013 W. The term decibel W = Sound power level, in W
that characterizes a dimensionless unit, was created to repre-
sent these large ranges conveniently. In the 1960s, Bell Labora- 7.1.2.4 Threshold of Hearing
tories coined the term “decibel.” The “deci” stands for the base Figure 7.5 reveals a map of the threshold of hearing in humans.
ten log scale on which the decibel is based. See Figure 7.3 for The y-axis represents SPL and the x-axis represents frequency.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Noise 227

130

Sound Pressure Level, dB

110
Sound pressure Threshold of pain
wave at a 90
distance r feet r 70
from the source.
Threshold of hearing
50

30
Noise source with a 10
The sound given sound power level
pressure level at (Lw) radiating outward. -10
this point can be
calculated as -30

16000

20000
1000

2000

4000

8000
31.5
1

125

250

500
16

63
L p = Lw + 10 log10 + 10.5
4π r 2
where r is in feet. Frequency, Hz

FIGURE 7.4 Sound pressure level at a distance r. FIGURE 7.6 Threshold of hearing and threshold of pain
in humans.
130
Sound Pressure Level, dB

110

70
Human threshold of hearing
50

-10

-30
63

500

2000
16

1000

4000

8000

20000
125

250

16000
31.5

Frequency, Hz

FIGURE 7.5 Threshold of hearing in humans. FIGURE 7.7 A-weighted scale for human hearing threshold.

Any SPL at a frequency that falls below the curve will be relatively flat. In general, a sound pressure level over 120 dB at
inaudible to humans. For example, a sound pressure level of 30 any frequency will cause pain. An important observation that
dB at 63 Hz will be inaudible; whereas an SPL of 70 dB at the can be derived from the two curves is that if a sound is audible
same 63 Hz will be audible. Humans are most sensitive to at very low or very high frequencies, persons subject to this
sounds in the so-called “mid-frequencies” from 1 kHz to about sound are very close to experiencing pain.
5 kHz. This is generally the region in which most of our every-
day hearing activities take place. Additionally, at a constant
7.1.2.6 Correction Scales
level, sound with a very low or very high frequency will not
have the same loudness sensation as that in the medium Sound meters are capable of measuring with equal sensitivity
frequency range. For example, a 100-Hz tone at L = 50 dB over the entire audible range. However, because humans do
gives the same loudness as a 1000-Hz tone at L = 40 dB.2 not hear with equal sensitivity at all frequencies, the sound
meter’s measurement needs to be modified to quantify what
really affects humans. This can be done using a correction
7.1.2.5 Threshold of Pain curve. The most common correction is the A-Scale. This is
Figure 7.6 shows the threshold of pain superimposed on the because, except for level, it resembles an idealized inverse
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

threshold of hearing. Fortunately, the threshold of pain is of the threshold of hearing curve (refer to Figure 7.7). An
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 228 Monday, March 21, 2005 2:33 PM

228 The John Zink Combustion Handbook

85 7.1.3 Measurements
Sound Pressure Level, dB

80 A simple schematic of a noise meter is shown in Figure 7.10.

75 The microphone is a transducer that transforms pressure
70
variations in air to a corresponding electrical signal. Because
65
the electrical signal generated by the microphone is rela-
60
A-weighted tively small in magnitude, a preamplifier is needed to boost
55
the signal before it can be analyzed, measured, or displayed.
50
Special weighting networks are used to shape the signal
45
spectrum and apply the various correction scales discussed
40

16000
above. The weighted signal then passes through a second
1000

2000

4000

8000
31.5

125

500
63

250

output amplifier into a meter. The meter and associated elec-

Frequency, Hz
tronic circuits detect the approximate rms value of the signal
FIGURE 7.8 A-weighted burner noise curve. and display it in dB.
Noise meters range from the simplest — microphone and
+20 needle gage — to sophisticated digital signal processing
(DSP) equipped analyzers. The more sophisticated analyzers
+10
are equipped with fast Fourier transform (FFT) capabilities
Relative Response, dB

0 that aid in accurate narrow band analysis. In general, spectrum

-10 analyzers allow the user to map the sound pressure level at
-20
different frequencies, in other words, generate a curve of the
A sound over different frequencies. However, there is a signif-
-30
B icant difference between instruments that make one measure-

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
-40 C ment per octave band and those that slice the octave band into
-50 D several intervals and make a measurement at each interval.
Typically, instruments are capable of:
-60

-70 1. octave band measurements

10 2 5 102 2 5 103 2 5 104 2 2. one-third octave band measurements
Frequency, Hz 3. narrow band measurements

FIGURE 7.9 Weighting curves A, B, C, and D. Table 7.2 shows the usual octave and one-third octave bands.
As the name suggests, a one-third octave band instrument
makes three measurements in each octave as opposed to the
A-weighted sound level correlates reasonably well with hear- single measurement of an octave band instrument. A narrow
ing-damage risk in industry and with subjective annoyance band instrument, on the other hand, uses digital signal process-
for a wide category of industrial and community noises. After ing (DSP) to implement fast Fourier transform analysis (FFT),
applying the A-scale correction, the unit of sound pressure and in the current state-of-the-art, FFT analysis allows the
level becomes dBA. Figure 7.8 shows a typical burner noise octave band to be sliced into as many as 128 intervals.
curve as measured by the noise meter (flat scale) and the Figure 7.11 provides a comparison of the same sound
result after applying the A-scale correction. spectrum as analyzed using three different frequency band
intervals: octave band, one-third octave band, and narrow
The other, less used correction scales are named, as might band. This comparison shows that the additional resolution
be expected, B, C, and D. Referring to Figure 7.9, one can provided by narrower band methods is of vital importance.
see that the C-scale is essentially flat over the range of interest In this example, the level at 1 kHz, as recorded by the octave-
and the B-scale lies somewhere between the A- and C-scales. band instrument, is 90 dB; on the one-third octave instru-
Given an understanding of the influence of low-frequency ment, it is 85 dB; and on the narrow band instrument, it is
sounds, one finds that the B- and C-scales do not apply ade- 70 dB. The lower resolution measurements produce higher
quate correction in the lower frequencies. Finally, the D-scale values due to the spill-over influence of the nearby peak at
is different from the others in that it has a pronounced cor- 1.8 kHz. In addition, in implementing noise control for this
rection in the range of 2 to 5 kHz. The D-scale was devised source, it is very valuable to know that it is the narrow peak at
for the aircraft industry and is rarely used otherwise. 1.8 kHz that is driving the maximum noise. This knowledge
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 229 Monday, March 21, 2005 2:33 PM

Noise 229

Weighting Networks
A
Microphone Rectifier Fast
B
Meter
C
Amplifier Amplifier Slow
D

Flat
Output

Band
filters

FIGURE 7.10 Block diagram of a sound level meter.

helps to zero-in on the source, which, for example may be an single number is to be used to represent the entire curve,
1800-rpm motor or pump. then it should adequately represent the peaks in the curve,
However, as with all things, there is a cost associated with because the peaks have the most influence on the listener.
high performance. For most purposes, a one-third octave anal- Consequently, it is not practical to use the average of the var-
ysis is usually quite adequate. The advantages of making ious levels in the octave bands because this number would
broad band analyses using octave or one-third octave band be less than the levels at the peaks. Therefore, one must not
filter sets are that less time is needed to obtain data and the confuse the average level with the overall sound level.
instrumentation required to measure the data is less expensive. The overall sound level is calculated by adding the individ-
In making sound measurements, several factors regarding ual levels in the various octave bands. In columns 1 and 2 of
the nature of the source should be considered. Whether the Table 7.3, the burner sound curve has been split up into
source is a true point source in space, radiating spherically, a its component levels in each octave band. In column 3, the
hemispherical source close to one flat surface, or a quarter A-weighted correction has similarly been split up and listed.
sphere between two flat surfaces, etc. will make a difference Column 4 gives the A-corrected values for the sound curve by
in the accuracy of the measurement. However, a detailed dis- simply subtracting column 3 from column 2. Now, the values
cussion of measurement issues is beyond the scope of this book, in column 4 must be added to obtain the overall sound level.
and the reader is encouraged to use some of the more compre- Because the decibel is based on a log10 scale, simple addi-
hensive works in the list of references at the end of this chapter. tion cannot be used. For example, if two values of equal
magnitude are added, say 100 dB and 100 dB, the result is
7.1.3.1 Overall Sound Level and How to 103 dB. The formula to be used is:
Add dB Values
As mentioned, most sounds are composites of several differ-
( )
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

L total = 10 log10 Σ i =1 to n 10 0.1L i (7.2)

ent levels at different frequencies. This is especially true of
industrial noise. A typical burner noise curve is shown in where
Figure 7.12. As can be seen, there are significant levels in Ltotal = Total level
two frequency zones, both of which will contribute to the
Li = each individual level i
apparent intensity experienced by a person working in the
N = Number of levels to be added
vicinity of the burner. It is difficult to describe this sound
without using either a diagram like the one shown or a table
Subtraction can be performed using
listing various SPLs occurring in the different octave bands.
The “overall sound level,” a single number, has been devised
to conveniently represent such composite sound curves. If a (
L diff = 10 log10 10 0.1L2 − 10 0.1L1 ) (7.3)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 230 Monday, March 21, 2005 2:33 PM

230 The John Zink Combustion Handbook

TABLE 7.2 Octave and One-Third Octave Bands

Octave One-Third Octave
Lower Band Upper Band Lower Band Upper Band
Band Limit Center Limit Limit Center Limit
12 11 16 22 14.1 16 17.8
13 17.8 20 22.4
14 22.4 25 28.2
15 22 31.5 44 28.2 31.5 35.5
16 35.5 40 44.7
17 44.7 50 56.2
18 44 63 88 56.2 63 70.8
19 70.8 80 89.1
20 89.1 100 112
21 88 125 177 112 125 141
22 141 160 178
23 178 200 224
24 177 250 355 224 250 282
25 282 315 355
26 355 400 447
27 355 500 710 447 500 562
28 562 630 708
29 708 800 891
30 710 1000 1420 891 1000 1122
31 1122 1250 1413
32 1413 1600 1778
33 1420 2000 2840 1778 2000 2239
34 2239 2500 2818
35 2818 3150 3548
36 2840 4000 5680 3548 4000 4467
37 4467 5000 5623
38 5623 6300 7079
39 5680 8000 11,360 7079 8000 8913
40 8913 10,000 11,220
41 11,220 12,500 14,130
42 11,360 16,000 22,720 14,130 16,000 17,780
43 17,780 20,000 22,390

The advantages of making broad band analyses of sound using octave or one-third octave band
filter sets are that less time is needed to obtain data and the instrumentation required to measure
the data is less expensive. The main disadvantage is the loss of detailed information about the
sound which is available from narrow band (FFT) analyzers.

However, some simple rules of thumb can be used to per- 1 W = 120 dB

form quick estimates. They are as follows: 1 W = 120 dB
1. When adding dB values of equal magnitude, the sum is 2 W = 123 dB
3 dB added to one of the numbers.
2. When the two values are different by 3 dB or less, the On the chart, 2 W registers 123 dB on the line. Similarly, the
sum is 2 dB added to the greater number. reason that numbers 10 dB or more in difference are
3. When adding two values that differ by 7 dB or less, the
neglected is because:
sum is 1 dB added to the greater number.
4. For values that differ by 8 dB or more, the sum is just the
larger number. 1.0 W = 120 dB
5. Always start with the smallest number in the list and add 0.1 W = 110 dB
it to the next larger number. 1.1 W = 120 dB
To understand why these rules work, refer to the chart in
Figure 7.3. From the chart it can be seen that 1 W is equal to Because the 110 dB contributes only 0.1 W power, it is
120 dB. neglected in the approximation. The example becomes more
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Noise 231

vivid when adding two numbers that differ by 20 dB or 100

more:
90
1.00 W = 120 dB 80
0.01 W = 100 dB 70
1.01 W = 120 dB 60
50
Rule 5 is especially necessary when adding a list that (a) Octave-band spectrum
contains several numbers that are almost equal in value and
one or more that are 10 dB greater, such as in a list that 100
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

contains six values of 90 dB and one value of 100 dB. If one 90

begins to add from the 100-dB value, one will obtain a wrong 80
result. Reader beware — the rules provided are approxima- 70
tions. For exact calculations, the formulae should be used. 60
Table 7.3 shows the effect of applying the addition rules 50
to the values generated by breaking up the burner noise curve. (b) Third-octave band spectrum
At the end of the addition list, 1 dB has been added to
100
compensate for any errors due to approximation.
90
7.1.3.2 Atmospheric Attenuation 80
When a sound wave travels through still air, it is absorbed or 70
attenuated by the atmosphere. Over a couple of hundred feet, 60
the atmosphere does not significantly attenuate the sound;
50
however, over a few thousand feet, the sound level can be 10 100 1k 10k
substantially reduced. The amount of sound attenuated in still (c) Narrow-band spectrum
air largely depends on the atmospheric temperature and rela-
tive humidity. Figure 7.13 depicts the atmospheric attenua- FIGURE 7.11 Same sound spectrum on three different
tion for aircraft-to-ground propagation in sound pressure intervals.
level per 1000 ft (300 m) distance for center frequencies of
500, 1000, 2000, 4000, and 8000 Hz.9 Notice that the atmos- 85
Sound Pressure Level, dB

pheric attenuation is more significant at higher frequencies 80

than at lower frequencies. For example, suppose that we are 75

1000 ft (300 m) away from a noise source and that the atmo- 70

spheric temperature and relative humidity are 80°F (27°C) 65

60
and 10%, respectively. The plots in Figure 7.13 show that the
55
atmospheric attenuation for 500 Hz is approximately 2 dB,
50
whereas for 8000 Hz, the attenuation is 55 dB.
45
Atmospheric attenuation, outdoors, can also be affected by
40
turbulence, fog, rain, and snow. Typically, the more turbulence
63

125

250

1000

8000
500

4000

16000
31.5

2000

present in the air, the more the sound is attenuated. There

appears to be conflicting evidence as to whether or not fog Frequency, Hz
attenuates sound. It is recommended that no excess attenua-
tion be assigned to fog or light precipitation. FIGURE 7.12 Typical burner noise curve.

7.2 INDUSTRIAL NOISE POLLUTION For example, it is not unusual for a person to encounter sound
Thus far, sound has been discussed. So what is noise? An all- pressure levels of 100 to 110 dB at a sporting event, in a stadium
encompassing definition would be that noise is any undesir- full of cheering fans, and yet not be perturbed by it. On the
able sound. By saying this, the concept is introduced that contrary, the barely 45-dB sound of a dripping faucet may cause
what is considered to be noise is somewhat relative and considerable annoyance in the quiet of the night. Table 7.4 gives
depends on several temporal and circumstantial factors. some typical noise levels for various scenarios.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 232 Monday, March 21, 2005 2:33 PM

232 The John Zink Combustion Handbook

TABLE 7.3 Addition Rules

What is the Overall dBA Level?

Frequency SPL A-scale CF SPL
Hz dB dB dBA

31.5 72 –39 33
63 75 –26 49 
→ 49
↓
125 79 –16 63  → 63
↓
250 79 –9 70  → 71
↓
500 72 –3 69  → 73
↓
1000 69 –0 69   → 75
↓
2000 68 –1 69   → 76
↓
4000 78 –1 79  → 80
↓
8000 83 –1 82  → 84
↓
16,000 80 –7 73   → 85

Notes: Overall sound level = 85 dBA.

Caution: Overall SPL does not = average SPL.

Industrial noise pollution is a major concern for society as from potentially hazardous noise. Table 7.5 shows OSHA
a whole. In a recent survey, the effects of exposure to noise in permissible noise exposure levels.
refinery workers was studied extensively. A cross-section of OSHA requires that the employer must provide protection
workers in different divisions/units was chosen. It was found against the effects of noise exposure when the sound levels
that noise levels averaged 87 to 88 dBA in aromatic and paraffin exceed those shown in Table 7.6. When the daily noise expo-
facilities and 89 dBA in alkylation facilities. In comparison, sure consists of two or more periods of noise exposure at
workers in warehouses, health clinics, laboratories, and offices different levels, their combined effect should be considered
were not found to be exposed to the same levels. rather than the individual effects of each. According to OSHA,
Noise can damage hearing and cause physical or mental the exposure factor (EF) is defined as:3
stress (increased pulse rate, blood pressure, nervousness, sleep
disorders, lack of concentration, and irritability). Irreparable
EF = C1 T1 + C 2 T2 + C 3 T3 + K + C n Tn (7.5)
damage can be caused by single transient sound events with
peak levels exceeding 140 dBA (e.g., shots or explosions).
Long-duration exposure to noise exceeding 85 dBA can lead where Cn is the total time of exposure at a specific noise level
to short-term reversible hearing impairment, and long-term and Tn is the total time of exposure permitted at that level
exposure to levels higher than 85 dBA can cause permanent (shown in Table 7.5). If the exposure factor exceeds 1.0, the
hearing loss. employee’s exposure is above OSHA limits. If OSHA identi-
The following is a mathematical model based on empirical fies such a situation, a citation may be issued and a grace
data (ISO 1999) used to calculate the maximum permissible period defined in which the employer must correct the viola-
continuous noise level at the workplace that will not lead to tion or face penalties as high as $10,000 per day.
permanent hearing loss:
7.2.2 International Requirements
L A,m < 85 + 10 log10 (24 Tn ), dBA (7.4)
Regulations aimed at protecting individuals from industrial
where Tn is the daily noise exposure time in hours. noise pollution have been enforced in almost all industrial-
Wearing ear protection devices at continuous noise levels ized countries. The noise caused in industry and the work-
greater than 85 dBA can prevent or reduce the danger of place is generally treated as a serious issue.
permanent hearing damage. Most countries have adopted 85 dBA as the limit for per-
missible noise. At any work place with sound levels exceeding
7.2.1 OSHA Requirements 85 dBA, ear protection devices must be worn, and workers
Title 29 CFR, section 1910.95 of the Occupational Safety and exposed to this level should have their hearing level checked
Health Act (OSHA) pertains to the protection of workers
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
periodically.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 233 Monday, March 21, 2005 2:33 PM

Noise 233

Temperature ºC
-10 0 10 20 30 40
40

dB/1,000 m
10

Atmospheric attenuation, dB/1,000 ft

355/710 Hz, 4th octave band, GMF 500 Hz 20

0 0
Relative 40

dB/1,000 m
10 Humidity, %
710/1,400 Hz, 5th octave band, GM F 1,000 Hz 10
20 20
30
0 50 0
70
20 90 60
1,400/2,800 Hz, 6th octave band, GMF 2,000 Hz

dB/1,000 m
40
10
20

0 0
0 10 20 30 40 50 60 70 80 90 100
Temperature, ºF

Temperature ºC
-10 0 10 20 30 40
Atmospheric attenuation, dB/1,000 ft

2,800/5,600 Hz
30 7th octave band 100
GMF 4,000 Hz
Relative
Humidity, % 80
10

dB/1,000 m
20 20
30
50
60
70
90
40
10

0 0
0 10 20 30 40 50 60 70 80 90 100

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Temperature, ºF

Temperature ºC

-10 0 10 20 30 40

50 5,600/11,200 Hz
Atmospheric attenuation, dB/1,000 ft

160
8th octave band
GMF 8,000 Hz
140
40 Relative
Humidity,
% 120
dB/1,000 m

10
20
30 100
30
50
60
80
70
90
20
60

40
10
20

0 0
0 10 20 30 40 50 60 70 80 90 100
Temperature, ºF
FIGURE 7.13 Same sound spectrum on three different intervals. (From Beranek.9).
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 234 Monday, March 21, 2005 2:33 PM

234 The John Zink Combustion Handbook

TABLE 7.4 Sound Levels of Various Sources 7.3 MECHANISMS OF INDUSTRIAL

Event dBA Level COMBUSTION EQUIPMENT NOISE
Threshold of hearing 0 dBA There are four major mechanisms of noise production in
Rustle of leaves 10 dBA
Normal conversation (at 1 m) 30 dBA
combustion equipment. They can be categorized as either
Minimum level in Chicago at night 40 dBA high-frequency or low-frequency sources. They are:
City street, very busy traffic 70 dBA
Noisiest spot at Niagara Falls 85 dBA 1. Low-frequency noise sources:
Threshold of pain 120 dBA a. Combustion roar and instability
Jet engine (at 50 m) 130 dBA
Rocket (at 50 m) 200 dBA b. Fan noise
2. High-frequency noise sources:
a. Gas jet noise
TABLE 7.5 OSHA Permissible Noise Exposures
b. Piping and valve noise
Duration per Day Sound Level, dBA
(hours) (slow response)
8.0 90 7.3.1 Combustion Roar and Combustion
6.0 92
4.0 95
Instability Noise
3.0 97 To understand combustion roar, the mixing process taking
2.0 100 place between the fuel and the oxidant on a very minute scale
1.5 102
1.0 105 must be considered. It is known that a well-blended mixture of
0.5 110 fuel and air will combust very rapidly if the mixture is within
≤ 0.25 115 the flammability limits for that fuel. On the other hand, a raw
Note: Exposure to impulsive or impact noise should not exceed 140 dBA. fuel stream that depends on turbulence and momentum to mix
in the ambient fluid, to create a flammable mixture, tends to
create a slower combustion process. In either case, when
regions in the mixing process achieve a flammable mixture and
7.2.3 Noise Sources and Environment encounter a source of ignition, combustion takes place. The
Interaction closer a mixture is to stoichiometry when it encounters the
ignition source, the more rapid will be the combustion. Com-
The predominant individual sources of noise in chemical and bustion occurring close to stoichiometry converts more of the
petrochemical plants are burners (process furnaces, steam energy release into noise. See the discussion on thermoacous-
boilers, and flares), fans, compressors, blowers, pumps, elec- tic efficiency in Section 7.3.1.1 for more details. The noise
tric motors, steam turbines, gears, valves, exhausts to open coming from each small region of rapidly combusting mixture
air, conveyors and silos, airborne splash noise from cooling adds up to create what is called combustion roar. Therefore,
towers, coal mills, and loading and unloading of raw and fin- combustion roar is largely a function of how rapidly the fuel is
ished materials. being burned. In addition, in the context of combustion equip-
ment like burners and flares, usually the larger the fuel release,
Although noise pollution caused by the industrial sector is the more the turbulence in the combustion process. Because
minor compared to that caused by road and rail traffic, indus- turbulence directly influences the mixing rate, high turbulence
trial noise receives more attention due to public representa- processes also produce more combustion roar. Thus, it is more
tion. ISO 1996 provides a set of international regulations for accurate to state that the level of combustion roar generated by
noise protection in residential neighborhoods located near a combustion process is a function of the amount of fuel being
industrial areas. burned and how rapidly one arranges to burn it.

National or local authorities must enforce noise limits that

7.3.1.1 Flare Combustion Roar
should not be exceeded in the neighborhood. The magnitude
It has been recognized for a long time that the noise emitted
of limiting values, additional fines for tonality and impulsive from a normal operating flare has two mechanisms at work:
noise, and the legalities change, not only from country to namely, combustion roar and gas jet noise. Combustion roar
country, but sometimes within different states and regions in typically resides in the lower frequency region of the audible
the same country. In general, nighttime noise limits are 10 to frequency spectrum, while gas jet noise occurs in the higher
15 dB lower than that for the daytime.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
frequencies, as illustrated in Figure 7.14.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 235 Monday, March 21, 2005 2:33 PM

Noise 235

As mentioned, the amount of combustion roar emitted from

a flare generally depends on how fast the waste gas stream
mixes with the ambient air. A waste gas stream that exits a
flare tip with low velocity and low levels of turbulence will
mix slowly with the ambient air and burn relatively quietly.
These types of flames are called buoyancy-dominated flames.
A waste gas stream that exits a flare tip with high velocity
and high levels of turbulence, however, will burn much faster
and create substantially more combustion noise for the same
heat release. These high-velocity flames are momentum-dom-
inated flames. Increasing the rate at which the waste gas burns
results in “bigger explosions” of the air/fuel mixture. These
“bigger explosions” create larger disturbances in the atmos-
phere, resulting in higher levels of combustion roar.
FIGURE 7.14 Typical noise signature emitted from a flare.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
High levels of turbulence in a flare flame are usually desir-
able because they help reduce radiation and increase the
TABLE 7.6 Overall Sound Pressure Level from Combustion
smokeless capacity of the flare. Unfortunately, such high
Frequency Resultant Noise Spectrum
turbulence levels also increase the combustion roar. Unlike (Hz) (dB)a
the solution for flare radiation reduction, it is not practical
31.5 OASPL - 5
to increase the height of a flare stack or boom to reduce 63 OASPL - 4
combustion noise. This is because combustion roar is low- 125 OASPL - 9
frequency sound and thus can travel a great distance without 250 OASPL - 15
500 OASPL - 20
being substantially attenuated by the atmosphere. The signa- 1000 OASPL - 21
ture of low-frequency combustion roar noise typically con- 2000 OASPL - 24
sists of a broadband spectrum with a single peak. 4000 OASPL - 28
8000 OASPL - 34
The combustion roar emitted from a stable burning flare
a OASPL = Overall sound pressure level.
typically peaks at a frequency of about 63 Hz. The combustion
noise spectrum can be estimated by adjusting the sound pres-
sure level emitted from combustion using the values in
Table 7.6.4 It is noted that, at frequencies above about 500 Hz,
the noise contribution from flare combustion is relatively
insignificant. A typical method for estimating the sound pres-
sure level (SPL) emitted from a flare flame is to relate the
energy released from the combustion of the waste gas stream
(chemical energy) to the noise energy liberated by the com-
bustion. The ratio of noise energy to chemical energy released
from the combustion is called the thermoacoustic efficiency
(TAE). For a stable-burning flare, the TAE typically varies
between 1 × 10–9 to 3 × 10–6. The value of the TAE largely
depends on the turbulent mixing of the waste gas with ambient
air and is usually determined experimentally.
In designing flares, the combustion noise emitted from
flares operating under various conditions is usually measured FIGURE 7.15 Engineer measuring flare noise level.
to determine the TAE. This information can then be used in
computer programs to model the level of combustion roar
emitted from a flare. Figure 7.15 is a photograph of a John order of 1 × 10–6. However, a flame with low levels of turbu-
Zink engineer collecting noise levels from a flare using a real- lence, such as the butane cigarette lighter shown in Figure 7.16,
time noise level meter. may have a TAE on the order of 1 × 10–9. For every order of
A flare flame that is highly turbulent, such as the high- magnitude that the TAE changes, the sound pressure level will
pressure flare shown in Figure 7.15, can have a TAE on the change by 10 dB.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 236 Monday, March 21, 2005 2:33 PM

236 The John Zink Combustion Handbook

Because the TAE is on the order of magnitude of

1 × 10–6, one would expect that the flame would
be highly turbulent and momentum-dominated.

The combustion roar emitted from a flare flame is not

highly directional and is considered a monopole source. That
is, it is analogous to a spherical balloon whose surface is
expanding and shrinking at various frequencies and emitting
uniform spherical waves.

7.3.1.2 Flare Combustion Instability Noise

If a flame lifts too far above a flare tip, it can become unsta-
ble. An unstable flame will periodically lift and then reattach
to the flare tip and create a low-frequency rumbling noise.
Typically, this low rumbling noise occurs in the frequency
range of 5 to 10 Hz and is usually called combustion instabil-
ity. Being as low in frequency as it is, combustion instability
noise is usually inaudible and can travel over several miles
without being substantially attenuated by the atmospheric air.
When there are reports of shaking the walls and windows of
buildings in the vicinity of a flare, it is usually due to com-
FIGURE 7.16 Shadow photograph of a burning butane bustion instability.
lighter. Combustion instability noise can occur if too much steam
or air is added to the base of the flame. Thus, over-aerating
the waste gas stream in a flare causes the flame to periodically
Example 7.1 lift from the flare tip. This periodic lifting and reattachment
Given: A flare burning a waste gas stream with a heat of the flame from the flare tip is the mechanism that drives
release of 5 × 10 9 BTU/h (1465 million W). Noise the low-frequency rumbling noise. Combustion instability
measurement shows that the sound pressure level noise can usually be reduced by lowering the steam flow rate
400 feet (120 m) from the flame is 100 dB. to a steam-assisted flare or by lowering the blower airflow
Find: The TAE of the flare flame. rate to an air-assisted flare. Figure 7.17 graphically depicts a
Solution: The sound power emitted from the flame, W, can typical steam-assisted flare operating under both normal con-
be determined as follows: ditions as well as over-steamed conditions.5 Note that the

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
combustion noise frequency shifts substantially to a lower
W (watts) = 1 × 10 −12 region and the level dramatically increases when the flare is
over-steamed.
 ( )
 L + 10 log10 4 π r 2 − 10.5  
anti log10  p  (7.6)
7.3.1.3 Burner Combustion Noise
  10  
 Like flares, burner combustion noise is an unwanted sound
associated with combustion roar and combustion instability.
where Lp is the sound pressure level in dB (100 dB
In many situations, the combustion noise can be the domi-
for this example), and r is the distance from the
nant source of noise emitted from a burner. Combustion roar
flame in feet (400 feet or 120 m for this example).
and combustion instability are quite complex by nature. The
Substituting these values into Eq. (7.6) gives
literature contains a variety of combustion noise and com-
W = 1792 watts. The TAE is then calculated to be:
bustion instability prediction techniques for burners operat-
ing in a furnace. Most of these prediction techniques are
Acoustical power
TAE = based on experimental studies that attempt to correlate the
Thermal power
acoustic power radiated by the burner/furnace geometry,
1792 watts laminar burning velocity of the air/fuel mixture, and various
= = 1.2 × 10 −6 (7.7)
1465 × 10 6 watts turbulence parameters such as the turbulent length scale and
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 237 Monday, March 21, 2005 2:33 PM

Noise 237

intensity. This section does not attempt to discuss these pre- 110

diction techniques in detail, but gives a broad and general Over-Steamed Flare
105
discussion of combustion roar and combustion instability
100
noise using some of the results from these studies.

Noise Level (dB)

Figure 7.12 is a plot showing a typical noise spectrum 95
emitted from a burner operating under normal conditions in
90
a furnace. Notice that the noise spectrum has two peak fre-
85
quencies associated with it; the high-frequency noise contri-
Normal Operating Steam Flare
bution is from the fuel gas jets, while the low-frequency 80
contribution is from the combustion roar. As with combustion
75
roar emitted from flares, burner combustion roar is associated
with a smooth broadband spectrum having relatively low con- 70
10 100 1000
version efficiency from chemical energy to noise — in the
range of 1 × 10–9 to 1 × 10–6. However, the combustion noise Frequency (Hz)
spectra associated with a burner and a flare are not similar.
FIGURE 7.17 A steam-assisted flare under normal and
The reason is that a flame burning in the open atmosphere over-steamed conditions.
will behave differently compared to a flame that is burning
in an enclosed chamber such as a furnace. Sound Pressure Level, dB 115
The combustion roar associated with flares typically peaks With instability
110
at a frequency of approximately 63 Hz, while the combustion 105
roar associated with burners can vary in the 200 to 500 Hz 100
range. Burner combustion roar can have a noise spectrum 95
shape and amplitude that can vary with many factors. These 90
factors include the internal shape of the furnace; the design 85

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
of the burner muffler, plenum, and tile; the acoustic properties 80
Normal operation
of the furnace lining; the transmission of the noise into the 75
fuel supply piping; and the transmissive and reflective char- 70
31.5

125

250

500

1000

2000

4000

8000

16000
acteristics of the furnace walls and stack.
Frequency, Hz
7.3.1.4 Burner Combustion Instability Noise
Combustion instability within a furnace is characterized by a FIGURE 7.18 Sound pressure level burner with instability.
high-amplitude, low-frequency noise resembling the puffing
sound of a steam locomotive. This type of noise can create
significant pressure fluctuations within a furnace that can instability. Some of these techniques include modifying the
cause damage to the structure and radiate high levels of noise (1) furnace stack height, (2) internal volume of the furnace,
to the surroundings. (3) acoustical properties of the furnace lining, (4) pressure
Figure 7.18 is a plot showing the SPL for a gas burner drop through the burner by varying the damper position,
operating under normal conditions and with instability. It is (5) fuel port diameter, (6) location of the pilot, and (7) flame
obvious that the sound pressure level increases substantially stabilization techniques.
when the operation is accompanied by instability. Combustion
instability noise has a high efficiency of conversion of chem-
ical energy to noise. Typically, the TAE from burner combus-
7.3.2 Fan Noise
tion instability is in the range of 1 × 10–4.6 The noise emitted from industrial fans typically consists of
The oscillations caused by combustion instability are nat- two noise components: broadband and discrete tones. Vortex
urally damped by pressure drop losses through the burner and shedding of the moving blades and the interaction of the
furnace, and therefore cannot be sustained unless energy is turbulence with the solid construction parts of the fan create
provided. These steady oscillations are sustained by energy the broadband noise. This broadband noise is of the dipole
extracted from the rapid expansion of the air/fuel mixture type, meaning that the noise is directional. On the other hand,
upon reaction. Over the years, furnace operators have used the discrete tones are created by the periodic interactions of the
several techniques in an attempt to eliminate combustion rotating blades and nearby upstream and downstream surfaces.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 238 Monday, March 21, 2005 2:33 PM

238 The John Zink Combustion Handbook

7.19. This orderly structure is known as the “global instabil-

ity” or “preferred mode” of the jet. The presence of both the
small-scale turbulent eddies within the jet and the large-scale
structure is responsible for the gas jet mixing noise.
The source of gas jet mixing noise begins near the nozzle
exit and extends several nozzle diameters downstream. Near
the nozzle exit, the scale of the turbulent eddies is small and
predominantly responsible for the high-frequency component
of the jet mixing noise. The lower frequencies are generated
further downstream of the nozzle exit where the large-scale
orderly pattern of the gas jet exists.
FIGURE 7.19 Development of orderly wave patterns.
Gas jet mixing noise consists of a broadband frequency
spectrum. The frequency at which the spectrum peaks
depends on several factors, including the diameter of the
nozzle, the Mach number of the gas jet, the angle of the
observer’s position relative to the exit plane of the jet, and
the temperature ratio of the fully expanded jet to the ambient
gas. In the flare and burner industry, gas jet mixing noise
typically peaks somewhere between 2000 and 25,000 Hz.
The overall sound pressure level created by gas jet mixing
depends on several variables, including the distance from the
gas jet, the angle of the observer relative to the gas jet cen-
FIGURE 7.20 Region of maximum jet mixing noise. terline velocity vector, the Mach number, the fully expanded
gas jet area, and the density ratio of the fully expanded jet to
the ambient gas. The maximum overall SPL of gas jet mixing
Discrete tonal noise is usually the loudest at the frequency at noise occurs at an angle between approximately 15° and 30°
which a blade passes a given point. The tonal frequency is relative to the centerline of the gas jet velocity vector, as
easily calculated by multiplying the number of blades times illustrated in Figure 7.20.7
the impeller rotation speed in revolutions per second. As one moves in either direction from this angle, the noise
The broadband and discrete tonal noise emitted from fans level (in some cases) can drop off significantly. For example,
can radiate from both the suction and pressure side of a fan the overall SPL created by gas jet mixing can be reduced as
and through the fan casing. The noise can radiate downstream much as 25 dB when one moves from an angle of maximum
through the ducting and discharge into the environment at an noise level to an angle directly behind the nozzle (180°).
outlet. Fan and duct systems should include provisions to
control this noise if residential areas are located nearby. 7.3.3.2 Shock-Associated Noise
Installation of mufflers and silencers on the suction and the When a flare or burner operates above a certain fuel pressure,
discharge sides of the fan, as well as wrapping of the casing a marked change occurs in the structure of the gas jet. Above
and the ducts, are common methods for reducing fan noise.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

a certain pressure called the critical pressure, the gas jet

develops a structure of shock waves downstream of the noz-
7.3.3 Gas Jet Noise zle, as shown in Figure 7.21. The critical pressure of a gas jet
Gas jet noise is very common in the combustion industry and typically occurs at a pressure of 12 to 15 psig (0.8 to 1 barg),
in many instances it can be the dominant noise source within a depending on the gas composition and temperature. These
combustion system. The noise created when a high-speed gas shock cells consist of compression and expansion waves that
jet exits into an ambient gas usually consists of two principal repeatedly compress and expand the gas as it moves down-
components: gas jet mixing noise and shock-associated noise.7 stream. Using Schlieren photography, several investigators
have seen as many as seven shock cells downstream of a noz-
7.3.3.1 Gas Jet Mixing Noise zle. These shock cells are responsible for creating two addi-
Studies have shown that a high-speed gas jet exiting a nozzle tional components of gas jet noise: screech tones and
will develop a large-scale orderly pattern, as shown in Figure broadband shock-associated noise.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 239 Monday, March 21, 2005 2:33 PM

Noise 239

Screech tones are distinct narrow-band frequency sounds Nozzle Exit

that can be described as “whistles” or “screeches.” The liter- Plane
ature reports that these tones are emitted from the fourth and
fifth shock cells downstream of the nozzle exit, as shown in
Figure 7.22.8
The sound waves from these shock cells propagate
upstream, where they interact with the shear layer at the
nozzle exit. This interaction then creates oscillating instability
waves within the gas jet. When these instability waves prop-
agate downstream, they interfere with the fourth and fifth
shock cells, causing them to emit the screech tones. Shock Waves
Broadband shock-associated noise occurs when the turbu-
FIGURE 7.21 Shock waves downstream of an air jet.
lent eddies within the gas jet pass through shock waves. The
shock waves appear to suddenly distort the turbulent eddies,
which creates a noise that can range over several octave are several techniques used in the flare and burner industry to
bands. The broadband, shock-associated peak frequency reduce the noise at the source, but these techniques have lim-
noise typically occurs at a higher frequency than the screech itations. Ear protection can reduce noise relative to the per-
tone peak frequency. sonnel using it; unfortunately, a plant operator cannot ask a
surrounding community or workers within a nearby office
7.3.4 Valve and Piping Noise building to wear ear protection when the noise levels become
When a gas flowing steadily in a pipe encounters a valve, a a problem. The most common method for reducing noise is in
change in the flow pattern and pressure will occur that can the path between the source and personnel, using silencers,
create turbulence and shock waves downstream of the valve. plenums, and mufflers. The purpose of this section is to dis-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Typically, when valves are partially closed, creating a reduc- cuss the most common and effective noise abatement tech-
tion in flow area, the small flow passage behaves much like an niques utilized in the flare and burner industry.
orifice and produces jet noise. As discussed above, turbulence
and shock waves create mixing noise and shock-associated
noise. This noise can radiate downstream through the pipe and 7.4.1 Flare Noise Abatement Techniques
exhaust into the environment at an outlet and/or radiate As previously discussed, the two principal sources of noise
through the pipe wall into the space near the valve itself, as emitted from industrial flares are combustion roar and gas jet
illustrated in Figure 7.23. noise. Inhibiting the rate at which the air and fuel streams mix
Usually, butterfly valves and ball valves are noisier than can reduce the level of combustion roar; however, this noise
globe valves. Butterfly valves and ball valves typically have abatement technique generally tends to reduce the smokeless
a smaller vena contracta than a globe valve operating at the performance and increase thermal radiation and flame length.
same pressure drop, which results in higher levels of mixing Reducing the mixing rate of the air and fuel stream in order to
and shock-associated noise. As a general guideline, when the lower combustion roar levels usually does not justify the
pressure ratio across a valve is less than approximately 3, the accompanying sacrifices in the performance of a flare.
mixing noise and shock-associated noise are within about the
In such cases, enclosed flares may provide one solution.
same order of magnitude. However, for pressure ratios greater
Enclosed flares are designed to completely hide a flare flame
than 3, shock noise usually dominates mixing noise.9 There
in order to reduce noise and thermal radiation levels. The
are several methods used for reducing the noise emitted from
design of these flare systems typically consists of an insulated
a valve. These include sound-absorptive wrapping of the pipes
enclosure with a wind fence around the perimeter, as shown
and the valve casings and the installation of silencers between
in the photograph in Figure 7.24. These types of flares can
the valve and connected pipes.
substantially reduce noise emissions as compared to open,
elevated flares.
7.4 NOISE ABATEMENT TECHNIQUES There are several abatement techniques commonly used to
There are three places noise can be reduced: at the source, in reduce the gas jet noise emitted from flares. Such techniques
the path between the source and personnel, and on the per- include mufflers, water injection, and modifications to the
sonnel.10 The ideal place to stop noise is at the source. There nozzle geometry. Mufflers are most commonly used on steam-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 240 Monday, March 21, 2005 2:33 PM

240 The John Zink Combustion Handbook

FIGURE 7.22 Location of screech tones emissions.

Sound Radiating from Flare

Tip to Surroundings

Sound Radiating from Valve

to Surroundings

Pressure
Relief or Steam
Valve

High Pressure Low Pressure

Side Side --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 7.23 Noise radiating from a valve.

assisted flares to abate the high-pressure steam jet noise, as shock-associated noise. A number of flare muffler styles have
shown in Figure 7.25. been used in the industry with varying degrees of noise abate-
In most flare systems, steam is supplied to nozzles at a ment performance. Many of these mufflers are designed with
pressure of 100 to 150 psig (7 to 10 barg). These high- a fiber material several inches thick placed on the inside. Muf-
pressure steam jets produce high-frequency mixing and flers usually do a good job of absorbing the high-frequency
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 241 Monday, March 21, 2005 2:33 PM

Noise 241
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 7.24 Two enclosed flares.

steam jet noise, as demonstrated by the data in Figure 7.26.

This plot shows the noise spectrum emitted from a steam-
assisted flare operating with and without a muffler on the lower
steam jets. The data clearly show that mufflers are more effi-
cient at absorbing higher noise frequencies than lower ones.
In high-pressure flaring applications, gas jet noise can be
the major source of noise. Recently, the John Zink Company
developed a unique method for reducing gas jet noise using
water injection.11 This method injects water into the waste gas
stream near the nozzle exit. The water injection method sub-
stantially reduces the shock-associated noise, as shown in
Figure 7.27. This plot depicts the noise spectrum emitted from
a John Zink high-pressure flare operating with and without
water injection. Schlieren photography shows that water injec-
tion does not eliminate the downstream shock cell structure,
but does appear to alter its appearance. This suggests that water
injection suppresses the feedback mechanism responsible for
growth of the gas jet instability that leads to screech tones.
Gas jet noise reduction using water injection is more
pronounced when flaring high-molecular-weight gases as
compared to low-molecular-weight gases at the same oper-
ating pressure. Test data and computer modeling show that
high-molecular-weight gases are more dominated by screech
tone noise than low-molecular-weight gases operating at the FIGURE 7.25 A steam-assisted flare with a muffler.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 242 Monday, March 21, 2005 2:33 PM

242 The John Zink Combustion Handbook

100
Sound-Pressure-Level (dB)
95

80
Without Muffler
75
With Muffler
70
100 1000 10000 100000
Frequency (Hertz)

FIGURE 7.26 Steam jet noise emitted with and without a muffler.

110

105

100
Noise Level (dB)

80
No Water Injection, 116 dB, 114 dBA
75
Optium Water Injection, 113 dB, 101 dBA
70
10 100 1000 10000 100000

Frequency (Hz)

FIGURE 7.27 Noise spectrum from a high pressure flare with and without water injection.

same pressure, which explains why gas jet noise reduction than a single larger nozzle operating at the same pressure and
using water injection is more pronounced when flaring high- mass flow rate, the primary reason being that the group of
molecular-weight gases. smaller nozzles will peak at a higher frequency, where the
It is very common in the flare industry to design a flare using human ear is less sensitive. Designing a flare with many small-
several small-diameter nozzles to reduce the A-weighted gas diameter nozzles is not always practical or economical to build.
jet noise level. Gas jet noise emitted from high-pressure flares Some large-capacity flare designs require several thousand noz-
usually peaks at a frequency between approximately 2000 and zles to substantially reduce the gas jet noise.
16,000 Hz. The peak frequency is a function of several vari-
ables, but is most affected by the diameter of the nozzle. For
example, a 4-in. gas jet nozzle will peak at a frequency between
7.4.2 Burner Noise Abatement Techniques
2000 and 4000 Hz, whereas a 1-in. gas jet nozzle will peak Burners used in industrial heaters and furnaces emit a broad-
between 8000 and 16,000 Hz. To the human ear, a group of band spectrum of noise. The broadband noise spectrum
several smaller-diameter gas jet nozzles will appear quieter --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
consists of (1) combustion roar, which resides in the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 243 Monday, March 21, 2005 2:33 PM

Noise 243

Sound Pressure Level, dBA

80 Unmuffled A-weighted
Burner Noise
75
70
65
60
55
50
Muffled A-weighted
45 Burner Noise
40
63

125

500

1000

2000

8000

16000
31.5

250

4000
Frequency, Hz

FIGURE 7.28 Sound pressure vs. frequency with and without a muffler.

frequency range of approximately 100 to 1000 Hz, and (2) The product of these two factors can cause larger-diameter
gas jet noise, which typically ranges between 4000 and jacketed pipes to radiate more noise than bare pipes.12
16,000 Hz. The mid-to-high-frequency noise is the most Piping requiring acoustical treatment in a typical petro-
annoying and damaging to the ear. Several techniques have chemical plant is often in cold service. These lagging systems
been used to suppress the noise emitted within the mid-to- have to be both thermal and acoustical insulators. For that
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

high frequencies. Four common techniques used to reduce reason, fibrous insulation followed by an outer leaded alumi-
noise in industrial burners are: num jacket is commonly used. Sometimes, very noisy pipes
need a layer of impregnated vinyl sandwiched between layers
1. sound insulation in the burner plenum
of fibrous insulation, called a septum system.12
2. mufflers at air inlets of natural-draft burners
3. acoustically optimized furnace wall construction 7.4.4 Fan Noise Abatement Techniques
4. acoustical treatment of the air ducts in forced-draft burners Fan noise can usually be addressed similar to valve and pip-
ing noise:
Figure 7.28 shows a plot of the sound pressure level vs.
frequency for a burner operating with and without a muffler. 1. Silencers can be installed at the suction and pressure sides
Clearly, without the muffler, the noise level is higher — of the fan, particularly for fans communicating with the
especially in the higher frequency region. atmosphere on either the suction or the pressure side, and
thereby cut down on noise coming out of these portals.
7.4.3 Valve and Piping Noise Abatement 2. Acoustically enclose the fan casing to address noise radi-
ated from or transmitted through the casing surface.
Techniques 3. Acoustically isolate the ductwork leading to and from a
Valve and piping noise abatements include sound-absorptive fan.
wrapping of the pipes and valve casings, and installation of At the design stage, one can consider the use of low-noise
silencers between the valves and the connecting pipes. Acous- motors (85 dBA or less) and the use of impellers with more
tical pipe lagging is similar to thermal pipe insulation. Acous- blades and reduced tip speed, etc.
tical pipe lagging also provides excellent thermal insulation,
but many thermal insulations provide poor noise control.
Rigid insulations for cold service (such as foam glass installed 7.5 ANALYSIS OF COMBUSTION
on smaller-diameter pipes) can actually aggravate the noise EQUIPMENT NOISE
situation by easily conducting the noise to the outer surface.
Although acoustical energy radiated per unit area of insulated 7.5.1 Multiple Burner Interaction
and jacketed pipe is less than for the same noninsulated pipe, A burner manufacturer will typically guarantee a burner
the surface area of an insulated and jacketed pipe is greater. noise level at a location 3 ft (1 m) directly in front of the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 244 Monday, March 21, 2005 2:33 PM

244 The John Zink Combustion Handbook

FIGURE 7.29 Burner noise example.

muffler. When several burners are installed in a furnace, how- be calculated at location 2 when burner A is
ever, the noise level 3 ft (1 m) from the burner may be higher operating alone by solving Eq. (7.9) for Lp:
than for a single burner, due to the noise contribution from
surrounding burners. The purpose of this section is to give an  1 
example that illustrates the noise level increase due to noise L p = Lw + 10 log10   + 10.5 (7.9)
 4π r 2 
emitted from surrounding burners.
For this case LwA = 95.03 and r = (52 + 32)0.5 =
Example 7.2
5.83 ft. Substituting these values into Eq. (7.10)
Given: Assume a furnace with a simple burner configu-
gives LpA = 79.2 dB. This is the sound pressure
ration, as illustrated in Figure 7.29, with burner
level contribution emitted from burner A mea-
B operating alone, and the noise level is 85 dB
sured at location 2. Because the distance from
at location 2.
burner C to location 2 is the same, the sound
Find: How is the noise level determined at location 2
pressure level contribution from burner C at loca-
when all burners are operating?
tion 2, LpC, is also 79.2 dB. The total sound pres-
Solution: First, find the sound power level, Lw , emitted from
sure level at location 2 can be determined by
each burner, assuming that the source is emitted
adding the sound pressure level contribution from
at the muffler exit at points 1A, 1B, and 1C.
each burner (79.2 dB + 79.2 dB + 85 dB). The
Assume that the noise spreads over a uniform
sound pressure levels can be added using the fol-
sphere from each of these points. The sound
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

lowing equation:
power level can be calculated as:

 1 
Lw = L p − 10 log10   − 10.5 (7.8)
L ptotal = 10 log10 10 ( Lp
A
10
+ 10
Lp
B
10
+ 10
Lp
C
10
)
 4π r 2 
= 86.8 dB (7.10)
where Lp is the sound pressure level and r is the
distance from the source (in feet). The noise level For this example, the noise level will be approx-
3 ft (1 m) from burner B (location 2) is 85 dB imately 1.6 dB higher when all the burners are
when it is operating alone. From Eq. (7.8), operating than if burner B is operating alone.
LwB = 95.0 dB. Assuming that all burners are
operating at the same conditions, the sound 7.5.2 High-Pressure Flare
power level must be 95.0 dB for each one. The Figure 7.30 is a plot showing the sound-pressure-level spec-
sound pressure level contribution, Lp, can now trum of a high pressure flaring event burning Tulsa natural gas
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 245 Monday, March 21, 2005 2:33 PM

Noise 245

120
Experiment Without Water Injection, 115 dB, 109 dBA
115
Model Without Water Injection, 114 dB, 110 dBA
110
Noise Level (dB)

105
100
95
90
85
80
75
70
10 100 1000 10000 100000

Frequency (Hz)

FIGURE 7.30 The sound-pressure-level spectrum of a high-pressure flare.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 7.31 The noise contributions separately based on the mathematical model.

in a 3.5-in (8.9-cm) tip. The symbols and the lines represent Figure 7.31 is a plot showing the noise contributions sep-
the noise spectrum gathered using a real-time sound-level arately based on the mathematical model. Notice that the gas
meter and mathematical modeling results, respectively. The jet mixing noise is a broadband frequency spectrum, while
sound pressure level spectrum consists of two major peaks: a the screech noise occurs over a fairly narrow bandwidth.
low-frequency peak that corresponds to the combustion roar The screech noise would not exist if the flare operated
and a high-frequency peak that corresponds to the gas jet below the critical gas pressure. Below the critical gas pres-
noise. The intermediate peak is a result of piping and valve sure, shock waves, which cause screech noise, do not form.
noise. Notice that the combustion roar peaks at a frequency of The summation of the combustion roar, gas jet mixing noise,
approximately 63 Hz, which is typical for a stable burning and screech noise provides the total sound pressure level
open flare. prediction emitted from the flare.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 246 Monday, March 21, 2005 2:33 PM

246 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 7.32 Effect of distance on flare noise.

TABLE 7.7 The Overall Sound Pressure Level (OASPL) noise level peaks at about 80 dBA, while at 3000 ft (910 m),
Determined Experimentally and Using the Mathematical the peak reduces to about 74 dBA. When the atmospheric
Model attenuation is taken into account, depending on the ambient
Jet Combustion temperature and humidity level at the time of measurement,
Mixing Noise Screech Noise Roar Total there is further reduction in noise levels. It is important to
Model note that the contribution in each case is significant. Given
the particular atmospheric conditions in this example, the
dB 105.7 105.1 113.0 114.3
dBA 105.2 105.3 97.4 108.6 attenuation has created a significant difference. The 10-dB
attenuation (from 74 dBA to 64 dBA) amounts to the sound
Experiment intensity reduction equal to one-tenth of its intensity at
dB — — — 113.7 3000 ft (910 m) without atmospheric attenuation. Hence, it
dBA — — — 109.2 should be noted that measurements may vary significantly on
different days for the same equipment if the atmospheric con-
ditions are significantly different.

The overall SPL (OASPL) determined experimentally and

calculated using the mathematical model is summarized in
Table 7.7. Notice that in this particular example, the OASPL,
7.7 GLOSSARY
on a dBA scale, is dominated by the gas jet noise. If this Absorption: Conversion of sound energy into another form of
3.5-in. (8.9-cm) diameter flare were designed with several energy, usually heat, when passing through an acoustical
smaller-diameter ports having the same total exit area, then medium.
the gas jet noise would shift to higher frequencies. If the Absorption coefficient: Ratio of sound absorbing effectiveness,
diameter of these ports were small enough to substantially at a specific frequency, of a unit area of acoustical absor-
shift the frequency of the gas jet noise, then the combustion bent to a unit area of perfectly absorptive material.
noise would dominate on the dBA scale. Acoustics: Science of the production, control, transmission,
reception, and effects of sound and of the phenomenon of
7.5.3 Atmospheric Attenuation hearing.
Figure 7.32 shows noise measurements emitted from a flare. Ambient noise: All-pervasive noise associated with a given
Notice that at a distance of 1500 ft (460 m) from the flare, the environment.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 247 Monday, March 21, 2005 2:33 PM

Noise 247

Anechoic room: Room whose boundaries effectively absorb all Phon: Loudness level of a sound, numerically equal to the
incident sound over the frequency range of interest, sound pressure level of a 1-kHz free progressive wave,
thereby creating essentially free field conditions. which is judged by reliable listeners to be as loud as the
Audibility threshold: Sound pressure level, for a specified fre- unknown sound.
quency, at which humans with normal hearing begin to Pink noise: Broadband noise whose energy content is inversely
respond. proportional to frequency (–3dB per octave or –10 dB per
Background noise: Ambient noise level above which signals decade).
must be presented or noise sources measured. Power spectrum level: Level of the power in a band 1 Hz wide
Decibel scale: Linear numbering scale used to define a loga- referred to a given reference power.
rithmic amplitude scale, thereby compressing a wide range Reverberation: Persistence of sound in an enclosure after a
of amplitude values to a small set of numbers. sound source has been stopped. Reverberation time is the
Diffraction: Scattering of radiation at an object smaller than time (in seconds) required for sound pressure at a specific
one wavelength and the subsequent interference of the frequency to decay 60 dB after a sound source is stopped.
scattered wavefronts. Root mean square (RMS): The square root of the arithmetic
Diffuse field: Sound field in which the sound pressure level is average of a set of squared instantaneous values.
the same everywhere, and the flow of energy is equally Sabine: Measure of sound absorption of a surface. One metric
probable in all directions. sabine is equivalent to 1 m2 of perfectly absorptive surface.
Diffuse sound: Sound that is completely random in phase; Sound: Energy transmitted by pressure waves in air or other
sound that appears to have no single source. materials which is the objective cause of the sensation of
Directivity factor: Ratio of the mean-square pressure (or inten- hearing. Commonly called noise if it is unwanted.
sity) on the axis of a transducer at a certain distance to Sound intensity: Rate of sound energy transmission per unit
the mean-square pressure (or intensity) that a spherical area in a specified direction.
source radiating the same power would produce at that
Sound level: Level of sound measured with a sound level meter
point.
and one of its weighting networks. When A-weighting is
Far field: Distribution of acoustic energy at a much greater used, the sound level is given in dB(A).
distance from a source than the linear dimensions of the
Sound level meter: An electronic instrument for measuring the
source itself. See also diffraction.
RMS of sound in accordance with an accepted national or
Free field: An environment in which there are no reflective
international standard.
surfaces within the frequency region of interest.
Sound power: Total sound energy radiated by a source per unit
Hearing loss: An increase in the threshold of audibility due to
time.
disease, injury, age, or exposure to intense noise.
Sound power level: Fundamental measure of sound power,
Hertz (Hz): Unit of frequency measurement, representing
defined as:
cycles per second.
Infrasound: Sound at frequencies below the audible range, that P
is, below about 16 Hz. Lw = 10 log10 , dB
P0
Isolation: Resistance to the transmission of sound by materials
and structures. where P is the RMS value of sound power in watts, and
Loudness: Subjective impression of the intensity of a sound. P0 is 1 pW.
Masking: Process by which the threshold of audibility of one Sound pressure: Dynamic variation in atmospheric pressure.
sound is raised by the presence of another (masking) The pressure at a point in space minus the static pressure
sound. at that point.
Near field: That part of a sound field, usually within about two Sound pressure level: Fundamental measure of sound pressure
wavelengths of a noise source, where there is no simple defined as:
relationship between sound level and distance.
Noise emission level: dB(A) level measured at a specified dis- P
L p = 20 log10 , dB
tance and direction from a noise source, in an open envi- P0
ronment, above a specified type of surface; generally
follows the recommendation of a national or industry stan- where P is the RMS value (unless otherwise stated) of
dard. sound pressure in pascals, and P0 is 1 µPa.
Noise reduction coefficient (NRC): Arithmetic average of the Sound transmission loss: Ratio of the sound energy emitted
sound absorption coefficients of a material at 250, 500, by an acoustical material or structure to the energy incident
1000, and 2000 Hz.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
on the opposite side.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 248 Monday, March 21, 2005 2:33 PM

248 The John Zink Combustion Handbook

Standing wave: A periodic wave having a fixed distribution in 11. W.R. Bussman and D. Knott, Unique concept for noise
space that is the result of interference of progressive waves and radiation reduction in high-pressure flaring, OTC
of the same frequency and kind. Characterized by the Conference, Houston, TX, 2000.
existence of maximum and minimum amplitudes that are
fixed in space. 12. L.D. Frank and D.R. Dembicki, Lower plant noise with
lagging, Hydrocarbon Processing, 71(8), 83-85, 1992.
Thermoacoustic efficiency: A value used to characterize the
amount of combustion noise emitted from a flame. Defined
as the ratio of the acoustical power emitted from the flame
to the total heat release of the flame. BIBLIOGRAPHY
Ultrasound: Sound at frequencies above the audible range, that
is, above about 20 kHz. Alberta Energy and Utilities Board, Calgary, Alberta, 1998.
Wavelength: Distance measured perpendicular to the wavefront American Petroleum Institute, 50, 125-146, 1972.
in the direction of propagation between two successive
points in the wave, which are separated by one period. R.S. Brief and R.G. Confer, Interpreting noise dosimeter
Equal to the ratio of the speed of sound in the medium to results based on different noise standards, Am. Indust.
the fundamental frequency. Hygiene J., 36(9), 677-682, 1975.
Weighting network: An electronic filter in a sound level meter S.C. Crow and F.H. Champagne, Orderly structure in jet tur-
that approximates, under defined conditions, the frequency bulence, J. Fluid Mech., 48(3), 547-591, 1971.
response of the human ear. The A-weighting network is
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

most commonly used. A.H. Diserens, Personal noise dosimetry in refinery and
White noise: Broadband noise having constant energy per unit chemical plants, J. Occupational Med., 16(4), 255-257,
of frequency. 1974.
A. Gharabegian and J.E. Peat, Saudi petrochemical plant noise
control, J. Environ. Eng., 112(6), 1026-1040, 1986.
REFERENCES HFP Acoustical Consultants, Effect of flow parameters on flare
stack generator noise, Proc. Spring Environmental Noise
1. A.P.G. Peterson, Handbook of Noise Measurement, 9th Conf.: Innovations in Noise Control for the Energy
ed., GenRad, Concord, MA, 1980. Industry, Alberta, Canada, April 19-22, 1998.
2. W. Daiminger, K.R. Fritz, E. Schorer, and B. Stüber, International Electrochemical Commission, IEC Standard,
Ullman’s Encyclopedia of Industrial Chemistry, Vol. B7, Publication 651, Sound Level Meters, 1979.
VCH, Weinheim, 1995, 384-401. ISO 1683 (E), Acoustics–Preferred Reference Quantities for
3. A. Thumann and R.K. Miller, Secrets of Noise Control, Acoustic Levels, 1983.
Fairmont Press, 1974. ISO 532 (E), Acoustics Method for Calculating Loudness
4. O.C. Leite, Predict flare noise and spectrum, Hydrocar- Level, 1975.
bon Processing, 68, 55, 1988. ISO 1996-1 (E), Acoustics: Description and Measurement of
5. W. Bussman and J. White, Steam-Assisted Flare Testing, Environmental Noise.
John Zink Co. Internal Report, September 1996. ISO 3744 (E), Acoustics: Determination of Sound Power
6. A.A. Putnam, Combustion Noise in the Handheld Levels of Noise Sources.
Industry, Battelle, Columbus Laboratories. Engineering Methods for Free Field Conditions over a
Reflecting Plane, 1081.
7. L.L. Beranek and I.L. Ve’r, Noise and Vibration Con-
trol Engineering, John Wiley & Sons, New York, 1992. ISO/DIS 8297, Acoustics: Determination of Sound Power
Levels of Multi-source Industrial Plants for the Evalua-
8. H. Shen and C.K.W. Tam, Numerical simulation of the
tion of Sound Pressure Levels in the Environment-
generation of axisymmetric mode jet screech tones,
Engineering Method, 1988.
AIAA Journal, 36(10), 1801, 1998.
ISO 9614-1 (E), Determination of Sound Power Levels of
9. L.L. Beranek, Noise and Vibration Control, McGraw-Hill,
Noise Sources Using Sound Intensity. Part I: Measure-
New York, 1971.
ment at Discrete Points; Part II: Measurement at
10. Allied Witan Co., Noise Facts and Control, 1976. Planned Points, 1993.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch07-Frame Page 249 Monday, March 21, 2005 2:33 PM

Noise 249

W.W. Lang, Ed., A commentary on noise dosimetry and stan- G. Seebold and A.S. Hersh, Control flare steam noise, Hydro-
dards, Proc. Noise Congress-75, Gaithersburg, MD, carbon Processing, 51, 140, 1971.
Sept. 15-17, 1975.
G.K. Selle, Steam-assisted flare eliminates environmental
J.C. Maling, Jr., Ed., Start-up silencers for a petrochemical concerns of smoke and noise, Hydrocarbon Process-
complex, Proc. Int. Conf. Noise Control Eng., Decem- ing, 73(12), 77-78, 1994.
ber 3-5, 1984.
B.N. Shivashankara, W.C. Strahle, and J.C. Henkley, Com-
A. Powell, On the noise emanating from a two dimensional
bustion noise radiation by open turbulent flames, Paper
jet above the critical pressure, The Aeronautical Quar-
73-1025, AIAA Aero-Acoustics Conference, Seattle,
terly, 4, 103, 1953.
WA, 1973.
A.A. Putnam, Combustion noise in the hand glass industry,
Tenth Annu. Symp. on the Reduction Cost in the Hand J.F. Straitz, Improved flare design, Hydrocarbon Processing,
Operated Glass Plants, 1979. 73(10), 61-66, 1994.
R. Reed, Furnace Operations, Gulf Publishing, Houston, TX, A. Thomas and G.T. Williams, Flame noise: sound emission
1981. from spark-ignited bubbles of combustion gas, Proc.
H.S. Ribner, Perspectives on Jet Noise, AIAA Journal, Roy. Soc., A294, 449, 1966.
19(12), 1513, 1981. E. Zwicker and H. Fastl, Psychoacoustics-Facts and Models,
J.P. Roberts, Ph.D. thesis, London University, 1971. Springer Verlag, Berlin, 1990.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 8

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Mathematical Modeling
of Combustion Systems
Lawrence D. Berg, Wes Bussman, Jianhui Hong,
Michael Henneke, I-Ping Chung, and Joseph D. Smith

TABLE OF CONTENTS

8.1 Overview ................................................................................................................................................. 252

8.2 Eduction Processes.................................................................................................................................. 252
8.2.1 Steam Flare Eduction Modeling ............................................................................................... 254
8.2.2 Eduction Processes in Pilots ..................................................................................................... 256
8.2.3 Eduction Processes in Premixed Burners ................................................................................. 257
8.3 Idealized Chemical Reactors and Combustion Modeling....................................................................... 258
8.3.1 Plug Flow Reactor..................................................................................................................... 258
8.3.2 Perfectly Stirred Reactor........................................................................................................... 259
8.4 Burner Pressure Drop.............................................................................................................................. 260
8.5 Flare Smokeless Operation ..................................................................................................................... 266
8.5.1 Predicting Flame Smoking Tendencies..................................................................................... 266
8.5.2 Application to Steam Flares...................................................................................................... 268
8.5.3 Modeling Air-Assisted Flares ................................................................................................... 270
8.5.4 Summary ................................................................................................................................... 273
8.6 Oil Gun Capacities .................................................................................................................................. 273
8.6.1 Summary of Two-Phase Flow Analytical Development ........................................................... 274
8.6.2 Results....................................................................................................................................... 277
8.7 Oil Gun Development ............................................................................................................................. 277
8.8 Regenerative Thermal Oxidizer (RTO) Performance ............................................................................. 280
8.8.1 Introduction............................................................................................................................... 280
8.8.2 RTO Model Development ......................................................................................................... 280
8.9 Conclusion .............................................................................................................................................. 282
References ................................................................................................................................................................ 283

251
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 252 Monday, March 14, 2005 2:54 PM

252 The John Zink Combustion Handbook

FIGURE 8.1 A typical eductor system.

FIGURE 8.2 Example eductor system.

8.1 OVERVIEW 8.2 EDUCTION PROCESSES

Over the course of time, the analytical modeling of combus- Some of the first theoretical and experimental studies on
tion processes has become increasingly more important. This eduction processes were performed in the early 1940s.1
trend is a result of two simultaneous events: (1) the prolifera- Since that time, a lot of work has been devoted to predicting
tion of powerful computers and (2) the development of a and understanding the mechanisms governing the perfor-
greater fundamental understanding of the underlying pro- mance of various eductor systems. This information is
cesses. The combination of these events has provided engi- important because these systems are widely used in industry
neers with opportunities to model and understand physical for many applications.
processes in much greater detail than previously possible. A typical eductor system — sometimes called a venturi,
This chapter highlights some of the applications developed inspirator, aspirator, ejector, or jet pump — is illustrated in
and utilized at the John Zink Company. Some applications Figure 8.1.
are straightforward extensions of previous techniques, while The objective of this device is to inspirate the surrounding
others are completely novel. The authors’ purpose is to (or secondary) gases into the mixer and then out through the
demonstrate how the practicing engineer can combine funda- tip. Mechanical energy is provided to the device by high-
mental knowledge and computational methods to analyze velocity gas (or motive gas) discharging from the orifice. An
combustion equipment and processes. eductor system consists of four fundamental parts:
This chapter discusses the following:
1. orifice
• modeling of eductors 2. mixer
• modeling of chemical reactions 3. downstream section
• modeling of burner pressure drop 4. tip
• modeling of flare smokeless rates
• modeling of oil gun performance and improvement
Each part plays a major role in the operating performance
• modeling of heat regenerator performance of the system, as follows. When the motive gas exits the
orifice, it will entrain the surrounding air by creating a low-
As appropriate, technical literature will be referenced in pressure zone in the area of the gas jet. The strength of this
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

order to provide the reader with additional details. low-pressure zone depends on the energy of the gas jet and
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 253 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 253

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 8.3 Experimental apparatus that has been successfully used to determine the flow coefficients and eduction
performance of various flare and burner eduction processes.

the resistance of the overall system. The amount of resistance and the conservation of energy can be written as:
or energy lost through the eductor system governs its ability
to entrain secondary gases. The four parameters that can Energy1 + Energy 2 − Energy 3 = Losses (8.3)
influence the performance of an eductor system include the
mass flow rate of the motive gas, pressure of the motive gas, In addition to the conservation equations, the following
pressure losses associated with the gas flowing through the constitutive equations and data are required:
system, and motive and secondary gas properties such as
specific gravity and ratio of specific heat. • equation of state (see Chapter 4.2.4) to relate densities to
For illustrative purposes, consider the control volume (see the mass flow rates
Chapter 4.3.3 for a discussion of control volumes) analysis • NASA nozzle performance data2
of an eductor system shown in Figure 8.2. The motive gas • saturated steam enthalpy and entropy
will expand fully somewhere downstream of the nozzle exit.
It is important that the losses (usually in the form of loss
The location of full expansion of the motive gas is designated
coefficients — see Chapters 4.3.2.2, 4.5.3, and 8.4 in this
as the inlet to the control volume. Location 1 represents the
book) be accurately determined because they can have a sig-
region where the motive gas and educted gas of location 2
nificant influence on the eduction performance of the system.
have the same pressure. Location 2 represents the annular
Typically, these losses must be measured experimentally. An
region of the secondary or educted gas within the eductor
experimental apparatus that has been successfully used to
tube. Somewhere downstream of the orifice is the outlet from
determine the loss coefficients and eduction performance of
the control volume, represented as location 3. From control
various flare and burner eduction processes is illustrated in
volume theory, one can write the conservation of mass for
Figure 8.3.
the process as:
As shown in Figure 8.3, the primary gas nozzle is located
inside a sealed chamber and the eduction system flows to the
Mass1 + Mass 2 = Mass3 (8.1) outside. A blower or compressed gas is used to deliver the
secondary gas into the chamber. The mass flow rate of the
Similarly, the conservation of momentum can be written as: secondary gas, entrained through the eduction system, is var-
ied until the pressure inside the chamber is zero while the
Momentum1 + Momentum 2 − Momentum 3 = Losses (8.2) eduction system is in operation. The mass flow rate of the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 254 Monday, March 14, 2005 2:54 PM

254 The John Zink Combustion Handbook

the equations discussed above, has been used successfully to

predict performance at the John Zink Company. Modeling
examples of several flare and burner eduction processes are
discussed and compared with experimental data.

Steam / Air / Waste Gas 8.2.1 Steam Flare Eduction Modeling

The ability to introduce air into the waste gas stream of a
flare, prior to combustion, is important for reducing the ther-
mal radiation and to improve the smokeless performance.
Some steam-assisted flares rely on the motive energy from
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

steam to inspirate air through several steam tubes. The steam-

air mixture is then delivered to the flare tip in the central core
of a waste gas stream. Figure 8.4 illustrates how this process
is accomplished, and Figure 8.5 shows the tip outlet side of a
large, steam-assisted flare tip equipped with eductor tubes.
To estimate and optimize the radiation and smokeless perfor-
mance of a steam-assisted flare, it is important to know the
eduction performance of the steam tubes.
The eductor model previously discussed was adapted to
Air Waste Air help predict the air entrainment rate into a steam flare. A full-
Gas scale tube was also set up in a test rig similar to Figure 8.3
Steam Steam to obtain verification data. Figure 8.6 shows a sample of the
data vs. model comparison. The vertical axis of the graph
FIGURE 8.4 Steam flare with steam eductor.
represents normalized air entrainment rates, and the horizon-
tal axis represents normalized steam pressure. Experimental
data vs. model predictions for sonic and supersonic nozzle
air entrainment rates are shown for a typical steam-assisted
flare eduction tube. As the steam pressure increases, the
steam-to-air mass ratio decreases. The decrease in air educ-
tion efficiency is due to the increase in pressure losses through
the steam tube. In general, it can be shown that these losses
increase with the gas velocity squared (see Eq. 8.17). The use
of a supersonic nozzle partially mitigates this effect. As
shown, at the maximum pressure, the increase in eduction
performance is about 15%.

8.2.1.1 Analysis of Supersonic Eductor Performance

The experimental and modeling results just presented are
somewhat surprising. Conventional wisdom would anticipate
FIGURE 8.5 Typical third-generation steam flare tube that a supersonic nozzle would inspirate more air than the
layout. incremental increase (~15%) observed. Steam at about
80 psig (a similar pressure to those reported) has an ideal
expansion velocity of about Mach 2 (twice the speed of
primary and secondary gas is measured using a flow mea- sound) and the velocity from a standard orifice exits at sonic
surement device such as a rotameter or orifice metering run. velocity (see Chapter 4.4.2.2). For the approximate condi-
The pressure inside the chamber can be measured using an tions at which many steam flares operate, it would seem that
inclined manometer or pressure transducer. a supersonic nozzle would double the momentum (the super-
Eduction processes are widely used in the flare and burner sonic nozzle has double the exit velocity of the sonic nozzle)
industry. Mathematical modeling of these processes, using and nearly double the entrainment rate would be anticipated.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 255 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 255

FIGURE 8.6 Normalized plot showing the sonic–supersonic eduction performance comparison in a single steam tube used
in a typical steam-assisted flare.

The following analysis will show that for steam at 100 psig, For a given inlet stagnation pressure Po and an ambient
an increase of only about 15% would be anticipated. It will pressure Pa , the maximum exit Mach number is achieved
also serve as a working example of how to apply control vol- when the exit pressure Pe is equal to the ambient pressure Pa .
umes and fundamental fluid mechanics (compressible flows) Assuming isentropic flow in the supersonic nozzle, the max-
to analyze real equipment. imum Mach number at the exit, Me , is related to Po and Pa by:40
Consider Figure 8.2 again. Instead of the control volume
γ −1 2 ( )
γ γ −1
inlet being set where the motive gas has the same pressure Po Po 
= = 1+ Me (8.5)
as the secondary gas, set the inlet at the outlet of the nozzle Pe Pa  2 
(see also McDermott and Henneke).32 Applying the integral
momentum balance yields the following: where γ is the specific heat capacity ratio of the motive gas.
Rearranging Eq. (8.5) gives the Mach number at the exit of a
well-designed supersonic nozzle:
G j + Ge = Gm + Losses (8.4)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

( γ −1) γ
2  Po  

where Me = −
γ − 1  Pa 
1 (8.6)

 
G – Momentum flux normal to control surface = ṁv + PA
The mass flow rate in a choked nozzle is:40
Notice that the momentum equation now has to specifically
( γ +1) ( γ −1)
include the pressure term, because the condition of equal Po γ  2 
m˙ = A*   (8.7)
pressure is no longer valid. Losses include viscous losses in To R  γ + 1
the tube, entry losses, and exit losses due to non-plug flow
velocity distributions. Since this section compares the driving The above equation is applicable to both sonic and super-
force (i.e., the momentum flux of the motive gas Gj ) from a sonic nozzles. The velocity v at the exit of a nozzle is:
converging–diverging (supersonic) nozzle to that from a con-
verging (sonic) nozzle, losses will be neglected. v = Me c = Me γRT (8.8)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 256 Monday, March 14, 2005 2:54 PM

256 The John Zink Combustion Handbook

TABLE 8.1 Exit Mach Number, Area Ratio, Driving Force ( γ +1) [ 2( γ −1)]
Ae 1  2   γ −1 2
Ratio, and Driving Force Percentage Increase for Various Gas =   1 + Me  (8.11)
Pressures (γ = 1.33 and Pa = 14.3 psiaa) A *
Me  γ + 1 2 

Po (psig) M Ae/A* Ratio Increase

15 1.096 1.0076 1.003 0.3% From Eqs. (8.9) and (8.10), the driving force ratio (Gj )ss /(Gj )s
20 1.221 1.0390 1.014 1.4% is:
30 1.410 1.1287 1.040 4.0%
40 1.550 1.2288 1.063 6.3%
50 1.663 1.3306 1.083 8.3% ( Pe Ae + mv
˙ e)
ss
=
( Pe Ae + mv
˙ e)
60 1.756 1.4315 1.100 9.2%
70 1.837 1.5308 1.114 11.4% s
80 1.907 1.6281 1.126 12.6%
 γ − 1 2  ( ) Ae
90 1.969 1.7234 1.137 13.7% − γ γ −1
100 2.026 1.8167 1.147 14.7% 1 + Me

 2 A*
a Local ambient pressure of Tulsa, Oklahoma.
(8.12)
−1 2 ( γ +1) ( γ −1) 
γ −1 2  2 
+ γ 1 + Me   Me 
For a nozzle with throat diameter A*, the driving force Gj is:  2   γ + 1 

( γ +1) ( γ −1) 
 γ − 1 ( )
− γ γ −1 −1 2
− γ ( γ −1) γ − 1  2 
˙ e = Po 1 +
Pe Ae + mv
γ −1 2
M Ae 1 + + γ 1 +  

 2   2   2   γ + 1 

( γ +1) ( γ −1)
Po γ  2 
+A *
  Me γRT The following procedure can be used to estimate the
To R  γ + 1 increase of driving force from the use of a supersonic nozzle
compared to a sonic nozzle:
 γ − 1 2  ( ) Ae
− γ γ −1
= Po A* 1 + Me
 
2 A* 1. Calculate exit Mach number Me from Eq. (8.6), assuming
isentropic flow in the supersonic nozzle.
( γ +1) ( γ −1)  2. Calculate the ratio of the exit area Ae to the throat area
γ T  2 
+   Me  from Eq. (8.11) (note: the nozzle needs to be designed to
To  γ + 1  have the right exit area and proper converging–diverging
 profile).
3. Calculate the driving force ratio from Eq. (8.12).
 γ − 1 2  ( ) Ae
− γ γ −1
= Po A* 1 + Me
 2  A* The driving force ratios calculated from the above procedure
are tabulated in Table 8.1 for a selected value of γ and ambi-
−1 2 ( γ +1) ( γ −1)  ent pressure Pa.
γ −1 2  2 
+ γ 1 + Me   Me  (8.9) It can be seen from Table 8.1 that a Mach number of
 2   γ + 1 
 over 2 corresponds, at best, to a 14.7 percent increase of driv-
ing force. The increase in the ṁv term is largely offset by the
For a choked sonic nozzle, the exit Mach number Me = 1, and
decrease in the PA term. As presented previously, experiments
the exit area is the throat area; therefore, the above equation
conducted at the John Zink Company (Tulsa, OK) showed
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

becomes
that the increase of air entrainment ratios roughly (within
− γ ( γ −1)
experimental errors) agree with the percentage increase of
γ − 1 
˙ e ) = Po A* 1 +
( Pe Ae + mv driving force listed in Table 8.1.
s  2  

( γ +1) ( γ −1) 
8.2.2 Eduction Processes in Pilots
−1 2
γ − 1  2 
+ γ 1 +  
 (8.10) The stability of the flame on a pilot is very critical. If the pre-
 2   γ + 1  mixed air-to-fuel ratio is not within an appropriate range,

then the flame could be easily extinguished.
For an isentropic flow, the exit area of a supersonic nozzle Figure 8.7 shows the normalized eduction performance of
needs to satisfy: a flare pilot operating on Tulsa Natural Gas (TNG) with two
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 257 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 257

FIGURE 8.7 Normalized eduction performance of a flare pilot operating on Tulsa natural gas with two different motive
gas orifice diameters.

FIGURE 8.8 Experimental and theoretical results of the eduction performance of a particular radiant wall burner firing
with two different orifice sizes and fuel gas compositions.

different motive gas orifice diameters. These results compare thorough understanding of the fundamental fluid mechanics
experimental data obtained from the test apparatus and theo- in the design of eductor-driven pilots.
retical modeling results. Here, the volumetric air-to-fuel ratio
remains fairly constant throughout the pressure range tested.
8.2.3 Eduction Processes in
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

A pilot that operates with a constant air-to-fuel ratio is desir-

able because it offers better stability over a range of operating Premixed Burners
pressures. As the motive gas orifice diameter increases, the Figure 8.8 is a plot showing experimental and theoretical
volumetric air-to-fuel ratio decreases for a given motive gas results of the eduction performance of a particular radiant
pressure. This demonstrates that slight modifications in the wall burner firing with two different orifice sizes and fuel gas
motive gas orifice diameter can significantly affect the educ- compositions. For this particular burner, the volumetric air-
tion performance. In addition, it highlights the need for a to-fuel ratio remains fairly constant over the range of heat
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 258 Monday, March 14, 2005 2:54 PM

258 The John Zink Combustion Handbook

releases tested. The slope of the operating curve for the volu- through the reactor from inlet to outlet. Chemical reactions
metric air-to-fuel ratio, however, can change significantly, are allowed to take place, but it is assumed that fluid advec-
depending on the fuel composition, mixer and tip design, and tion is the only transport process occurring. This also means
ambient conditions. that the mixing processes and recirculations are ignored. A
Eductor models have been developed and successfully plug flow model makes a number of assumptions. It is
applied to different types of combustion equipment at the important that the user understand the limitations imposed
John Zink Company. The experimental and theoretical meth- by these assumptions.
ods described inevitably led to both greater insight into the
operation of the equipment and better performance. The abil- 8.3.1.1 Assumptions
ity to measure and predict the eduction performance is crucial
• The kinetic energy of the fluid flow is neglected. This
in the flare and burner industry. Without this ability, it is
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

assumption is normally quite reasonable for burners and

difficult — if not impossible — to consistently estimate
flares because in most gas flows, changes in the thermo-
burner and flare performance. dynamic enthalpy due to chemical reactions and heat
transfer processes are much larger than changes in the
kinetic energy of the fluid flow.
8.3 IDEALIZED CHEMICAL • Pressure is assumed constant. Again, this is normally
REACTORS AND COMBUSTION quite reasonable. In flows with moderate Mach numbers
(e.g., 0.3 or above), this assumption becomes erroneous
MODELING
• No concentration gradients are present normal to the
Over the last 20 years, utilizing laminar flames and plug flow
flow.
reactors, researchers have made substantial progress in
detailing the chemical reaction pathways for combustion pro-
8.3.1.2 Governing Equations
cesses. Some of this progress has been summarized in Bartok
The mathematical description of the plug flow reactor
(1991).3 The comprehensive understanding of combustion
requires solving the conservation of energy and the conserva-
chemistry has proven to be a daunting task. A simple meth-
tion of species equations. Because the flow is one-dimen-
ane–oxygen–nitrogen system requires on the order of
sional and steady-state, conservation of mass is trivial and
50 species and 300 chemical reactions. In general, a system
simply requires that the mass flux is constant. Energy conser-
of 50 species requires simultaneous solution of 50 ordinary
vation requires:
differential equations. During the 1980s, a general-purpose
FORTRAN package for calculation and standardized storage
of chemical kinetic data, called CHEMKIN, was developed (Energy change)
at Sandia National Laboratory (Kee et al.4–6).
= (Heat release due to chemical reactions)
For practical combustion equipment, reaction chemistry
is closely coupled to the turbulent mixing occurring in the + (Heat transferred into system) (8.13)
flame. Unfortunately, to effectively manipulate the large
number of chemical reaction equations and constraints, it is The rate of production or destruction of a species can be
very difficult to include accurate fluid mechanics. To cir- computed by summing its production or destruction rate in
cumvent this difficulty, CHEMKIN assumes extremely sim- all the chemical reactions in which it participates. This over-
plified fluid flows (perfectly stirred or plug flow), coupled all rate is a function of all 300+ rate equations. The conserva-
to the detailed combustion reaction kinetics. This chapter tion statement for chemical species is:
summarizes some of these classic “reactor” assumptions,
followed by subsequent utilization of these simplifications
to assist in the performance analysis of real equipment. For (Change of species mass fraction)
additional details of these concepts, the interested reader is = (Overall reaction rate) (8.14)
directed to Turns.7
If there are N chemical species, then N–1 ordinary differen-
8.3.1 Plug Flow Reactor tial equations (with independent variable x = distance along
A true plug flow reactor is characterized by one-dimensional, reactor axis) must be solved given an inlet condition at x =
diffusion-free flow. A fluid enters the reactor at a known 0. These equations are frequently stiff. A system of ordinary
temperature, pressure, and composition and flows axially differential equations is called stiff when the eigenvalues of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 259 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 259

the Jacobian matrix span several orders of magnitude. Burner Primary Burner Secondary
Press8 gives a good introduction to the solution of these Secondary Fuel
Primary Fuel and Air
types of equations.
Ambient Flue Gas
Mix Mix
8.3.2 Perfectly Stirred Reactor
The perfectly stirred reactor (PSR), sometimes called a con-
tinuously stirred tank reactor, is a vessel of spatially uni- PSR PFR

form composition and temperature, and was first described Mix

by Longwell.9 Experimentalists frequently approximate a

PSR using high-velocity jets to introduce the unreacted
mixture into a reaction vessel. In reality, since the entering PFR NOx
mixture has a different composition than the mixture in the
Prediction
tank, the notion of a spatially uniform composition is some-
what idealistic. However, this setup closely approximates a FIGURE 8.9 Simplified reactor modeling of a staged
PSR, which approximates the behavior of many real sys- fuel burner.
tems where fluid mixing rates are much larger than chemi-
cal reaction rates.
The PSR is characterized by either a volume or a residence efficient, allowing the solution error to be reduced very
time. These are related by: quickly, it is not very robust and requires an initial solution
estimate that lies within its domain of convergence. For readers
Residence time = (Volume)(Density) (Mass flow rate) who might consider using the CHEMKIN PSR model, the
following numerical scheme is recommended.
The assumptions made in formulating the PSR model are Instead of Newton iteration, use only time-stepping to
numerous and impractical to list here, but, on a practical level, satisfy the governing equations. To optimize the solution
are the same as used for the plug flow reactor (PFR). The algorithm, an adaptive time-step size to control discretization
most important assumption is that the mixing rate is much errors as outlined in Press8 is recommended. The resulting
faster than the reaction rate. algorithm, while not as efficient in some cases as the
CHEMKIN PSR model, is more robust and rarely requires
8.3.2.1 Governing Equations any tuning by the user.
The PSR model, like the PFR model, requires that conserva-
tion of energy and species be satisfied. Conservation of mass 8.3.3 Systems of Reactors
is satisfied by noting that the PSR has one inlet and one
The above idealized reactors, either PSRs or PFRs, are
outlet, and these have the same mass flow rates.
rarely ever encountered in practical combustion systems.
The energy equation for the perfectly stirred reactor is:
However, starting with Beer10 and more recently Lutz et
al.,11,12 it has been shown that appropriately arranged sys-
(Energy out ) − (Energy in) tems of reactors can be utilized to simulate combustion pro-
cesses. The basic concept is that a combustion system may
= ( Energy exchanged with surroundings) (8.15)
be modeled as a series of parallel sequences of PSRs, PFRs
and mixing regions. A mixing region provides for interac-
Similarly, the species conservation requirement is:
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

tion with the environment by allowing additional material

to be added to the composition and “freezing” the reactions
(Species out) − (Species in) = (Species produced) (8.16) during the process. Figure 8.9 schematically represents how
this concept can be implemented for a simple staged fuel
The equations governing the PSR are nonlinear algebraic burner. Similar flow sequences have also been made for a
equations. When PSR calculations are made using detailed radiant wall burner (shown in Figure 8.10) and NOx reduc-
kinetics, the stiffness problem must again be considered. The tion thermal oxidizers (or NOxIDIZERS©‚ shown in Figure
PSR model supplied with the CHEMKIN4,33 package uses a 8.11). Sample results are shown in Figures 8.12 and 8.13. In
hybrid Newton/time-stepping algorithm to solve the set of Figure 8.12, trends from a PSR are compared to NOx from
nonlinear algebraic equations. While the Newton iteration is a premixed burner; and in Figure 8.13, predicted vs. actual
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 260 Monday, March 14, 2005 2:54 PM

260 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 8.10 Picture of a radiant wall burner.

FIGURE 8.11 Picture of a thermal oxidizer.

NOx for NOx reduction thermal oxidizers is shown. 8.4 BURNER PRESSURE DROP
Although the method is simplistic, trends are correctly pre- Figure 8.14 shows a typical capacity curve plot that many
dicted for the burner and excellent predictions for NOx burner manufacturers use for sizing burners. Capacity curves
reductions are possible. Because of the possibility of describe the airside pressure drop through various burner
improved, fundamentally based performance predictions, sizes at different heat releases. The curves shown in this par-
improvement of this technique is an active research topic ticular example are based on burners operating in the natural-
for many academic and industrial workers. draft mode with 15 percent excess air in the furnace at an
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 261 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 261

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 8.12 Sample results of simplified modeling.

262 The John Zink Combustion Handbook

Predicted vs Measured Exit NO Levels

400
Predicted (ppmv dry) pt 2
300 pt 2/3 pt 1/3

200

100 pt 1

0
0 100 200 300 400
Measured (ppmv dry)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 8.13 Sample results of simplified modeling for a thermal oxidizer.

Excess Air = 15%, Air Temp = 60 F, Altitude = Sea Level

100
Net Heat Release (MM BTU/Hr)

19
10
18

1
0.1 1 10
Air Pressure Drop (Inches H2O)
FIGURE 8.14 Capacity curves that many burner manufactures use for sizing burners.

atmospheric temperature and pressure of 60°F (16°C) and the velocity pressure of the air and can be written as (see
14.7 psia (1 Bar), respectively. Chapter 4.5.3):

When burners operate at different ambient conditions and

∆P ∝ ρV 2 (8.17)
excess air levels, the airside pressure drop obtained from the
capacity curves must be corrected. The equation used to cor- where ∆P is the pressure drop through the burner, ρ is the
rect for the airside pressure drop can be derived as follows. density of the combustion air, and V is the mean velocity of
The airside pressure drop through a burner is proportional to the air at a particular location in the burner. The density of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 263 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 263

the combustion air can be related to the combustion air tem- be the pressure drop if the burner is operating at 15 percent
perature, T, and atmospheric pressure, P, using the ideal gas excess air with the combustion air temperature at 60°F and
law (Chapter 4.2.4): an atmospheric pressure of 14.7 psia. To correct for the actual
firing conditions we use Eq. (8.21) to obtain:
P
ρ∝ (8.18)
T
100 + 12  460 + 100   14.7 
2
∆P1 = 0.5 ×  × × = 0.54
The velocity, V, of the air through the burner is proportional  100 + 15   460 + 60   14.0 
to the mass flow of air going through the burner and the den-
sity. This can be written as:
Although the percent excess air is reduced from 15 to 12 per-
cent, the pressure drop through the burner has increased due
m˙ (100 + EA) to the reduced density of the combustion air.
V∝ ∝ (8.19)
ρ ρ
Capacity curves are convenient for helping engineers size
where ṁ represents the mass flow and EA represents the per- burners to be used for a specific burner design operating under
cent excess air in the furnace. Substituting Eqs. (8.18) and certain conditions. However, capacity curves may not provide
(8.19) into Eq. (8.17) gives: accurate estimates of airside pressure drop if a burner design
or operation is modified. For example, several variables that
can affect the airside pressure drop include: (1) fuel and
T
∆P ∝ (100 + EA) ×
2
(8.20) atomizing gas tip drill angles, (2) fuel and atomizing gas tip
P
position, (3) burner tolerances, (4) fuel and atomizing gas
Equation (8.20) can be used to write the following equation to pressure and temperature, (5) flue gas recirculation, and (6)
correct for the airside pressure drop at actual firing conditions: fuel splits in the primary and secondary combustion zones.

2
 100 + EAActual 
∆PActual = ∆PCC ×  
 100 + EACC 

T   P 
× Actual  ×  CC  (8.21)
 CC   PActual 
T

where the subscript Actual represents the actual firing condi-

tions and CC represents the variables from the capacity
curves (i.e., TCC = 460 + 60, PCC = 14.7 psia, EACC = 15).
Notice that as the temperature of the combustion air
increases, the airside pressure drop through the burner also
increases. This occurs because increasing the temperature
reduces the density of the combustion air. A reduction in the
density of the combustion air requires a higher volumetric air
flow rate through the burner, which results in a pressure drop
increase. Similarly, if the atmospheric pressure is reduced,
the combustion air density is reduced and, hence, the airside
pressure drop is increased.
As an example, consider the pressure drop through a size
15 burner with a 3 MMBtu/h heat release with 12 percent
excess air, operating at a combustion air temperature of 100°F
and an atmospheric pressure of 14.0 psia.
From the capacity curves (see Figure 8.14), for a heat FIGURE 8.15 Burner designs typically consist of a
release of 3 MMBtu/h at standard conditions, the pressure muffler, damper, plenum, throat, and tile section.
drop will be approximately 0.5 in. water column. This will
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

264 The John Zink Combustion Handbook

fuel nozzles positioned inside the tile section that can help
aspirate the air through the burner but may tend to either
increase or reduce the pressure drop depending on the nozzle
position and orientation. The advantage of using semi-empir-
ical modeling is that it can take into account the effect that
all of these variables have on the airside pressure drop.
John Zink engineers generally use a cold flow test rig
specifically designed to study the airside pressure drop
through various burner designs. The cold flow test rig consists
of a test chamber on support legs located above the ground,
as shown in the photograph in Figure 8.16. The test chamber
is designed so that various sizes and types of burners can be
mounted to the bottom. With this apparatus, burners can be
tested in either the forced- or natural-draft mode. In the nat-
ural-draft mode, air flows through the burner and out of the
test box. At the top, the test box transitions to a duct with
flow straighteners, an orifice metering run, and dampers for
controlling the air flow rate. The flow straighteners provide
a uniform flow distribution of the air before it enters the
orifice metering run. The data obtained from the cold flow
test apparatus provide the necessary details to develop semi-
empirical models used for estimating the airside pressure drop
through burners.
A discussion of the methodology employed to develop the
necessary information follows. For this discussion, Figure 8.15
will be utilized as the typical process burner. Details will vary
from burner to burner, but the method will stay the same.
Starting with the muffler, as combustion air accelerates from
FIGURE 8.16 Cold flow furnace consists of a test a zero velocity to a given velocity at the muffler inlet, energy
chamber (8 × 8 × 8 ft) supported on legs approximately 7 ft is lost. Energy is also lost due to the vena contracta created
above the ground. as the air enters the muffler. As the air flows through the vena
contracta, it decelerates and loses additional energy. The
acceleration and deceleration of the air as it moves through
A method useful for capturing the possible variations of the muffler inlet has a pressure loss associated with it that can
burner design is burner semi-empirical modeling, or just be related to the mean velocity pressure (see Chapter 4.5.3) as
semi-empirical modeling. Burner designs typically consist of
a muffler, damper, plenum, throat, and tile section, as illus- ρ air Vmuffler
2
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

∆Pmuffler = kmuffler
inlet
trated in Figure 8.15. The strategy for determining the airside (8.22)
inlet inlet 2 gc
pressure drop, using the semi-empirical modeling technique,
is to relate the airside pressure drop to the velocity pressure
where ∆Pmuffler inlet = Pressure drop associated with the
of the air going through each section of the burner. This
combustion air entering the muffler
strategy can usually be applied to all sections of a burner ρair = Combustion air density
except the tile section. In the tile section, other factors can Vmuffler inlet = Mean velocity of the combustion air at
affect the airside pressure drop, depending on the design of the muffler inlet
the burner. For example, some burner designs use the pressure kmuffler inlet = Loss coefficient into the muffler
energy of the fuel to aspirate flue gas into the tile section gc = gravitational constant
where it mixes with the air and burns. If combustion occurs
within the tile section, it can significantly increase the overall The loss coefficient will vary depending on the muffler inlet
pressure drop through a burner. Other burner designs have design. For example, a square-edged inlet will have a loss
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 265 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 265

coefficient equal to approximately 0.5. However, the loss TABLE 8.2 Values for a 90° Mitered Elbow
coefficient can be as small as 0.02 if the inlet is well-rounded. C′0
For more information on loss coefficients through various fit-
W1/W0
tings, see Chapter 4.4.5.3, the ASHRAE Handbook,13 or
H0/W0 0.6 0.8 1.2 1.4 1.6 2
Idelchik.14 Independent of the reference, experimental valida-
tion of the loss coefficient will be accomplished in the cold 0.25 1.8 1.4 1.10 1.10 1.10 1.10
flow test rig. 1 1.7 1.4 1.10 0.95 0.90 0.84
As the combustion air flows past the muffler inlet, it takes 4 1.5 1.1 0.81 0.76 0.72 0.66
infinite 1.5 1.1 0.69 0.63 0.06 0.55
a sudden 90° turn through the muffler elbow. The pressure
loss associated with the flow through an elbow can also be
loss coefficient = C′0 Rec
related to the mean velocity at the muffler inlet similar to the
equation described in Eq. (8.17). The loss coefficient through where
the elbow is also available in the literature, with values shown C′0 = values from table above
as flow through mitered elbows provided in Table 8.2.13 The Rec = Reynolds number correction factor (listed below)
values can also be determined/validated experimentally with Re = Reynolds number at inlet to elbow = V × D / ν
V = mean velocity into elbow
the cold flow test rig, or determined using computational fluid
D = hydraulic diameter = 2 × H0 × W0/(H0 + W0)
dynamics (CFD, see Chapter 9).
ν = kinematic viscosity
Notice that the loss coefficient through a mitered elbow is
a function of the Reynolds number. This suggests that the loss
coefficient through the muffler can change with burner turn- Re × 10–4 1 2 3 4 6 8 10 >14
down conditions or draft levels. Capacity curves, as discussed Rec 1.40 1.26 1.19 1.14 1.09 1.06 1.04 1.0
earlier, assume a constant loss coefficient throughout the
burner, regardless of turndown conditions or draft levels. That
Reynolds dependence may be a hidden source of error for
capacity curves is something to be aware of when sizing
combustion equipment.
At this time, it is important to note a couple of points. First,
as the combustion air flows through each burner component,
the density will decrease due to the reduction in pressure. As
discussed earlier, a reduction in the combustion air density
will increase the pressure drop. The pressure drop through a to take as much pressure drop across the burner throat as
burner system typically varies between 0.2 to 0.7 in. water possible to achieve good mixing between the air and fuel.
column, which corresponds to an air density variation of about With the damper blades positioned fully open, the pressure
only 0.1 percent. This variation in density is not significant drop through the damper can usually be neglected. The pres-
enough to be considered in the pressure drop calculations. sure drop across the damper blades for various damper set-
However, if draft levels are high, the density variation through tings can be determined using the same technique as discussed
the burner should be considered. Second, as the combustion above. The loss coefficient for each damper setting would
air approaches each section of the burner, the air velocity need to be determined experimentally or by using CFD and
profile may not be fully developed. Most sources provide loss based on the approach velocity of the combustion air in the
coefficients based on the assumption of a fully developed damper section.
velocity profile of the gas upstream of the fitting. For practical
As the air enters into the plenum section, it turns 90°. The
combustion equipment, uniform velocity distributions may
loss coefficient can be determined experimentally, using CFD
not exist, causing the loss coefficient reported in the literature
modeling, or approximated using the loss coefficients given
to be lower than reality. When using literature loss coeffi-
in the ASHRAE Data Handbook for a 90° mitered elbow. Just
cients, the reader should be aware of this possible difference
downstream of the elbow, the combustion air enters into the
and, to the maximum extent possible, utilize experimental
throat section of the burner.
validation for individual burner components.
In addition to the major components, dampers add possible The pressure loss through this section can be approximated
pressure drop to a burner. Burners are typically sized with the as a sudden contraction. The loss coefficient for a sudden
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

damper blades positioned fully open because one would like contraction depends on the design of the entrance and the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 266 Monday, March 14, 2005 2:54 PM

266 The John Zink Combustion Handbook

throat-to-plenum area ratio. Table 4.4 shows typical loss coef- flares and high-pressure flares were also developed and intro-
ficient information, generated from experimental data with a duced into the industry.
fully developed approach velocity profile. Again, due to the The smokeless operating capacity of these early flares rep-
possibility of nonuniform velocity distributions, experimental resented an improvement over a continuously smoking flare.
validation is preferred. Historically, the smokeless performance of a flare was esti-
When the air mixes with the fuel and reacts inside the tile mated by vendor performance testing, field tests, or “rules of
section, the products of combustion expand to several times thumb” based on operating experience.
the volume of the original air/fuel mixture. The rapid increase As smokeless flare capacity increased, vendor performance
in volume flow rate can have a significant effect on the total testing became impractical and performance predictions relied
pressure drop through the burner. The John Zink engineers more heavily upon field data. Unfortunately, field experience

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
have correlated the pressure drop through the tile section was unreliable because flares were typically not monitored.
based on the expansion and density of the products of com-
More recently, the environmental and regulatory agencies
bustion. These correlations are developed from both hot and
have required a minimum smokeless rate as part of plant
cold flow furnace data and CFD analysis.
operating permits. Customers have also begun installing flow
Fuel nozzles located around the tile section can either
meters and video cameras to monitor flare performance. This
increase or decrease the overall pressure drop through the
additional focus has resulted in better field data to assess
burner, depending of the location, orientation, and size of the
actual flare performance. In many instances, the initial “rules-
fuel ports. Cold flow testing is required to provide the infor-
of-thumb” or other estimation methods have failed to accu-
mation necessary to correlate the effects of burner pressure
rately predict smokeless operation. Figure 8.17 highlights
drop for various nozzle configurations.
some of the difficulties of predicting flare performance. This
Semi-empirical modeling does require an attention to
shows that for the same flare, with the same gas, at nearly
detail, but, as it turns out, it only requires about the same
the same operating conditions, one flare smokes while the
amount of time to accomplish as the capacity curve method.
other does not.
However, it has several advantages over the traditional
method of using capacity curves. These models can (1) pro- To address previous shortcomings, a more fundamental
vide a consistent way of determining airside pressure drop technique for determining flare smokeless performance for
for design engineers, (2) give better insight into what vari- commercial flare equipment has been developed. This tech-
ables affect airside pressure drop, and (3) be used as a tool nique is based on the hypothesis that some combination of
to help improve future burner designs. As a first approxima- nondimensional parameters can describe the smokeless burn-
tion, capacity curves are an excellent engineering tool and ing of turbulent diffusion flames of known initial conditions
are easily generated from semi-empirical models. Semi- (i.e., fuel type, diameter, exit velocity, etc.). This method
empirical modeling, however, takes burner capacity and siz- divides the flame into two sections (see Figure 8.18) and
ing to the next level and provides the burner design engineer focuses on the main body of the flame. Effects due to tip
unprecedented flexibility in meeting the needs of a diverse geometry are represented by other proprietary models.
customer base.
8.5.1.1 Industrial Experience
Over time, John Zink engineers have identified several major
8.5 FLARE SMOKELESS OPERATION factors that affect the smokeless operation of a flare: (1) fuel type,
(2) tip diameter, (3) inerts, (4) flow velocity, (5) ambient con-
8.5.1 Predicting Flame Smoking Tendencies ditions (e.g., wind speed, relative humidity, and temperature),
Prior to 1947, venting of unburned hydrocarbons to the atmo- and (6) total mass flow rate.
sphere was standard industry practice. After 1947, regula- These factors, from a qualitative viewpoint, have a signifi-
tions required hydrocarbons to be burned (or “flared”) due to cant impact on the smokeless capacity, but the effect of each
serious health and safety hazards. Initially, flares burned the parameter is difficult to quantify. For example, the tendency
hydrocarbon waste gas stream directly at the vent exit. This to smoke was found15 to roughly correlate with the fuel’s
method of flaring, however, often produced large clouds of hydrogen-to-carbon ratio (H:C) and lower heating value
black smoke that could be seen from miles away. In 1952, the (LHV). For years, the H:C ratio and the LHV of the fuel were
John Zink Company patented and built the first smokeless used to analyze the smoking tendency of a hydrocarbon fuel.
flare (see Chapter 20). This flare eliminated smoke by inject- This information, coupled with experience, were utilized to
ing steam into the waste gas stream. Several years later, air estimate the smokeless rate of a particular flare.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 267 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 267

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 8.17 (a) Smoking and (b) non-smoking flares.

8.5.1.2 Combustion Literature burning off at a faster rate than they were being produced.
The orange and yellow color observed in a flame is produced They concluded that the aeration rate of the flame increased
by light emitted from glowing carbon particles, or soot, with exit velocity. This qualitative observation implies that
inside the flame. When these carbon particles cool, they turn flame length correlations might also be used for predicting
black and are seen as soot or smoke. To eliminate soot, the carbon burn-off rates (Becker17 and Blake18).
particulate carbon must burn off at a faster rate than that at Numerous options have been identified for scaling in-flame
which it is produced. Currently, no directly applicable work soot formation rates, as summarized by Glassman19 and Bar-
has been documented that addresses the smokeless rates of tok.3 A partial list includes: (1) sooting equivalence ratios,
industrial-scale flare flames. However, there is a substantial (2) number of C–C bonds, (3) critical oxygen partial pres-
body of literature in related areas of flame lengths and soot sures, (4) smoke heights, (5) maximum soot volume fraction,
formation rates. (6) H:C ratio (Reed15), and (7) maximum radiant fractions.
To characterize the length of a flame, Hottel and Hawthorn16 Considering the number of scaling parameters, several
first noted that as the exit velocity of an external hydrocarbon models are possible. All of the effects initially identified as
jet increased, the flame length initially increased, while the important have been incorporated into John Zink’s Flare Per-
flame color became more transparent (less yellow). The more formance Model. In general, however, the model scales the
translucent flame indicated that the carbon particles were smoke evolution from turbulent diffusion rates by taking the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 268 Monday, March 14, 2005 2:54 PM

268 The John Zink Combustion Handbook

mass flow rate of hydrocarbon and lowest nozzle exit velocity.

As the mass flow of hydrocarbon is increased, the amount of
diluent also increases. Flames on this part of the graph would
be characterized by fairly low velocities, and smoke suppres-
sion is achieved by dilution. As the hydrocarbon flow contin-
ues to increase, a point is reached where a rapid decrease in
the amount of diluent is required. Flames in this region are
characterized by high velocities, and smoke suppression is
achieved by increasing aeration rates. Thus, the model pre-
dicts that there are two different regimes that flare equipment
could successfully operate in: (1) the dilution region and (2)
the high-velocity or momentum region.
It has been historically observed that steam-assisted flares
and air-assisted flare equipment (see Chapter 20) follow the
general trends of the dilution region. It has also been observed
that high-pressure flares (JZ LRGO flare and hydra flare)
operate in the high-velocity region. Thus, the model as
derived is capable of capturing historical knowledge and
observations of flare operations. This matches previous qual-
itative observations for small flames as reported by Gollahalli
and Parthasarathy.20 In addition, the model provides insight
into the picture sequence shown in Figure 8.17. This sequence
is of a John Zink Company high-pressure flare. The model
qualitatively predicts that for high-pressure flares near the no
dilution required limit, small changes in operation might
result in substantial changes in performance. This prediction
is also verified by historical flare operations.
Figure 8.21 shows the same graphs as discussed above but
with experimental data points superimposed. The comparison
FIGURE 8.18 Flame of a flare is divided into two parts: between the predicted and measured results is excellent. Thus,
near tip interactions and the far field flame region. the model not only predicts historical flare operations accu-
rately, but also quantitative data. The above documented val-
idation has demonstrated that the new model can, with a high
ratio of carbon burn-off, represented by flame length scaling, degree of confidence, predict whether a diffusion flame with
to the soot formation rate, which was represented by an appro- known initial conditions will or will not smoke.
priate set of scaling parameter requirements.

8.5.1.3 Calibration and Validation 8.5.2 Application to Steam Flares

To calibrate and validate the scaling of the performance model, Steam-assisted flares were first introduced in 1952 to provide
the experimental setup shown in Figure 8.19 was used. Experi- smokeless combustion for small to moderate hydrocarbon
ments consisting of five different nozzle sizes, three different gas flow rates during a flaring event. These early steam flare
diluents, and two different fuel types were conducted. The designs were unable to provide good smokeless performance
largest nozzle utilized had nearly 200 times the diameter of the as relief capacity requirements increased and drove tip
smallest nozzle. Results are shown in Figures 8.20 and 8.21. designs to larger diameters. During the 1960s and 1970s, a
Figure 8.20 is a log-log plot of the predicted normalized new steam flare design was introduced to improve smokeless
hydrocarbon flow rate vs. the predicted normalized diluent performance. Equipped with steam eductor tubes, and shown
flow rate. The lines on the graph represent the amount of in Figures 8.4 and 8.5, this design provided substantial bene-
diluent required to prevent the hydrocarbon stream from fits. Recently, the new prediction method for calculating
smoking. The “left” side of the graph represents the lowest smokeless rates has been applied to steam flares.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Mathematical Modeling of Combustion Systems 269

FIGURE 8.19 Illustration showing the experimental setup utilized to obtain calibration and validation data.

8.5.2.1 Predicting Steam Flare Smoking Tendencies steam-to-hydrocarbon ratio (on a mass basis) for various
As previously discussed, John Zink engineers have devel- molecular weights of paraffinic hydrocarbons. The “data”
oped a new predictive method to determine whether a flame points in the plot are from Leite.22 These data points, origi-
will smoke or not. This method relates fundamental measures nally published several years ago, are currently used in the
of the momentum of the air, steam, and flared gas, at the base industry as guidelines to estimate the steam-to-hydrocarbon
of the flare tip, to smokeless capacity. This method can be ratio required for smokeless flaring of paraffinic hydro-
used, in general, to predict the smokeless performance of any carbons. The information likely represents a combination of
flare, as long as the mass flow rates and momentum of the small-scale testing and field observations. Assuming this data
various gas constituents can be accurately determined. was gathered on flares with diameters between 16 and 24 in.
To apply the predictive tool to steam-assisted flares, one (41 and 61 cm) and steam pressures of approximately
must first determine the initial conditions at the base of the
100 psig (6.8 barg), the data may be compared to predictions
flame. In this case, only the velocity and the air/steam mixed
from the steam flare predictive tool. This comparison is also
into the hydrocarbon stream are unknown.
shown in Figure 8.23. The comparison between Leite’s data
For basic steam flares, the combined velocity of the hydro-
and the predicted performance from the steam model, based
carbon and steam mixture is determined from an overall
on the most likely operating conditions, shows good agree-
momentum balance, and entrained air is estimated using free
ment. However, experience suggests that for large-diameter
jet entrainment laws.21
flare tips, the steam-to-hydrocarbon ratios required for
For more complex steam flares, the same procedure is
followed, except the air mixed into the hydrocarbon stream smokeless flaring can increase dramatically above the values
is enhanced by steam eduction tubes. The eductor model provided in Figure 8.22. Figure 8.23 compares predictions
discussed in Chapter 8.2 is used to determine the additional from the new model and field data from flares with over three
air available at the base of the flame. times the diameter. In this case, the graph clearly shows that
These predictive tools have been applied to estimating the there is excellent agreement between the model predictions
smokeless performance of both basic and complex steam and quality field data. The model, however, also dramatically
flares with good success. Such tools also indicate the possible demonstrates a limitation when using steam-to-hydrocarbon
error associated with using standard industry design guide- ratios for estimating steam flare smokeless performance —
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

lines for steam-assisted flares. Figure 8.22 is a plot of the they (steam-to-hydrocarbon ratios) are a function of diameter.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 270 Monday, March 14, 2005 2:54 PM

270 The John Zink Combustion Handbook

106 106

105 Nozzle "D" 105 Predicted Nozzle "D"

Normalized flow rate

Predicted

Normalized flow rate

vs
104 104
Actual
103 Nozzle "C" Nozzle "C"
103

102 102
Nozzle "B" Nozzle "B"
1
10 101

100 Nozzle "A" 100 Nozzle "A"

0 0
100 101 102 103 104 105 106 100 101 102 103 104 105 106
Normalized Propylene Flow Normalized Propylene Flow
=
=

106 106
Predicted Predicted
105 vs
Normalized dilluent flow
105 Nozzle Size "D"
Normalized dilluent flow

Nozzle Size "D"

Actual
104 104

Nozzle Size "C" 103 Nozzle Size "C"

103

102 102

Nozzle Size "B" Nozzle Size "B"

101 101
Nozzle Size "A" Nozzle Size "A"
100 100

0 0
100 101 102 103 104 105 106 100 101 102 103 104 105 106
Normalized Propane Flow Normalized Propane Flow
> >

FIGURE 8.20 Prediction of diluent flow for smokeless FIGURE 8.21 Figure 8.20 with data points added.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

operation.

Figure 8.24 shows a steam flare in operation that was designed mix into the hydrocarbon as the fan comes up to speed; and
by this new method. the third is steady-state operation with a highly enhanced
air mixing rate. Figure 8.26 also highlights the enhanced
8.5.3 Modeling Air-Assisted Flares aeration effects on the flame; flame length decreases as
An air-assisted flare uses air supplied from a high-pressure smoking is reduced.
fan as a supplemental energy source to achieve smokeless To apply the predictive tools to air-assisted flares, hydro-
combustion of a hydrocarbon stream. This is achieved by carbon and air velocity, along with a rate of aeration must be
increasing the overall exit velocity of the stream and determined. The velocity increase of a hydrocarbon stream
through increased aeration rates. A typical schematic is due to interactions with the supplied air from the blower is
shown in Figure 8.25. Figure 8.26 shows the effect of the calculated from a momentum balance. The rate at which the
high-velocity air on an unsaturated hydrocarbon. The first supplied air is mixed into the hydrocarbon stream is more
photo is with the fan off; the second is the air starting to difficult, but may be assumed to be a turbulent diffusion
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 271 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 271

FIGURE 8.22 Steam-to-hydrocarbon ratios (per Leite).22

FIGURE 8.23 Steam-to-hydrocarbon ratios (large-diameter flares).

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

272 The John Zink Combustion Handbook

process. Turbulent diffusion for this system is a function of

(1) velocity of fuel gas, (2) air velocity, (3) circumference
perimeter, (4) characteristic dimension, and (5) turbulent mix-
ing length.21
The turbulent mixing length, according to Rokke23 and Schli-
chting,21 is a function of (1) exit velocity, (2) reaction rate, (3)
density, and (4) viscosity. These parameters can be combined
into a “scaling functional,” which allows the aeration rate of
a hydrocarbon stream to be calculated.
A series of calibration tests were performed to evaluate this
scaling functional. Tests were conducted on air flares with tip
areas of 30 and 500 square in. For each test, the aeration rate
was determined and plotted against the scaling functional (see
Figure 8.27). The relationship shown for aeration rate was
developed for a variety of hydrocarbon fuels (propane, pro-
pylene, natural gas, and mixtures) and for the various tip
outlet areas. Combining the aeration parameter with the pre-
viously described turbulent flame model has improved our
ability to calculate the smokeless performance for a wider
variety of air flare designs, flow rates, and hydrocarbon types.
This model has been successfully utilized to design numerous
air-assisted flares worldwide (Figure 8.28 shows an air-
assisted flare firing 190,000 lbs/hr of propane). Application
of the model has also provided valuable insight into the oper-
ation and design of air flares as discussed below.
Leite22 describes an alternative method for sizing air flares.
FIGURE 8.24 Smokeless steam flare designed with This method predicts smokeless performance for an air flare
new method.
if a specified fraction of the stoichiometric air requirement
for a hydrocarbon fuel is supplied at the flare tip. No mini-
mum velocity requirement is provided, but common practice
is to provide an exit velocity of 120 ft/sec. This industry “rule
of thumb” has been used for many years. Similar to the
original steam flares, the vast majority of air flares (until
recently) had relatively small exit areas (100 to 150 in.2),
enabling most vendor test facilities to evaluate air flare
designs. However, as plant relief capacities and customer
requirements have increased over the last several years, full
vendor testing is not normally possible.
As demonstrated earlier, the standard rules of thumb were
not applicable for large-diameter steam tips. The logical
question that must be asked is whether the standard rule of
thumb for air flare sizing applies to larger tip sizes as well
as small tips.
Figure 8.29 shows the results generated using the air flare
predictive tool with the aeration parameter included. Items
in the box represent variables held constant. As the tip area
is reduced, the exit velocity increases for a given hydrocar-
FIGURE 8.25 Typical air flare. bon flow rate. The model clearly highlights the real effects
of exit velocity on the smokeless rate of an air-assisted flare.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

The two cross symbols reflect actual experimental data gath-

Mathematical Modeling of Combustion Systems 273

ered by John Zink R & D test engineers. The square symbols

represent anticipated smokeless performance as predicted
by Leite’s method (assuming that the calibration of the Leite
method was accomplished on a 125 in.2 air flare tip). This
graph shows that the John Zink model accurately predicts
the variation in performance trends for the various tip sizes
tested. For a given hydrocarbon flow rate, the velocity
decreases with an increase in tip diameter. As discussed
earlier, the reduction in hydrocarbon velocity decreases the
amount of air mixed with the hydrocarbon stream and
decreases the overall stream velocity, resulting in decreased
smokeless performance.
In Figure 8.29, all geometric variables were held constant
(a)
except for the outlet area. Flare vendors typically fabricate
larger tips to enhance air/hydrocarbon mixing. Although this
helps mitigate the mixing limitation, it does not eliminate the
problem, and serious performance problems can result. A
variation of as much as a factor of three is observed when
comparing the predicted smokeless performance using the
Leite method vs. an advanced predictive tool.

8.5.4 Summary
An analytical model to predict smoke evolution from a tur-

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
bulent diffusion flame has been developed and implemented
by John Zink Company engineers. Recent applications of
the method have resulted not only in optimized equipment
but greater insights into equipment operation and possible (b)
failure modes as well. In every case, the most significant
problems were observed when attempting to “scale-up”
existing equipment to larger sizes. Even though the model
predicted the possible performance problems with larger
equipment, validation with large-scale data was still
required. As described at the beginning of this section, reli-
ance on field data may be problematic; ultimately, only
larger test facilities can provide the data required to prevent
serious under-sizing of equipment and verify the accuracy of
predictive models. Having this experimental and analytical
capability is a requirement for supplying reliable, high-
capacity industrial flare equipment.

(c)
8.6 OIL GUN CAPACITIES FIGURE 8.26 Effect of high velocity air: (a) blower off,
Oil firing of process heaters is common in most of the world. (b) commence blower and (c) blower on (Courtesy of John
To efficiently combust oil, it must be “broken up” into very Zink Co., LLC., Tulsa, OK).
small droplets, or atomized (Figure 8.30 shows a typical oil
flame). Typical applications utilize steam as an atomizing
agent and require some type of equipment to effect mixing of Zink oil gun, and Figure 8.32 shows a typical oil gun sche-
the two streams. This oil/steam mixer or atomizer is normally matic. Steam and oil are supplied separately, then mixed in a
referred to as an oil gun. Figure 8.31 shows the standard John manner to enhance atomization. The equipment is operated
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 274 Monday, March 14, 2005 2:54 PM

274 The John Zink Combustion Handbook

30 square inch tip, C3H6 & Natural Gas

12 500 square inch tip, C3H8
500 square inch tip, C3H6

10
Airation Rate Parameter
8

0
0 500 1000 1500 2000 2500
Scaling Functional
FIGURE 8.27 The aeration rate was determined and plotted against the scaling functional.

by setting the steam and oil pressures, with the steam having cating issues in developing an analytical model of the system;
a higher pressure than the oil. Typically, these pressures will two-phase flows, several choked flows, two-phase choked
have a fixed differential; for example, setting the steam to flow regions, atomization efficiency, and accurate determina-
50 psig (3.4 barg) and the oil to 35 psig (2.4 barg) would cor- tion of flow rates are just a few of the issues. This section
respond to a 15 psig (1 barg) differential. summarizes the model formulation and results obtained.
For a given application, the customer is interested in know- The problem formulated and solved by this flow model is
ing what heat release (or capacity) and steam usage to expect that of determining the required oil pressure to flow a
from given pressure settings. Typical oil gun capacity curves specified amount of oil, atomized by steam, through a partic-
will show heat release as a function of available oil pressure ular size proprietary gun. The required user inputs to this
only. These curves are based on limited experimental data problem are the mass flow rate of oil (or, alternatively, the
from a specific oil, oil temperature, steam temperature, and heat release), the steam differential (typical value ~30 psi),
pressure differential. For these types of curves, the only prop- the grade of oil fired, and the steam and oil temperatures.
erty that can be considered is the oil heating value. All other This effort has resulted in a model that predicts the two-phase
factors (e.g., oil temperature, specific gravity, steam temper- flow rates with greater accuracy than previously possible.
ature, or differential) cannot be considered because there is
typically no basis to correct the capacity curves for these 8.6.1 Summary of Two-Phase Flow
factors. In addition, it would be expected that steam usage Analytical Development
(expressed as a fraction of the oil mass flow rate) would vary The description below follows the general development out-
as a function of absolute oil pressure. But again, there is no lined in Kaviany,24 Kaviany,25 Chislom,26 and Chislom.27
basis for adjustment of anticipated steam flow rate based on
a different operating pressure. Clearly, the need for a more 8.6.1.1 Conservation of Mass
thorough analytical model existed. In the model equations, the three components (oil, gaseous
A recent characterization effort at the John Zink Company water, and liquid water) are considered to have different
has focused on utilizing fundamental principles to develop a velocities. This complicates the description of the local com-
more general capacity prediction method for these devices. position because, as shown below, allowing each phase to
As can be seen from the schematic, there are several compli- have different velocities requires the definition not only of
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Mathematical Modeling of Combustion Systems 275

FIGURE 8.28 Annular air-assisted flare; 190,000 lb/hr propane.

conventional mass and volume fractions, but also the defini- 8.6.1.3 Momentum Conservation
tion of mass flux fractions. A mass flux fraction is defined as For an idealized, one-dimensional inviscid flow, the differen-
the mass flux of an individual component divided by the total tial momentum equation can be written as in Anderson.28
mass flux. Mathematical description also requires the quality
of the steam. dP + ρudu = 0 (8.24)

In the separated flow model,27 the various phases are postu-

8.6.1.2 Choked Flow lated to flow with separate velocities, but at any location, it is
From Chisolm,28 the mass flux through a sonic orifice of a assumed that the pressure of each phase is the same. Follow-
two-phase mixture is the square root of the pressure to spe- ing this line of reasoning, the above differential conservation
cific volume derivative. Functionally, this is: equation is split into liquid and gas phases for the velocity
and density terms. Under this model, this differential
∂( Pressure) momentum equation can be written:
Choked flow mass flux = (8.23)
∂(Specific volume)
dP + (ρudu) Steam + (ρudu) Liquid water + (ρudu)Oil = 0 (8.25)
The model further assumes that the choked flow process is
isentropic and that there is no condensation, boiling, or 8.6.1.4 Conservation of Energy
evaporation. For these assumptions, the only variable In an incompressible, constant-density flow, conservation of
through the orifice that is a function of pressure is the specific mass and momentum are sufficient to analyze the flow. The
volume of the steam. The above differential is then evaluated term incompressible does not imply constant density, but
explicitly in terms of the steam specific volume. rather that the pressure forces arising within a fluid flow are
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

276 The John Zink Combustion Handbook

Air / Pressure insufficient to change its density. Formally, an incompressible

Assisted Flare 30 in2 flow is one with Mach number M2 << 1. When flow velocities
16
are large enough that the arising pressures can change the den-
14
sity appreciably, it becomes necessary to consider energy con-
Normalized Smokless Rate

servation in analyzing a flow. Another way of thinking about

Constant SCFM
12 Constant Air Velocity
this is that M2 represents the ratio of the kinetic energy to the
Geometric Similarity thermal energy. As the Mach number gets large, the conver-
10 Propane as the Hydrocarbon sion of thermal energy into kinetic energy becomes important.
Outlet Area was Varied
For an idealized, one-dimensional compressible flow,
8 energy conservation can be written as in Anderson:28
6
O.C. Leite Guidelines, d (Thermal energy + Kinetic energy) = 0 (8.26)
125 in2 Tip
4
In the three-component flows considered here, the kinetic
2 500 in2 Tip energy flux (KE) of the mixture is computed by calculating
the kinetic energy in the mass flow of each phase (oil, steam,
0 and water). In a similar manner, the thermal energy (TE) is
0 250 500 750 1000 1250 1500 calculated by summing the thermal energy in each phase.
Tip Open Area (square inches) This can be summarized as follows:

FIGURE 8.29 Comparison of three different air flares KE = KE Oil + KE Steam + KE Liquid water
to prediction.
TE = TE Oil + TE Steam + TE Liquid water (8.27)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 8.30 Typical process heater oil flame.

Copyright CRC Press

Mathematical Modeling of Combustion Systems 277

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
The reference condition chosen is pure liquid at 373 K and
1 atm. Pressure changes and interfacial heat transfer will
induce phase changes to the steam/water system. These
effects are modeled by appropriate latent heat terms, thus
accounting for condensation and evaporation effects. To con-
serve energy, the sum of the kinetic energy and the thermal
energy must remain constant throughout the oil gun. Thus, a
total energy (enthalpy) was defined as the sum of the thermal FIGURE 8.31 Standard John Zink oil gun.
and kinetic energies

h t ≡ TE + KE (8.28)

This results in the following conservation equation:

d( h t ) = 0

The above system of coupled differential equations along

with flow coefficient information for our proprietary equip-
ment were then coded into a Visual Basic program utilizing a
“shooting” Runge-Kutta solver.

8.6.2 Results
Typical results are shown in Figure 8.33. As can be seen, FIGURE 8.32 Schematic of a typical oil gun.
excellent agreement is achieved between experimental and
theoretical flow rates for different oil pressures. Similar The performance of atomizers depends on their design, the
results were observed for different oils and pressure differen- physical properties of liquid fuel and atomizing medium, and
tials. These efforts have resulted in a capacity prediction the operating conditions (i.e., pressure and temperature). The
technique of unprecedented flexibility. John Zink application liquid properties include viscosity, surface tension, and den-
engineers now have the ability to optimize oil gun applica- sity. The effect of fluid properties on atomization can be found
tions to meet a wider variety of customer conditions. in Chung and Presser.34 General atomizer design, atomizing
medium properties, and operating conditions have been dis-
cussed in detail by Lefebvre.35 Here, the focus will be on
8.7 OIL GUN DEVELOPMENT steam-assist oil gun characterization and improvement.
Following on the successful development of an oil gun Oil gun design has a significant influence on the spray
model, it was decided to further optimize oil gun design by combustion performance and exhaust pollutant emissions. A
studying atomization. There are many ways to atomize a bulk good oil gun should generate a good flame shape, consume
liquid into small droplets. Normally, a high relative velocity minimal atomization medium (steam or compressed air), and
between the liquid to be atomized and the surrounding air or exhaust limited particulate and NOx emissions. The major
gas is created. The high shear force between the liquid and factor that determines such performances is the atomization
the gas disrupts the liquid into droplets. Some atomizers quality of the oil gun. The parameters that characterize the
accomplish this by discharging the high-velocity liquid into a atomization quality include the mean droplet size, droplet size
relatively slow-moving stream of air or gas. Examples of this distribution, spray cone angle, and liquid distribution.
technique include pressure atomizers and rotary cup or disk A fuel spray is usually composed of a wide range of droplet
atomizers. An alternative approach is to introduce a high- sizes. The biggest droplets in the spray may be 50 or 100
velocity gas stream into the liquid to assist atomization. This times the size of the smallest droplet. The actual sizes of the
is generally known as a twin-fluid, air-assist, or airblast droplets represent the degree of spray fineness. When com-
atomizer. In industrial furnaces, the steam-assist or air-assist paring the fineness of different sprays, it is useful to introduce
atomizers are most common. some “mean droplet size.” In spray combustion, the Sauter
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 278 Monday, March 14, 2005 2:54 PM

278 The John Zink Combustion Handbook

FIGURE 8.33 Comparison of predicted vs. actual oil and steam flow rates.

mean diameter (SMD) is usually used to represent the mean well understood that the finer the droplets, the better the
droplet size. The SMD is the diameter of the droplets whose evaporation. As for droplet size distribution, a more compli-
ratio of volume to surface area is the same as that of the entire cated correlation exists. If one defines the initial size distri-
spray and defined as: bution before evaporation as qo, then after sprays injecting
into the hot furnace, q will vary with evaporation time. The

SMD =
∑n d i i
3

(8.29)
general trend is for q to increase with evaporation time, and
the effect is more significant for a spray having a low value
∑n d i i
2
of qo. Usually, for the ignition of a fuel spray, the time required
to vaporize 20% of the spray mass is important; whereas for
where ni is the number of droplets in size di . The SMD is well combustion efficiency, the time required for vaporization of
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

accepted because total droplet volume and surface are most 90% of the spray mass is crucial. Chin et al.37 reported that
significant in connection with the combustion process. with a given mean droplet size, a spray of large qo would have
The range of droplet size variation (or size distribution) a low 90% evaporation time and a high 20% evaporation time.
indicates the degree of spray uniformity. At present, the most Thus, from a combustion efficiency viewpoint, it is desirable
widely used expression for the droplet size distribution is the to have a fuel spray with a wide size distribution, but for good
Rosin–Rammler relationship.36 It is expressed as: ignition performance, it is better to have a narrow size distri-
bution. It is not possible to know a priori what is the optimal
1 − Q = exp − ( D X ) compromise between combustion efficiency and ignition;
q
(8.30)
experimental performance will always be required.
where Q is the fraction of the total volume contained in drop- The cone angle of a spray is usually defined as the angle
lets of diameter less than D, and X and q are constants. The between tangents to the spray envelope at the oil gun tip. The
constant X represents the droplet diameter, and the exponent value to be selected for the cone angle of a spray will depend
q provides a measure of the spread of droplet size. The higher on the shape of the furnace and the conditions controlling the
the value of q, the more uniform is the spray. If q is infinite, mixing of air and fuel. For furnaces with a high degree of air
the droplets in the spray are all the same size. For most movement (i.e., swirl or forced-draft), sprays with a wide
sprays, the value of q lies between 1.5 and 4. cone angle will give good results. On the other hand, furnaces
For a given liquid fuel and a fixed environmental condition, with limited air movement (i.e., natural draft) will require
the mean droplet size (SMD) and the size distribution (q) are sprays of narrow cone angle. If a short period of time between
two major parameters that affect the evaporation rate. It is the beginning of injection and the beginning of combustion
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 279 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 279

is desired, a compact and penetrating spray (i.e., narrow spray

cone) should be used. However, for a short flame, a soft and
well-dispersed spray (i.e., wide spray angle) is preferred.
The liquid fuel distribution within a spray has an important
effect on pollutant generation. The technologies used to mea-
sure the liquid distribution can be found in Chung et al.38
Nonuniformity in liquid fuel distribution can give rise to local
pockets of fuel-lean or fuel-rich mixtures. The fuel-lean pock-
ets have low burning rates and thereby produce high concen-
trations of carbon monoxide and unburned hydrocarbons. The
fuel-rich mixtures are characterized by high soot formation
and lead to high particulate emissions. The best oil guns
exhibit excellent radial symmetry and circumferential mald-
istributions of less than 10%.
As can be readily understood from this discussion, steam-
assist oil guns have numerous competing characteristics, all FIGURE 8.34 John Zink Co. LLC (Tulsa, OK) Spray
of which can affect performance. For instance, a well-atom- Research Laboratory. (Courtesy of John Zink Co. LLC.,
ized and well-dispersed spray generates a high-temperature Tulsa, OK.)
flame. It is known that NOx formation strongly depends on
the flame temperature. The high flame temperature has a

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
tendency to produce high thermal NOx. On the other hand, 65
large droplets have a poor evaporation rate and provide a 60 Standard Oil Gun
favorable condition for soot formation. An inadequate atom- 55 New Oil Gun
ization system can even reduce combustion efficiency and 50
result in a low combustion turndown ratio. Therefore, actual 45
SMD ( mm)

spray combustion experiments are always required to deter- 40

mine the best compromise among the various parameters. 35
Prior to optimizing steam-assist oil gun characteristics, a
30
method of characterizing the atomization was required. Sev-
25
eral methods have been employed to measure the droplet sizes
20
in sprays, for example, droplet collection on slides, molten-
15
wax and frozen-drop techniques, cascade impactors, charged-
wire and hot-wire techniques, high-speed photography, light 10

scattering, etc. The most advanced and popular method in 5

current use is the Phase Doppler Particle Analyzer (PDPA). 0
-1 0 1 2 3
The PDPA can simultaneously measure the droplet size and
the velocity in situ. The combined information of droplet size
Radial Position (in)
and velocity is useful for calculating the spray cone angle and FIGURE 8.35 Droplet size comparison between a stan-
the liquid distribution. The John Zink PDPA instrument setup dard and a newer oil gun.
is shown in Figure 8.34. The PDPA uses the Doppler signal
and its phase shift to simultaneously characterize the particle the ordinate represents the droplet size measurements and the
velocity and the droplet size. It counts every droplet that abscissa is the radial position of the spray.
passes through the sample volume. The sample volume is the Utilizing the John Zink Company PDPA, the oil gun oper-
cross-section of two laser beams. Every measurement con- ating parameters were characterized. During the character-
tains 10,000 sample points and statistically represents the ization efforts, it was found that the liquid jet diameter, or
mean droplet size, size distribution, standard deviation, and film thickness, was also an important parameter affecting
velocity distribution. The detailed theoretical description of atomization quality. To improve atomization, a pre-filming
PDPA can be found in Bachalo.39 technique was developed and introduced into the design.
An example of the use of a PDPA to compare two different The new design was then optimized by combustion testing
designs of oil guns is provided in Figure 8.35. In the figure, in the John Zink Company test furnaces. The optimized gun
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 280 Monday, March 14, 2005 2:54 PM

280 The John Zink Combustion Handbook

FIGURE 8.36 High-efficiency new oil gun.

showed a high turndown ratio, low steam consumption, short numerous cycle variations. Thermal performance is mea-
flame length, and low NOx and particulate emissions. The sured in terms of heat recovery efficiency (HRE), and is
pre-filming technique makes the best use of steam momen- defined as shown below:
tum, which reduces the steam consumption. The pre-filming
technique improved the atomization control and generated TCombustion Chamber − TOutlet
optimal droplets. The droplets have a fast evaporation rate, HRE = (8.31)
TCombustion Chamber − TInlet
resulting in high turndown ratios and low particulate emis-
sions, and appropriate size distribution shortened the flame
The units typically can obtain an HRE of about 95 percent.
length and reduces the conversion of fuel-bound nitrogen to
For an industry standard 1500°F (820°C) combustion cham-
NOx, resulting in the reduction of total NOx emissions. The
ber temperature and 100°F (38°C) process inlet temperature,
net result of this program was a major improvement in oil
a 95 percent HRE would result in an exhaust temperature of
gun performance. The new gun is shown in Figure 8.36.
about 170°F (77°C). One result of a high HRE is that the
additional cost of regeneration equipment over the cost of a
standard incinerator is typically recovered in about nine
8.8 REGENERATIVE THERMAL months due to lowered fuel consumption.
OXIDIZER (RTO) PERFORMANCE
8.8.2 RTO Model Development
8.8.1 Introduction

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Figure 8.37 is a schematic of heat transfer in a packed bed
For over 100 years, regenerators have been utilized in various between the ceramic material and the air stream. Neglecting
process industries (steel and glass manufacturing are typical heat release effects in the air stream, a first law of thermo-
applications). Applications and performance solution dynamics analysis for the air results in:
methodologies are documented in Schmidt.31 More recently,
in response to air quality concerns, regenerators have been
a − w = mcp ( dTa )
˙
dQ ˙ (8.32)
applied to incineration of high flow rate, low contamination
(VOC) streams. Called regenerative thermal oxidizers
(RTOs), they were first introduced in the early 1970s. Typical where
applications have VOC concentrations that vary from the d Q̇a-w= Heat transferred from the ceramic to the air
ppm level to about 1 percent, or have a heating value from in distance “dx”
about 0 to 10 Btu/scf. These streams are typically produced ṁ = Mass flow of air
by processes requiring ventilation, such as paint booths, cp = Specific heat of air
printing, paper mills, and others. Many units are also Ta = Air temperature
equipped with a third bed to increase destruction and removal dTa = Change of air temperature during distance
efficiency (DRE). When the inlet stream is switched to a dif- “dx”
ferent bed, the VOC-laden air is purged out of the bed with
clean air for a period of time. Depending on design, the purge Similarly, a first law analysis of the ceramic results in the fol-
period may be shorter than the cycle time, allowing for lowing:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 281 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 281

FIGURE 8.37 Heat transfer in a packed bed between

the ceramic material and the air stream.
FIGURE 8.38 Installing small, type K thermocouple
“pairs” into numerous ceramic saddles.

dQ̇ w−a = mc(∂Tw ) (8.33)

where

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
d Q̇w-a= Heat transferred from the air to the ceramic
in distance “dx”
m = Mass of ceramic
c = Specific heat of ceramic
Tw = Ceramic temperature
∂Tw = Change of ceramic temperature during time
“dt”

Utilizing Newton’s law of cooling for the heat transfer rate,

and appropriate expressions for (1) mass of ceramic and
(2) surface area of ceramic, the following coupled differential
equations can be developed:

dTa hSA(Tw − Ta )
=
dx ˙ p
mc
FIGURE 8.39 Summary of saddle data.
dTw hSA(Ta − Tw )
= (8.34)
All of the above parameters are readily known, except for
dt mc
the convective heat transfer coefficient. Calibration informa-
where tion for this parameter was obtained by installing small, type K
Tw = Ceramic temperature thermocouple “pairs” into numerous ceramic saddles as illus-
Ta = Air temperature trated in Figure 8.38. One of the thermocouples is inside the
ṁ = Mass flow of air ceramic, and the other is directly outside the saddle. Temper-
cp = Specific heat of air ature variations with time in an actual RTO bed were measured
m = Mass of ceramic and deconvoluted to obtain the value of the Nusselt number
c = Specific heat of ceramic (nondimensional convective heat transfer coefficient) as a
SA = Specific surface area of ceramic (area/volume) function of the local Reynolds number. Typical results are
h = Convective heat transfer coefficient shown in the graph in Figure 8.39. The differential equations,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 282 Monday, March 14, 2005 2:54 PM

282 The John Zink Combustion Handbook

Symbols = Experimental
1600 Solid Lines = Numerical
30 seconds
60 seconds
150 seconds
Temperature (deg F)

1200 270 seconds

800

400

0
0 10 20 30 40 50
Height (inches)

FIGURE 8.40 Rock temperature distribution with time.

along with a best fit for the Nusselt correlation, were then
coded into a FORTRAN program, utilizing a Bulirsch-Stoer
technique to solve the differential equations. This model was
then compared to various scenarios; first, the bed with the
thermocouple pairs was subjected to cooling and heating
cycles. In-bed temperatures were measured as a function of
time and position. The above model was then utilized to pre-
dict in-bed temperatures and compared to experimental data.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Typical graphical results are shown in Figure 8.40. As can be
seen, excellent agreement was obtained between the model
and the experimental data. Figure 8.41 shows the John Zink FIGURE 8.41 John Zink RTO test unit.
Test RTO Test Unit.
Finally, the model was compared to heat recovery efficien- This model is currently managed by Koch-Knight Division
cies (HREs) measured in the field, with results as follows. (a supplier of ceramic packing materials), and has been suc-
cessfully utilized for numerous RTO bed retrofit applications.
Efficiency Comparison It is employed to estimate HRE changes for different flow
1. 8.5 ft of 1.5-in. Berl saddles rates and packing materials when customers want to change
Inlet: 450°F, .25 Btu/scf,
Combustion chamber: 1600°F
process conditions.

Predicted Actual

Outlet temperature 602°F 595°F

HRE 0.868 0.874
8.9 CONCLUSIONS
Numerous analytical techniques have been utilized to success-
2. 36 in. Type 28H, 24 in. Type 48, fully model actual equipment performance by John Zink engi-
42 in. (1.5 in.) Berl saddles
neers. The following applications were reported in this chapter:
Inlet: 450°F, .25 Btu/scf,
Combustion chamber: 1600°F
1. control volume analysis (eductor modeling — 8.2)
Predicted Actual
2. compressible flow analysis (eductor — 8.2, and oil gun
Outlet temperature 538°F 549°F modeling — 8.6)
HRE 0.923 0.914 3. detailed chemical kinetics (combinations of chemical
reactors — 8.3)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 283 Monday, March 14, 2005 2:54 PM

Mathematical Modeling of Combustion Systems 283

4. extensions to classical fluid mechanics (burner pressure 9. Longwell, J.P. and Weiss, M.A., High temperature
drops — 8.4) reaction rates in hydrocarbon combustion, Ind. Eng.
5. novel scaling methodologies (flare smokeless operation Chem., 47, 1634–1643, 1955.
— 8.5)
10. Beer, J.M. and Lee, K.B., The effect of the residence
6. classic turbulence closure methods (flare smokeless oper-
time distribution on the performance and efficiency of
ation — 8.5)
combustors, Tenth Symposium on Combustion, X,
7. two-phase flow analysis (oil gun modeling — 8.6)
1187–1202, 1965.
8. advanced experimental techniques (oil gun development
— 8.7) 11. Lutz, A.E. and Broadwell, J.E., Simulation of Chemical
9. numerical ordinary differential equation solution tech- Kinetics in Turbulent Natural Gas Combustion, GRI
niques (oil gun modeling — 8.6, and RTO modeling — 8.8) Report 92-0315, Gas Research Institute, 1992.
12. Lutz, A.E. and Broadwell, J.E., Simulation of Chemical
Combustion systems can be, and have been, modeled with Kinetics in Turbulent Natural Gas Combustion, GRI
good success. The broad range of applications and techniques Report 94-0421.1, Gas Research Institute, 1994.
reported in this chapter serve to illustrate the power of some
13. ASHRAE Handbook, 1985 Fundamentals, published by
of the various methods. Practicing engineers no longer have
the American Society of Heating, Refrigeration and
to be limited by narrow “rules of thumb.” Rigorous applica-
Air-Conditioning Engineers, Inc., Atlanta, GA.
tion of fundamentals married to the power of modern com-
puters can provide insight into equipment operation and yield 14. Idelchik, I.E., Handbook of Hydraulic Resistance,
performance predictions of greater precision than previously Hemisphere, New York, 1986.
thought possible. 15. Reed, R.D., Furnace Operations, 3rd ed., Gulf Publishing,
Houston, TX, 1981.
16. Hottel, H.C. and Hawthorne, W.R., Diffusion in laminar
REFERENCES flame jets, Third Symp. Combustion and Flame and
Explosion Phenomena, Williams and Wilkins Company,
1. Keenan, J.H. and Neumann, E.P., A simple air ejector, Baltimore, MD, 1949.
J. Appl. Mech., A-75, 1942. 17. Becker, H.A. and Liang, D., Visible length of vertical free
2. Ames Research Staff, Equation, Tables and Charts for turbulent diffusion flames, Comb. Flame, 32, 115-137,
1978.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Compressible Flow, NACA Report 1135, 1953.
3. Bartok, W. and Sarofim, A., Fossil Fuel Combustion, 18. Blake, T.R. and McDonald, M., An Examination of
A Source Book, John Wiley & Sons, New York, 1991, Flame Length Data from Vertical Diffusion Flames,
291-320. Comb. Flame, 94, 426-432, 1993.
4. Kee, R.J., Miller, J.A., and Jefferson, T.H., CHEMKIN: 19. Glassman, I., 22nd Int. Symp. Combustion, Combus-
A General-Purpose, Problem-Independent, Transport- tion Institute, Pittsburgh, PA, 1988, 295.
able, FORTRAN Chemical Kinetics Code Package, 20. Gollahalli, S.R. and Parthasaranthy, R.P., Turbulent
Sandia Report SAND80-8003, 1980. Smoke Points in a Cross-Wind, Research Testing
5. Kee, R.J., Rupley, F.M., and Miller, J.A., The Services Agreement No. RTSA 3-1-98, University of
CHEMKIN Thermodynamic Data Base, Sandia Report Oklahoma, Norman, OK, August 1999.
SAND87-8215B, 1990. 21. Schlichting, H., Boundary-Layer Theory, 7th ed.,
6. Kee, R.J. and Miller, J.A., A Structured Approach to McGraw-Hill, New York, 1979.
the Computational Modeling of Chemical Kinetics and 22. Leite, O.C., Smokeless, efficient, nontoxic flaring,
Molecular Transport in Flowing Systems, Sandia Hydrocarbon Process., March 1991, 77-80.
Report SAND86-8841, 1991. 23. Rokke, N.A., Hustad, J.E., and Sonju, O.K., A study of
7. Turns, S.R., An Introduction to Combustion, McGraw-Hill, partially premixed unconfined propane flames, Comb.
New York, 1996. Flame, 97, 88-106, 1994.
8. Press, W.H., Teukolsky, S.A., Vetterling, W.T., and 24. Kaviany, M., Principles of Heat Transfer in Porous
Flannery, B.P., Numerical Recipes in FORTRAN — Media, Springer-Verlag, Berlin, 1991.
The Art of Scientific Computing, Cambridge University 25. Kaviany, M., Principles of Convective Heat Transfer,
Press, Cambridge, UK, 1992. Springer-Verlag, Berlin, 1994.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch08-Frame Page 284 Monday, March 14, 2005 2:54 PM

284 The John Zink Combustion Handbook

26. Chisolm, D., Flow of compressible two-phase mixtures 33. Glarborg, P., Kee, R.J., Grear, J.F., Miller, J.A., PSR: A
through sharp-edged orifices, J. Mech. Eng. Sci., 23, FORTRAN Program for Modeling Well-Stirred Reac-
45-48, 1981. tors, Sandia Report SAND86-8209, 1986.

27. Chisolm, D., Gas–liquid flow in pipeline systems, in 34. Chung, I.P. and Presser, C., Fluid properties effects on
Handbook of Fluids in Motion, Cheremisinoff, N. P. sheet disintegration of a simplex pressure-swirl atom-
and Gupta, R., Eds., Ann Arbor Science, Ann Arbor, izer, AIAA J. Propulsion Power, in press.
MI, 1983.
35. Lefebvre, A.H., Atomization and Sprays, Hemisphere,
28. Anderson, J.D., Modern Compressible Flow: With 1989.
Historical Perspective, McGraw-Hill, New York, 1982. 36. Rosin, P. and Rammler, E., The law governing the fine-
29. Chung, I.P., Dunn-Rankin, D., and Ganji, A., Character- ness of powdered coal, J. Inst. Fuel, 7(31), 62-67, 1933.
ization of a spray from an ultrasonically modulated 37. Chin, J.S., Durrett, R., and Lefebvre, A.H., The inter-
nozzle, Atomization and Sprays, 7, 295-315, 1997. dependence of spray characteristics and evaporation
30. Bachalo, W.D., Method for measuring the size and history of fuel sprays, ASME J. Eng. Gas Turbine
velocity of spheres by dual-beam light scatter inter- Power, 106, 639-644, 1984.
ferometry, Appl. Opt., 19, 363-370, 1980. 38. Chung, I.P., Dunn-Rankin, D., and Ganji, A., Charac-
31. Schmidt, F.W. and Willmott A.J., Thermal Energy terization of a spray from an ultrasonically modulated
Storage and Regeneration, Hemisphere/McGraw-Hill, nozzle, Atomization and Sprays, 7(3), 295-315, 1997.
Washington, D.C., 1981. 39. Bachalo, W.D., Method for measuring the size and
velocity of spheres by dual-beam light scatter inter-
32. McDermott, R. and Henneke, M.R., “High Capacity,
ferometry, Appl. Opt., 19(3), 363-370, 1980.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Ultra Low NOx Radiant Wall Burner Development,”
12th Ethylene Forum, May 11–14, 1999, The Wood- 40. Saad, M.A., Compressible Fluid Flow, Prentice-Hall,
lands, TX. Englewood Cliffs, NJ, 1993.

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 9
Computational Fluid Dynamics (CFD)
Based Combustion Modeling
Michael Henneke, Joseph D. Smith, Jaiwant D. Jayakaran, and Michael Lorra

TABLE OF CONTENTS

9.1 Overview ................................................................................................................................................. 288

9.2 Introduction............................................................................................................................................. 288
9.2.1 CFD Model Background........................................................................................................... 291
9.2.2 The CFD Simulation Model ..................................................................................................... 291
9.3 CFD-based Combustion Submodels ....................................................................................................... 296
9.3.1 Solution Algorithms .................................................................................................................. 297
9.3.2 Radiation Models ...................................................................................................................... 297
9.3.3 Combustion Chemistry Models ................................................................................................ 299
9.3.4 Pollutant Chemistry Models ..................................................................................................... 301
9.3.5 Turbulence Models.................................................................................................................... 301
9.4 Solution Methodology ............................................................................................................................ 302
9.4.1 Problem Setup: Preprocessing .................................................................................................. 302
9.4.2 Solution Convergence ............................................................................................................... 303
9.4.3 Analysis of Results: Post-processing ........................................................................................ 303
9.5 Applications: Case Studies ..................................................................................................................... 305

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
9.5.1 Case 1: Ethylene Pyrolysis Furnace.......................................................................................... 305
9.5.2 Case 2: Xylene Reboiler ........................................................................................................... 306
9.5.3 Case 3: Sulfur Recovery Reaction Furnace .............................................................................. 307
9.5.4 Case 4: Incineration of Chlorinated Hydrocarbons .................................................................. 309
9.5.5 Case 5: Venturi Eductor Optimization ...................................................................................... 319
9.6 Future Needs ........................................................................................................................................... 319
9.7 Conclusion .............................................................................................................................................. 319
9.8 Nomenclature .......................................................................................................................................... 321
References ................................................................................................................................................................ 322

287
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 288 Tuesday, March 22, 2005 7:36 AM

288 The John Zink Combustion Handbook

9.1 OVERVIEW TABLE 9.1 Current CFD Applications in the Chemical

Process Industry
If you want to win big, then you have to define the * Rotary kiln incinerator * Gas scrubber towers
rules of the game. * Drying ovens * Thermal oxidation
* Packed catalysis beds * Ventilation of foam scrap bin
— M. Stahlman1 * Crystallization * Polymer extrusion
* Mixing tanks (liquid/liquid, etc.) * Dust separation systems
This concept has been used by companies and countries to * Impeller design * Caustic evaporators
control markets and economies. Changes to the rules are gen- * Ceramic production * Retention basin flow
* Membrane flow * Liquid migration in diapers
erally caused by significant technological advances. The * Caustic degasser inlet * Nozzle design
invention of the telescope, the microscope, and the atom
smasher are all examples of such paradigm shifting techno-
logical advances. Development of, and increased access to,
supercomputers represents a significant advance that has communicate the power and potential of a cutting-edge
opened new scientific frontiers and again changed the rules. 21st century technology. In writing about CFD it is difficult
not to give the appearance of “wall to wall” equations. This
Several years ago, Nobel laureate and physicist Kenneth
is because, as the name implies, it is a computationally inten-
Wilson identified a collection of “grand challenges” to be
sive science. In this chapter we limit the discussions to out-
addressed by researchers using supercomputers.2 These chal-
lining the major topics, but even this necessitates providing
lenges included designing efficient aircraft, simulating semi-
several equations.
conductor materials, rationally designing drugs, understanding
catalytic phenomena, studying air pollution, petroleum First, an overview of what CFD is presented, followed by
exploration, and analyzing fuel combustion. The chemical pro- a discussion of CFD software, with comments on the under-
cess industry (CPI) as a whole faces several of these “Grand lying foundations and assumptions. Next, the subprocesses
Challenges.” Future success of businesses in the process indus- involved in combustion simulation will be explained. In addi-
try will depend on making advances in these “Grand Chal- tion, the advantages and limitations of CFD in simulating
lenges.” fluid flows as well as combustion are also discussed. The
Computational fluid dynamics (CFD) has been used in the latter part of the chapter is dedicated to case studies that will
CPI to solve various types of flow problems.3,4 Krawczyk5 illustrate the problem-solving power of CFD. Finally, conclu-
has reviewed the applications and limitations of CFD mod- sions and a summary of issues that are important to the
eling in the hydrocarbon processing industry (HPI). Accord- process industry regarding the future use of CFD as a process
ing to Krawczyk, the use of CFD in the HPI is relatively improvement tool will be presented.
unknown because of a general lack of awareness of what CFD
is and where it can be applied.
Past work by researchers in the chemical process industry6 9.2 INTRODUCTION
and the electric power industry7 demonstrates the role that Knowledge in the field of fluid dynamics has evolved in
CFD can play in addressing important process questions three areas. First, experimental fluid dynamics emerged in
(see Table 9.1). Specific work discussed in this chapter further France and England in the seventeenth century. Next, in the
illustrates how CFD has been used to investigate and solve eighteenth and nineteenth centuries, theoretical fluid dynamics
important technical issues related to combustion processes. developed. For most of the twentieth century, fluid dynamics
The cases described here are a small sampling of the many was practiced in these two realms. The advent of the high
CFD simulations performed at the John Zink Company. The speed digital computer combined with the development of
examples in this chapter deal with petrochemical process accurate numerical algorithms for solving physical problems
furnace optimization, venturi eductor design, sulfur recovery has added a third approach to the study of fluid dynamics.
furnace simulation, and chlorinated hydrocarbon incineration. This is computational fluid dynamics, or CFD. Today,
As with any written work, it is a challenge to attempt to computational fluid dynamics is an equal partner to the
address a wide audience. Such is certainly the case for this experimental and theoretical approaches. CFD synergisti-
chapter. In order to stay within the scope of this book, we have cally complements the other two approaches nicely, but it
written this chapter to speak to that vital component of today’s will not replace them.64
high-tech society, the engineer working in industry. The pur- CFD is based on the fundamental governing equations of
pose of this chapter is to give the reader a concise introduction fluid dynamics — the continuity, momentum, and energy
to both the science and application of CFD and to thereby
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- equations. These equations are physics. So it is worthwhile
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 289 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 289

to note at this point that much of CFD is based on fundamental tools developed by the intermediate user to perform calcula-
physics and not on empirical functions. Therein lies CFD’s tions but does not add to the software.
power to extend to solving new flow problems. Simply stated, These three levels of users may also be identified in CFD.
the fundamental physical principles underlying CFD, and all The development engineer is highly specialized and must
of fluid dynamics, are as follows: have several years of education in fluid dynamics and CFD.
One may call this first level the CFD specialist. Next, the
1. Mass conservation
intermediate user in CFD not only uses the software, but has
2. Newton’s second law: F = ma
to have a sound understanding of CFD to extend its capabil-
3. Energy conservation
ities and critically evaluate the results it is providing. This
These fundamental physical principles can be expressed in second level may be termed the CFD engineer. The third level
terms of mathematical equations, which in their most general is the end-user, really the customer requiring the CFD. The
form are either integral or partial differential equations. CFD customer knows enough about the capabilities of the technol-
is the art of replacing the integrals or the partial derivatives ogy to be able to identify when a problem may be a candidate
in these equations with discretized algebraic forms. These for CFD and is able to consult with the CFD engineer about
discrete equations are solved to obtain magnitudes for the the technical details as well as the benefits and costs. In
different variables of interest (pressure, velocity, temperature, dealing with combustion simulation, the added complexity
etc.) in the flow field at discrete points in time and space. requires the CFD engineer (intermediate level in the example
As the following sections will indicate, the equations are above) to have specialized knowledge to set up complex
complex because they have to simultaneously deal with sev- three-dimensional CFD simulations and interpret the results.
eral interrelated transport mechanisms in great detail. Need- As such, the skill set required for CFD combustion modeling
less to say, they are also difficult to solve. Leading scientists leaves little or no distinction between a CFD specialist and a
have spent the last four decades developing models and CFD engineer.
approximations in order to quantify some of the terms in the CFD first evolved, and still is strongest, in its ability to

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
equations that cannot be solved for directly. These models solve non-reacting flows. A wide variety of practical engi-
use a combination of fundamental physics, empirical func- neering problems can be solved by analyzing non-reacting
tions and proven approximations to both reduce computa- flows. For example, the ability to predict the behavior of fluid
tional intensity and simplify the task at hand. One such model flows through various flow passage configurations and con-
is the well-known k-ε turbulence model. The k-ε model is ditions has promoted the rapid progress in the automotive and
discussed in further detail later. Figure 9.1 is a graphical aircraft industries. However, in more recent decades, model
representation of the various elements that constitute CFD. development work in CFD has naturally extended into react-
The CFD solution results in a collection of numbers that ing flows because of the undeniable need.
describe the flow field quantitatively. This matrix of values Combustion is a complex reacting flow. Due to the inher-
can then be queried for the values of interest, at the locations ent complexities of combustion and the relatively recent
of interest, or, more commonly, represented graphically in development of the combustion models, there are some
contour or vector plots using color scales. limitations in the current state-of-the-art of CFD combus-
At this point the reader may have acquired the impression tion modeling. These limitations are discussed later. Limi-
that because of its complexity, CFD is the realm of pure tations and youth notwithstanding, currently available
research and academics. It is true that this was once the case, combustion modeling tools can provide tremendous, real-
but not any longer. The following analogy is given in an world problem-solving power.
attempt to illustrate the current status of commercial imple- The main intent of this chapter is to describe combustion
mentation of CFD. modeling, and so, for brevity, the discussions of basic CFD
Let us consider the analogy of a piece of software, say, a flow modeling have been kept to a minimum. However, it must
spreadsheet software package. One may say that there are three be kept in mind that reacting flow modeling is merely a super-
levels of engineers involved with this spreadsheet software. set of basic non-reacting flow modeling. In performing a CFD
First is the developer who develops the algorithms, software simulation of a reacting flow, the computations pertaining to
architecture, and basic code. Second is the intermediate user reaction induced changes to fluid properties and composition
who uses the basic package to build sophisticated calculations are performed in conjunction with the same calculations that
for his or her specific needs, and in addition, can add modules would be performed for any non-reacting flow simulation.
of custom code or macros to the basic package to extend its Consequently, reacting flow modeling is considerably more
abilities. Third would be the end-user who uses the spreadsheet resource intensive and challenging. In many cases it is possible
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 290 Tuesday, March 22, 2005 7:36 AM

290 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 9.1 Elements in CFD modeling.

to use non-reacting flow simulation results to infer the reacting The reader will find that the emphasis of this chapter is
flow result. Therefore, frequently, in planning the solution gas-fired furnaces, with only brief discussion of oil-fired
approach, the CFD engineer must make a value vs. effort furnaces. This bias is reflective of the bias in the open liter-
decision to choose between full combustion modeling and ature where little discussion of modeling full-scale oil-fired
non-reacting flow modeling. furnaces can be found. Petrochemical furnaces are typically
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 291 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 291

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
gas- or oil-fired. In the United States, most petrochemical “stretched out” by fuel staging strategies, it becomes more
furnaces are gas-fired, but in other parts of the world, various and more difficult to control the mixing between the fuel, air,
grades of oil are used. The examples discussed here rely on and furnace gases (products of combustion). In addition, the
the ability to accurately simulate turbulent reacting flow with stability of these ultra low NOx burner flames is problematic.
radiant heat transport in an enclosed space. With regard to full scale single burner development, it should
The reader will quickly observe that industrial-scale com- be kept in mind that the dominant flow currents in an operating
bustion modeling is an immature field (even after years of environment may be very different from those in a single
scientific effort), and may wonder why industries invest in burner furnace where the burner was developed. These oper-
such efforts. The simple answer is that there is no better ational problems, which otherwise could only be tackled by
alternative. In the past, experimentalists have tried to simulate trial and error, can be systematically analyzed and solved
the flow patterns inside industrial furnaces using scaled-down using CFD.
plexiglas representations of the geometry. Although the com- Herein lies ample justification for the petrochemical indus-
plex physics of the real furnace (with combustion, radiation, try to pursue CFD technology to better understand the per-
buoyancy, etc.) are completely neglected, these experimen- formance of furnaces. Hopefully, this chapter will illustrate
talists have met with some success by being able to relate some of the potential value of CFD technology and stimulate
cold flow phenomena observed in the plexiglas model to the further studies to better understand and extend the perfor-
combustion phenomena in a furnace. However, this technique mance of CFD models in these applications.
has its limitations because a good portion of the expected
result is inferred and not based on detailed knowledge of the 9.2.1 CFD Model Background
actual physics and chemistry. On the other hand, even with
In the work described below, two CFD software packages
its numerous approximations about turbulence, radiation, and
have been used, Fluent and PCGC3. Both codes have evolved
chemical reactions, a CFD model provides much more
from similar backgrounds, having been based on the SIM-
detailed information about the furnace process than a nonre-
PLE (Semi-Implicit Method for Pressure Linked Equations)
acting, scaled-down plexiglas model.
algorithm. SIMPLE solves a set of nonlinear coupled partial
Other possible options available to the engineer include
differential equations describing the conservation of mass,
either scaled reacting flow or full scale reacting flow experi-
momentum, and energy, as described in more detail below.
mentation. Full scale research is very expensive because
Originally, Fluent was designed to simulate basic fluid
industrial, multiple burner furnaces cost millions of dollars,
dynamics, while PCGC3 was designed to model pulverized
and reacting flow experimentation on a scaled-down model
coal combustion systems. More recently, both codes have
has its share of disadvantages. Scaling combustion systems
been extended to address general combustion phenomena
from the laboratory scale to the industrial scale is very diffi-
occurring in several diverse systems (e.g., pulverized coal
cult. No change scaling is well-understood theoretically. To
combustor, hazardous waste incinerators, petrochemical pro-
have “similarity” as a combustion system is scaled, the Rey-
cess heaters, etc.). To illustrate the basis from which CFD
nolds number and Damköhler numbers must remain
codes are derived and to illustrate the capabilities and limita-
unchanged. This is practically impossible, so some sort of
tions, a brief review of the CFD code PCGC3 is given.
incomplete scaling is used. Typically, combustion systems are
scaled using a constant-velocity or constant-momentum flux
method which does not provide predictable results in all cases. 9.2.2 The CFD Simulation Model
Development of burners for petrochemical applications is Fluent is based on the original work done by Swithenbank
usually done experimentally using a single burner at full scale. and co-workers at the University of Sheffield.65 For more
This method frequently produces a burner that performs well than fifteen years, this code has been developed and extended
in the industrial setting; however, multiple burner firing typ- by engineers at Fluent, Inc. PCGC3 was developed in the
ically produces higher NOx emissions than single burner Advanced Combustion Engineering Research Center (ACERC)
firing.8 Other flame interaction problems can arise as well. at Brigham Young University in the same time frame by vari-
For example, flames from individual burners are occasionally ous researchers.7,9–18 PCGC3 describes a variety of reactive
observed to merge together, causing flame length to increase and nonreactive flow systems, including turbulent combus-
significantly when burner spacing is not sufficient. These tion and gasification of pulverized coal. To illustrate the
problems are especially significant for ultra-low-NOx burners model capabilities and limitations, a limited discussion of
because these burners typically produce flame lengths signif- PCGC3 will be presented. A more detailed description is
icantly longer than a conventional burner. As the flame is given by Smoot and Smith.19
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 292 Tuesday, March 22, 2005 7:36 AM

292 The John Zink Combustion Handbook

9.2.2.1 Transport Equations first decomposing the variable into a mean and a fluctuating
Nonreactive turbulent gas flow is modeled using the steady- term and then time-averaging the resulting equation set:
state form of the Navier-Stokes equations by assuming a con-
tinuous flow field described locally by the general conserva- Φ=Φ
˜ + φ ′′ (9.7)
tion of mass and momentum:20
where Φ̃ = ρΦ ρ and ρφ′′ = 0, but φ ′′ ≠ 0. Applying this
Dρ v decomposition to all variables except density and pressure, the
= −ρ(∇ ⋅ u ) (9.1) conservation equations [Eqs. (9.3) through (9.6)] are trans-
Dt
formed into the mass-averaged or Favre-averaged transport
r equations:
Du
ρ = −∇p − ∇ ⋅ τ + ρg (9.2)
Dt
( )=0
∂ ρ ũ j
(9.8)
where D/Dt is the total (or substantial) derivative. These ∂x j
equations can be simplified by assuming steady-state flow of
a Newtonian fluid. If the fluid is assumed to be incompress-
r
ible, the dilatation (∇ ⋅ v ) can also be used to further simplify (
∂ ρ u˜ i u˜ j ) = − ∂P +
the equations. An incompressible assumption implies that the ∂x j ∂x i
∂
(
τ − ρui′′u ′′j + ρ fi
∂x j ij
) (9.9)
code is applicable to low Mach number flows (i.e., Mach
number < ~ 0.3). Given these simplifying assumptions, the
general conservation or transport equations for mass and (
∂ ρ Φu j )= ∂  ∂Φ
Γφ

− ρφ ′′u ′′j  + SΦ
momentum can be written in Cartesian tensor form as: ∂x j 
∂x j  ∂x j
(9.10)


( )=0
∂ ρu j
(9.3)
These constitute the turbulent transport equation set for
nonreacting flow. However, as a result of the averaging pro-
∂x j
cedure, several additional variables called Favre stresses
( ρui′′u ′′j ) and Favre fluxes ( ρφ′′u ′′j ) have been introduced.
(
∂ ρui u j ) = − ∂P + ∂τ ij
+ ρfi (9.4)
These stresses and fluxes represent the mean-momentum
transport and the mean-scalar transport by turbulent diffusion.
∂x j ∂xi ∂x j
Additional ancillary equations are required to solve for these
new turbulent transport variables. These extra equations make
 ∂u ∂u j   ∂u up the “turbulence” model.
+ µ B − µ k δ ij
2
τ ij = µ i +  (9.5)
 ∂x j ∂x i   3  ∂x k
9.2.2.2 Turbulence Equations
Although several turbulence models have been proposed
Similarly, a transport equation can be written for a con- (Nallasamy21), the k-ε turbulence model, originally proposed by
served scalar, Φ, as: Harlow and Nakayama,22 remains the most widely used model to
describe practical flow systems (Speziale23). The k-ε turbulence
(
∂ ρΦu j )= ∂  ∂Φ  model employs a modified version of the Boussinesq hypothesis

∂x j ∂x j  Γφ ∂x  + SΦ (9.6) (Smoot and Smith19):

 j

 ∂u˜ ∂u˜ j  2 r
Turbulent transport is characterized by both time and length −ui′′u ′′j = − vt  i +  − (vt ∇ ⋅ v + k )δ ij (9.11)
scales, the smallest scales being too small to numerically  ∂x j ∂xi  3
resolve for practical problems. Therefore, Eqs. (9.3) through
(9.6) are not solved directly, but are transformed by one of where vt is known as the eddy diffusivity or turbulent viscos-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

two averaging techniques: Reynolds averaging (most common ity. This approach allows the molecular viscosity to be
operation applicable to incompressible flows) or Favre aver- replaced with the eddy diffusivity, which allows the instanta-
aging (mass-weighted averaging applicable to compressible neous transport equations [Eqs. (9.3) through (9.6)] to be
flows). Favre or mass-weighted averaging is accomplished by modeled using the mean-value equations [Eqs. (9.8) through
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 293 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 293

(9.10)]. A disadvantage to this approach is the need to assume TABLE 9.2 Universal “Empirical” Constants Used in k-ε
isotropic eddy diffusivity. However, given this assumption and Turbulence Model
the specific velocity and length scales u′ and l′: Constant: Cµ C1 C2 σk σe κ
Value: 0.09 1.44 1.92 0.9 1.22 0.42

k3 2
u′ ≈ k l ′ ≈ Cµ (9.12)
ε
reduce computer storage and runtimes the k-ε turbulence
where the turbulent kinetic energy can be defined as: model uses wall functions instead. The Van Driest hypothesis
on turbulent flow near walls is used to derive wall functions

( )
1 1 consistent with the logarithmic law of the wall.24 These func-
k= u u = u u + u2 u2 + u3u3 (9.13)
2 i i 2 1 1 tions relate the dependent variables near the wall to those in
the bulk flow field. Boundary conditions used in the turbu-
With these definitions, additional transport equations for lence model are discussed at length by Gillis and Smith.18
the turbulent kinetic energy and the dissipation rate of turbu-
Given the turbulence model with the necessary boundary
lent kinetic energy, ε, can be written:
conditions, the full equation set can be written (see Tables 9.3
and 9.4). As shown, each equation is conveniently cast into
∂k 1 v  a general convection-diffusion form with the off-terms col-
+ u • ∇k = ∇ •  t ∇k  + G − ε
∂t ρ  σk  lected on the right-hand side, and the specific terms that
(9.14) depend on the coordinate system selected. Examining the
 3 3
 ∂ui ∂u j   θ-momentum equation (see Table 9.4) helps illustrate the
∑∑
v
G= t 2 ∇ 2 u +  ∂x + ∂x  
σk   j i 
meaning of each term:
i =1,i ≠ j j =1 

∂ε r v  ε  ε2  ∂(ρ˜ uw
˜ ˜) ∂(rρ˜ vw
˜ ˜) ∂(ρ˜ ww
˜ ˜)
+ u • ∇ε = ∇ •  t ∇ε + f1c1G  − f2 c2   (9.15)
1 r + +
∂t ρ  σε   k  k ∂x ∂r ∂θ
∂  ∂w˜  ∂  ∂w˜  ∂  µ e ∂w˜ 
and the eddy diffusivity is defined as: −r  µe  −  rµ e −  =
∂x  ∂x  ∂r  ∂r  ∂θ  r ∂θ 

fµ cµ k 2 ∂p ∂  µ ∂u˜  ∂  ∂v˜ 
vt = − +r  e  + µ − µ e w˜  +
ε
(9.16) ∂θ ∂x  r ∂θ  ∂r  e ∂θ 

Several key “empirical” constants are required by the k-ε ∂  µ e   ∂w˜ ∂v˜  
  + 2r   +
turbulence model. The values used by PCGC3 are shown in ∂θ  r   ∂θ ∂r  
Table 9.2.
 ∂w˜ 1 ∂v˜ w˜  ˜
The values shown in Table 9.2 were determined through µe  + −  − ρvw
˜ ˜ + rρ˜ gθ (9.17)
 ∂r r ∂θ r 
comparison of turbulence and numerical optimization as dis-
cussed by Sloan et al.12 These values are similar to those
originally proposed by Launder and Spalding,24 but differ The first three terms of Eq. (9.17) represent the net rate of
slightly from those reported by other researchers (Nallasamy,21 momentum addition to a volume element by convection from
Lilleheie et al.,25 Jones and Whitelaw,26). This may be due to the three direction components. The fourth, fifth, and sixth
the fact that these “empirical” constants are based on simple terms represent the corresponding diffusion terms. When the
two-dimensional flows and adjustment may be required to turbulence model solves for the individual Reynolds stresses,
simulate more complex flows. Regardless, this fact and the the diffusion terms do not strictly represent molecular diffu-
other simplifying assumptions suggest that the flow results sion, but also include momentum contributions due to the
be closely scrutinized when applying any CFD code using turbulent motion of the fluid. The first term on the right-hand
this turbulence model to simulate complex flow systems. side (RHS) of Eq. (9.17) represents the pressure force on the
Application of the k-ε turbulence model requires boundary volume element. All other terms on the RHS of the equation
conditions for both k and ε. Boundary layer theory could be represent either a source or sink term for momentum (e.g.,
used to derive the equations for flow near the wall, but to gravity force, centripetal forces, etc.).
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

294 The John Zink Combustion Handbook

TABLE 9.3 Cartesian Differential Equation Set

∂( ρ u˜ φ) ∂( ρ v˜ φ) ∂( ρ w˜ φ) ∂  ∂(φ)  ∂  ∂(φ)  ∂  ∂(φ) 
+ + − Γ − Γ − Γ = Sφ
∂x ∂y ∂z ∂x  φ ∂x  ∂y  φ ∂y  ∂z  φ ∂z 
Equation φ Γφ Sφ
Continuity 1 0 0
∂p ∂  ∂u˜  ∂  ∂v˜  ∂  ∂w˜  2 ˜
X-momentum ũ µe − + µ + µ  + µ  + ρgx − ρk
∂x ∂x  e ∂x  ∂y  e ∂x  ∂z  e ∂x  3
∂p ∂  ∂u˜  ∂  ∂v˜  ∂  ∂w˜  2 ˜
Y-momentum ṽ µe − + µ + µ  + µ  + ρgy − ρk
∂y ∂x  e ∂y  ∂y  e ∂y  ∂z  e ∂y  3
∂p ∂  ∂u˜  ∂  ∂v˜  ∂  ∂w˜  2 ˜
Z-momentum w̃ µe − + µ + µ  + µ  + ρ gz − ρ k
∂z ∂x  e ∂z  ∂y  e ∂z  ∂z  e ∂z  3
µe
Mixture fraction f̃ 0
σf

µe Cg1µ e  ˜  2  ˜  2  ˜  2 
∂f ∂f ∂f ε˜
Mixture fraction variance g̃ − +   +   +    − Cg 2 ρ g˜
σg σg  ∂x   ∂y   ∂z   k˜
 
µe
Turbulent energy k̃ G − ρε̃
σk
µe  ε˜ 
Dissipation rate ε̃
σε k
(
 ˜  c1 G − c2 ρε˜ )
  ∂u˜  2  ∂v˜  2  ∂w˜  2   ∂u˜ ∂v˜  2  ∂u˜ ∂w˜  2  ∂v˜ ∂w˜  2 
 
Note: G = µ e 2   +   +   + +  + +  + +  
  ∂x   ∂y   ∂z    ∂y ∂x 

 ∂x ∂x   ∂x ∂y  


TABLE 9.4 Cylindrical Differential Equation Set

∂( ρ u˜ φ) ∂(r ρ v˜ φ) ∂( ρ w˜ φ) ∂  ∂(φ)  ∂  ∂(φ)  ∂  Γφ ∂(φ) 
r + + −r Γ − rΓ − = Sφ
∂x ∂r ∂θ ∂x  φ ∂x  ∂r  φ ∂r  ∂θ  r ∂θ 
Equation φ Γφ Sφ
Continuity 1 0 0
∂p ∂  ∂u˜  ∂  ∂v˜  ∂  ∂w˜  2 ˜
X-momentum ũ µe −r +r µ  +  rµ + µ  + rρgx − rρk
∂x ∂x  e ∂x  ∂r  e ∂x  ∂θ  e ∂x  3
∂p ∂  ∂u˜  ∂  ∂v˜  ∂  ∂w˜ w˜  2µ e ∂w˜ 2v˜ µ e 2
R-momentum ṽ µe −r +r µ  +  rµ + µ − − − + ρ w˜ 2 + r ρ gr − ρ k˜
∂r ∂x  e ∂r  ∂r  e ∂r  ∂θ  e ∂r r r ∂θ r 3
∂ρ ∂  µ ∂u˜  ∂  ∂v˜ 
− +r  e  + µ − µ e w˜ 
∂θ ∂x  r 2θ  ∂r  e ∂θ 
θ-momentum w̃ µe
∂  µ e   ∂w˜   ∂w˜ 1 ∂w˜ w˜  2
− ρ vw
˜˜ +   − 2v˜   + µ e  + −  + r ρ gθ − r ρ k̃
∂θ  r   ∂θ   ∂r r ∂θ r  3
µe
Mixture fraction f̃ 0
σf

 ˜  2  ˜  2  2

µe Cg1µ e r ∂f ∂f 1 ∂f˜   ε˜
Mixture fraction variance g̃ − +   +   +   − Cg 2 r ρ g˜
σg σg  ∂x  ∂
 
r  r ∂θ   k˜
 
µe
Turbulent energy k̃ σk
(
r G − ρε̃ )
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

µe  ε˜ 
Dissipation rate ε̃
σε  k
(
 r ˜  c1 G − c2 ρε˜ )
  ∂u˜  2  ∂v˜  2  1 ∂w˜ v˜  2   ∂u˜ ∂v˜  2  1 ∂u˜ ∂w˜  2  1 ∂v˜ ∂w˜ w˜  
2

Note: G = µ e 2   +   +  +  + +  + +  + + − 2 
 ∂x   ∂r   r ∂θ r    ∂r ∂x   r ∂θ ∂x   r ∂θ ∂r r  
   
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 295 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 295

9.2.2.3 Solution Technique residual or relative error in the equation. As the solution con-
The equation set shown above is composed of several steady- verges, the residual is forced to zero and the convergence cri-
state, second-order, nonlinear, elliptic partial differential teria is satisfied. Comparison of errors from each of the seven
equations (PDEs). Each of these continuous PDEs is trans- equations is difficult because of the relative magnitude of the
formed into a discrete finite difference equation (FDEs). coefficients (Ai) for each equation. Normalization is also dif-
Examining the equation set, seven equations (mass (1), ficult due to the range of variable and source term magni-
momentum (3), turbulence (2), and conserved scalar (1)) tudes within each equation. Without comparison of the
with six unknowns (P, u, v, w, k, ε, and f) describe the turbu- convergence of each equation, it is impossible to determine
lent flow system. Typically, the momentum equations are when “overall” convergence is achieved or which equation is
solved for each velocity component, the turbulence equations slowing the convergence process.
are solved for the respective turbulence variables, and the PCGC3 uses the largest term found in each variable’s FDE
continuity equation is left for the pressure field. A key issue to normalize the respective residual. This truncation term,
in CFD is solving for the pressure gradient source terms, defined as:
found in the momentum equations, because the pressure
fields for enclosed flows are usually unknown. PCGC3 uses ψ p = Ap φ p (9.19)
variations of the SIMPLE (Semi-Implicit Method for Pres-
sure Linked Equations) algorithm to solve the equations of is guaranteed to exceed the magnitude of the other terms in
motion and continuity in a decoupled fashion, by transform- the FDE because of the requirement of diagonal dominance
ing the continuity equation into a pressure correction equa-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

in the coefficient matrix (satisfied by the TDMA algorithm)

tion. A detailed description of the various forms of SIMPLE and because of the manipulation of the source terms (RHS
found in PCGC3 is given by MacArthur.27 of equation). The final normalized equation error is thus
In PCGC3, all equations are cast into one standard form calculated as:
so that a single solver is required. PCGC3 uses the tri-diag-
onal algorithm to solve the FDEs for each variable along a
∑
N
line on each plane of the computational space. The variables Rφo
ψ n =1
R = (9.20)
∑
are solved for in succession, starting with a velocity compo- φ N

nent and ending with one of the turbulence variables. Because ψφ

n =1
the FDEs are solved in a decoupled fashion, only four to five
“micro-iterations” are required per variable. A complete cycle This normalization allows for the comparison of equation
through the equation set, termed a macro-iteration, resolves error from the different equations and measures the close-
the nonlinear coupling between equations to a prespecified ness of computer round-off error to equation error. The total
convergence criterion. Overall convergence typically requires equation error, Rφψ, ranges from approximately 1 to 10–nd,
between 200 and 1000 macro-iterations. where nd is the number of digits of computer accuracy.
When the equation error reaches 10–nd, computer round-off
9.2.2.4 Convergence Criteria error prevents further reductions in equation error. Having
Although PCGC3 iterates on each equation individually, the determined the machine accuracy, PCGC3 determines and
equation coupling necessitates simultaneous convergence of prints out the difference between this and the number of
the entire equation set. Various methods have been used to digits of computer accuracy, nd, for each equation and uses
measure convergence, compare convergence rates of each the term representing the greatest inaccuracy to judge when
equation, and determine when the required level of conver- a calculation is converged.
gence is obtained. Typically, the error used to track conver-
gence represents the residual for each FDE as shown: 9.2.2.5 Model Validation
The PCGC software has been extensively validated by direct
R = AE φ E + AW φ W + AN φ N + AS φ S + AT φ T
0
φ
comparison between model predictions and experimental data
for several systems.14 Fletcher 9 shows a comparison between
+ AB φ B + SU − AP φ P (9.18) mean velocity values obtained using laser Doppler velocime-
try (LDV) techniques and predictions. This comparison shows
where Apφp represents the computational node and the other reasonable agreement for a hydrogen/air diffusion flame for
Aiφi represent the neighboring nodes, Su represents the source two different flow rates. Similarly, measured values of gas-
term (RHS of equations shown above), and Rφ represents the phase concentrations of H2, H2O, and O2 agree well with
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 296 Tuesday, March 22, 2005 7:36 AM

296 The John Zink Combustion Handbook

prediction. Measured RMS velocities show some disagree- to mixing effects. Conversely, the second limit, referred to as
ment with predicted values. This discrepancy illustrates the the “fast-chemistry” limit occurs when:
nonisotropic nature of the fluid mechanics, which are
assumed isotropic in the turbulence model. However, the Da → ∞ (9.23)
overall agreement between predictions and measurements
demonstrates the applicability of the PCGC approach to simu- In this case, the reaction time is very short (fast reactions)
lating high-temperature reacting flow systems such as com- relative to the mixing time. Many diffusion flames are
bustion. approximated well by the latter limit.
Given the large disparity between short reaction time scales
9.2.2.6 Modeling Basis and long mixing time scales, chemical activity may be con-
The simulation results of reacting flow systems using PCGC fined to an infinitesimally thin layer, commonly referred to
software are based on two main assumptions: as a “flamelet” or “flamesheet.” 29 This assumption allows
1. uniform heat loss from the reactor (no heat loss indicates flame chemistry to be approximated using local thermody-
adiabatic operation) namic equilibrium without significant error. Well-known
2. local gas-phase chemistry is micro-mixing-limited and com- exceptions include NOx and CO chemistry where the reaction
position is determined using thermodynamic equilibrium. time scale and the mixing time scale are of similar magnitude:
The first assumption, based on the Crocco similarity,7 is
valid for cases where thermal diffusion and mass diffusion are Da ≈ O(1) (9.24)
equal. Some knowledge relative to the total reactor heat load
and the related heat loss is required. The user must specify Here, finite-rate chemistry must be coupled with the tur-
reactor heat loss as a fraction of the total energy in the system. bulent fluid mechanics calculations. Because turbulent effects
This fraction is then extracted equally from each discrete com- must be included in the kinetic scheme, global mechanisms
putational cell in the overall simulation. Thus, this assumption are generally used to avoid solving individual transport equa-
neglects the effect of the temperature gradient that can produce tions for each specie in a detailed kinetics mechanism.19
artificially low local temperature regions while yielding appro-
priate exit temperatures.
The second assumption, generally valid for high-temper- 9.3 CFD-BASED COMBUSTION
ature combustion chemistry, suggests that the homogeneous SUBMODELS
kinetics are sufficiently fast so that gas mixing is controlling. The current state-of-the-art approaches in modeling petro-
This is commonly referred to as the “mixed-is-burnt” chemical furnaces are described in this section. This is not a
assumption. If this assumption applies, the local chemistry comprehensive discussion of all of the available models and
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

can be calculated from general thermodynamic calculations. algorithms available to the CFD analyst because not all of the
The validity of this assumption is determined by the relative models are commonly used in modeling furnaces. While a
reaction and mixing (convective/diffusive) time scales, as large number of algorithms and models are studied in aca-
expressed by the Damköhler number, Da: demic circles, the majority of CFD analyses of large-scale fur-
naces are done with commercial CFD packages. There are
t flow lt v ′ currently several commercial CFD packages with very similar
Da = = (9.21)
trxn lF SL modeling capabilities. This section discusses the modeling
approaches used by these codes, as well as several models
where sL is the burning velocity, lF is the reaction zone thick- currently being studied at the research level. These “under
ness, v′ is the turbulence intensity, and lt is the turbulence development” models will lead to improved combustion
length scale. Using this relationship, two physical limits have modeling capability in the future for industrial users.
been identified.28 The first, referred to as the “frozen” limit There are a large number of approximations involved in mod-
occurs when: eling combustion processes in furnaces. Even in a gas-fired
furnace, the multitude of important physics is daunting. The
Da → 0 (9.22) flow in the furnaces is turbulent flow with a very large integral
length scale (the characteristic dimension of the furnace). The
In this case, the reaction time (trxn) is much larger than the combustion chemistry in the furnace involves tens to hundreds
flow time (tflow), and kinetic effects are negligible compared of chemical species reacting with time scales from less than a
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 297 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 297

microsecond to several seconds. Radiative transport from a non- fiable. Pope32 notes that the appeal of LES in nonreacting
gray gas (the products of combustion) to the furnace walls and flows is the expectation that the small scales of turbulence are
tubes (with the process fluid flowing inside) whose emissivity universally related to the large scales. In a reacting flow, there
is temperature dependent is the primary mode of heat transfer. is no similar expectation. LES does have the advantage of
The interaction between these physical processes is of consid- resolving the large-scale structures that challenge the Favre-
erable importance. The turbulence/chemistry interaction has averaged models, but many difficult problems remain to be
been well-studied for many years, particularly for non-premixed addressed before LES will be a useful tool. Bray33 notes that
systems. More recently, the interaction between turbulence and despite the fact that LES faces current difficulties, it will be
radiative emission from a non-gray gas has been studied.30 successfully developed and will be a useful tool for the com-
The geometries of the burners used in these furnaces are bustion modeler. DNS can be categorically neglected for this
becoming increasingly complex. The dominant driver in most class of problems because the computational demands are far
burner designs is NOx reduction, and this leads to burners in excess of current computational resources.33,34
that are more and more geometrically complex. Added to the
geometric complexity is the chemical complexity. The pre- 9.3.2 Radiation Models
ferred strategy for reducing NOx emissions from gas-fired Typical petrochemical furnaces consist of a radiant section
burners is to use staged fuel systems and to use the fuel jets and a convection section. These regions are so named
to entrain large amounts of the products of combustion into because of the dominant mode of heat transfer. In the radiant
the flame zone. This means that to make a burner with lower section, refractory surface temperatures can be higher than
NOx emissions, one has to make a turbulent flame that is 2200°F (1200°C). Radiant heat is incident on the process
comparatively less stable. For the CFD analyst, modeling the tubes, both from the high-temperature surfaces and directly
stability of a flame or its lift-off height is at present a very from the flame. Accurate modeling of the heat delivered to
imposing problem. For this reason, CFD predictions of NOx the process fluid requires an accurate prediction of the radi-
emissions from ultra-low-NOx burners are typically poor. ant intensity inside the furnace. In addition, accurate predic-
However, even if the quantitative NOx predictions are poor, tion of radiation from the flame is necessary to accurately
the qualitative information from a CFD calculation can often predict emissions. For example, Barlow35 notes that the dif-
be very useful. For example, the results of a CFD analysis ferent radiation models can affect NOx predictions just as
can be used to study the entrainment of cooled furnace gases much as the different turbulence/chemistry interaction mod-
by fuel jets as well as the mixing of the fuel jets with com- els that were evaluated by him.
bustion air. When used by an analyst who has a thorough Thermal radiation transport presents a difficult problem
knowledge of what real flames do, these quantitative mea- because of the number of independent variables. The radiation
sures, in conjunction with experimentation, can be used to transport equation (RTE) describes radiation transport in
solve equipment problems. As an example, McDermott and absorbing, emitting, and scattering media. The equation is:36
Henneke31 used an axisymmetric CFD model to design turn-
ing vanes in a premixed burner. The problem being addressed ∂Iλ σ r r r
was flashback, which is very difficult to model in a CFD
study. However, by combining CFD analysis with knowledge
∂s
= κ λ Ibλ − β λ Iλ + sλ
4π ∫ I (s )Φ(s , s)dΩ
4π
η i i i (9.25)

of what conditions allow flashback, the authors were able to

design a series of turning vanes that eliminated flashback for Radiant intensity is a function of location (three coordinates
a wide range of operating conditions. in a three-dimensional problem), direction (two angular inde-
pendent variables), and wavelength (one independent variable
varying from 0 to infinity). This means that the problem of
9.3.1 Solution Algorithms radiative transport is a six-dimensional problem. A common
This section discusses the algorithms used to solve the Reynolds- approach is to remove the wavelength dependence by making
(or Favre-) averaged Navier-Stokes equations, the Reynolds- gray media approximations discussed below. The angular
(or Favre-) averaged energy and species equations, and the dependence can be treated by angular discretization, such as
radiative transfer equation. No discussion is given here of in the P-1 method or the discrete ordinates method. Methods
large eddy simulations (LES) or direct numerical simulations such as the finite-difference and finite-volume methods dis-
(DNS). LES may become a viable option for CFD modeling cussed above treat the spatial dependence. Both the P-1 and
of furnaces in the near future, but at present the increase in discrete ordinates methods approximate the angular depen-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

computational cost of an LES calculation is usually not justi- dence of the equation of transfer. The Monte Carlo method
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 298 Tuesday, March 22, 2005 7:36 AM

298 The John Zink Combustion Handbook

takes a much different approach. In the Monte Carlo method, as a yellow-red brightness in the flame. C2 hydrocarbons and
individual photons of radiant energy are emitted, reflected, higher have more tendency to soot, while methane does not
and absorbed by both opaque surfaces and participating media normally produce a sooty flame. Soot has a strong impact
using ray tracing algorithms. This method provides a very ele- on flame radiation. Emission from soot in flames is fre-
gant approach to treating non-gray radiation as well as the quently much larger than the gas radiation emitted by the
directional dependence of radiation. Its use is limited by its flame.36 In some applications (oil-firing, in particular), soot
computational cost. emissions from the flame are regulated by environmental
Siegel and Howell37 and Modest36 provide extensive dis- agencies. In flaring applications, smokeless (smoke results
cussion of the solution methods for radiation in participating from unoxidized soot particles leaving the flame) operation
media. These texts discuss the accuracy, computational is frequently guaranteed by the flare vendor for some range
effort, and limitations of the various models. The reader of conditions. In petrochemical applications, the gases flared
should consult these books for further discussion of these are a wide range of hydrocarbons, typically ranging in
solution methods. molecular weight from 16 to 40. These gases have compo-
nents such as ethylene and acetylene, which are known pre-
9.3.2.1 Gas Radiation Properties cursors to soot formation. Current CFD codes (limited by
Molecular gas radiation is an important mode of heat transfer physical model availability) cannot predict smoking from
in gas-fired furnaces. Radiative emission from nonluminous these large, buoyant flare fires, but current LES work in this
hydrocarbon flames is mostly due to the H2O and CO2 spe- area appears promising.
cies present in the products of combustion. Radiation from
these gases is fairly well-understood, but a rigorous treatment 9.3.2.3 Weighted Sum of Gray Gases
of this radiation requires significant computational resources. The weighted-sum-of-gray-gases (WSGG) model 37 provides
For example, Mazumder and Modest 30 considered ten radia- formulae for computing the emissivity of a gas volume as a
tive bands in modeling emission from a hydrocarbon flame. function of its temperature and partial pressures of CO2 and
This means that they solved the RTE for ten different intensi- H2O. The model assumes the gas is a mixture of radiating
ties. In a large-scale furnace calculation, such a model would gases that is transparent between the absorption bands. The
be extremely computationally demanding. WSGG model is probably the most widely used method to
Quantum mechanics postulates that molecular gases emit calculate radiation within combustion gases. Alternatives
and absorb gases only at distinct wavenumbers, called spec- include band models discussed below. The computational cost
tral lines. However, in reality, these distinct lines are broad- of radiation transport can be very high compared to the flow
ened by several mechanisms, including collision broadening, solver portion of a simulation because of the large number of
natural line broadening, and Doppler broadening. These indi- independent variables in the RTE. In practice, it is usually rea-
vidual lines are characterized by a line strength and a line sonable to lag the calculation of the RTE for a number of flow
width. These lines are caused by quantum transitions in the solver iterations, with the actual number dependent on the
vibrational or rotational state of a molecule. Frequently, vibra- solver in use and stability requirements.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
tional and rotational transitions occur simultaneously, leading
to a tightly clustered array of lines around a given vibrational 9.3.2.4 Turbulence/Radiation Interaction
transition. This subject is beyond the scope of the present The turbulence/radiation interaction plays an important role in
chapter. The intent here is to illustrate the complexity of predicting the radiative emission from a flame. Unfortunately,
modeling a radiating gas. none of the available commercial products that we are aware
of attempt to model this interaction. To appreciate the signifi-
9.3.2.2 Soot cance of this issue, consider the time-averaged radiative trans-
The presence of soot in a flame can significantly increase the port equation in an absorbing/emitting media:
flame emissivity. Predicting soot formation within a flame is
very difficult because soot is formed in fuel-rich regions of a
∂Iλ
flame when the temperature is high. Models such as those of
∂s
(
= κ λ I bλ − I λ ) (9.26)
Khan and Greeves38 and Tesner et al.39 allow the prediction
of soot concentrations, but these models are very empirical
and cannot be expected to provide quantitative results. Frequently, the time-averaged emission is computed as κ λ Ibλ
Soot within a flame is caused by the combustion of hydro- ≈ κ λ Ibλ , which neglects correlations between κλ and T as well
carbons under fuel-rich conditions. Soot is visually observed as the effect of temperature fluctuations on the time-averaged
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 299 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 299

4
emission (i.e., T 4 ≠ T ). Mazumder and Modest 30 discuss the than 1, the time scale of larger turbulent eddies has become
history of the turbulence/radiation interaction. smaller than the chemical time scale. In these conditions, the
combustion process is described as a well-stirred reaction
zone. For intermediate values of Da and Ka, combustion is
9.3.3 Combustion Chemistry Models said to occur in distributed reaction zones. This term indicates
This section discusses the modeling of combustion chemistry that the turbulent flow can affect the structure of the reaction
in petrochemical applications. The focus of this section is on zone, in contrast to the flamelet regime, but the turbulent
methods for modeling the interaction of turbulence with com- mixing is not so fast that the reaction can be considered to
bustion chemistry. This is an area of intense current research, occur under well-stirred conditions.
and some of this research is briefly discussed as it pertains to
current CFD calculations as well as near-future CFD calcula- 9.3.3.2 Non-premixed Combustion
tions. There are several relatively new turbulence/chemistry
This section discusses modeling of non-premixed combus-
interaction models (such as CMC and joint-pdf transport
tion systems. It is implicitly assumed that non-premixed
models) that are not currently available for use in any of the
combustion is being discussed, but the notion of non-pre-
commercial CFD packages. One can hope that this situation
mixed combustion is an idealization. In real combustion sys-
will change soon and these models will be available for more
tems, mixing occurs simultaneously with combustion, and to
widespread use.
call the combustion process non-premixed implies that the
combustion takes place much faster than the mixing and that
9.3.3.1 Regimes of Turbulent Combustion the flame is not lifted off or near extinction at any location.
Damköhler numbers are ratios of a fluid dynamical time scale Although it is an idealization, the assumption that combus-
to a chemical time scale.40 In a turbulent flow, there are a vari- tion is non-premixed provides very useful insight into the
ety of time scales, such as the integral scale (a convective scale) combustion processes occurring in real systems.
and the Kolmogorov scale (a viscous scale). There are also a
There are a multitude of computational models for non-
variety of chemical time scales because of the many chemical
premixed (also called diffusion) flames. One of the earliest
reactions that accompany the combustion of even a simple
models to appear is the eddy-breakup model of Spalding.41
molecule such as CH4. Frequently, combustion problems are
The model of Magnussen and Hjertager 42 limits the reaction
described as being in the high Damköhler or flamelet regime.
rate according to the local mass fractions of the reactant con-
The term “flamelet” is used because of the notion that within
centrations or product concentrations. The ratio of the turbu-
a turbulent non-premixed flame, the actual combustion reac-
lent kinetic energy k to the dissipation rate ε is used as the
tions take place within small layers termed “flamelets.” These
time scale of the turbulent eddies controlling mixing. These
flamelets are so thin that they are not affected by the turbu-
models give physically reasonable predictions of species con-
lent motions within the fluid; instead, molecular diffusion
centrations in non-premixed systems, but do not consider the
effects dominate and the structure of the reaction zone is that
important effect of turbulent fluctuations on reaction rates. The
of a laminar flame (albeit a strained laminar flame).
model can be extended to consider finite-rate chemistry, but
Following Bray,33 the Damköhler number is defined as the model is a moment model, using the time-averaged tem-
tT ku L0 perature in the Arrhenius rate expression. This limitation is
Da = =
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

where the subscript L and superscript 0 refer

t L0 εlL0 severe in light of the large temperature fluctuations observed
to an unstretched laminar flame, and the subscript T refers to in flames.
the scale of the turbulence. In cases where non-premixed com- The mixture fraction concept plays a central role in reduc-
bustion is studied, it is common to use the velocity and length ing a turbulent non-premixed flame to a mixing problem. The
scales (the laminar premixed flame speed and thickness) as mixture fraction is a conserved scalar, meaning that it is
representative of the relevant chemical scales. The Karlovitz convected and diffused by fluid motions and gradients, but it
t L0 t is neither created nor destroyed. The mixture fraction, f, rep-
number is Ka = = L where the subscript K refers to
tk 0
νε resents the mass fraction of fluid at a particular location that
the Kolmogorov time scale. When the laminar flame time is originated with the fuel stream. The pure fuel stream then
less than the Kolmogorov scale (i.e., Ka < 1), the flame is will have f = 1, while the oxidant stream will have f = 0.
considered to be a laminar flame convected and stretched by In a turbulent flow, the mixture fraction f fluctuates at a
a turbulent flow. Combustion in this regime is referred to as given point with time. A probability density function (pdf)
flamelet combustion. When the Damköhler number is less for these fluctuations can be defined so that the probability
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 300 Tuesday, March 22, 2005 7:36 AM

300 The John Zink Combustion Handbook

where

 f (1 − f ) − 1
< f > = 0.1, < f ‘2> = 0.1 α= f  (9.31)
 f ′2 
p( f )

< f > = 0.3, < f ‘2> = 0.1  f (1 − f ) − 1

β = (1 − f ) (9.32)
 f ′2 
< f > = 0.3, < f ‘2> = 0.4
Other pdf shapes, such as a clipped Gaussian function and a
double delta-function, are discussed in Jones and Whitelaw.44
0 0.2 0.4 0.6 0.8 1 The equilibrium chemistry assumption is poor in flames that are
f lifted or flames near extinction. Figure 9.2 shows the shape of
the β-pdf for several values of f and f ′ 2 .
FIGURE 9.2 Plot of the β-pdf for several values of f
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

and f ′ 2 . An alternative to using the equilibrium chemistry assump-

tion is to use the laminar flamelet model. In this model, the
relationship between the state of the mixture and the mixture
of f lying between x and (x + dx) is P(x)dx. The pdf has some fraction f is determined by a laminar diffusion flame calcu-
additional properties:43 lation. Peters 45 introduced this idea, which assumes that the
reaction length scale, LR, is much smaller than the Kolmog-
1 orov length scale, LK. Bilger 46 has criticized the classical

∫ P( f )df = 1
0
(9.27) flamelet method, claiming that for most non-premixed flames
of interest, the flamelet criterion, LR < LK, is violated. Bish
and Dahm (1995)47 discuss the concept further and attempt
1 to eliminate what they view as a key limitation of the method:
f =
∫ fP( f )df
0
(9.28) its assumption that the reaction layers are bounded by pure
fuel on one side and pure oxidizer on the other. Their SDRL
model is based on the one-dimensionality of the reaction
1 layer, but does not assume the reaction layer to be thin relative
∫(f − ) P( f )df
2
f ′2 = f (9.29) to the dissipative scales.
0 The classical flamelet model’s assumption that the reaction
zones are bounded by pure fuel on one side and pure oxidizer
where the f notation indicates the expectation value on the other is severe in light of the NOx control strategies
(or ensemble average, equivalent to the time average in a used in practical combustion systems. NOx control is pred-
statistically stationary flow) of f, and f ′ is the turbulent fluc- icated on entraining cooled combustion products into the
tuation of f. f ′ 2 is the variance of f. reaction zone, and the proportion of these gases entrained
In CFD calculations of large-scale furnaces with non- varies along the length of the flame. The effects of this flue
premixed burners, the most common combustion model used gas entrainment are to reduce flame temperatures and dilute
is the assumed-pdf model with equilibrium chemistry. In this the reactants. Both of these effects are effective in reducing
model, the shape of the pdf of f is assumed. The β-pdf is a NOx formation.
commonly used function that describes the probability of
Research of models of non-premixed combustion continues
finding the instantaneous fluid to have a specific mixture
at a fervent pace. Pope’s 32 joint pdf methods appear promising
fraction. The β-function is given by:
because they have the ability to treat finite-rate kinetics and
eliminate the closure problems. Bilger’s 48 conditional
(1 − f )
α −1 β −1
f moment closure (CMC) method is also a promising model
P( f ) = 1 (9.30) for non-premixed combustion modeling. Both of these mod-
∫ (1 − f )
α −1 β−1
f d f els are applicable to premixed combustion as well.33 These
0 models are still subjects of active research and academic
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 301 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 301

debate and are not implemented in any of the commercial (selective catalytic reduction) and SNCR (selective noncata-
packages of which these authors are aware. lytic reduction) to deal with NOx in the stack. Sulfur scrub-
bers are used to reduce SOx levels after the combustion
9.3.3.3 Premixed Combustion process.
Most petrochemical applications use non-premixed combus- A recent paper by Barlow et al.35 evaluated NOx predic-
tion because of safety issues in premixed combustion. There tions using two different models for the turbulence/ chemistry
are some important exceptions to this statement, however. interaction: the probability density function (pdf) model of
One important class of premixed burners in the petrochemi- Pope (see Pope,32 for example) and the conditional moment
cal industry are venturi-based radiant wall burners. These closure (CMC) model of Bilger.48 The pdf model here is not
burners use high-pressure fuel to educt combustion air from the assumed-pdf discussed above. Instead, the method used
the ambient environment. The fuel and combustion air are solves for the transport and production of the scalar joint pdf
then mixed in a tube prior to the combustion zone. and is extremely computationally expensive because a Monte
Turbulent premixed flames have proven to be much more Carlo solution algorithm must be used. The particularly inter-
difficult to model than their non-premixed counterparts.32 In esting thing about this article is the comment in the introduc-
a turbulent, mixing-limited, non-premixed flame, the flame tion that “a realistic target for agreement between experiment
structure is governed by turbulent mixing, a reasonably well- and prediction might be ±20 to ±30%.” The article goes on
understood phenomenon. The ideal turbulent premixed flame to discuss how sensitive NOx predictions are to the radiation
consists of a flame sheet propagating at some flame speed with model used. The flame studied in this paper is a simple diluted
respect to the fluid around it, which is itself undergoing tur- hydrogen jet flame. If the most sophisticated turbu-
bulent motions. The consequence of superposing flame prop- lence/chemistry models currently under research applied to a
agation and turbulent fluid motions is that premixed flame very simple flame in a very simple geometry can only be
modeling is much more challenging than modeling non-pre- expected to yield an accuracy of ±30%, then how accurately
mixed flames.34 For this reason, most commercial CFD codes can one realistically expect to predict NOx emissions in more
only include limited support for premixed flame modeling. complex flames?
The model of Magnussen and Hjertager 42 can be used to
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

simulate a premixed flame. The model is unsatisfactory in 9.3.5 Turbulence Models

many ways, however. It has no means of modeling the effect
A turbulent flow is a flow with a wide range of temporal and
of temperature fluctuations on the reaction rate and no descrip-
length scales. Figure 9.3 is an example of a typical point mea-
tion of the turbulent flame as an ensemble of premixed flame-
surement (e.g., pressure or velocity) within a turbulent flow.
lets.
Within a turbulent flow, the quantities of interest such as pres-
A significant limitation of many of the flamelet models for
sure and velocity fluctuate in an apparently random fashion.
premixed combustion (see, e.g., Warnatz et al.34) is that they
Analysis reveals that these quantities are not truly random.43
assume that the combustion process is adiabatic. In operating
Information revealed by spectral analysis of point measure-
furnaces, heat losses from the flame to the load are an integral
ments reveals that there are ranges of temporal and length
part of the process. The inability to adequately model pre-
scales that contain significant energy (the large or integral
mixed flames is a significant limitation. Pdf methods such as
scales) and smaller scales where this turbulent energy is dissi-
those discussed by Pope32 may allow improved simulations
pated by viscous processes. The energy cascade is the mecha-
in the future. In addition, it may be possible to include heat
nism by which energy is moved from the large scales to the
losses in the flamelet models of premixed combustion.
small scales. For more information on the physics of turbulent
flows, the reader should refer to Libby.43
9.3.4 Pollutant Chemistry Models Prediction of turbulent flow from the Reynolds- or Favre-
Pollutant emissions are among the most important drivers in averaged conservation equations requires closure approxima-
the petrochemical industry, especially in the United States tions. This is because the time-averaged conservation equa-
(see Chapter 6). The U.S. EPA allowed levels of NOx and tions contain terms that are not known. In the case of the
SOx emissions from petrochemical plants and refineries con- momentum equations, the time-averaging of the convection
tinue to decrease. To respond to this challenge, burner manu- terms leads to the following Reynolds stresses:
facturers strive to develop burners that produce less and less
emissions. In addition, furnace manufacturers and other ven-
dors develop post-combustion technologies such as SCR ρu ′ 2 , ρv ′ 2 , ρw ′ 2 , ρu ′v ′, ρu ′w ′, ρv ′w ′
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 302 Tuesday, March 22, 2005 7:36 AM

302 The John Zink Combustion Handbook

9.4 SOLUTION METHODOLGY

Measured quantity in a turbulent flow

There are many schemes used to discretize the partial differ-

ential equations of fluid flows onto different types of meshes.
Because the primary focus of this chapter is applied CFD
where mesh types, by necessity, include tetrahedra, the two
important discretization schemes are the finite-volume
method and the finite-element method. The finite-volume
Time-averaged value method is clearly the method of choice in the industry today
for large-scale computations of turbulent flows. The dominant
software products commercially available for these problems
almost exclusively use the finite-volume method. There are
Time occasions when other methods, such as the finite-difference
method, are used. Most advanced combustion models are first
FIGURE 9.3 Point measurement of scalar in a turbulent implemented in academic CFD codes, which, as mentioned in
flow.
the introduction, are rarely intended to model complex geom-
etry. For this reason, industrial CFD analysts rarely have
Closure approximations are required to solve the Reynolds- access to advanced combustion models.
averaged conservation equations.
The workhorse turbulence model used in furnace simula- 9.4.1 Problem Setup: Preprocessing
tions is the k-ε model.24 The popularity of this model can be The preprocessing phase of a problem includes all the steps
ascribed to its relative simplicity (compared to a Reynolds from the initial problem definition through the beginning of
stress model, for example) and its good performance in a computations. In typical problems, this includes geometry
variety of engineering flows. Its weaknesses include its per- creation, mesh generation, model selection, fluid property
specification, and enabling and setting up the models (such as
formance in unconfined flows, in rotating and swirling flows,
k-ε for turbulence).
and in flows with large strains, such as curved boundary layers.
The Reynolds stress model (RSM) addresses some of these Problem definition is critical. At this phase, the scope of
performance issues. RSM is much more computationally the geometry to be studied should be considered. In many
cases, it is difficult to determine where to place the outside
demanding because it involves seven extra partial differential
boundaries of a CFD model. Flow condition must be known
equations rather than the two of the k-ε model. However, in a
at inlets. Thus, for example, putting a flow inlet just down-
typical combustion calculation, the number of PDEs solved is
stream of an elbow would probably be a poor choice because
typically quite large, so adding five more can be easily justified it would be difficult to know the velocity and pressure profiles
if the quality of the predictions improve. at such a location. This issue is particularly important if heat
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

A number of variants of the classical k-ε model exist. The transfer is to be considered. Thermal boundary conditions are
classical k-ε model uses a single eddy viscosity in all direc- typically difficult to specify, requiring considerable physical
insight into a problem.
tions. The nonlinear k-ε model by Speziale49 addresses this
assumption, which is known to be poor even in relatively It is also important to consider the capabilities of the soft-
simple flows. Another development in k-ε modeling is the ware and computer hardware to be used when specifying a
Renormalization Group (RNG) k-ε model of Yakhot et al.50 problem. For example, if the software’s only turbulence model
is the k-ε model, then studying a highly swirling flow (where
Its performance in complex flows has been promising, so
k-ε is known to perform poorly) may generate useless results.
much so that several of the commercial CFD code vendors
On the other hand, if one is aware of this limitation and recog-
have implemented the RNG k-ε model. The Realizable k-ε nizes that the turbulence model will not accurately predict the
model51 represents yet another variant recently introduced. decay of the swirl, a conscious decision can be made to neglect
The advantages and limitations of these turbulence models the portions of the solution that are expected to be poor and
are discussed in more detail in Veersteeg and Malalasekera.52 only use the results that are expected to be meaningful.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 303 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 303

9.4.1.1 Complex Geometry

The capability of a CFD package to treat complex geome-
tries is an important consideration for industrial applica-
tions. The geometries encountered in low emissions burners
frequently employ complicated shapes and jet angles. The
purpose of these geometries is to precisely control when and
where the fuel is oxidized. These combustion control strate-
gies are critical to the performance of the equipment. CFD
models must be able to accurately capture the effect of these
complex geometries in order to be useful. This means, as a
practical matter, that cell types other than hexahedra are
required. All current-generation commercial codes are com-
patible with a variety of cell types, including hexahedra,
prisms, pyramids, and tetrahedra. Figure 9.4 shows a ren-
dered view of a CFD model of a hearth burner. The burner
shown in Figure 9.4 is vertically fired, and the view shown is FIGURE 9.4 Rendered view of a CFD model of a John
looking down at the burner from above. There are four gas Zink Co. burner. This view illustrates the complex geometry
tips, two in the primary combustion zone (inside the tile) and that necessitates a variety of cell types. This mesh consists of
two firing secondary fuel (these tips are sitting on the tile). hexahedral, pyramidal, and tetrahedral cell types. (Courtesy
Each primary tip has five small orifices firing fuel in differ- of John Zink Co.)
ent directions, as shown in Figure 9.5. The orifice diameters
range from 0.0625 in. (1.59 mm) to about 0.25 in. (6.4 mm).
Each secondary tip has three fuel orifices of similar size. The vergence problems are encountered. Normally, all that is
figure illustrates the necessity of a CFD package to treat required of the analyst is to observe the progress of the CFD
complex geometry to accurately model the performance of code toward convergence and perhaps adjust under-relax-
an industrial burner. ation factors and adapt the grid. It is always the stated goal to
Generating a computational mesh for these geometries is a obtain a solution that is grid-independent, but in practice it
well-known bottleneck in CFD analysis. Mesh generation can usually is too time-consuming to refine the mesh such that
consume well over half of the time budgeted for a CFD project. the obtained solution can be proven grid-independent.
Improvements in mesh generation technology greatly benefit
industrial CFD users as they allow more and more of the actual 9.4.3 Analysis of Results: Post-processing
geometry to be included in the CFD model. In addition, mesh
generation improvements frequently simplify the process of A typical CFD simulation provides on the order of 106 to 108
modifying an existing geometry. Making geometric modifica- discrete numerical outputs. For example, a simulation with
tions to a geometry after the initial meshing can be nearly as 500,000 nodes and 11 variables per node (pressure, density,
time-consuming as generating the initial mesh. three velocity components, k, ε, temperature, mixture fraction,
variance of mixture fraction, and irradiation) would generate
New combustion and turbulence models are frequently
5,500,000 numbers. If the various chemical species are con-
developed only for simple Cartesian grids. Only much later,
sidered as well as the detailed results of a discrete ordinates
after considerable proof of concept in the simple geometries,
model, the number of variables per node could easily exceed
are these models adapted for use on mesh types used in indus-
50, leading to 25,000,000 numerical results. The generation of
trial analysis. In the present authors’ opinions, the ability to
x-y plots (e.g., temperature vs. position along the burner cen-
treat complex geometry is the main reason that industries turn
terline), contour plots, velocity vector plots, streamline plots,
to commercial CFD vendors instead of using codes developed
and combinations and animations of these outputs are neces-
at national laboratories and universities.
sary for the analyst to understand the results of a simulation.
The production of these different sorts of outputs becomes
9.4.2 Solution Convergence very important in communicating the results of a simulation.
After generating the mesh and setting up the problem for This is especially true when the intended audience is not com-
solution, the calculations begin. This phase of the CFD anal- posed of CFD specialists. Current post-processing packages
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

ysis does not usually require much effort unless severe con- have the ability to add lighting to a model, which makes the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 304 Tuesday, March 22, 2005 7:36 AM

304 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 9.5 Close-up view of primary tip. This view reveals the five fuel jets (indicated by the arrows on the image)
issuing from the primary tip.

FIGURE 9.6 Rendered view inside an ethylene pyrolysis furnace showing flow patterns near the premixed radiant wall
burners. (Courtesy of John Zink Co.)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 305 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 305

images more realistic to the viewer. Figure 9.6 shows an

example of using the rendering capabilities of a CFD package
to generate an image with photo-realistic qualities. Images
such as Figure 9.6 can take anywhere from several seconds to
several minutes for current generation scientific workstations
to render, depending on the number of lights applied, the
number of surfaces in the scene, and the complexity of these
surfaces. High-performance virtual reality environments must
be able to regenerate these scenes many times per second.
In addition to still images, animations can be effectively
used to illustrate CFD results. Animated velocity vectors
and streamlines accurately illustrate the path of fluid flow
in internal and external flow problems. Sweeping planes
showing either velocity vectors or filled contour maps of a
scalar result can quickly present information about an entire
three-dimensional simulation.
Generating effective presentations, including still images
and animations, is a time-consuming task. Creating a suitable
image to make a specific argument frequently requires the
analyst to look at and reject a large number of candidate images.
It also requires significant expertise from the CFD analyst. It
is certainly true that CFD results can be misinterpreted or
misapplied to lead to an incorrect conclusion. In addition, in FIGURE 9.7 CFD model of an ethylene pyrolysis fur-
an industrial setting, the audience will frequently not have the nace. There are six burners shown in each row at the bottom
expertise required to assess the quality of a simulation. of the furnace, and the tubes are approximately 35 feet long.
The endwalls are not shown in this image.

9.5 APPLICATIONS: CASE STUDIES trates this geometry. In the figure, the process fluid tubes
This section describes several applications of CFD in the extend from the floor of the furnace to the roof of the radiant
petrochemical industry. CFD can address a wide variety of section. In the image, only the radiant section is shown
problems in this industry. The applications discussed here because the radiant section is where the combustion occurs.
relate to fired heaters and incinerators. Many other applica- In a production furnace, the products of combustion would
tions (e.g., flare systems) exist in petrochemical plants where leave the radiant section and enter a convection section where
CFD analysis is valuable. heat is recovered from the products of combustion.
Figure 9.4 shows a view of the burner geometry. The burner
9.5.1 Case 1: Ethylene Pyrolysis Furnace is a staged-fuel gas burner. This example illustrates the dis-
Ethylene pyrolysis furnaces produce ethylene and propylene parity in scales in a furnace analysis. The furnace has a height
from feedstock containing ethane, propane, butane, and of approximately 30 ft (9 m), while the fuel orifices can be
hydrocarbons including naphtha. The process entails rapidly as small as 0.0625 in. (1.59 mm) in diameter. The ratio from
heating the feedstock for a short time (less than 1 second is the largest dimension to the smallest is then greater than 5000.
typical) to a temperature of about 1600°F (870°C). The feed In this example, a nonconformal mesh interface was used to
gases are then rapidly cooled and subjected to a number of reduce the cell requirements.
separation processes. The CFD model of the ethylene pyrolysis furnace includes
This section focuses on the modeling of the pyrolysis fur- detailed information about all of the fuel jets in the burner.
nace. Typical pyrolysis furnaces are approximately 10 ft (3 m) In this particular burner, there are five fuel jets on each of the
wide, 30 ft (9 m) long, and 40 ft (12 m) tall. There are two two primary tips (Figure 9.5) and four fuel jets on each
rows of “flat flame” burners that directly fire onto the walls secondary tip.
of the furnace. These fired walls then radiate heat to the Figure 9.8 shows the predictions of heat flux to the process
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
process tubes in the center of the furnace. Figure 9.6 illus- tubes as a function of height above the furnace floor. These
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 306 Tuesday, March 22, 2005 7:36 AM

306 The John Zink Combustion Handbook

0.8
Height (normalized)

0.6

0.4

0.2

0
0 0.25 0.5 0.75 1
Heat flux (normalized)

FIGURE 9.8 Plot showing heat flux to the process tubes in the modeled ethylene furnace as a function of height above
the furnace floor.

heat flux profiles drive the design of modern ethylene pyrol- slice of pie. The images shown here are created by rotating
ysis furnaces. CFD is being used in these designs more and the results about the vertical axis of the furnace.
more as the results of the model become better validated.
The combustion model used in these calculations is an
These results have not been validated, and it seems unlikely
assumed pdf of mixture fraction. Because heat transfer to
that such data will become available, given the difficulty of
the tubes and furnace temperature are known to be impor-
acquiring data in operating furnaces. Availability of data is a
tant, a nonadiabatic mixture fraction table was constructed.
significant limitation in further use of CFD in petrochemical
The independent variables in the lookup table are mixture
applications, as discussed in the introduction.
fraction, variance of mixture fraction, and enthalpy. Radia-
tion was modeled using the discrete ordinates model with
9.5.2 Case 2: Xylene Reboiler 32 ordinates. All solid surfaces were assumed to be radia-
This study involves an operating furnace in a refinery. The tively black. Gas radiation properties were computed using
problem observed in the furnace was that the flames from the the weighted-sum-of-gray-gases method.
ultra-low-NOx burners were very long and had the potential Figure 9.10 shows a CFD simulation of the burners as they
to damage process tubes in the top of the furnace. The were originally installed. The figure shows an iso-surface of
authors have observed this phenomena in several vertical/ OH, which is a good indicator of flame shape in this case.
cylindrical furnaces with ultra-low-NOx burners. The prob- The results reveal that the flames from adjacent burners merge
lem is related to the flow pattern within the furnace as it does together to produce a single long flame, which is confirmed
not allow complete mixing of the combustion air with the by observations of the operating furnace. This burner has two
fuel, but rather distorts the flame prior to burnout. primary fuel tips that fire fuel inside the tile and four second-
The geometry of the vertical cylindrical furnace is shown ary fuel tips. The solution to this flame interaction problem
in Figure 9.9. The small wall around the burners is a reed was to change the burner so that only three of the secondary
wall and is used to heat the cold flue gases coming from the tips actually fired. The CFD results for that configuration are
tubes. The periodicity of the furnace was used to simplify the shown in Figure 9.11. This solution was implemented and
model. The computational model of the vertical/cylindrical tested in the operating furnace and found to yield qualitatively
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

heater consisted of only one burner with periodic boundary the same result: the flames became distinct and burned out at
conditions applied. This model has the shape of a very tall the appropriate height.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 307 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 307

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 9.9 Geometry of a xylene reboiler. This view shows half (sliced vertically) of the furnace. This view shows only
three of the six burners at the bottom of the image.

9.5.3 Case 3: Sulfur Recovery Reaction and density at any location in the domain as a function of
mean mixture fraction, mixture fraction variance, and
Furnace
enthalpy. Radiation heat transfer was not considered.
This case study considers incineration of an acid gas to pro- Figure 9.12 shows the geometry of the reaction furnace.
duce atomic sulfur. The reaction furnace is part of a Klaus The exterior of the burner is shown in the lower left corner
process for sulfur recovery. The composition of the acid gas of the figure. The furnace is approximately 5 m (16 ft) in
is given in Table 9.5. During the course of the study, the diameter and 16 m (52 ft) in length. It is operated at about
burner geometry was modified in order to shorten the flame 60 kPa of positive pressure.
and create better mixing within the reaction furnace. Figure 9.13 shows the geometry of the burner from the
In the present study, the “realizable k-ε” turbulence model furnace looking at the burner throat. The green swirl vanes
is used to simulate the effects of turbulence on transport swirl the combustion air, while the acid gas is carried in the
within the domain. The combustion process in this study is light blue tube and swirled in the opposite direction by the
modeled using the assumed pdf (probability density function) red swirl vanes. The initial burner geometry did not have the
of the mixture fraction model. In this model, a transport red acid gas swirl vanes or the yellow bluff body in the acid
equation for a conserved scalar, called the mixture fraction, gas passageway.
is solved. The effect of turbulence on the chemistry is simu- The simulations predict all the species concentrations
lated by solving a transport equation for the variance of mix- throughout the burner and furnace. Figure 9.14 shows the pre-
ture fraction. An assumed form for the pdf (a beta-pdf, which dicted O2 mass fractions in the furnace. The plot shows a top-
is a common choice in combustion simulations) allows for down view of the furnace with the O2 mass fractions contoured
the creation of a lookup table by assuming that chemical at the mid-plane. This figure shows that the oxygen does not
equilibrium exists. The lookup table gives the composition penetrate through the combustion zone, but is consumed near
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 308 Tuesday, March 22, 2005 7:36 AM

308 The John Zink Combustion Handbook

FIGURE 9.10 This view shows half of the furnace with unmodified burners. The “blob” in the furnace is the 50-ppm OH
mole fraction iso-surface. This surface is colored according to its temperature (°F).

FIGURE 9.11 This view shows half of the furnace with the modified burners firing. The 50-ppm OH mole fraction iso-
surface is shown as an indicator of the flame shape. This surface is colored according to its temperature (°F).
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Computational Fluid Dynamics (CFD) Based Combustion Modeling 309

the burner. The asymmetry observed is due to the swirling flow. predicted HCl, Cl2, O2, Cl, H outlet concentrations and average
The swirling flow creates a swirling flame in the furnace. exit temperature for each case were recorded and evaluated.
Figure 9.15 shows the same view of the furnace, only the Although performing actual test burns for the cases iden-
mid-plane is colored according to the mole fraction of H2S. tified during this study would be prohibitively expensive and
Figures 9.16 and 9.17 show the stoichiometric iso-surfaces for time-consuming, the CFD-based incineration model was able
the initial and final burner design, respectively. The initial to assess the impact that various operating conditions might
design did a poor job of mixing the acid gas and the combustion have on HCl/Cl2 products in the effluent gas. From these
air. This resulted in a long corkscrewing flame that did not
completely burn out, even by the end of the reaction furnace. TABLE 9.5 Composition of Acid Gas
Used in CFD Study
Figure 9.18 shows a close-up view of the temperature results
Mole
near the burner quarl. Temperature contours in the mid-plane Component (%)
of the burner are also shown, clearly showing the reaction and
N2 0.005
mixing regions. The figure also shows velocity vectors, which CH4 0.194
reveal the swirl both in the combustion air and acid gas. CO2 17.981
C2H6 0.035
Figure 9.19 shows temperature profiles exiting the reaction H2S 75.223
furnace for the initial and final design. The figure shows that COS 0.001
H2O 6.500
the final design produces much better temperature uniformity
C3H8 0.009
exiting the furnace. In this case study, uniformity of the exit- C4H10 0.003
ing temperature profile was critical because this furnace had
a waste heat recovery boiler directly downstream of the reac-
tion furnace. Nonuniformity in the exiting temperature profile
would have produced significant deterioration of the boiler
tube metal. simulations, the estimated range of expected HCl and Cl2
production was:
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

9.5.4 Case 4: Incineration of Chlorinated

Expected Cl2 concentration range: 0.01–221 ppmv
Hydrocarbons
Research performed by VanDell and co-workers53–55 and by Expected HCl concentration range: 3.9–6.4 mol%
workers at other research laboratories (Choudhry et al.;56
Having performed the matrix of incineration simulations,
Shaub and Tsang;57 Graham et al.;58 Senkan;59 Taylor and
two cases were identified as representative of the limiting
Dellinger;60 Altwicker et al.;61 Young and Voorhees 62) have
scenarios for HCl/Cl2 production. To better understand the
investigated the chemical processes leading to products of
causes for the predicted dramatic impact on Cl2 production
incomplete combustion (PICs) formation. Although earlier
rates and on exit gas temperature, these cases were further
work has identified specific mechanisms leading to PIC for-
investigated. One advantage of the CFD-based incineration
mation, the quantity of PICs formed depends greatly on local
model is its ability to generate local profiles of velocity,
process conditions.63 Thus, developing and implementing a
temperature, and species (e.g., CH4, CO, HCl, and Cl2). With
methodology to predict local conditions inside a thermal oxi-
these profiles, results from the limiting cases were closely
dizer may significantly enhance the ability to predict PIC
examined to better understand the differences between each.
formation during incineration. This example describes a
study where the PCGC code was applied to a thermal oxi-
dizer to study the formation of HCl/Cl2 resulting from the 9.5.4.1 Excess Oxygen Condition
combustion of several RCl feed streams. The first case considered represented an excess oxygen sce-
Given the necessary inputs to the CFD model (e.g., reactor nario. Important input data for this case are shown in Table 9.6.
geometry (see Figure 9.20), inlet flowrates and compositions, Both the axial velocity and local gas temperature are shown
boundary conditions, etc.), a base case was established. Next, first (Figure 9.21(a) and (b)). A maximum velocity of nearly
several separate simulations were performed for various com- 60 m/s is predicted near the reactor entrance. This is likely
binations of the feed stream flow rates. Each case represented caused by the temperature driven gas expansion associated
a combination of the nominal and/or maximum flow rate of with combustion (Figure 9.21). A weak recirculation zone is
fuel gas, of combustion air, and of organic feed streams. The predicted near the reactor wall in the quarl section (lower left-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 310 Tuesday, March 22, 2005 7:36 AM

310 The John Zink Combustion Handbook

FIGURE 9.12 Exterior geometry of furnace included in model. The surface mesh is also shown.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 9.13 Burner geometry. The acid swirl vanes are shown in red; the air swirl vanes are shown in green; and the
start-up fuel tip is shown in purple.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 311 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 311

FIGURE 9.14 Oxygen mass fractions viewed from the above the furnace. The contour scale is logarithmic. The mass
fractions are contoured on the mid-plane of the furnace.

FIGURE 9.15 H2S mole fractions contoured on the mid-plane of the furnace.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

312 The John Zink Combustion Handbook

FIGURE 9.16 Stoichiometric iso-surface colored by temperature (°C) for the initial burner design.

FIGURE 9.17 Stoichiometric iso-surface colored by temperature (°C) for the final burner design.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Computational Fluid Dynamics (CFD) Based Combustion Modeling 313

FIGURE 9.18 Mid-plane of geometry colored by temperature (°C). This view shows the burner quarl and the mixing

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
regions of acid gas and air.

FIGURE 9.19 Temperature profiles (°C) exiting the reaction furnace for the initial (left) and final (right) burner geometries.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 314 Tuesday, March 22, 2005 7:36 AM

314 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 9.20 Geometric information describing the thermal oxidizer examined during this study.

TABLE 9.6 Limiting Cases Considered During the quarl region. This may be caused by the recirculation of
RCl Combustion Study the cooler gas fed with the secondary feed stream. Also, the
Excess Oxygen Case abnormally low prediction for gas temperature may be a result
Organic feed (g/s): 390 Inlet temperature (K): 298 of using the uniform reactor heat loss. Finally, a second high-
Combustion air feed (g/s): 1955 Swirl number (–)a: 0.313 temperature region (>1600°C or 2900°F) is predicted near the
Fuel gas feed rate (g/s): 1.4 Heat loss (%): 35%
reactor exit. This is caused by the combustion of remaining
Stoichimetric Case fuel (organic vents) and by the CO oxidation reaction that
Organic feed rate (g/s): 390 Inlet temperature (K): 298 occurs in this portion of the reactor. CO formation and oxida-
Fuel gas feed rate (g/s): 1987 Swirl number (–)a: 0.313 tion (CO2 formation) are shown in Figure 9.20(b). The spatial
Fuel gas feed rate (g/s): 69.4 Heat loss (%): 35%
nature of these predictions is illustrated by the high CO levels
a The swirl number is a measure of tangential velocity in the secondary inlet near the reactor centerline (>4000 ppmv) that decreases toward
stream. (–), time-averaged value.
the reactor walls to essentially zero.
Finally, predicted values of HCl and Cl2 concentrations
hand section of plot). Toward the reactor exit, the gas flow is are shown in Figure 9.23(a) and (b). Figure 9.23(a) clearly
fully developed with an exit velocity of about 15 m/s (49 ft/s). shows the localized nature of the HCl/Cl2 chemistry in the
The predicted local gas temperature is shown in Figure combustion zone. The maximum predicted HCl concentra-
9.21(b). The quarl region, clearly shown in this plot, has a wall tion (20.6 mol%) is located early in the reactor (high-tem-
temperature of 1000°C (1800°F), which is a preset boundary perature region), compared to a predicted exit HCl
condition. A high-temperature envelope (>1000°C or 1800°F) concentration of 5.3 mol%. Also, the maximum Cl2 level
is predicted near the reactor entrance, associated with the initial (3600 ppmv) occurs just beyond the reactor quarl wall (low-
combustion zone. Methane is consumed in this region (see temperature region), compared to an exit Cl2 concentration
Figure 9.22(a)). A cooler region exists near the reactor wall in of 75 ppmv. This indicates how dramatically the local con-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 315 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 315

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 9.21 Predicted centerline profiles for excess air case: (a) axial velocity (m/s), and (b) gas temperature (K) for
the furnace section of the thermal oxidizer shown in Figure 9.20. Two distinct combustion zones are illustrated, with an exit
temperature of about 1600 K (2400°F).

ditions affect the HCl/Cl2 levels. This indicates that those quarl section. The flame also expands to the reactor wall and
process variables affecting local regions of the reactor can has a maximum temperature of approximately 1450°C
be used to optimize emissions levels. (2640°F). The low-temperature zone (~400°C or 750°F) in the
quarl section of the reactor and the weak recirculating flow
9.5.4.2 Stoichiometric Oxygen Condition zone (~ –1 m/s or –3 ft/s) from the earlier case have also
The second case considered here represents an overall sto- changed. The gas temperature has increased (>1200°C or
ichiometric condition inside the reactor. Predictions from this 2200°F), as has the recirculating velocity (~ –5 m/s or –16 ft/s).
case indicated a much lower Cl2 emission and a lower exit The expansion of the high-velocity region near the reactor
temperature. Therefore, this case was carefully analyzed to entrance, the stronger recirculation zone in the quarl region,
understand the reasons for the apparent difference. Important and the increased gas temperature in the quarl region appear
input data for this case are also shown in Table 9.6. to correlate with the relative amounts of natural gas and organic
Although the axial velocity profiles for this case are similar feeds considered in the respective cases. Two factors can help
to those of the first case, there are some interesting differences. explain the different results from these cases.
First, the velocity region near the reactor inlet is higher than First, while this case represents a near-stoichiometric reac-
before, and the recirculation zone near the wall in the quarl tant mixture, the previous case represents a reactant mixture
region is larger than before (see Figure 9.24(a)). Both differ- with excess air. Examining Figure 9.25(b), the CO oxidation
ences are caused by dissimilar temperature profiles throughout that occurs midway down the reactor (due to the excess oxy-
the reactor (see Figure 9.24(b)). First, the early centerline gas gen and turbulent mixing) is not observed here (see Figure
temperature appears to be about the same (~950°C or 1740°F) 9.25(b)). Thus, higher exit CO levels (>8000 ppmv) are pre-
as before. Instead of two distinct high-temperature zones (near dicted, which results in less energy release from the exother-
the entrance and near the exit), in this case a single flame zone mic oxidation reaction of CO (∆Hr = –282 KJ/mol), which
extends from the centerline reactor entrance to just beyond the leads to a lower exit gas temperature (1398°C or 2548°F).
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 316 Tuesday, March 22, 2005 7:36 AM

316 The John Zink Combustion Handbook

FIGURE 9.22 Predicted centerline profiles for excess air case: (a) methane concentration (ppmv) and (b) carbon monoxide
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

concentration (ppmv) for the furnace section of the thermal oxidizer shown in Figure 9.20. These predictions depict the CO
formation and oxidation zones common to most combustion processes.

206000 76000
178000 146000
106000
70000
68000

50000 58000
34000
20000 48000
10000 42000
0 2000

HCl concentration profile (ppmv)

100
200
600
2300 400
600 800
2400
1300 1300
2500 3200 2000
3600

Cl2 concentration profile (ppmv)

FIGURE 9.23 Predicted centerline profiles for excess air case: (a) HCl concentration (ppmv) and (b) Cl2 concentration
(ppmv) for the furnace section of the thermal oxidizer shown in Figure 9.20. The predicted maximum Cl2 concentration, nearly
3200 ppmv, occurs in the cooler reactor regions, while an exit Cl2 concentration of about 100 ppmv is predicted.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 317 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 317

FIGURE 9.24 Predicted centerline profiles for stoichiometric case: (a) axial velocity (m/s), and (b) gas temperature (K)
for the furnace section of the thermal oxidizer shown in Figure 9.20. A single combustion zone with the local maximum
temperature of 1450 K (2150°F) and an exit temperature of about 1350 K (1970°F).

Another possible cause for the different temperature pro- 1. HCl is favored over Cl2 at higher temperatures
files might be related to how and where the organic vents are 2. Lower O2 concentrations favor Cl2
burned. In the first case, the organic vents appear to be burned
after the methane is consumed. This would lead to two reac- For the initial case, the early maximum of Cl2 can also be due
tion zones: one where fuel gas is burned (early flame zone), to poor mixing of O2 with the fuel and the accompanying low
and one where organic vents ignite, resulting in a secondary gas temperatures (see Figure 9.21(b)). Similarly, for the present
flame. These two high-temperature regions, observed in case, the global maximum is in the same region, but now the
Figure 9.21(b), would lead to a higher exit temperature. In effect of higher gas temperature (favors HCl formation) results
either case, the relative amounts of fuel gas and oxidizer are in significantly less Cl2 formation. Thus, both the tempera-
critical to the predicted combustion characteristics inside the ture effect and the oxygen effect are important.
reactor. More importantly, they dramatically affect the
This prediction is most interesting when considering the
HCl/Cl2 chemistry in the reactor, as seen in Figure 9.25(a)
slight increase in HCl production accompanied by the dra-
and (b).
matic decrease in Cl2 production. In an attempt to validate this
In both cases, the predicted exit HCl concentration is predicted behavior, the two conditions were reproduced in the
greater than 5 mol%. However, the predicted exit Cl2 concen- field by adjusting the fuel gas and the organic feed rates
tration in Figure 9.26(b) is nearly 2 orders of magnitude less accordingly. The flame inside the thermal oxidizer was visu-
than that for the initial case. The same general trends show ally monitored along with the exit gas temperature during the
up in both cases: maximum Cl2 concentrations near the outer two tests. For the high RCl feed rate/low fuel-gas feed rate
wall just past the quarl section that is reduced to uniform exit scenario, the flame appeared to nearly fill the entire combus-
concentrations toward the reactor exit. Two factors explain tion zone of the thermal oxidizer. As the fuel gas was increased
the significant differences in the HCl/Cl2 concentrations for while holding the organic feed rate constant, the visible flame
the different cases: --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
front appeared to retreat toward the front of the burner. Also,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 318 Tuesday, March 22, 2005 7:36 AM

318 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 9.25 Predicted centerline profiles for stoichiometric case: (a) methane concentration (ppmv), and (b) carbon
monoxide concentration (ppmv) for the furnace section of the thermal oxidizer shown in Figure 9.19. Here, the post-flame CO
oxidation zone, shown in the first prediction, is not present; this results in a predicted exit CO concentration of 9000 ppmv.

HCl concentration profile (ppmv)

195000 175000 150000 100000 70000

55000
25000
50000
20000 35000
5000

0.3
1.5 0.9
3.0 0.6 0.3
0.9
0.3 3.6
3.0
1.5 5.4 2.7
1.5 0.3
0.3 6.9

Cl2 concentration profile (ppmv)

FIGURE 9.26 Predicted centerline profiles for stoichiometric case: (a) HCl concentration (ppmv), and (b) Cl2 concentration
(ppmv) for the furnace section of the thermal oxidizer shown in Figure 9.20. Dramatically less Cl2 formation is predicted (local
maximum of 7 ppmv and exit concentrations less than 1 ppmv) in this case due to excess H+ radical present from the increased
fuel gas.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 319 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 319

the exit temperature decreased by nearly 200°C (400°F) during

this test. Although this is only a limited amount of data, it
provides some degree of validation of the predictions.

9.5.5 Case 5: Venturi Eductor Optimization

Natural draft, premixed, radiant wall burners use a venturi
to educt air and mix it with the fuel, leading to a combusti-
ble mixture which will be burned in a furnace. This exam-
ple illustrates the use of CFD in optimizing the performance
of venturis.
New burner design technologies use the motive energy of
fuel gas injection to entrain flue gas for the purpose of diluting
the combustible mixture to reduce NOx emissions. Using an
eductor, these designs entrain flue gas from above the con-
vection section of a furnace, as shown in Figure 9.27. A large
pressure drop through the upstream and downstream piping FIGURE 9.27 Illustration showing entrainment of flue
of the eductor system can create a recirculation pattern in the gas using the motive energy of the fuel gas.
eductor throat. This re-circulation results in a reduction of the
flue gas entrainment rate. Semi-empirical models, based on
the conservation of mass, momentum and energy, have been single most important improvement in the CFD analysis of
used in the past to estimate the entrainment performance of industrial systems. Accurate NOx predictions will continue
an eductor system. These models, however, cannot reliably to be a challenge due to the nature of the physical problem
estimate the entrainment performance of an eductor system (i.e., the strong coupling between the relatively slow chemical
when a re-circulation zone occurs in the throat. The eduction reactions and the turbulence in the flame).
performance is dependent on the size of the re-circulation The ability of CFD models to treat stability problems in
zone in the eductor throat as illustrated in Figure 9.28. non-premixed combustion is another area needing improve-
The dimensions of the recirculation zone are a function of ment. As discussed above, the presumed PDF model with
the ratios of fuel orifice-to-eductor throat diameter, motive and equilibrium chemistry is probably the most accessible model
educted gas densities, and flow rates. Numerical modeling has to the industrial analyst, yet significant departures from equi-
been used to study the effects of entrainment performance librium are observed in flames near extinction.34 Given the
based on these parameters. Figures 9.29 and 9.30 show the chemical complexity of the problem, finding a model that can
results of a numerical model with boundary conditions leading be valid for a range of fuels will be a significant challenge,
to a re-circulation pattern in the eductor throat. Several calcu- but such a model would benefit the industry. It seems likely
lations performed for a variety of different boundary condi- that a model that can capture these effects is a prerequisite
tions such as fuel pressures, eductor and orifice diameters, and to improving NOx predictions due to the strategies employed
backpressures. These results are shown as a series in Figure in industrial combustion systems for reducing NOx.
9.31. The inability of current-generation commercial products to
model premixed flames with heat losses is a significant lim-
itation. Premixed combustion is significantly more challeng-
ing than non-premixed combustion because of the coupling
9.6 FUTURE NEEDS of flame propagation and turbulent motions in the fluid.
It seems fitting to summarize the needs for future develop- Improvement in models of premixed combustion that are
ment. Throughout this chapter there have been discussions of applicable to petrochemical burners would benefit the indus-
model limitations and the need for better basic physical mod- trial CFD user.
els. These needs are relatively clear and easily understood.
The focus of this section is on issues that hinder CFD from
becoming a valued design and troubleshooting tool in the 9.7 CONCLUSION
petrochemical industry. CFD modeling of industrial furnaces is a valuable tool that can
Given the importance of NOx emissions in the installation be used profitably. CFD modeling can help identify the cause of
of new combustion equipment, NOx predictions may be the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
problems and it can be used to test solutions. In addition, CFD
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 320 Tuesday, March 22, 2005 7:36 AM

320 The John Zink Combustion Handbook

FIGURE 9.28 Illustration showing re-circulation region in the eductor throat.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 9.29 Re-circulation zone starting to occur in eductor throat.

Computational Fluid Dynamics (CFD) Based Combustion Modeling

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
321

FIGURE 9.30 Re-circulation zone developing in eductor throat.

modeling can be a valuable design tool for combustion equip- Da (-) Damköhler number: ratio of reaction time to
ment in the petrochemical industry. It is also clear that CFD has flow time
not achieved the status of stress analysis in terms of ease of use. f (-) General body force in momentum equation
In many cases, engineers without advanced understanding of g (m/s2) Gravity
the physics do stress analysis of mechanical designs and obtain I (W/m2 sr) Radiation intensity
reasonable results. With CFD analysis, especially the study of k (m2/s2) Turbulent kinetic energy
combustion systems, this is not the case. In a typical furnace lF (m) Reaction zone thickness used to define
model, the science involved is multi-disciplinary, involving heat Damköhler number
transfer, fluid flow, and combustion kinetics. Understanding lt (m) Turbulent length scale used to define the
and interpreting the results of a CFD model require a thorough Damköhler number
understanding of the underlying physics. p (N/m2) Pressure
r (m) Cylindrical coordinate position variable
R (-) Residual or relative equation error
9.8 NOMENCLATURE Sφ Source term in conservation equations for
general property
Symbol (Units) Description
SL(m/s) Laminar flame speed
A (-) Difference coefficient composed of convection/ tflow (s) Characteristic time for flow to adjust to
diffusion terms imposed shear
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 322 Tuesday, March 22, 2005 7:36 AM

322 The John Zink Combustion Handbook

FIGURE 9.31 Plots of contours of streamfunction with increasing backpressure at the burner tip (left to right).

trxn (s) Characteristic time for chemical species to ψ (-) Dimensionless position variable, difference
react with each other equation truncation ψ (-); Truncation error
u (m/s) Axial gas velocity from difference equation
v (m/s) Radial gas velocity
Overlines
v′ (-) Turbulence intensity used to define Damköhler
- Time-averaged value
number
~ Favre- or mass weighted-averaged value
w (m/s) Tangential gas velocity
→ Vector quantity
x (m) Cartesian coordinate position variable
y (m) Cartesian coordinate position variable Superscripts
″

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
z (m) Cartesian coordinate position variable Fluctuating portion of instantaneous value
δij (-) Kronecker delta o Initial value
ε (m2/s2 s) Kinetic energy dissipation rate p Center point in Difference scheme
µ (kg/m s) Viscosity
Subscripts
µb (kg/m s) Bulk viscosity
i,j,k Indices representing coordinate directions in
µe (kg/m s) Eddy viscosity 3-space
ν (m2/s) Eddy diffusivity used in the k-ε turbulence model φ General transport property
ρ (kg/m3) Density E,W,N,S,T,B East, West, North,South, Top, Bottom —
τ (kg/m s2) Viscous stress tensor relative directions in grid
Γφ(-) General transport coefficient for transport
property φ
Φ (-) Conserved scalar REFERENCES
φ (-) General transport property
Ω (-) Represents a solid angle in radiation transport 1. M. Stahlman, Timeline: ten years of successes, SunWorld,
equation Special Commemorative Issue, Fall 1992.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 323 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 323

2. M. Grossman, Supercomputers/emerging application: 14. M.W. Rasband, PCGC-2 and the data book: a concur-
modeling reality, IEEE Spectrum, 56-60, Sept. 1992. rent analysis of data reliability and code performance,
3. E.A. Foumeny, and F. Benyahlya, Application of com- M.S. thesis, Department of Chemical Engineering,
putational fluid dynamics for process equipment Brigham Young University, Provo, UT, 1988.
design, Heat Exchanger Engineering, Vol. 2, Foumeny 15. L.L. Baxter, and P.J. Smith, Turbulent dispersion of
and Heggs, Eds., Ellis Horwood Ltd., 1991, 155-164. particles, Western States Section, The Combustion
Institute, Salt Lake City, UT, 1988.
4. E.A. Foumeny, and F. Benyahlya, Can CFD improve
the handling of air, gas, and gas-liquid mixtures?, 16. R.D. Boardman, Development and evaluation of a
Chem. Eng. Progress, 89(2), 21-26, 1993. combined thermal and fuel nitric oxide predictive
model, Ph.D. dissertation, Department of Chemical
5. J.R. Krawczyk, Applications and limitations of CFD in
Engineering, Brigham Young University, Provo, UT,
the HPI, Hydrocarbon Processing, 76(7), 53-56, 1997.
1990.
6. J.D. Smith, Application of Computational Fluid 17. J.D. Smith, A detailed evaluation of comprehensive
Dynamics to Industrial “Opportunities” at The Dow simulation software describing pulverized-coal com-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Chemical Company, invited lecture at the Basel World bustion and gasification using advanced sensitivity
CFD User Days: Second World Conference in Applied analyses techniques, Ph.D. dissertation, Department of
Computational Fluid Dynamics, Basel, Switzerland, Chemical Engineering, Brigham Young University,
May 1–5, 1994. Provo, UT, 1990.
7. P.J. Smith, T.H. Fletcher, and L.D. Smoot, Model for 18. P.A. Gillis, and P.J. Smith, An evaluation of 3-D com-
pulverized coal fired reactors, 18th Symp. (Int.) Com- putational combustion and fluid-dynamics for indus-
bustion, The Combustion Institute, Pittsburgh, PA, trial furnace geometries, Twenty-Third Symp. (Int.)
1981, 1285. Combustion, The Combustion Institute, Pittsburgh, PA,
8. K.A. Davis, M.J. Bockelie, P.J. Smith, M.P. Heap, July 1990.
R.H. Hurt, and J.P. Klewicki, Optimized fuel injector 19. L.D. Smoot, and P.J. Smith, Coal Combustion and
design for maximum in-furnace NOx reduction and Gasification, Plenum Press, New York, 1985.
minimum unburned carbon, presented at the First Joint 20. R.B. Bird, W.E. Stewart, and E.N. Lightfoot, Transport
Power and Fuel Systems Contractor Conference, Pitts- Phenomena, John Wiley & Sons, New York, 1960.
burgh, July 1996.
21. M. Nallasamy, A Critical Evaluation of Various Turbu-
9. T.H. Fletcher, A two-dimensional model for coal gasifi- lence Models as Applied to Internal Fluid Flows, NASA
cation and combustion, Ph.D. dissertation, Department TP - 2474, Marshall Space Flight Center, Alabama,
of Chemical Engineering, Brigham Young University, May 1985.
Provo, UT, 1983.
22. F.H. Harlow, and P.I. Nakayama, Transport of turbu-
10. S.C. Hill, L.D. Smoot, and P.J. Smith, Prediction of lence energy decay rate, Rpt. # LA-3854, Los Alamos
nitrogen oxide formation in turbulent coal flames, 20th Scientific Laboratory, Los Alamos, NM, January 1968.
Symp. (Int.) Combustion, The Combustion Institute, 23. C.G. Speziale, On Nonlinear k-l and k-ε models of
Pittsburgh, PA, 1984, 1391. turbulence, J. Fluid Mech., 178, 459, 1987.
11. L.L. Baxter, P.J. Smith, P.J., and L.D. Smoot, Compari- 24. B.E. Launder, and D.B. Spalding, Mathematical Models
son of model predictions and experimental data for coal- of Turbulence, Academic Press, London, 1972.
water mixture combustion, Western States Section, The
25. N.I. Lilleheie, I. Ertesvag, T. Bjorge, S. Byggstoyl, and
Combustion Institute, San Antonio, TX, April 1985.
B.F. Magnussen, Modeling and Chemical Reactions:
12. D.G. Sloan, P.J. Smith, and L.D. Smoot, Modeling of Review of Turbulence and Combustion Models, NEI-
swirl in turbulent flow systems, Prog. Energy Combust. DK--286, DE90 760249, SINTEF/The Norwegian
Sci., 12, 163, 1986. Institute of Technology, Division of Thermodynamics,
13. A.S. Jamaluddin, and P.J. Smith, Prediction of radiative July 11, 1989.
heat transfer in cylindrical furnaces, Western States Sec- 26. W.P. Jones, and J.H. Whitelaw, Calculation methods
tion, The Combustion Institute, Banff, Alberta, Canada, for reacting turbulent flows: a review, Combustion and
April 1986. Flame, 48, 1, 1982.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 324 Tuesday, March 22, 2005 7:36 AM

324 The John Zink Combustion Handbook

27. J.W. MacArthur, Development and implementation of on soot formation and combustion, 16th Symp. (Int.) on
robust direct finite-difference methods for the solution Combustion, The Combustion Institute, 1976.
of strongly coupled elliptic transport equations, Ph.D.
dissertation, University of Minnesota, May 1986. 43. P.A. Libby, Introduction to Turbulence, Taylor and
Francis, New York, 1996.
28. C.K. Law, Heat and mass transfer in combustion: fun-
damental concepts and analytical techniques, Progress 44. W.P. Jones and J.H. Whitelaw, Calculation methods for
in Energy and Combustion Science, 10, 295-318, 1984. reacting turbulent flows: a review, Combustion and
29. N. Peters, Length scales in laminar and turbulent Flame, 48, 1-26, 1982.
flames, Prog. Astro. Aero., 35, 155-183, 1991.
45. N. Peters, Progress in Energy and Combustion Science,
30. S. Mazumder and M.F. Modest, Turbulence radiation 10, 319, 1984.
interactions in nonreactive flow of combustion gases,
ASME J. Heat Transfer, 121, 726-729, 1999. 46. R.W. Bilger, The structure of turbulent nonpremixed
flames, 22nd Symp. (Int.) on Combustion, The Com-
31. R. McDermott and M.R. Henneke, High capacity, ultra
low NOx radiant wall burner development, 12th Ethyl- bustion Institute, 1988.
ene Forum, May 11-14, 1999, The Woodlands, TX. 47. E.S. Bish and W.J.A. Dahm, Strained dissipation and
32. S.B. Pope, Computations of turbulent combustion: reaction layer analyses of nonequilibrium chemistry in
progress and challenges, 23rd Symp. (Int.) on Combus- turbulent reaction flows, Combustion and Flame, 100,
tion, The Combustion Institute, 1990. 3, 1995.
33. K.N.C. Bray, The challenge of turbulent combustion,
48. R.W. Bilger, Conditional moment closure for turbulent
26th Symp. (Int.) on Combustion, The Combustion
reacting flow, Physics of Fluids A, 5, 436, 1993.
Institute, 1996.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
34. J. Warnatz, U. Mass, and R.W. Dibble, Combustion: 49. C.G. Speziale, On non-linear k-l and k-ε models of tur-
Physical and Chemical Fundamentals, Modeling and bulence, J. Fluid Mechanics, 178, 459-478, 1987.
Simulation, Experiments, Pollutant Formation, Springer-
50. V. Yakhot, et al., Development of turbulence models for
Verlag, Berlin, 1995.
shear flows by a double expansion technique, Physics of
35. R.S. Barlow, Nitric oxide formation in dilute hydrogen Fluids A, 4(7), 1510-1520, 1992.
jet flames: isolation of the effects of radiation and tur-
bulence-chemistry submodels, Combustion and Flame, 51. T.H. Shih, W.W. Liou, A. Shabbir, and J. Zhu, A new
117, 4, 1999. k-ε eddy-viscosity model for high Reynolds number
36. M.F. Modest, Radiative Heat Transfer, McGraw-Hill, turbulent flows — model development and validation,
New York, 1993. Computers Fluids, 24(3), 227-238, 1995.
37. R. Siegel and J.R. Howell, Thermal Radiation Heat 52. H.K. Versteeg and W. Malalasekera, An Introduction to
Transfer, Hemisphere, Washington, D.C., 1992. Computational Fluid Dynamics, The Finite Volume
38. I.M. Khan, and G. Greeves, A method for calculating Method, Addison-Wesley Longman Limited, England.
the formation and combustion of soot in diesel engines, 1995.
in Heat Transfer in Flames, N.H. Afgan and J.M. Beer,
Eds., Scripta, Washington D.C., 1974, chap. 25. 53. R.D. Van Dell and L.A. Shadoff, Relative rates and par-
39. P.A. Tesner, T.D. Snegiriova, and V.G. Knorre, Kinetics of tial combustion products from the burning of chloro-
dispersed carbon formation, Combustion and Flame, benzenes and chlorobenzene mixtures, Chemosphere,
17, 253-260, 1971. 13, 177, 1984.
40. F.A. Williams, Combustion Theory, Addison-Wesley, 1985. 54. R.D. Van Dell and N.H. Mahle, The role of carbon par-
41. D.B. Spalding, Mixing and chemical reaction in steady ticle surface area on the products of incomplete com-
confined turbulent flames, 13th Symp. (Int.) on Com- bustion (PICs) emissions, in Emissions from
bustion, The Combustion Institute, 1970. Combustion Processes: Origin, Measurement, Control,
42. B.F. Magnussen and N.H. Hjertager, On mathematical Clement, R.E. and Kagel, R.O., Eds., Lewis, Boca
modeling of turbulent combustion with special emphasis Raton, FL, 1990, 93-107.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch09-Frame Page 325 Tuesday, March 22, 2005 7:36 AM

Computational Fluid Dynamics (CFD) Based Combustion Modeling 325

55. R.D. Van Dell and N.H. Mahle, A study of the products 60. P.H. Taylor and B. Dellinger, Thermal degradation
of incomplete combustion and precursors produced in the characteristics of chloromethane mixtures, Environ.
flame and their post flame modification from the combus- Sci. Technol., 22, 438.
tion of o-dichlorobenzene in a high resolution laboratory 61. E.R. Altwicker, R. Kumar, N.V. Konduri, and M.S. Milli-
thermal oxidizer, Comb. Sci. Tech., 85, 327, 1992. gan, The role of precursors in formation of polychloro-
dibenzo-p-dioxins and polychloro-dibenzofurans during
56. C.G. Choudhry, K. Olie, and O. Hutzinger, Mecha-
heterogeneous combustion, Chemosphere, 20(10-12),
nisms in the thermal formation of chlorinated com-
1935, 1990.
pounds including polychorinated dibenzo-p-dioxins,
in Chlorinated Dioxins and Related Compounds: 62. C.M. Young and K.J. Voorhees, Thermal decomposi-
Impact on the Environment, Hutzinger, O., Frei, R.W., tion of 1,2-dichlorobenzene. II. Effect of feed mixtures,
Merian, E., and Pocchiari, F., Eds., Pergamon Press, Chemosphere, 24(6), 681, 1992.
Oxford, 1982, 275-301. 63. D.P.Y. Chang, W.S. Nelson, C.K. Law, R.R. Steeper,
M.K. Richards, and G.L. Huffman, Relationships
57. W.M. Shaub and W. Tsang, Dioxin formation in incin-
between laboratory and pilot-scale combustion of some
erators, Environ. Sci. Technol., 17, 721, 1983.
chlorinated hydrocarbons, Env. Prgs., 8(3), 152, 1989.
58. J.L. Graham, D.L. Hall, and B. Dellinger, Laboratory 64. J.D. Anderson, Computational Fluid Dynamics: The
investigation of thermal degradation of a mixture of Basics with Applications, McGraw-Hill, New York,
hazardous organic compounds, Environ. Sci. Technol., 1995.
20, 703, 1986.
65. J. Swithenbank, A. Turan, and P.J. Felton, Three-
59. S.M. Senkan, Thermal destruction of hazardous dimensional, two-phase mathematical modeling of gas
wastes: the need for fundamental chemical kinetic turbine combustors, paper presented at Project SQUID,
research, Environ. Sci. Technol., 22, 368, 1988. Purdue University, 1978, 1-89.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 10
Combustion Safety
Terry Dark and Charles E. Baukal, Jr.

TABLE OF CONTENTS

10.1 Introduction............................................................................................................................................. 328

10.2 Overview ................................................................................................................................................. 329
10.2.1 Definitions................................................................................................................................. 329
10.2.2 Combustion Tetrahedron........................................................................................................... 330
10.2.3 Fire Hazards .............................................................................................................................. 331
10.2.4 Explosion Hazards .................................................................................................................... 334
10.2.5 Process Hazard Analysis (PHA) ............................................................................................... 337
10.2.6 Codes and Standards ................................................................................................................. 338
10.3 Design Engineering................................................................................................................................. 339
10.3.1 Flammability Characteristics .................................................................................................... 339
10.3.2 Ignition Control......................................................................................................................... 342
10.3.3 Fire Extinguishment.................................................................................................................. 344
10.3.4 Safety Documentation and Operator Training .......................................................................... 344
10.4 Sources of Further Information............................................................................................................... 346
References ................................................................................................................................................................ 347

327
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

328 The John Zink Combustion Handbook

10.1 INTRODUCTION activity.12 There are also many potential dangers caused by
Fires and explosions are a major concern in hydrocarbon and fires and explosions: flying shrapnel, pressure waves from a
petrochemical plants as the consequences can be very severe blast, high heat loads from flame radiation,13–15 and high
and very public because of the high volume of flammable temperatures. All of these can have severe consequences for
liquids and gases handled in those plants.1,2 The process both people and equipment and may need to be considered
industries have invested much money in equipment, instru- in minimizing the potential impact of an incident. Fry16
mentation, training, and procedures to enhance safety. Con- showed how computer models can be used to simulate fires
siderable progress has been made in the recent past in and explosions in the chemical process industry to help
improving the safety of their operations. Unfortunately, the design appropriate measures to prevent these incidents and
industry is not immune to accidents, as evidenced by explo- how to respond if they should occur. Ogle17 presented a
sions that have been documented.3,4 Loss Prevention Bulletin method for analyzing the explosion hazard in an enclosure
has listed all the major incidents worldwide that occurred that is only partially filled with flammable gas. Ogle showed
from 1960 to 1989 in the hydrocarbon chemical process that an explosion pressure at the stoichiometric condition is
industries, including refineries, petrochemical plants, gas approximately 50 times greater than the failure pressure of
processing plants, and terminals.5 Some of these involved most industrial structures. This obviously can have cata-
large property losses and deaths. These types of events have strophic results.
heightened the safety consciousness of these industries to A number of good books are available on safety in com-
both prevent such incidents and to effectively handle them if bustion systems and in the chemical and petrochemical indus-
they should occur.6 The moral, social, economic, environ- tries.18–28 Crowl and Louvar have written a textbook designed
mental, and legal ramifications of an accident make combus- to teach and apply the fundamentals of chemical process
tion safety a critical element in plant design and operation. safety.29 King has written a large book on safety for the
Preventing an incident is definitely preferred to protecting process industries, including the chemical and petrochemical
people and equipment from the consequences of an incident industries, with specific emphasis on U.K. and European stan-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

if it occurs.7 While fires and explosions can occur at many dards and regulations.30 Kletz has written an encyclopedia-
different processes in a plant, this chapter deals specifically format book on safety and loss prevention, containing small
with the fired heaters section. articles on about 400 different topics.31 Nolan has written an
There are many factors that can contribute to an accident:8 extensive guide to understanding and mitigating hydrocarbon
fires and explosions.32 Nolan characterized accidents or fail-
• human error,9,10 ures into the following basic areas: ignorance, economic con-
• equipment malfunction siderations, oversight and negligence, and unusual
• upset plant conditions occurrences. However, it is noted that nearly all incidents are
• fire or explosion near the apparatus preventable. Nolan listed the following principles as the gen-
• improper procedures eral philosophy for fire and explosion protection for oil, gas,
• severe weather conditions and related facilities:

In a report prepared by the American Petroleum Institute,11 1. Prevent the immediate exposure of individuals to fire and
the following causes were noted for 88 incidents that explosion hazards.
occurred in refining and chemical unit operations from 1959 2. Provide inherently safe facilities.
to 1978: 3. Meet the prescriptive and objective requirements of gov-
ernmental laws and regulations.
• 28% equipment failures 4. Achieve a level of fire and explosion risk that is acceptable
• 28% human error to the employees, the general public, the petroleum and
• 13% faulty design related industries, the local and national governments, and
• 11% inadequate procedures the company itself.
• 5% insufficient inspection 5. Protect the economic interest of the company for both
• 2% process upsets short- and long-range impacts.
• 13% education 6. Comply with a corporation’s policies, standards, and
guidelines.
Uehara (1991) analyzed the risks to Japan’s petrochemical 7. Consider the interest of business partners.
plants in the event of a large earthquake, which has a stron- 8. Achieve a cost-effective and practical approach.
ger likelihood in Japan due to the high frequency of seismic 9. Minimize space (and weight if offshore) implications.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 329 Tuesday, March 22, 2005 8:18 AM

Combustion Safety 329

10. Respond to operational needs and desires. Detonation: An explosion that results in a shock wave that
11. Protect the reputation of the company. moves at a speed greater than the speed of sound in the
12. Eliminate or prevent the deliberate opportunities for unreacted medium.
employee- or public-induced damages. Deflagration: An explosion that results in a shock wave that
moves at a speed less than the speed of sound in the
unreacted medium.
10.2 OVERVIEW Explosion: A rapid expansion of gases that results in a
It is the intent of this section to provide a general education rapidly moving shock wave.
regarding the many hazards associated with the unsafe
Fire: The generic term given to the combustion process.
operation of combustion equipment, as well as to discuss the
evaluation tools and regulations used to eliminate hazards Flame: A controlled fire produced by a burner.
and unsafe practices from combustion system operation. Flammable: In general, a material that is capable of being
easily ignited and burning rapidly.
10.2.1 Definitions Flammable liquid: A liquid having a flash point below
Crowl and Louvar,29 Nolan,32 and NFPA 86 33 all provide 140°F (60°C) and having a vapor pressure below 40
extensive definitions of common combustion safety psia (2000 mmHg) at 100°F (38°C).
vocabulary. Some of the most commonly used definitions Flare: A device incorporating a large burner, typically on top of
related to fire and explosion phenomena are provided below. a large exhaust stack, used for the burning of combustible
exhaust gases vented from an industrial process.
Autoignition: The process through which a flammable Flash point (FP): The lowest temperature of a liquid at which
liquid’s vapors are capable of extracting enough energy it gives off enough vapor to form an ignitable mixture
from the environment to self-ignite, without the with air immediately over the surface of the liquid.
presence of a spark or flame.
Ignition: The process of initiating the combustion process
Autoignition temperature: The minimum temperature at through the introduction of energy to a flammable
which a flammable liquid is capable of autoignition. mixture.
Autooxidation: The process of slow oxidation, resulting in Lower flammability limit (LFL): The minimum concentration
the production of heat energy, sometimes leading to of a combustible gas or vapor in air, below which

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
autoignition if the heat energy is not removed from the combustion will not occur upon contact with an ignition
system. source; sometimes referred to as the lower explosive
Burner: A device or group of devices used for the limit (LEL).
introduction of fuel and oxidizer into a furnace at the Minimum ignition energy (MIE): The minimum energy
required velocities, turbulence, and mixing proportion to required to initiate the combustion process.
support ignition and continuous combustion of the fuel. Overpressure: The pressure generated by an explosive
Combustion: A chemical process that is the result of the blast, relative to ambient pressure.
rapid reaction of an oxidizing agent and a combustible Shock wave: A pressure wave moving through a gas as the result
material. The combustion reaction releases energy of an explosive blast. The generation of the shock wave
(in the form of heat and light), part of which is used to occurs so rapidly that the process is primarily adiabatic.
sustain the combustion reaction. Spontaneous combustion: The combustion process
Combustible: In general, a material capable of undergoing resulting from autooxidation and subsequent
the combustion process in the presence of an oxidation autoignition of a flammable liquid.
agent and a suitable ignition source. Upper flammability limit (UFL): The maximum concentration
Combustible liquid: A liquid having a flash point at or of a combustible gas or vapor in air, above which
above 140°F (60°C) and below 200°F (93°C). A combustion will not occur upon contact with an ignition
combustible liquid basically becomes a flammable source; sometimes referred to as the upper explosive
liquid when the ambient temperature is raised above limit (UEL).
the combustible liquid’s flash point. Vapor pressure: The pressure exerted by a volatile liquid as
Confined explosion: An explosion occurring within a determined by the Reid Method (ASTM D-323-58);
confined space, such as a building, vessel, or furnace. measured in terms of pounds per square inch (absolute).
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 330 Tuesday, March 22, 2005 8:18 AM

330 The John Zink Combustion Handbook

fuel source is removed, then the fire goes out. For example, in
a fired heater, the flame goes out when the fuel supply to the
burner is shut off.
The second face of the fire tetrahedron is the oxidizer.
Oxidizers are also present in solid, liquid, and gaseous forms.
Solid oxidizers include metal peroxides and ammonium
nitrate. Liquid oxidizers include hydrogen peroxide, nitric
acid, and perchloric acid. Gaseous oxidizers include oxygen,
fluorine, and chlorine. Oxygen (contained in air) is the oxi-
dizer used in industrial combustion. A fire will be extin-
guished if the oxidizer is removed. This can be done, for
example, by smothering the fire with a blanket or by injecting
an inert gas like N2 or CO2 in and around the fire to displace
the oxidizer. In nearly all cases, the oxidizer is oxygen that
is present in normal air at about 21% by volume. Higher
concentrations of oxygen can cause a flame to burn more
rapidly and violently.34 For example, pure oxygen is used to
enhance many high-temperature industrial combustion pro-
cesses.35 Even metals can burn in pure oxygen. In reality, a
fuel and oxidizer can be in the presence of an ignition source
without combusting, if the mixture is not within the flamma-
bility limits. Table 10.1 provides the flammability limits for
a few common fuels.36 For example, the lower flammability
limit for methane (CH4) in air is 5.0% CH4 by volume, with
the balance being air. The upper flammability limit for CH4
FIGURE 10.1 Fire tetrahedron.
in air is 15.0% CH4 by volume. If the mixture contains less
than 5.0% or more than 15.0% CH4 by volume, then the
mixture is outside the flammability limits and will not com-
10.2.2 Combustion Tetrahedron bust at standard temperature and pressure.
Four basic elements must be present for a combustion The third face of the tetrahedron involves an energy source
process: to both initiate and sustain the combustion reactions.
1. fuel Table 10.2 provides the minimum ignition temperature
2. oxidizer required to initiate the combustion reaction of various gaseous
fuels and oxidizers at stoichiometric conditions and standard
3. heat
temperature and pressure. For example, the minimum ignition
4. reaction chain
temperature for an air/CH4 mixture is 1170°F (632°C). In
This is usually referred to as the “fire tetrahedron” as shown normal burner operation, the flame is ignited with either a pilot
in Figure 10.1. Fire can be defined as a rapid chemical or a spark igniter. A fire or explosion can be initiated if the
reaction between a fuel and an oxidant, where there is fuel/oxidizer mixture contacts a hot surface, an unintended
sufficient heat to both initiate and sustain the reaction. The spark, or static electricity.37 A common fire prevention step is
typical distinction between a flame and a fire is that a flame is to eliminate all ignition sources from an area containing
controlled and desirable, while a fire is uncontrolled and known fuel sources. After the fire has been initiated, energy
undesirable. To prevent or extinguish a fire, one or more of is still required to sustain the flame. That source is normally
the four legs of the combustion tetrahedron must be removed. the exothermic heat release from the reaction itself that makes
Fuel may be in the form of a solid, liquid, or gas. Typical most flames self-sustaining. A common method of extinguish-
solid fuels include coal, wood dust, fibers, metal particles, ing fires is to deluge the area with water, which eventually
and plastics. Typical liquid fuels include gasoline, acetone, cools all surfaces below the ignition point. Water is inert,
ether, fuel oil, and pentane. Typical gaseous fuels include inexpensive, and has a very high heat capacity relative to other
natural gas, propane, hydrogen, acetylene, and butane. If the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
liquids, which makes it a good extinguishing agent.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 331 Tuesday, March 22, 2005 8:18 AM

Combustion Safety 331

TABLE 10.1 Flammability Limits for Common Fuels at Standard Temperature

and Pressure
Flammability Limits
(% fuel gas by volume)
Fuel Lower, in air Upper, in air Lower, in O2 Upper, in O2

Butane, C4H10 1.86 8.41 1.8 49

Carbon monoxide, CO 12.5 74.2 19 94
Ethane, C2H6 3.0 12.5 3 66
Hydrogen, H2 4.0 74.2 4 94
Methane, CH4 5.0 15.0 5.1 61
Propane, C3H8 2.1 10.1 2.3 55
Propylene, C3H6 2.4 10.3 2.1 53

Source: Reed, R.J., North American Combustion Handbook, Vol. I, 3rd ed., North American

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Manufacturing Company, Cleveland, OH, 1986. With permission.

TABLE 10.2 Minimum Ignition Temperatures for Common Fuels at Standard

Temperature and Pressure
Minimum Ignition Temperature
Fuel In air (°F) In air (°C) In O2 (°F) In O2 (°C)

Butane, C4H10 761 405 541 283

Carbon monoxide, CO 1128 609 1090 588
Ethane, C2H6 882 472
Hydrogen, H2 1062 572 1040 560
Methane, CH4 1170 632 1033 556
Propane, C3H8 919 493 874 468

Source: Reed, R.J., North American Combustion Handbook, Vol. I, 3rd ed., North American Man-
ufacturing Company, Cleveland, OH, 1986. With permission.

The fourth face of the tetrahedron is a reaction chain to safest, and least costly mode of operation is to prevent the
sustain combustion. If any of the steps in the chemical chain occurrence in the first place.
reaction are broken, the flame can be extinguished. This is Furnace operators should be carefully trained to monitor
the principle behind certain types of fire extinguishers, such and assess the state of tubes as a regular part of the operational
as dry chemical or halogenated hydrocarbon. These extin- routine. Every occasion of tube overheating, whether it is
guishing agents inactivate the intermediate products of the caused by excessive firing or inadequate cooling action by the
flame reactions. This reduces the combustion rate of heat process fluid, should be recorded and assessed to determine
evolution, which eventually extinguishes the flame by remov- the likelihood of tube failure. Tubes are normally designed to
ing the heat source that makes the flame self-sustaining. last 10 or more years under normal operation, but excessive
temperatures can shorten the life of the tubes to a few days or
less. Coking in the tubes, if not identified and removed expe-
10.2.3 Fire Hazards
diently, can cause hot spots on the tubes that will result in
10.2.3.1 Heat Damage premature failure. Another mechanism for tube failure is
One of the most devastating causes of fires in process plants trapped liquid that freezes and expands (see Figure 10.3).
is process tube rupture (see Figure 10.2). A furnace tube Prevention of tube failure requires adequate instrumenta-
rupture feeds the furnace firebox with an uncontrolled tion and proper furnace design to allow the tube condition to
amount of fuel, usually resulting in enormous damage and, be monitored continuously and accurately. Accurate tempera-
sometimes, loss of life. As with all safety issues, prevention ture measurement of tubes can be a difficult problem. Ade-
is preferred to remediation. Although there are numerous quate instrumentation is expensive, but inadequate
safety features of process equipment that can minimize monitoring of tube condition can be dangerous and much
damage when a tube failure occurs, the most desirable, more expensive. Temperature measurements of the process
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 332 Tuesday, March 22, 2005 8:18 AM

332 The John Zink Combustion Handbook

FIGURE 10.2 Tube rupture in a fired heater. (Courtesy of Butterworth Heinemann.2)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Combustion Safety 333

fluid should be possible at several points in the tube layout.

Furnace wall temperatures should be measurable at several
locations in the firebox, as well as the convection section.
Interlocks should be capable of shutting down the fuel supply
if the process fluid temperature is too high or if the process
fluid pressure rises above safe levels.
Furnaces should be provided with adequate viewing ports
and operators should be trained to visually assess tube condi-
tion during furnace operation. Flame impingement on tubes
is the most common cause of tube coking and failure, and is
usually discovered only by visual observation. When impinge-
ment occurs, adjustments in furnace operations should be
made immediately. Tubes intended to last 10 years can fail in
a matter of hours if they are heated more than 100°F (38°C)
above their designed operating temperature. It is important to
be able to assess flame impingement visually, because even
well-instrumented systems cannot be certain of measuring the
tubes at their point of highest temperature. If the furnace is FIGURE 10.3 Trapped steam in a dead-end that can
shut down for maintenance, the opportunity should be used to freeze and cause pipe failure. (Courtesy of Gulf Publishing.46)
make a careful, close-up inspection of the tubes. Alterations
in the furnace structure, layout, or operating procedure should
only be undertaken with the advice and consent of an engineer fluid had frozen in the piping inside the furnace. The start-
qualified to assess the proposed changes. up procedure called for the operator to establish a flow of the
If, despite careful precautions, tube failure occurs, adequate heat-transfer fluid through the piping in the furnace, light the
safety features should be in place to minimize the hazard and furnace to heat the transfer fluid, then switch the flow of the
the damage. Process liquid flow should be controllable from heated fluid through the piping in the reactor.
a location that is an adequate distance from the furnace. When the circulation pump was started, a flow could not
Remote isolation valves should be included in the furnace be established through the furnace. Correctly assuming that
design and should be inspected and tested regularly. Remote the fluid in the furnace had frozen, the operator started a small
stop buttons should be located at a sufficient distance from fire in the furnace in the hopes of thawing the fluid and
the furnace to allow circulation pumps and any other equip- establishing a flow. This procedure had been used success-
ment that feeds any form of fuel or oxidizer to the furnace fully in the past. About 1 hour after the operator had ignited
area to be shut down, even in a worst-case scenario. the furnace and established a low fuel rate to thaw the pipes,
Sanders38 reported two cases of tube failures in process the operations foreman entered the control room and noticed
heaters that illustrate the necessity of the above guidelines. that the natural gas flow to the furnace was much higher than
In one case, a fire occurred during the start-up procedure of it should be for the thawing operation. The rate was imme-
a plant during a cold winter morning. A natural gas-fired diately cut to about one fourth of the existing flow.
heater was used to provide the heat energy to a heat-transfer Approximately 15 minutes later, black smoke and fire were
fluid. The combustible heat-transfer fluid was to be circulated reported coming from the heater stack. The fire was quickly
through piping in a gaseous phase reactor until a start-up extinguished, but inspection of the heater revealed that two
temperature of 500°F (260°C) for the process was reached. tubes had ruptured. Fire investigators could find no one who
Once the start-up temperature was reached, the exothermic would admit to increasing the fuel flow to the furnace. The
nature of the reaction was used to sustain the reaction and flow recorder for the natural gas supply to the heater was not
the furnace was taken offline. working properly, so neither the exact flow rate nor the times
After start-up, the process fluid was switched to circulate of flow rate change could be ascertained.
through the reactor and a cooler to remove excess heat from An investigation revealed that there were no written
the reactor. On this occasion, the weather had been unseason- instructions for thawing frozen pipes in the heater. The opera-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

ably cold. All of the piping in the system, except for the piping tions crew was not fully aware of the hazards of lighting the
that was inside the furnace, was steam traced and insulated. furnace before a flow was established in the piping. The heater
Apparently, while the system was not in use, the heat-transfer tubes had carbon buildup, which restricted the flow of the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 334 Tuesday, March 22, 2005 8:18 AM

334 The John Zink Combustion Handbook

heat-transfer fluid. There was evidence of thinning of the vided for automatic shutdown due to high tube-wall temper-
tubes in the higher heat flux zones. Because the same tech- ature was equipped with a setpoint adjustment that allowed
nique for thawing the tubes had been used in the past, damage the operator to set the temperature at which alarms would
to the tubes may have accumulated over time. sound and automatic shutdown would occur. For some reason,
In this case, several common safety precautions were not the setpoint temperature had been set to 1600°F (870°C). The
observed, including: alarm temperature should have been set at 830°F (440°C),
and the shutdown temperature should have been set at 850°F
• Adequate instrumentation for measuring the temperature (450°C). No reason was ever discovered for the unreasonable
of the process fluid in the heater tubes was not available. and clearly dangerous temperature setting.
• The flow recorder for the fuel to the heater was not in
Accidents occur due to both human error and equipment
working order.
failure. Safety instrumentation is designed to prevent human
• An alternative startup procedure, which had apparently
error from creating a dangerous situation. In this case, the
been used on more than one occasion, was not properly
documented.
safety equipment that would have prevented the human error
from producing a destructive incident had been defeated by
• The operators and foreman had failed to adequately
oversee the furnace operation during a time in which the improper management of the equipment. An appropriate
furnace was being used for a purpose for which it was safety program that included routine inspection and testing
not designed. of the safety equipment would have caught the improper
setting and prevented a costly accident.
In the second incident reported by Sanders, a similar tube
failure in a furnace caused about $1.5 million in property 10.2.3.2 Smoke Generation
damage and over $4 million in business interruption. The
Smoke is produced in most uncontrolled fires. Smoke is
furnace in this incident performed a function similar to the
generated by incomplete oxidation of the fuel, caused by
furnace in the aforementioned incident. A combustible heat-
insufficient mixing of air and the combustible materials.
transfer fluid was heated in the furnace, then circulated through
Smoke contains fine particles made primarily of solid carbon.
gas-phase reactors where solvents were produced. Because the
In many fires, more people die from smoke inhalation than
gas-phase reaction is self-sustaining once the operating temp-
from the heat produced in the fire. There are several potential
erature is reached, a single furnace was used to provide a start-
problems with smoke. One problem is elevated temperatures,
up supply of heated fluid to each of five reactors, one at a
which can damage the lungs upon inhalation. Another
time. The piping and operational procedures for the system
problem is the deposition of smoke particles on the lungs,
were complex. When a reactor was brought online, the oper-
which can hinder breathing. Smoke can also block or impair
ator was to align the valves to allow circulation of the heat-
vision, which can hinder escape from the fire. Only trained
transfer fluid from the heater to the reactor being started, start
personnel with adequate breathing and eye protection should
the circulation pump, and ignite the heater. On this occasion,
ever deliberately enter smoky conditions produced by a fire.
the operator erred by starting the heater while the heater tubes
Another product of incomplete combustion is carbon mon-
were isolated from the circulation pump by closed block
oxide (CO), which is an extremely toxic gas that can quickly
valves. About 30 minutes after the heater was fired, a water
kill humans via respiratory failure. Smoke generation is an
sprinkler system tripped, followed shortly thereafter by a
indicator of the probability of the presence of CO. CO kills
heater-flame failure alarm and the rupture of a heat-transfer
by blocking the ability of hemoglobin in the blood to carry
fluid pipe in the heater. Heat from the resulting 50-foot flames
oxygen to the cells in the body. CO has a 300 times greater
was spread throughout the unit by 10 to 12 mph winds.
affinity for hemoglobin than does oxygen. Unfortunately, CO
The 15-year-old heater had been relatively well-designed
is colorless and odorless, so one must be careful to avoid
to prevent an accident of the type that occurred. Automatic
possible situations where it may be present, such as in smoky
shutdown equipment and alarms were provided to respond to
fires. Fortunately, inexpensive CO detectors are available to
flame failure, high tube-wall temperature, low fuel supply
warn of its presence (see Figure 10.4).
pressure, and high heat-transfer fluid pressure. Although
records indicated that some maintenance of the shutdown
equipment had been performed, there was no systematic pro- 10.2.4 Explosion Hazards
gram of inspection and testing that would have ensured that Danger of explosion may come from many sources, but explo-
the equipment was properly adjusted and would operate sions most often occur when the equipment involved is in a state
dependably in an emergency. The specific system that proof change such as start-up, shutdown, or maintenance. Because
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Combustion Safety 335

(a)

(c)

also in any connecting plumbing or equipment. Second, test-

ing should be done immediately before the work begins and
periodically during the time the ignition source is in use.

10.2.4.1 Explosions in Tanks and Piping

For example, liquid storage tanks should be thoroughly
flushed, cleaned, and tested before any maintenance begins.
Yet, even storage tanks that have been thoroughly flushed can
have vapor buildup over time from the evaporation of liquid
residue in the cracks, seams, and structural members of the
tank. It is almost impossible to completely clean tanks that
have contained heavy oils or polymers. Such tanks can test
completely clean with a combustible gas detector, yet fill
(b) with explosive vapors when maintenance activities heat the
residues to the vapor point.
FIGURE 10.4 CO detector: (a) permanent, (b) portable,
(c) area monitor. Pipes containing a number of bends, low points, or attached
equipment may test clear of flammable materials in a pro-
posed work area, yet contain significant amounts of explosive
furnaces are made primarily of metals, maintenance often materials that can be vaporized by welding or migrate to the
involves welding. The welding process provides an ignition work area from elsewhere in the piping. Low points or dips
source for any combustible gases or liquids that might remain in in piping are particularly dangerous because they may store
the work area. Typically, tests for the presence of flammable liquids that can vaporize and be ignited by maintenance work.
gases or liquids are conducted before any maintenance is Storage tanks, furnaces, pipes, or other metallic containers
allowed to commence. However, thorough testing to ensure the that have been in contact with acid can have hydrogen buildup
absence of flammable materials can be complex and difficult. due to the action of the acid on the metal.

Two principles should be followed when testing for com- 10.2.4.2 Explosions in Stacks
bustibles in a planned, enclosed maintenance area. First, test- Stacks are designed to vent exhaust gases from industrial pro-
ing should be done not only in the immediate work area, but
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
cesses. If the gases are combustible, flare stacks are used to
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 336 Tuesday, March 22, 2005 8:18 AM

336 The John Zink Combustion Handbook

occurred in which lightning has ignited stack gases over a

distance of 100 to 200 ft (30 to 60 m) above the top of a
furnace stack. As long as there is no oxygen in the stack itself
and the gases are ignited above the top of the stack, such
incidents, although disquieting, are not particularly danger-
ous. Other incidents have occurred in flare stacks in which
fuel gases have reached combustion temperature and
exploded inside the stack. The only way to ensure that stack
explosions do not occur is to prevent oxygen from infiltrating
and mixing with the stack gases.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

10.2.4.3 Explosions in Furnaces

The most common source of furnace explosions is the use of
an improper lighting procedure. Even procedures that have
been used for years without incident can cause explosions if
conditions change. For example, if the fuel source is not
completely isolated from the furnace before the ignition
source is inserted, a leaky valve can allow enough gas into
the furnace to cause an explosion. A safe furnace lighting
procedure for a furnace containing piloted burners should
FIGURE 10.5 Flarestack explosion due to improper include the following generalized steps:
purging. (Courtesy of Gulf Publishing.46) 1. Confirm that the burner fuel lines are completely isolated
from the furnace (either by disconnection, blind flange,
or a double block and bleed valve assembly).
burn the exhaust gases as they are vented into the atmosphere. 2. Confirm that all auxiliary furnace equipment is functioning
In the case of combustible gases, operational procedures properly, including all instrumentation and measurement
should be used that prevent the infiltration of oxygen into the devices. Open both the burner air inlet register/damper and
stack where it could mix with the exhaust gases and cause an the furnace stack damper to the fully open position.
explosion. Figure 10.5 shows the results of an explosion 3. Purge the furnace of any combustible or flammable sub-
caused by improper purging of a flare stack, allowing air to stances. Follow the NFPA recommendation of purging the
leak in the stack, and leading to the explosion when fuel gas furnace with four furnace volumes of fresh air or inert gas
was introduced back into the stack. Stacks should be built with (N2 or CO2).34
4. Test the atmosphere inside the furnace to ensure that there
welded seams rather than bolted joints. Bolted joints, espe-
are no combustibles present.
cially joints having surfaces that are not machined, can leak air
5. Connect the pilot fuel line(s) to the burner. Activate the
into the stack in areas of negative pressure. Operational proce-
permanent pilot igniter or insert a portable pilot igniter or
dures should be used that ensure that a continuous flow of gas premixed ignition torch.
is moving up the stack. A continuous flow of gas will sweep 6. Slowly open the pilot fuel control valve to the manufac-
away small leaks of air and prevent air from moving down into turer-specified pilot fuel pressure; visually confirm stable
the stack from the top. If the industrial process does not pro- ignition of the pilot flame(s).
vide a continuous flow of gas, purge gases should be used to 7. Reestablish the main burner fuel supply (either by con-
ensure a flow velocity of 1 to 3 in./s (8 cm/s). Oxygen sensors necting the main fuel line(s), removing a blind flange, or
should be used to ensure that the oxygen content of the exhaust reversing the double block and bleed valve assembly) to
gases does not rise above 5%. A lower percentage is safer if provide a fuel source to the furnace.
hydrogen is present in the stack gases. 8. Slowly open the main burner fuel control valve to supply
Flare stacks always include some form of pilot burner or the burner with the manufacturer-specified ignition fuel
other ignition source. However, other ignition sources exist pressure. Visually confirm stable ignition of the main
burner flame.
even in vent stacks where ignition of the stack gases is not
intentional. For example, in furnaces that are improperly If the burner ignition attempt is unsuccessful, or if the
operated with too little oxygen for combustion, the stack procedure is aborted prior to successful burner ignition, the
gases can contain unburned fuel. Spectacular incidents have burner fuel supply should be immediately disconnected.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 337 Tuesday, March 22, 2005 8:18 AM

Combustion Safety 337

Subsequent attempts to successfully light the burner must 10.2.5 Process Hazard Analysis (PHA)
begin at step 1. The increasing quantities of hazardous materials used at a
In multi-burner furnaces, the operator must be certain to given plant and the increasing complexity of the plants have
follow the designed lighting procedures. In some furnaces, made safety analyses both more difficult and more important
one burner may be cross-ignited by another burner until all because of the possibility of catastrophic accidents.39 There
burners are lit. In others, however, the burners are too far are various types of analyses used for a process hazard analy-
apart for one to be safely lit from another. Although the eight sis (PHA) of the design of equipment and processes, includ-
steps given are a general procedure for a typical furnace, the ing the effects of human error. Qualitative methods include
designed lighting procedure provided by the manufacturer of checklists, what-if reviews, and HAZOP. Quantitative meth-
the furnace should always be followed. The manufacturer of ods include event trees, fault trees, and failure modes and
the furnace should preapprove any deviation from the effect analysis (FMEA). All of these methods require rigor-
designed procedure. ous documentation and implementation to ensure that all
potential safety problems and the associated recommenda-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Another source of explosions in furnaces is improper air

tions are addressed.
management. If the furnace is starved for air, a pulsating
HAZOP (hazard and operability) and what-if reviews are
huffing sound may result. The flame will be unsteady, chang-
two common qualitative methods used to conduct process
ing from long to short or wide to narrow. The variations are
hazard analyses in the chemical and petrochemical indus-
the result of the available air being completely consumed
tries.29,40–43 These methods are used to review equipment and
without burning all of the fuel. The air-starved flame will then
process designs to identify potential hazards44 and to mini-
be reduced in volume until more air is available. It will then
mize the risk of accidents.45
increase in size until the available air is again consumed. This
In a what-if review, many questions are asked to determine
alternating cycle causes the huffing sound.
what the consequences might be in the event of a particular
The correct action, when a furnace is huffing, is to reduce incident. Some of the parameters commonly assessed include
the flow of fuel until there is enough air for full combustion. pressure, temperature, and flow. Questions are then asked as
If the operator incorrectly increases the air without first reduc- to what happens if that parameter is too high or too low, for
ing the fuel flow, the increased air supply can mix with the example, to see what the consequence would be. If the conse-
large volume of unburned fuel already in the furnace, causing quence could cause an accident, then preventative actions can
an explosion. be taken either to prevent that parameter from ever reaching
If low fuel pressure causes the flame to be extinguished in that state, or to add a safety system to respond in the event
a furnace that burns fuel oil, careful tests for combustible that the parameter does reach that state. For example, if a
vapors should be made immediately before attempting to furnace temperature exceeds a predetermined level indicating
relight the furnace. Although most furnaces are equipped with a problem, then the burners can automatically be shut off to
automatic valves that shut off the fuel flow when the flame prevent damaging the furnace. The system is designed to
is lost, the piping between the valve and the furnace can still prevent the parameter, temperature in this case, from ever
contain fuel that can be vaporized into an explosive source. reaching a critical level. Another example is causing a relief
The vaporization process can take time. Fuel that is still in valve to vent to reduce the pressure, if the pressure becomes
its liquid state at the time of a test can cause a furnace to test too high in a system.
clear of combustibles. Yet, the vaporization process can cause Kletz46 lists four circumstances that are frequent causes of
the furnace to be filled with explosive vapors 10 minutes later. accidents or dangerous conditions:

Obtaining an accurate test with a gas detector can be prob- 1. performing or preparing for maintenance
lematic when oil is being used as fuel. Vapors in the furnace 2. making modifications to furnace design
3. human error
can condense back into their liquid state in the tube of the
4. labeling errors or labeling omissions
gas detector, preventing the vapors from reaching the detector
head and being recorded. If the vaporization temperature of When preparing for maintenance, it is important to remove
the fuel oil being used is near the ambient temperature, an hazards from the maintenance area, isolate the area and/or
accurate reading may be difficult to obtain. Where inaccurate equipment from operational equipment, and carefully follow
readings are suspected, the furnace should be purged until the maintenance procedures. When modifying the furnace
operator is certain that all unburned fuel oil has evaporated design, even when the modification seems minor, the pro-
and has exited the furnace. posed modification should go through design procedures
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 338 Tuesday, March 22, 2005 8:18 AM

338 The John Zink Combustion Handbook

similar to those used for the original installation of the equip- as all associated piping and electrical wiring. The
ment. Without careful, detailed analysis, it is often difficult to standard provides the NFPA’s rules on the design and
determine how a seemingly small change will affect the selection of burners, fuel piping, fittings, valves, and flue
entire process. ventilation devices.
Human error is sometimes caused by inattention or poor • Chapter 5: Safety Equipment and Application — the
training, but is frequently caused by a deliberate attempt to NFPA’s detailed guidance concerning the design of auto-
shortcut a cumbersome procedure or to make an inconvenient mated safety systems (burner management systems) and
the process conditions (safety interlocks) that should trig-
piece of equipment more convenient to use. Accidents caused
ger an automated emergency shutdown (ESD). Require-
by labeling are frequently the result of out-of-date labeling,
ments regarding the design and placement of automated
incorrect labeling, or no labeling at all, thus resulting in the fuel gas safety shutoff valves, high and low fuel pressure
incorrect operation of equipment. switches, flame supervision, excess temperature limit con-
trollers, burner pilots, and flame proving devices are
10.2.6 Codes and Standards included. Guidelines for burner preignition procedures and
There is often confusion regarding the differences between ignition trials are also well-documented.
codes and standards. The National Fire Protection Associa- 2. NFPA 70: National Electric Code (NEC), updated annually.49
tion (NFPA)47 defines codes and standards in the following NFPA 70 provides “practical safeguarding of persons and
manner: property from hazards arising from the use of electricity.”
The NEC covers the installation of electric conductors
Code: A standard that is an extensive compilation of and associated equipment in both the public and private
provisions covering broad subject matter or that is sectors, including all electrical wiring associated with
suitable for adoption into law independently of other fired equipment. The NEC is accepted as law throughout
codes and standards. the United States.
Standard: A document, the main text of which contains 3. NFPA 497: Classification of Flammable Liquids, Gases,
only mandatory provisions using the word “shall” to or Vapors and of Hazardous (Classified) Locations for
Electrical Installations in Chemical Process Areas, 1997
indicate requirements and which is in a form generally
Edition.50 NFPA 497 recommends steps to determine the
suitable for mandatory reference by another standard or
location, type, and scope of hazards presented by electri-
code or for adoption into law. cal installations in operations where flammable or com-
bustible liquids, gases, or vapors are processed or handled.
10.2.6.1 NFPA Codes and Standards NFPA 497 can be considered a companion standard to
The National Fire Protection Association (NFPA) publishes a NFPA 70: National Electrical Code (NEC).
variety of codes and standards that address key safety issues 4. NFPA 54: National Fuel Gas Code, 1999 Edition.51 NFPA
related to fire protection. The NFPA Web site48 contains a 54 sets minimum safety requirements for fuel gas piping
complete listing and description of all available codes and systems, fired equipment, flue-gas ventilation systems,
standards. However, the following NFPA codes and and related equipment. The NFPA considers fuel gas to
standards are essential to the safe operation of combustion include natural gas fuel, manufactured gas, and liquefied
equipment. petroleum gas (propane/butane). NFPA 54 is an American
National Standard, appearing as designation Z223.1.
1. NFPA 86: Standard for Ovens and Furnaces, 1999 Edition.33 5. NFPA 58: Liquefied Petroleum Gas Code, 1998 Edition.52
NFPA 86 is the primary standard that addresses fire and NFPA 58 provides minimum safety requirements for the
explosion hazards related to the operation and design of
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

design and operation of liquefied petroleum gas (LPG)

fired equipment used for heat utilization. Many of the com- facilities, including fuel tank locations, piping, etc. NFPA
ponents of Chapters 3 to 5 can be directly applied to process 58 is the basis of LPG law for the United States.
furnaces in the hydrocarbon and petrochemical industries:
6. NFPA 30: Flammable & Combustible Liquids Code, 1996
• Chapter 3: Location and Construction — the NFPA’s Edition.53 NFPA 30 provides the minimum safety require-
recommendations regarding fired equipment and its ments for liquid fuel installations, including requirements
proximity to personnel, buildings, and external combus- for bulk storage tanks, spill control, emergency relief ven-
tible materials. Also addressed are design considerations tilation, etc. NFPA 30 is accepted as law in 35 U.S. states.
such as structural integrity, explosion relief, observation 7. NFPA 921: Guide for Fire and Explosion Investigations,
port locations, skin temperature restrictions, etc. 1998 Edition.54 NFPA 921 provides guidance for the safe
• Chapter 4: Furnace Heating System — the furnace and systematic investigation and analysis of fires and
heating system refers to both the heating source as well explosions. Topics include fire science, fire patterns, plan-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 339 Tuesday, March 22, 2005 8:18 AM

Combustion Safety 339

ning and execution of investigations, origin and root cause 10.3.1 Flammability Characteristics
determination, etc. There are three primary parameters used to measure the
relative flammability of a substance: the flash point (FP), the
10.2.6.2 Additional Standards and Guidelines upper flammability limit (UFL), and the lower flammability
In addition to the aforementioned NFPA codes and standards, limit (LFL). The flash point is the lowest temperature of a
several voluntary standards and guidelines address the design liquid at which it evaporates enough vapor to form an
and operation of combustion devices. These standards and ignitable mixture with the air immediately over the surface of
guidelines include: the liquid. The upper and lower flammability limits bracket
1. European Committee for Standardization (CEN). The the ignitable concentration range of a gas or vapor mixed
multi-national European organization develops standards with air. Table 10.3 contains experimentally determined flash
addressing industrial safety concerns (including fuel han- point temperatures, as well as upper and lower flammability
dling and combustion) for its 19 national member countries. limits for a wide range of pure component substances in air. 59
2. CSA International. The independent, not-for-profit organi-
zation is the largest standards development organization in 10.3.1.1 Liquids
Canada. The CSA has published many standards addressing The flash point of pure component liquids is usually
combustion and the petroleum refining industry. determined experimentally. However, flash point estimates
3. American Petroleum Institute (API). The API publishes a can be obtained for multi-component mixtures containing a
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

wide variety of standards applicable to combustion pro- single combustible species if both the flash point and the
cesses. API Publication 535: Burners for Fired Heaters
molar concentration of the combustible component are
in General Refinery Service provides guidelines for the
known. Raoult’s law is used to determine the vapor pressure
selection and/or evaluation of burners installed in fired
of the pure component in the diluted mixture (PSAT), based
process heaters.
upon the vapor pressure of the combustible species at its
flash-point temperature (p):
10.2.6.3 Industrial Insurance Carriers
1. Industrial Risk Insurers (IRI). The IRI provides compre- p = xP SAT (10.1)
hensive insurance protection for industrial losses due to
fire, explosion, hail, lightening, windstorm, smoke, etc. where
The IRI generally requires adherence to NFPA codes and PSAT = Vapor pressure of the combustible component
standards. However, the IRI often supplements the NFPA present within the mixture
with its own requirements.55
x = Mole fraction of the combustible component
2. Factory Mutual (FM). The Factory Mutual consists of
present within the mixture
three insurance firms, as well as the Factory Mutual
p = Vapor pressure of the pure combustible
Research Corporation. The FM Research Corporation
conducts reliability and efficiency testing on a variety of
component at its flash point
equipment. The “FM Approval” label provides consumers
with the confidence that equipment bearing that label has Once the vapor pressure of the combustible component
been rigorously tested and found worthy of use in a fire present within the mixture (PSAT) has been calculated, the
protection system. The Factory Mutual Approval Guide resulting flash-point temperature of the mixture can be
contains a listing of FM-approved items, as well as the determined using a vapor pressure vs. temperature diagram.29
details regarding the application and installation criteria Figure 10.6 is a vapor pressure vs. temperature diagram for
for which the equipment is approved. Similar to the IRI, light hydrocarbon fuels.60
Factory Mutual will often supplement the NFPA codes Experimental methods are recommended for flash-point
and standards with its own requirements.55 determination of multi-component mixtures involving two or
more combustible components.29

10.3 DESIGN ENGINEERING 10.3.1.2 Vapors

Several resources are available with general discussions of Similar to the flash point for pure component liquids, the
safety equipment used for industrial combustion applica- upper and lower flammability limits are also determined
tions.36,56–58 The intent of this section is not to detail how experimentally. For multi-component gas mixtures, the
equipment should be designed, but to point out the factors Le Chatelier equation61 is used to estimate the upper and
that should be considered in the design. lower flammability limits of gaseous fuels:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 340 Tuesday, March 22, 2005 8:18 AM

340 The John Zink Combustion Handbook

TABLE 10.3 Flammability and Ignition Characteristics of Liquids and Gases

Normal Boiling Point Autoignition
at 14.695 psia Flash Point LFL AFL Temperature
Compound (°F) (°F) (vol% in air) (vol% in air) (°F)

Acetone 133 –4 2.5 12.8 869

Acetylene –120 — 2.5 100 581
Acrolein 127 –15 2.8 31 428
Aniline 363 158 1.3 11 1139
Benzene 176 12 1.2 7.8 928
Butane 31 –76 1.9 8.5 549
Carbon monoxide –313 — 12.5 74 1128
Chlorobenzeze 269 82 1.3 9.6 1099
Cyclohexane 177 –4 1.3 8.0 473
Ethane –127 — 3.0 12.5 882
Ethyl alcohol 173 55 3.3 19 685
Ethylene –155 — 2.7 36 842
Ethyl oxide 51 –4 3.0 100 804
Ethyl ether –13 –42 3.4 27 662
Formaldehyde –2 185 7.0 73 795
Heptane 209 25 1.1 6.7 399
Hexane 156 –8 1.1 7.5 437
Hydrogen –423 — 4.0 74 —
Isopropyl alcohol 180 54 2.0 12.7 750
Isopropyl ether 45 –35 2.0 10.1 374
Methane –259 — 5.0 15 999
Methyl acetate 134 14 3.1 16 849
Methyl alcohol 148 52 6.0 36 867
Methyl chloride –11 — 8.1 17.4 1170
Methyl ethyl ketone 175 16 1.4 11.4 759
Methyl isobutyl ketone 242 54 1.2 8.0 838
Methyl propyl ketone 216 45 1.5 8.2 846
Napthalene 424 174 0.9 5.9 979

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Octane 258 55 1.0 6.5 403
Pentane 97 –40 1.4 8.0 500
Phenol 359 174 1.8 8.6 1319
Propane –44 –155 2.1 9.5 842
Propylene –54 — 2.0 11.1 851
Propylene dichloride 206 70 3.4 14.5 1035
Styrene 293 88 0.9 6.8 914
Toluene 231 39 1.1 7.1 896
o-Xylene 292 90 0.9 6.7 865
m-Xylene 282 81 1.1 7.0 981
p-Xylene 281 81 1.1 7.0 982

Source: Lide, D.R., Ed., CRC Handbook of Chemistry and Physics, 80th ed., CRC Press, Boca Raton, FL, 1999.

1 UFL = Upper flammability limit for component i

LFLMIX = n (10.2) (vol%)
∑ LFL
yi
yi = Mole fraction of component i on a combustible
i =1 i
basis
n = Number of combustible species present within
1
UFLMIX = n (10.3) the fuel mixture

∑
yi
i =1
UFLi Flammability limit data is often provided at process con-
ditions of 77°F (25°C) and 14.695 psia (1 atm). However, the
where flammability limit ranges increase dramatically with temper-
LFL = Lower flammability limit for component i ature. The following empirical equations describe the tem-
(vol%) perature dependency of the flammability limits in air62:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
Copyright CRC Press
10,000.0

Provided by IHS Markit under license with CRC Press

Combustion Safety

1000.0

No reproduction or networking permitted without license from IHS

n e
Ethyle
ne e
pan
Etha Pro ane
ent
2337-ch10-Frame Page 341 Tuesday, March 22, 2005 8:18 AM

n-P
100.0
ylene
Prop

tan e utane
i-Bu n-B
10.0

Vapor pressure, psia

a ne e
ent n
i-P exan ctio
n-H a
1.0 Fr
e n e 0 °F

Not for Resale, 01/15/2018 21:50:21 MST

an ca 40
ne N on De
ne ta n- n-
pta Oc
e n-
n-H

0.1
0 20 40 60 80 100 200 300 400 500 600

Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher

Temperature, °F
FIGURE 10.6 Vapor pressures for light hydrocarbons. (Adapted from GPSA Engineering Data Book, Vol. II, 10th ed., Gas Processors and Suppliers
Association, Tulsa, OK, 1994.)
341

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
2337-ch10-Frame Page 342 Tuesday, March 22, 2005 8:18 AM

342 The John Zink Combustion Handbook

TABLE 10.4 Ignition Sources of Major Fires Variation in pressure does not significantly affect the LFL
Electrical (wiring of motors) 23% except at very low pressures (below 27 in. w.c. absolute
Smoking 18% [50 mmHg absolute]). However, increases in pressure can
Friction (bearings or broken parts) 10% significantly raise the UFL. The following empirical expres-
Overheated materials (abnormally high temperatures) 8%
sion describes the pressure dependence of the UFL in air63:
Hot surfaces (heat from boilers, lamps, etc.) 7%
Burner flames (improper use of torches, etc.) 7%
Combustion sparks (sparks and embers) 5%
UFL P = UFL1 atm + 20.6(log10 P + 1) (10.6)
Spontaneous ignition (rubbish, etc.) 4%
Cutting and welding (sparks, arcs, heat, etc.) 4%
Exposure (fires jumping into new areas) 3% where
Incendiarism (fires maliciously set) 3%
UFLP = Upper flammability limit at pressure P
Mechanical sparks (grinders, crushers, etc.) 2%
Molten substances (hot spills) 2% (vol%)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Chemical action (processes not in control) 1% UFL1 atm = Upper flammability limit at 1 atm
Static sparks (release of accumulated energy) 1% (vol%)
Lightning (where lightning rods are not used) 1%
P = Actual process pressure (MPa, absolute)
Miscellaneous 1%

Source: National Safety Council, Accident Prevention Manual for Industrial The presence of pure oxygen as the oxidizing agent (as
Operations, National Safety Council, Itasca, IL, 1974.
opposed to air) has very little effect on the LFL, as the oxygen
TABLE 10.5 Minimum Ignition Energies Required for concentration of air is in excess of that required for combus-
Common Fuels tion at the LFL. However, the UFL of most hydrocarbon fuels
Pressure Minimum Ignition Energy in pure oxygen is increased by approximately 45 to 55% when
Compound (atm) (mJ) compared to the equivalent UFL in air.64 Table 10.1 compares
Methane 1 0.29 flammability limits of several common fuels using both air
Propane 1 0.26 and pure oxygen as the oxidizing agent.36
Heptane 1 0.25
Hydrogen 1 0.03
Propane (mol%) 10.3.2 Ignition Control
[O2/(O2 + N2)] × 100% Ignition is the process through which combustion is
1.0 1 0.004
initiated, and occurs when a flammable mixture of fuel and
0.5 1 0.012
0.21 1 0.15 oxidizer comes in contact with a suitable ignition source.
1.0 0.5 0.01 The minimum ignition energy is the minimum energy
Source: Zabetakis, M.G., AIChE-Inst. Chem. Engr. Symp., Ser. 2, Chem Engr. required to initiate combustion, and can be obtained through
Extreme Cond. Proc. Symp., American Institute of Chemical Engineers, New a variety of sources: direct contact with a spark or flame,
York, 1965, 99–104. static electricity, autoignition, autooxidation, and adiabatic
compression. Table 10.4 lists the ignition sources tabulated
from over 25,000 fires by the Factory Mutual Engineering
1 − 0.75(T − 25)  Corporation.65 Table 10.5 contains the minimum ignition
LFL T = LFL 25   (10.4)
 ∆HC 
energy (MIE) required for several common fuels.66
In general, the MIE decreases with pressure and increases
1 + 0.75(T − 25)  with inert gas concentration. MIEs for hydrocarbon fuels are
UFL T = UFL 25   (10.5) relatively low when compared to ignition sources. Walking
 ∆HC 
across a rug can produce an electrostatic discharge of 22 mJ.
where An internal combustion engine’s spark plug can generate an
∆HC = the heat of combustion (kcal/mole) electrical discharge energy of 25 mJ.29
LFLT = lower flammability limit at temperature, T Direct contact with a spark or flame is a very common
(vol%) energy source often used for the intentional ignition of indus-
LFL25 = lower flammability limit at 25°C (vol%) trial combustion equipment. For process burners in fired heat-
UFLT = upper flammability limit at temperature, T ers, the ignition source may be in the form of a small premixed
(vol%) pilot burner, a portable electrostatic ignitor, or a portable
UFL25 = upper flammability limit at 25°C (vol%) premixed gas torch. Flare burners typically use a continuous
T = actual process temperature (°C) flare pilot that is ignited by a flame front generator (FFG)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 343 Tuesday, March 22, 2005 8:18 AM

Combustion Safety 343

FIGURE 10.7 Ethylene oxide plant explosion caused by autoignition. (Courtesy of Gulf Publishing.46)

(see Chapter 20). An FFG is a custom-built device that sends Kletz46 discusses several case histories of fires and explosions
a small burst of flame to the top of a tall flare stack, allowing ignited by electrostatic discharge.
the operator to ignite the pilot burner from ground level. FFG Autoignition is the process through which a flammable
devices have been proven to safely ignite flare pilot burners liquid’s vapors are capable of extracting enough energy from
a distance of 1 mile away from the flame front generator. the environment to self-ignite, without the presence of a spark
Regardless of the source, all ignition devices should be or flame. The ability of a flammable liquid to autoignite is
designed for the particular equipment and the specific set of characterized by the liquid’s autoignition temperature. Table
process conditions for which they will be used. 10.3 contains common autoignition temperatures for a variety
of flammable liquids. Autoignition temperatures depend on a
Static electricity is a common ignition source of fires and number of factors, including fuel vapor concentration, fuel
explosions in chemical processing plants. An electrostatic volume, system pressure, presence of catalytic material, and
charge is formed whenever two dissimilar surfaces move rel-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
flow conditions. Because the autoignition temperature is a
ative to each other. A relevant example is liquid flowing function of so many process variables, it is important that the
through a pipeline, moving past the walls of the pipe. In this autoignition temperature is determined experimentally under
example, one charge is formed on the pipe surface, while the conditions that most closely simulate actual process con-
another equal but opposite charge is formed on the surface of ditions.29 Figure 10.7 shows the wreckage of an ethylene
the moving liquid. When the voltage becomes strong enough, oxide plant explosion caused by autoignition leading to fire
the static electricity will discharge in the form of an electrical and explosion.46
spark. The spark can ignite combustible and flammable mate- Autooxidation is the process of slow oxidation, resulting
rials present. Crowl and Louvar29 and the NFPA67 present in the production of heat energy, sometimes leading to autoi-
detailed explanations of design fundamentals for the preven- gnition if the heat energy is not removed from the system.
tion of fires and explosions due to electrostatic discharge. The most common example of this potential ignition process
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 344 Tuesday, March 22, 2005 8:18 AM

344 The John Zink Combustion Handbook

TABLE 10.6 Benefits of a Successful Process Knowledge flame speed (speed with which the flame front burns back
and Documentation Program towards the fuel source). To prevent flashback in premixed
• Preserves a record of design conditions and materials of construction for burners, flame arrestors are commonly used.68 The primary
existing equipment, which helps ensure that operations and maintenance applications for flame arrestors are to protect people and
remain faithful to the original intent
equipment from flashbacks, fires, and catastrophic explosions.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
• Allows recall of the rationale for key design decisions during inception,
design, and construction of major capital projects, which is useful for a A flame arrestor is designed to stop (extinguish) a flame or
variety of reasons (i.e., an aid in future projects and modifications) explosion from further propagation past the arrestor. It is a
• Provides a basis for understanding how the process should be operated and
why it should be run in a given way
special type of heat exchanger that cools the flame, thus
• Offers a “baseline” for use in evaluating a process change removing one of the legs of the fire tetrahedron. Time is
• Records accident/incident causes and corrective actions and other operating required to dissipate the heat, so the design and construction
experience for future guidance
of the quenching media are important.69 In fuel piping sys-
• Protects the company against unjustified claims of irresponsibility and
negligence tems, the arrestor must be perforated or porous to allow gas
• Retains basic research and development information on process chemistry flow through it. The technique often used in flame arrestors is
and hazards to guide future research effort to cool the propagating flame or explosion enough to extin-
Source: Center for Chemical Process Safety, Plant Guidelines for Technical guish the fire. Thermal mass, usually in the form of metal, is
Management of Chemical Process Safety, rev. ed., American Institute of
used to extract enough energy from the reacting gases that the
Chemical Engineers, New York, 1995. With permission.
flame can no longer be supported and is extinguished. Many
different arrestor designs are available, including gauzes,
is when rags saturated with oils are discarded or stored in a perforated plates, expanded metal, sintered metal, metal foam,
warm area. If allowed to autooxidize, the increased temper- compressed wire wool, loose filling, hydraulic arrestors,
atures can result in autoignition of the rags, and a damaging stacked plate, and crimped ribbon.70
fire or explosion can result. Relatively high-flash-point mate-
rials are the most susceptible to the autooxidation process,
10.3.4 Safety Documentation and
while low-flash-point materials can often evaporate without Operator Training
ignition. Fuel leaks that saturate thermal insulation or other Safety documentation and operator training provide the back-
absorbent materials should be isolated immediately, and the bone of a strong safety program, and are absolutely essential
contaminated absorbent should be removed promptly and in order to maintain a safe combustion working environment.
discarded in a suitable manner.64 Table 10.6 illustrates some of the benefits of a successful pro-
Adiabatic compression of combustible or flammable mate- cess documentation program.71
rials can result in high temperatures, which in turn may result The AIChE Center for Chemical Process Safety publishes
in autoignition of the compressed fuel. Examples of adiabatic several titles that address implementation of process safety
compression include internal combustion engines and gas documentation.71–74 Safety documentation for combustion-
compressors. The temperature rise associated with the adia- related processes includes design information, process hazard
batic compression of an ideal gas can be determined using analysis (PHA) reports, standard operating procedures, and
thermodynamic relationships: training documentation. Feedback from each of theses docu-
mentation elements are linked together as part of a plant’s
( γ −1)
T2  P2  γ overall process safety program. Figure 10.8 visually describes
=
T1  P1 
(10.7) the documentation feedback linkage suggested by the AIChE.72

where 10.3.4.1 Design Information

T2 = Final absolute temperature Design information, or process knowledge, refers to all of
T1 = Initial absolute temperature the documents that pertain to the safe design of the
P2 = Final absolute pressure combustion system. This set of information can include, but
P1 = Initial absolute pressure is not limited to:71,72
γ = Specific heat ratio, Cp /Cv • process information:
– detailed information regarding the design criteria
10.3.3 Fire Extinguishment of the combustion system (i.e., heat load, process
In premixed burners, there is a chance that the fire can flash flow rates, temperatures and pressures, fuel com-
back into the burner. Flashback can occur when the velocity position, etc.)
of the fuel mixture leaving the combustor is exceeded by the – process chemistry
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 345 Tuesday, March 22, 2005 8:18 AM

Combustion Safety 345

Feedback

Design PHA Operating Training

Documents Reports Procedures Documentation

Feedback

FIGURE 10.8 Safety documentation feedback flow chart. (Adapted from Center for Chemical Process Safety, Guidelines
for Process Safety Documentation, American Institute of Chemical Engineers, New York, 1995.)

– safe operating ranges for process conditions (flow, 10.3.4.3 Standard Operating Procedures (SOPs)
temperature, pressure, compositions, etc.) Day in and day out, standard operating procedures (SOPs)
– the known hazardous effects of deviation from the provide operators with clear, detailed, sequenced instructions
stated safe operating conditions regarding the safe operation and maintenance of combustion
• equipment information: equipment. In addition to providing detailed instructions for
– process flow diagrams (PFDs) the operation of a combustion system, SOPs also assist in the
training of both new and existing plant personnel. The AIChE
– piping and instrumentation diagrams (P&IDs)
provides guidance regarding the preparation, revision,
– detailed equipment drawings (i.e., heater, flare, or content, and distribution of SOPs.72
incinerator drawings illustrating details such as tube
As with any set of SOPs, written procedures for combustion
locations, burner orientation, etc.)
systems should include pre-startup, start-up, normal operation,
– electrical wiring schematics and electrical classifi- shutdown, and emergency shutdown procedures, as well as
cation data
additional procedures for preventative maintenance opera-
– manufacturer’s equipment manuals (including design tions. The equipment manufacturer should always be con-
criteria, safe operation recommendations, etc.) sulted regarding the proper ignition and operation of individual
– equipment, valve, and instrumentation specification combustion devices. Figure 10.9 shows the charred wreckage
sheets of a refinery where an accident was caused due to improper
– maintenance manuals maintenance procedures.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

– automated safety interlock system details

10.3.4.4 Operator Training and Documentation
10.3.4.2 Process Hazard Analysis (PHA) Reports Training is essentially the successful communication and
Regardless of the evaluation technique chosen, the PHA transfer of knowledge and skills. Applied to combustion
process (see Section 10.2.5) puts the combustion system (as systems, training is the successful communication and trans-
well as the entire processing unit) under the microscope, fer of knowledge regarding basic combustion science, safety
systematically analyzing the entire system piece by piece. To hazards associated with fires and explosions, and skillful
ensure that all process safety recommendations are executed, execution of standard operating procedures. In general, train-
thorough documentation must be conducted in an organized ing exercises should emphasize the need to complete all tasks
manner. The PHA process is conducted both prior to initial in a safe manner. Training is not a substitute for poor com-
start-up of the system, and periodically (usually every 5 years) bustion system design. Rather, the design, documentation,
during the system’s lifetime. This cyclic approach guarantees and training should complement each other to provide a safe
continuous safety feedback for the operators and engineers working environment.
entrusted with the safe operation of the combustion system. Employees should be selected for a level of training that
The AIChE provides detailed guidance on the selection and is commensurate with the level of exposure they will encoun-
execution of various PHA evaluation techniques.74 ter on the job. New employees should successfully complete
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 346 Tuesday, March 22, 2005 8:18 AM

346 The John Zink Combustion Handbook

FIGURE 10.9 Refinery damaged due to improper maintenance procedures. (Courtesy of Butterworth-Heinemann.2)

their training regimen before being allowed to perform train- 10.4 SOURCES OF FURTHER
ing-required tasks unsupervised. Training may involve formal INFORMATION
(attendance and examination required), informal (safety dis-
In addition to the references cited in this chapter, there are
cussions, demonstrations, seminars), self-study, and on-the-
many organizations that have good information on safety.
job training methods.72 Training sessions can include but are

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Many of these have general safety information, but some also
not limited to the following topics:
have specific information related to combustion. Typical
• start-up operations (including combustion device ignition) organizations are listed below:
• normal operations American Society of Safety Engineers (ASSE)
• shutdown operations (including combustion device 1800 E. Oakton Street
extinguishment) Des Plaines, IL 60018
• maintenance (847) 699-2929
www.asse.org
• OSHA HAZWOPER (Hazardous Waste Operations and
Emergency Response) American National Standards Institute (ANSI)
• fuels handling 11 West 42nd Street, 13th Floor
• emergency procedures and evacuation New York, NY 10036
(212) 642-4900
• confined space entry
www.ansi.org
• lock-out and tag-out (hazardous energy sources)
• hazard communication Board of Certified Safety Professionals (BCSP)
208 Burwash Avenue
• blood-borne pathogens
Savoy, IL 61874
• fire extinguishment (217) 359-9263
www.bcsp.com
The safety training program must be thoroughly docu-
mented to ensure that it is conducted in an acceptable and Center for Chemical Process Safety (CCPS)
timely manner. Management and trainers should solicit con- American Institute of Chemical Engineers
tinuous feedback in order to evaluate and improve program Three Park Avenue
effectiveness. The AIChE provides guidance on the successful New York, NY 10016-5991
management and implementation of a comprehensive safety (212) 591-7319
training program.71,72 www.aiche.org/cps
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 347 Tuesday, March 22, 2005 8:18 AM

Combustion Safety 347

DuPont Safety Resources 7. F.K. Crawley and G.A. Dalzell, Fire and explosion
Christiana Executive Campus hazard management in the chemical and hydrocarbon
131 Continental Drive, Suite 307 processing industry, IMechE Conf. Trans.:
Newark, DE 19713 Management of Fire & Explosions, 5, 61-72, 1997.
(800) 532-SAFE
8. T. Kletz, Learning from Accidents, 2nd ed.,
www.dupont.com/safety
Butterworth-Heinemann, Oxford, 1994.
Occupational Safety & Health Administration (OSHA)
9. T. Kletz, An Engineer’s View of Human Error, 2nd ed.,
U.S. Department of Labor
Institution of Chemical Engineers, Rugby, U.K., 1991.
Office of Public Affairs - Room N3647
200 Constitution Avenue 10. Center for Chemical Process Safety, Guidelines for
Washington, D.C. 20210 Preventing Human Error in Process Safety, American
(202) 693-1999 Institute of Chemical Engineers, New York, 1994.
www.osha.gov
11. American Petroleum Institute, Safety Digest of Lessons
National Fire Protection Association (NFPA) Learned: Section 2, Safety in Unit Operations,
1 Batterymarch Park Publication 758, Washington, D.C., 1979.
P.O. Box 9101
12. Y. Uehara, Fire safety assessments in petrochemical
Quincy, MA 02269-9101
plants, Fire Safety Sci., Proc. 3rd Int. Symp., Int’l Assoc.
(800) 344-3555
Fire Safety Sci., Edinburgh, U.K., 1991, 83-96.
www.nfpa.com
13. W.P. Crocker and D.H. Napier, Thermal radiation
Society of Fire Protection Engineers (SPFE)
hazards of liquid pool fires and tank fires, IChemE
7315 Wisconsin Avenue, Suite 1225 W
Symp. Series No. 97, Institute of Chemical Engineers,
Bethesda, MD 20814
Pergamon Press, Oxford, U.K., 1986, 159-84.
(301) 718-2910
www.sfpe.org 14. Center for Chemical Process Safety, Thermal
Radiation 1: Sources and Transmission, American
Institute of Chemical Engineers, New York, 1989.
REFERENCES 15. Center for Chemical Process Safety, Thermal
Radiation 2: The Physiological and Pathological
1. Center for Chemical Process Safety, Explosions in the Effects, American Institute of Chemical Engineers,
Process Industries, American Institute of Chemical New York, 1996.
Engineers, New York, 1994.
16. M.A. Fry, Benefits of fire and explosion computer
2. R.E. Sanders, Chemical Process Safety: Learning from modeling in chemical process safety, Proc. of
Case Histories, Butterworth-Heinemann, Boston, 1999. American Chemical Industries Week ’94, Oct. 18-20,
3. T. Richardson, Learn from the Phillips explosion, Philadelphia, PA, 1994.
Hydrocarbon Proc., 70(3), 83-84, 1991. 17. R.A. Ogle, Explosion hazard analysis for an enclosure
4. Workplace Health, Safety and Compensation partially filled with a flammable gas, Process Safety
Commission of New Brunswick, Explosion and Fire at Progress, 18(3), 170-177, 1999.
the Irving Oil Refinery in Saint John, New Brunswick: 18. D.R. Stull, Fundamentals of Fire and Explosion,
Interim Report - EDB 98-22, Fredericton, New AIChE Monograph Series, No. 10, Vol. 73, American
Brunswick (Canada), published by Government of Institute of Chemical Engineers, New York, 1977.
New Brunswick, 1998.
19. W. Bartknecht, Explosions, Springer-Verlag, New
5. Anonymous, A thirty year review of large property
York, 1980.
damage losses in the hydrocarbon chemical process
industries, Loss Prevention Bulletin, Part 99, 3-25, 1991. 20. F.T. Bodurtha, Industrial Explosion Prevention and
Protection, McGraw-Hill, New York, 1980.
6. Fire Protection Association, The hydrocarbon
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

processing industry: fire hazards and precautions, Fire 21. F.P. Lees, Loss Prevention in the Process Industries,
Protection Assoc. J., No. 82, 252-264, 1969. Vol. 1, Butterworths, London, 1980.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 348 Tuesday, March 22, 2005 8:18 AM

348 The John Zink Combustion Handbook

22. G.L. Wells, Safety in Process Design, George Goodwin, 38. R.E. Sanders, Chemical Process Safety: Learning from
London, 1980. Case Histories, Butterworth-Heineman Publishers,
Boston, 1999.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
23. H.H. Fawcett and W.S. Wood, Safety and Accident
Prevention in Chemical Operations, 2nd ed., John 39. J.H. Burgoyne, Reflections on process safety, IChemE
Wiley & Sons, New York, 1982. Symp. Series No. 97, Inst. of Chem. Engineers,
24. Center for Chemical Process Safety, Guidelines for Pergamon Press, Oxford, U.K., 1986, 1-6.
Engineering Design for Process Safety, American 40. Center for Chemical Process Safety (CCPS),
Institute of Chemical Engineers, New York, 1993. Guidelines for Chemical Process Quantitative Risk
25. Center for Chemical Process Safety, Guidelines for Analysis, 1st ed., AIChE, New York, 1989.
Safe Process Operations and Maintenance, American 41. Center for Chemical Process Safety (CCPS),
Institute of Chemical Engineers, New York, 1995. Guidelines for Hazard Evaluation Procedures, 2nd ed.,
26. Center for Chemical Process Safety, Guidelines for AIChE, New York, 1992.
Process Safety Fundamentals in General Plant 42. D.P. Nolan, Application of the HAZOP and What-If
Operations, American Institute of Chemical Engineers, Safety Reviews to the Petrochemical and Chemical
New York, 1995. Industries, Noyes Publications, Park Ridge, NJ, 1994.
27. American Petroleum Institute (API), RP 750, 43. T. Kletz, Hazop and Hazan, 4th ed., Taylor & Francis,
Management of Process Hazards, First Edition, API, Philadelphia, PA, 1999.
Washington, D.C., 1990, reaffirmed 1995.
44. G. Wells, Hazard Identification and Risk Assessment,
28. R. Skelton, Process Safety Analysis, Gulf Publishing,
Gulf Publishing, Houston, TX, 1996.
Houston, TX, 1997.
45. G. Wells, Major Hazards and Their Management, Gulf
29. D.A. Crowl and J.F. Louvar, Chemical Process Safety:
Publishing, Houston, TX, 1997.
Fundamentals with Applications, Prentice-Hall,
Englewood Cliffs, NJ, 1990. 46. T. Kletz, What Went Wrong: Case Histories of Process
Plant Disasters, 4th ed., Gulf Publishing, Houston TX,
30. R. King, Safety in the Process Industries, Butterworth-
1998.
Heinemann, London, 1990.
47. From the NFPA website: http://www.nfpa.org/Codes/
31. T.A. Kletz, Critical Aspects of Safety and Loss
Background/background.html.
Prevention, Butterworths, London, 1990.
32. D.P. Nolan, Handbook of Fire and Explosion 48. From the NFPA website: http://www.nfpa.org/.
Protection Engineering Principles for Oil, Gas, 49. National Fire Protection Association, NFPA 70:
Chemical, and Related Facilities, Noyes Publications, National Electric Code (NEC), 2000 edition, NFPA,
Westwood, NJ, 1996. Quincy, MA, 2000.
33. National Fire Protection Association, NFPA 86 50. National Fire Protection Association, NFPA 497:
Standard for Ovens and Furnaces, 1999 edition, Classification of Flammable Liquids, Gases, or Vapors
NFPA, Quincy, MA, 1999. and of Hazardous (Classified) Locations for Electrical
34. M.A. Niemkiewicz and J.S. Becker, Safety overview, Installations in Chemical Process Areas, 1997 Edition,
in Oxygen-Enhanced Combustion, C.E. Baukal, Ed., NFPA, Quincy, MA, 1997.
CRC Press, Boca Raton, FL, 1998, 261-278. 51. National Fire Protection Association, NFPA 54:
35. C.E. Baukal, Ed., Oxygen-Enhanced Combustion, National Fuel Gas Code, 1999 edition, NFPA, Quincy,
CRC Press, Boca Raton, FL, 1998. MA, 1997.
36. R.J. Reed, North American Combustion Handbook, 52. National Fire Protection Association, NFPA 58:
Vol. I: Combustion, Fuels, Stoichiometry, Heat Liquefied Petroleum Gas Code, 1998 edition, NFPA,
Transfer, Fluid Flow, 3rd ed., North Amer. Mfg. Co., Quincy, MA, 1998.
Cleveland, OH, 1986. 53. National Fire Protection Association, NFPA 30:
37. T.H. Pratt, Electrostatic Ignitions of Fires and Flammable & Combustible Liquids Code, 1996
Explosions, Burgoyne Inc., Marietta, GA, 1997. edition, NFPA, Quincy, MA, 1996.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch10-Frame Page 349 Tuesday, March 22, 2005 8:18 AM

Combustion Safety 349

54. National Fire Protection Association, NFPA 921: 65. National Safety Council, Accident Prevention Manual
Guide for Fire and Explosion Investigations, 1998 for Industrial Operations, National Safety Council,
edition, NFPA, Quincy, MA, 1998. Itasca, IL, 1974.

55. J.W. Coons, Fire Protection Design Criteria, Options, 66. M.G. Zabetakis, Flammability Characteristics of
Selection, R. S. Means Company, Kingston, MA, 1991. Combustible Gases and Vapors, U.S. Bureau of Mines
Bulletin 627, USNT AD 701, 576, 1975.
56. J.R. Cornforth, Ed., Combustion Engineering and Gas
67. National Fire Protection Association, NFPA 77:
Utilisation, E&FN, London, 1992.
Recommended Practice on Static Electricity, 1993
57. IHEA, Combustion Technology Manual, 5th ed., edition, NFPA, Quincy, MA, 1993.
Industrial Heating Equipment Assoc., Arlington, VA, 68. V.A. Mendoza, V.G. Smolensky, and J.F. Straitz,
1994. Don’t detonate — arrest that flame, Chem. Eng.,
103(5), 139-142, 1996.
58. M.A. Niemkiewicz and J.S. Becker, Equipment design,
in Oxygen-Enhanced Combustion, C.E. Baukal, Ed., 69. V.A. Mendoza, V.G. Smolensky, and J.F. Straitz,
CRC Press, Boca Raton, FL, 1998, 279-313. Understand flame and explosion quenching speeds,
Chem. Eng. Prog., 89, 38-41, 1993.
59. D.R. Lide, Ed., CRC Handbook of Chemistry and
Physics, 80th ed., CRC Press, Boca Raton, FL, 1999. 70. H. Phillips and D.K. Pritchard, Performance
requirements of flame arresters in practical applications,
60. GPSA Engineering Data Book, Vol. II, 10th ed., Gas IChemE Symp. Series No. 97, Inst. of Chem. Engineers,
Processors and Suppliers Association, Tulsa, OK, Pergamon Press, Oxford, U.K., 1986, 47-61.
1994. 71. Center for Chemical Process Safety, Plant Guidelines
61. H. Le Chatelier, Estimation of firedamp by flammability for Technical Management of Chemical Process Safety,
limits, Ann. Mines, 19(8), 388-395, 1891. rev. ed., American Institute of Chemical Engineers,
New York, 1995.
62. M.G. Zabetakis, S. Lambiris, and G.S. Scott, Flame
72. Center for Chemical Process Safety, Guidelines for
temperatures of limit mixtures, Seventh Symp.
Process Safety Documentation, American Institute of
Combustion, Butterworths, London, 1959, 484.
Chemical Engineers, New York, 1995.
63. M.G. Zabetakis, Fire and explosion hazards at 73. Center for Chemical Process Safety, Guidelines for
temperature and pressure extremes, AIChE-Inst. Chem. Technical Management of Chemical Process Safety,
Engr. Symp., Ser. 2, Chem. Engr. Extreme Cond. Proc. American Institute of Chemical Engineers, New York,
Symp., 1965, 99-104. 1989.
64. R.H. Perry, D.W. Green, and J.O. Maloney, Eds., 74. Center for Chemical Process Safety, Guidelines for
Perry’s Chemical Engineers’ Handbook, 7th ed., Hazard Evaluation Procedures, 2nd ed., American
McGraw-Hill, New York, 1997. Institute of Chemical Engineers, New York, 1992.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 11
Burner Design
Richard T. Waibel and Michael Claxton

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
TABLE OF CONTENTS

11.1 Introduction............................................................................................................................................. 352

11.2 Combustion ............................................................................................................................................. 352
11.3 Burner Design ......................................................................................................................................... 353
11.3.1 Metering: Fuel........................................................................................................................... 353
11.3.2 Metering: Air (Combustion O2) ................................................................................................ 356
11.3.3 Air Control ................................................................................................................................ 358
11.3.4 Mixing Fuel/Air ........................................................................................................................ 359
11.3.5 Maintain (Ignition).................................................................................................................... 360
11.3.6 Mold (Patterned and Controlled Flame Shape) ........................................................................ 361
11.3.7 Minimize (Pollutants) ............................................................................................................... 362
11.4 Burner Types ........................................................................................................................................... 362
11.4.1 Premix and Partial Premix Gas ................................................................................................. 363
11.4.2 Raw Gas or Nozzle Mix............................................................................................................ 364
11.4.3 Oil or Liquid Firing................................................................................................................... 364
11.5 Configuration (Mounting and Direction of Firing) ................................................................................. 367
11.5.1 Conventional Burner, Round Flame.......................................................................................... 367
11.5.2 Flat Flame Burner ..................................................................................................................... 368
11.5.3 Radiant Wall.............................................................................................................................. 369
11.5.4 Downfired ................................................................................................................................. 369
11.6 Materials Selection.................................................................................................................................. 369
References ................................................................................................................................................................ 370

351
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 352 Tuesday, March 22, 2005 10:08 AM

352 The John Zink Combustion Handbook

11.1 INTRODUCTION oxidant. The oxidant is usually atmospheric air. However,

What is a burner? In its simplest form, a burner is a device the combustion O2 can be from a number of alternative
used to provide heat. Specifically, it is a device used to sources, including:
provide a controlled exothermic oxidation reaction.
• Ambient atmospheric air is the predominant source.
Using this definition, one could argue that a wooden torch
• Preheated atmospheric air is commonly used to
is a burner — even if there is some question as to the use of
improve the thermal efficiency and typically requires a
the exothermic reaction condition being met. However, it is
forced-draft application.
also beneficial to assume that the device is not itself consumed
• Turbine exhaust gas (TEG) is the high-temperature,
by the reaction. As a more reasonable approach, a burner is
reduced O2 content exhaust gas from a gas turbine. In
a device that provides three basic design functions: general, this is a good source for combustion O2; however,
1. A burner must provide for controlled mixing of the reac- low oxygen content with low-temperature exhaust gas can
tant, fuel, and the oxidizing agent, in most cases air. result in an oxidant that cannot support combustion.
2. The burner must provide a stable and self-renewing igni- • Diesel engine exhaust is used periodically as an
tion source. alternative to TEG. The major problem with this as a
3. The burner should provide for a controlled region of reac- source of combustion O2 is in the low temperatures
tion or controllable flame shape. associated with lowered oxygen levels and the pulsing
flow associated with internal combustion exhaust.
The modern concept of patterning the reaction flame is a
• Kiln and drier off-gas is seldom seen, but can be a
direct outgrowth of continuous hydrocarbon and petrochem- usable source for combustion O2. Again, attention must
ical processes. Processing of naturally occurring crude oils be paid to the O2/temperature relationship.
and organic by-products was initially accomplished through • Oxygen-enriched streams are not currently common in
“batching.” In this method, the necessity of controlling the typical industrial combustion applications; however,
actual flame to precise dimensions was not critical. If, at the they do exist in specialized processes.1
end of a single batch process, there were carbon residues or
undesirable deposits, the fire was extinguished and the vat Each of these sources for combustion O2 has been designed
cleaned. With the advent of tubed, continuous throughput by burner designers and utilized by industry. For the purposes
process furnaces, it was no longer economically desirable, of this chapter, the majority of the discussion is limited to the
and in some cases physically possible, to perform mechanical use of ambient atmospheric air. “Special” considerations
cleaning. This made it necessary to reduce or eliminate the within a burner design that are particular to one or more of
carbon residues and deposits generated by localized over- the alternative sources are periodically noted.
heating of the fluid being processed. Direct conductive heat The temperature vs. %O2 (gross) relationship with respect to
transfer from flame impingement is a major source of this sustainable combustion is particularly unforgiving. Figure 11.1
localized overheating. shows the theoretical limitations of methane combustion with
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

a combustion O2 stream varying in oxygen and temperature.

Practical application does not allow for operation near the
11.2 COMBUSTION theoretical limit.
Combustion as a controlled process is often considered a Practical writings on combustion often include discussions
“black art.” While there are substantial amounts of of the 3-Ts of combustion. The 3-Ts are simply a condensa-
experience-based information and design “rules of thumb” tion of fluid flow and chemical reaction principles:
involved, there are very real chemical and physical laws
1. Turbulence is the interaction between two fluid streams
involved. The layman’s view of flames and combustion
required to achieve intermixing of the two.
stability are actually soundly based in the principles of
2. Temperature is the required energy for the initiation of a
chemistry, chemical reactions, and fluid flow.
chemical reaction — oxidation.
Combustion has been defined as a relatively fast exother-
3. Time is the period for the reaction to reach completion.
mic gas-phase chemical reaction. It can occur in either flame
or non-flame mode. For the purposes of discussion in burner In other words, mix the fuel with the air and ignite the
design, only flame mode will be considered and is defined mixture by heating it to sufficient temperature, and the result
as a dual reactant flame. This flame is an exothermic reaction is a flame with a specific volume dictated by the time for the
propagating subsonically through the mixture of two reac- reaction to complete. The principles applied to burner design
tants. The two reactants are, of course, the fuel and the are consistent.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 353 Tuesday, March 22, 2005 10:08 AM

Burner Design 353

• maximum, normal, and minimum heat release rates

• maximum fuel pressures available
• maximum atomizing medium pressures available for
liquid firing
• fuel temperature
• oxidant source, either ambient air, exhaust gases, or
enriched
• available combustion air pressure, whether forced (positive)
or induced (negative)
• combustion air temperature (ambient or preheated)
• furnace firebox temperature
• furnace dimensions for flame size restrictions
• type of flame (configuration or shape)

To provide acceptable operation, the burner must be designed

to perform the 5-Ms:
1. Meter the fuel and air into the flame zone.
2. Mix the fuel and air to efficiently utilize the fuel.
FIGURE 11.1 Graph of sustainable combustion for 3. Maintain a continuous ignition zone for stable operation
methane. over the range.
4. Mold the flame to provide the proper flame shape.
5. Minimize pollutant emissions.
The major differences in burner design come from the
specific requirements addressed by the design. Reaction 11.3.1 Metering: Fuel
rates, or flame lengths and diameters, are directly related Typically, the furnace operating system is able to monitor
to the 3-Ts. They are also dependent on the fuels them- only the total flow of fuel to a furnace. A typical process
selves. The composition, distribution, and port velocity of heater has multiple burners installed to provide the proper
the fuel can and will produce different results. The flame heat distribution. The fuel system must then be designed to
generated by a combustion system on one fuel will not ensure that the fuel is properly distributed to all burners.
duplicate the flame generated by that same system on Uniform fuel pressures to each burner are critical to the
another fuel. Their general flame shape, length, and diam- proper operation of the burners. The burner designer then
eter or width, may be within acceptable tolerances; how- ensures that each burner takes the correct amount of fuel
ever, they will not be completely identical. The wider the from the system. Controlling the proper amount of fuel
variance in the physical properties of the fuels, the greater flow is accomplished through a system of metering orifices
the deviation in flame properties. designed for each burner. These ports are specifically
designed to act as metering and limiting orifices, passing a
specified and known amount of fuel at a given fuel
11.3 BURNER DESIGN pressure.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Specialization of processes, and furnace designs to meet the
11.3.1.1 Gas Fuel
demands of those processes, has resulted in the necessity of
The system of ports provided by the burner designer allows
specialized burners. A burner is designed to provide stable
him to provide the operator with a capacity curve that
operation and an acceptable flame pattern over a specific set
specifies heat release vs. pressure for a given fuel
of operating conditions. In addition, there may be a specified
composition and temperature. For gaseous fuels, in
maximum level of pollutant emissions that can be generated
compressible or incompressible flow, the calculations for the
through the combustion process. The American Petroleum
mass flow through a given orifice are dependent on:
Institute gives some guidelines for burners used in fired
heaters.2 Specifications of operating conditions include: • Po, the fuel pressure immediately upstream of the orifice
• Pa, the downstream pressure (generally atmospheric
• specific types of fuels pressure)
• specific range of fuel compositions • To, the fuel temperature upstream of the orifice
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 354 Tuesday, March 22, 2005 10:08 AM

354 The John Zink Combustion Handbook

FIGURE 11.2 Typical raw gas burner tips.

• K, the fuel’s ratio of specific heats, which is dependent 11.3.1.2 Liquid Fuel
on the composition of the fuel (this is a factor used in Liquid fuels must be vaporized in order to burn. Burners
calculating the compressibility of the fuel gas) designed for firing liquid fuels include an atomizer designed
• A, the area of the port to produce a spray of small droplets, which enhances the
• Cd, the discharge coefficient, which depends on the design vaporization of the fuel. The design of the atomization system
of the orifice port will have a significant impact on the liquid fuel flow metering
of the burner. With liquid fuels, the metering design is more
The fuel metering orifices for gaseous fuels in raw gas complicated because of the need to “mix” the oil with an
burners are typically installed at the point where the flame is atomizing medium. This results in fuel metering orifices and
formed. This is generally located in a region of the burner atomizing media (typically steam or air) metering orifices in
tile often termed the “burner throat.” The fuel injector tips combination with orifices designed to flow the mixture.
with the fuel orifices can be centered in the burner throat or Figure 11.5 depicts typical liquid fuel atomizer/spray tip
located on the periphery. Figure 11.2 shows typical raw gas configurations for various liquid fuel systems. Based on the
burner tips. burner designer’s knowledge of the oil metering and
atomization system, the designer is capable of providing the
For premix or partial premix burners, the metering orifice operator with a capacity curve that specifies heat release vs.
serves dual functions. It is generally a single port located at fuel pressure for a given liquid fuel.
the entrance of a venturi eductor. The fuel gas discharging
Many factors are important in the operation of a liquid fuel
from the orifice is used to entrain air for combustion. A typical
atomization system. Variations in any of a number of factors
premix metering orifice spud and air mixer assembly is shown
must be considered when designing the system of ports for
in Figure 11.3.
the metering of flow for a liquid fuel, including the:
Capacity curves are generally presented in terms of heat
release vs. fuel pressure. The heat release is calculated from • temperature/viscosity relationship of the fuel
the mass (or volume) flow of the fuel, multiplied by the heating • atomization medium temperature
value per unit mass (or volume) of the fuel. Figure 11.4 pre- • fuel/atomizing medium pressure ratio
sents a typical gas fuel capacity curve. • temperature/vaporization relationship of the fuel
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Burner Design 355

Primary air
Venturi throat
control

Orifice

FIGURE 11.3 Typical premix metering orifice spud and air mixer assembly.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 11.4 Typical gas fuel capacity curve.

Copyright CRC Press

356 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
(a)
(c)

FIGURE 11.5 Typical liquid fuel atomizer/spray tip

configurations.

Because capacity curves are generally presented in terms

of heat release vs. fuel pressure, the heat release is calculated
from the mass/volume capacity of the system, multiplied by
the heating value per unit mass of the fuel. Figure 11.6 pre-
sents a typical liquid fuel capacity curve.

11.3.2 Metering: Air (Combustion O2)

The burner designer must ensure that the required air flow for
the design operating conditions can be achieved with the air
pressure that is available. The design of the burner throat is
critical for two independent reasons. First, there is the
requirement of achieving the proper flow of combustion O2 to
meet the demands of the fuel. Second, the burner throat is
important in controlling the pattern of the flame.
The primary air flow metering point and the point at which
the bulk of the air pressure loss is utilized is designated as
the “throat.” In most burners, especially raw gas or nozzle
mix and liquid fuel burners, it is the point at which the fuel
is first injected into the air stream. In a premix or partial
(b) premix burner, it is the point at which the fuel/air mixture is
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 357 Tuesday, March 22, 2005 10:08 AM

Burner Design
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- 357

FIGURE 11.6 Typical liquid fuel capacity curve.

injected and, if necessary, secondary or completion air is

introduced. In all cases, it is the location at which the initiation
of combustion — or ignition — occurs. Figure 11.7 illustrates
the throat of a typical raw gas burner.

11.3.2.1 Natural Draft

Many burners installed in refinery or petrochemical process
heater service utilize the natural draft available in the furnace
as the motive force for drawing combustion air through the
burner. This negative pressure is typically less than 1 in.
water column (<1.0 in. w.c.) at the burner location. This
pressure is dependent on the height of the radiant firebox, the
temperature of the radiant firebox, and the level of draft
(or negative pressure) maintained at the top of the radiant
firebox. For burners required to operate under a natural-draft
application, the burner throat is designed to pass the correct
amount of combustion air required at the design heat release,
utilizing only the available negative pressure provided by the
furnace. In some cases, the natural draft can be enhanced by
an induced draft fan; however, the available pressure still FIGURE 11.7 Typical throat of a raw gas burner.
rarely exceeds 1 in. (1 in. w.c.) of draft.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 358 Tuesday, March 22, 2005 10:08 AM

358 The John Zink Combustion Handbook

11.3.2.2 Forced Draft Normal duty: 8.0 MMBtu/h (2.344 MW, 2.016 MMkcal/h)
Some burner applications provide the combustion O2 stream at Minimum design duty (turndown): 2.0 MMBtu/h (0.586 MW,
a positive pressure. These are generally applications either 0.504 MMkcal/h)
with preheated air, turbine exhaust gas, or another alternative Draft: 0.5 in. w.c. (124.5 Pa, 1.25 mbar, 12.7 mm H2O)
oxidant source. Applications requiring smaller flame volumes 20% Excess air design, 10% excess air operation
can also utilize increased air pressure drop and enhanced Design to meet API 560 and 535
mixing. If the application operates on forced draft at all times, Reference: API 560 Section 10.1.12
the available pressure loss is often in the range of 4 to 10 in. “The burner shall be selected to use no less than 90% of
water column (4 to 10 in. w.c.). With this level of available the maximum draft available for the maximum specified heat
pressure loss, the flame dimensions are considerably smaller release.”
due to the increased turbulent mixing in the flame. In addition,
Reference: API 560 Appendix A — Equipment Data Sheets
it is often possible to control the combustion O2 stream flow
Section b. Burner data sheets, Sheet 3 of 3
over a wider heat release range. This enhanced mixing and
Note 1
improved control allows operation with low excess O2 over a
wider range of fuel input, resulting in improved combustion “At design condition, minimum of 90% of the available
efficiencies across the heat release range. draft with the air register fully open shall be utilized across
In some forced-draft applications, it is specified that the the burner. In addition, a minimum of 75% of the airside
burner must also operate under an ambient air, natural-draft pressure drop with the air registers full open shall be utilized
operating mode. For such a dual operational mode specifica- across burner throat.”
tion, it is most common that the natural-draft case is the Reference: API 535 Section 8.2 — Dampers and Registers
limiting design. Burners designed for natural-draft and “Dampers and burner registers shall be sized such that the
forced-draft operation seldom require forced-draft pressures air rate can be controlled over a range of at least 40 to 100%
greater than 2 in. water column (<2 in. w.c.). The exception of burner capacity.”
to this is when the forced-draft operation uses the lowered
oxygen concentrations of some of the alternative oxidant Assume that, through the efforts of the burner designer,
sources. For this design condition, the increased mass flow the burner actually requires between the specified 90 to 100%
of the combustion O2 source, with oxygen content lower than of the available draft, and that the specified 75% is measur-
air, along with the increased temperature of that stream, gen- able as static loss across the throat. The following tables
erally increase the required burner pressure to levels greater indicate the:
than 2 in. water column (>2 in. w.c.).
• operating heat release (column 1, Heat Release)
11.3.3 Air Control • desired operational excess air (column 3, Percent X-Air)
• required pressure loss across the throat (column 4,
As a primary metering restriction, a majority of the pressure
Throat Drop)
loss is expected to be utilized within the burner throat. • required pressure loss across the register or air control
Engineering groups such as the American Petroleum Institute device (column 5, Control Drop)
(API),2,3 process design companies, and some furnace • ratio percentage of register or damper opening required
manufacturers provide burner design guidelines that define (column 7, Percent Control Open)
the use of the available airside pressure drop at that point.
Depending on the level of accuracy desired, there are limi- Table 11.1 shows the theoretical percentage of air control
tations to the practical range of control, especially when applied opening based on no change in control flow coefficient and
to burners designed for natural-draft applications. The percent- 3% leakage at the full closed position. Table 11.2 is based on
age of available draft utilized for throat metering can leave little 6% leakage. These tables assume that the variable orifice, the
room for the design of an adequate control mechanism. air control mechanism, has a constant discharge coefficient
throughout its full range of operation. This is a simplification
Example 11.1 and is not totally correct in practical application.
Natural-draft burner These charts (Tables 11.1 and 11.2) provide insight into
Refinery fuel gas: 1200 Btu/scf, 0.70 sp.gr. (47.27 MJ/Nm3, another common fallacy in burner specification — the error
11288.8 kcal/Nm3) in excess design capacity. Many who specify combustion
Maximum design duty: 10.0 MMBtu/h (2.93 MW, 2.52 equipment seek to provide excess capacity as a cushion
MMkcal/h) --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- against reduced operation with burners out of service, upset
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 359 Tuesday, March 22, 2005 10:08 AM

Burner Design 359

TABLE 11.1 Theoretical Air Control Opening, Based on No Change in Control Flow Coefficient and 3% Leakage
at the Full Closed Position
Heat Percent Percent Throat Drop Control Drop Percent Control Open
Release Design X-Air 90% Total 100% Total 90% Total 100% Total Total dP 90% Total 100% Total
(MMBtu/hr) (%) (%) (in. w.c.) (in. w.c.) (in. w.c.) % (w/3% leakage)

10.0 100 15 0.338 0.375 0.162 0.125 0.50 86.3 Full open
10.0 100 10 0.309 0.343 0.191 0.157 0.50 75.9 83.8
9.0 90 10 0.250 0.278 0.250 0.222 0.50 59.4 63.1
8.0 80 10 0.198 0.220 0.302 0.280 0.50 47.7 49.6
7.0 70 10 0.151 0.168 0.349 0.332 0.50 38.6 39.6
6.0 60 10 0.111 0.124 0.389 0.376 0.50 31.0 31.6
5.0 50 10 0.077 0.086 0.423 0.414 0.50 24.5 24.8
4.0 40 20 0.059 0.065 0.441 0.435 0.50 20.7 20.9
3.0 30 40 0.045 0.050 0.455 0.450 0.50 17.6 17.7
2.0 20 80 0.033 0.037 0.467 0.463 0.50 14.7 14.8
0 0 3 0.00031 0.00034 0.49969 0.49966 0.50 Full closed Full closed

TABLE 11.2 Theoretical Air Control Opening, Based on No Change in Control Flow Coefficient and 6% Leakage
at the Full Closed Position
Heat Percent Percent Throat Drop Control Drop Percent Control Open
Release Design X-Air 90% Total 100% Total 90% Total 100% Total Total dP 90% Total 100% Total
(MMBtu/hr) (%) (%) (in. w.c.) (in. w.c.) (in. w.c.) % (w/6% leakage)

10.0 100 15 0.338 0.375 0.162 0.125 0.50 84.8 Full open
10.0 100 10 0.309 0.343 0.191 0.157 0.50 74.4 82.3
9.0 90 10 0.250 0.278 0.250 0.222 0.50 57.9 61.6
8.0 80 10 0.198 0.220 0.302 0.280 0.50 46.2 48.1
7.0 70 10 0.151 0.168 0.349 0.332 0.50 37.1 38.1

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
6.0 60 10 0.111 0.124 0.389 0.376 0.50 29.5 30.1
5.0 50 10 0.077 0.086 0.423 0.414 0.50 23.0 23.3
4.0 40 20 0.059 0.065 0.441 0.435 0.50 19.2 19.4
3.0 30 40 0.045 0.050 0.455 0.450 0.50 16.1 16.2
2.0 20 80 0.033 0.037 0.467 0.463 0.50 13.2 13.3
0 0 6 0.0012 0.0014 0.4988 0.4986 0.50 Full closed Full closed

operation, process design contingencies, and future desired exceptions to this rule. Intimate mixing does not always
capacity. The fallacy is in the ability to control excess air and produce the most desirable results.
to utilize the available pressure drop for the normal operation. Designing a combustion system for special applications
The air and fuel pressures are both reduced in the mixing (e.g., high inert composition fuels) can also set limitations on
zone and the resulting flame quality will suffer. the desired level of mixing. All combustion reactions have a
rate at which they will proceed and a minimum temperature
that is required to initiate or sustain that reaction. The intro-
11.3.4 Mixing Fuel/Air duction of inert components to the fuel can generate two
Mixing is a general term used to describe the function of conditions that will affect burner stability. First, the inert
bringing the fuel (reactant) into close molecular proximity components slow the reaction. This slowing of the flame
with the air (oxidant). The higher the level of turbulence and speed can result in the stabilization point being translated out
shear between the streams, the more uniform the fuel and air of the desired position. Second, the inerts introduce a heat
mixture and the more rapid and complete the combustion absorption component that narrows the flammability limits
reaction. It is generally accepted that the higher the level of and reduces the flame temperature. Quenching of the flame,
mixing, the more intense and more complete and better the to extinction, is an important consideration when inert com-
combustion. However, there are a number of conditions and ponents are present in the fuel or the combustion O2 stream.
factors that should be considered, including flame shaping Many combustion O2 streams with greater inert composi-
and control of pollutant emissions that lead to some tions than ambient air are the result of a combustion process.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 360 Tuesday, March 22, 2005 10:08 AM

360 The John Zink Combustion Handbook

As such, they are typically available at an elevated tempera- region in which the components of the streams become
ture. This elevated thermal energy will often partially offset intermixed. However, in low-velocity differentials, as in all
the potential quenching effect of the elevated inert levels. transitional or laminar boundary layers, the relative thickness of
However, there are limits, as discussed in Section 11.1. this interface is small in comparison with the total cross-
An effective form of emissions control is through the section. Co-flow mixing with low-velocity differential
delaying of the combustion process. Highly mixed fuel and conditions is — by its dynamic configuration — slow mixing.
combustion O2 in the proper proportions will generate the
maximum flame temperature. This will result in the formation 11.3.4.3 Cross-flow
of a large quantity of nitrogen oxides (NOx), which is a highly The shear energy generated between two flowing streams is
regulated pollutant (see Chapter 6). Certain burner designs greater anytime the streams are intersecting. The work
will utilize a reduced level of fuel and air mixing, or a delayed required to redirect the combined flow of the streams results
mixing, extending the reaction zone to achieve reductions in in turbulent intermixing of those streams. The included angle
these emissions. of intersection of the streams, the relative differential
In the case of a dual reactant flame, there are only two velocity of the streams, and the mass densities of the streams
sources from which to obtain the energy required for mixing. are all factors in the resulting direction and the turbulence
First is the relatively high-mass, low-velocity combustion O2 generated for the mixing of the streams.
stream, and second is the low-mass, relatively high-velocity Included angles of intersection closer to perpendicular
fuel stream. Each stream will contribute energy to the work result in higher levels of mixing. However, quick mixing is
required to mix in proportion to its mass and its velocity. not always the most important function being sought. Flame
The intimate mixing of two or more dissimilar fluid streams shape, stability, and emissions are all primary functions that
under flowing conditions occurs in a turbulent shear zone can be affected by the rate and direction of combustion.
defined by the intersection of the two streams. This region
can be described as the dynamic interaction of each stream’s 11.3.4.4 Flow Stream Disruption
mass and velocity, or momentum. This surface of intersection Additional turbulence can be developed through the disruption
will vary in its turbulence in proportion to the magnitude of of the “path” of a flowing fluid. Strategically located obstruc-
the shear forces developed. tions (e.g., bluff bodies) and sudden expansions (e.g., tile
Four basic mechanisms are available for the development ledges) provide forced changes in flow streams, generating
of fuel and air mixing: turbulence. If these disruptions are located in a region where
both fuel and air streams are present, mixing will occur.
1. entrainment
2. co-flow mixing
3. cross flow mixing 11.3.5 Maintain (Ignition)
4. flow streamline disruption or eddy formation The most important function that a burner performs is to
provide for the continuous and reliable ignition of the fuel
11.3.4.1 Entrainment and air passing through the burner over a specified range of
Entrainment is an effective demonstration of the law of conser- operating conditions. Each burner is designed to provide a
vation of momentum. One stream, usually the fuel, is utilized specific location in which a portion of the fuel and air is
to inspirate the other. This motive fluid’s energy and momen- continuously introduced in near-stoichiometric proportions at
tum are conserved. As the velocity of the fuel jet is dissipated, velocities at or below the mixture flame speed, thereby
the mass of the fuel stream must increase by the entrainment of allowing for a continuous “ignition zone.” At this location, a
ambient air to satisfy the conservation of momentum. continuous flame is maintained and ignites the fresh fuel/air
A well-designed premix burner primarily utilizes entrain- mixture as it is introduced. The ignition zone is designed to
ment as its fuel/air mixing mechanism. Intimately mixed fuel operate over the specified range of operation of the burner.
and oxidant prior to the distribution tip results in the highest The flame from this ignition zone is then used to ignite the
possible rate of reaction. remainder of the fuel and air mixture.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

In natural-draft burners, the ignition zone is often situated

11.3.4.2 Co-flow in the eddy formed downstream of a step or ledge in the
When the streams are effectively parallel in directional flow, the burner tile (see Figure 11.8), or in the wake of a bluff body
amount of shear force developed is proportional to only the “flame stabilizer or flame holder” (see Figure 11.9) located
differential mass velocity of the streams. At high-velocity in the center of the combustion air stream. Forced-draft burn-
differentials, the interface of these streams becomes a turbulent ers can utilize similar techniques or other aerodynamic meth-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 361 Tuesday, March 22, 2005 10:08 AM

Burner Design 361

FIGURE 11.8 Ledge in the burner tile. FIGURE 11.9 Flame stabilizer or flame holder.

ods, such as a swirler (see Figure 11.10), to generate reverse

flow zones that recirculate hot combustion products that assist
in providing continuous ignition zones.
Safe ignition of the fuel/air mixture depends on the fuel and
combustion O2 source composition. These factors are primary
in affecting the flammability limits, ignition temperatures, and
stoichiometry of the mixture. The rate of reaction (flame speed)
is controlled by the stoichiometry of the mixture and combus-
tion characteristics of the fuel. The ability to sustain a single
point of ignition for a fuel/air mixture is controlled by both
the stoichiometry and the velocity of the mixed stream. Wide
ranges in capacity, excess air requirements, and fuel compo-
sitions make strict dependence on the mixed stream for sta-
bility unreliable. The development of low-velocity zones and
recirculating eddies ensures a single point for the kindling of
the flame over a broad range of operating conditions. Each
burner is designed for a limited range of fuel (and “air”)
compositions and rates, and one should not attempt to operate
the burner outside this range.
FIGURE 11.10 Swirler.
11.3.6 Mold (Patterned and Controlled
Flame Shape) all burners providing energy to a process will be restricted in
The design of combustion equipment, in the form of burners, flame qualities by the design of the chamber, or furnace, into
for the hydrocarbon processing industry (HPI), the chemical which it is fired.
processing industry (CPI), and to a major extent the power The furnace chamber and burner flame in every process
generation industry (PGI) has critical restrictions on the furnace are each designed to provide for the efficient transfer
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

shape, size, and consistency of the flame generated. In fact, of heat to the process load. Flame size and shape for the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 362 Tuesday, March 22, 2005 10:08 AM

362 The John Zink Combustion Handbook

HPI is especially critical due to the sensitivity of the hydro- require a fan-shaped flame, often termed a flat flame. In
carbons being processed to overheating. The rate of heat that case, the burner tile is generally rectangular and the
transfer to the process tubes must be limited to prevent fuel is injected in a manner producing a flame that is essen-
overheating of the process tubes leading to the formation of tially rectangular or “flat” rather than round (see
carbon or coke inside the process tubes. As a result, there Figure 11.12).
are generally strict guidelines for the flame dimensions.
Typical specifications for flames include maximum flame 11.3.7 Minimize (Pollutants)
lengths and widths. The number, heat release, and layout of Most societies have come to the point at which the environ-
the burners in the furnace are designed to provide the proper ment is of foremost concern. Therefore, the governments of
heat transfer pattern. most nations and localities are very critical of any source that
Patterning of the air flow through approach distribution, puts certain undesirable materials into the air, soil, or water.
tile throat sizing and shape, and the tile exit configuration Environmental regulations limiting air pollution have direct
provides the most reliable method for flame pattern control. impact on the design of combustion equipment.
Introduction of the fuel into the established air flow streams The challenge presented to burner design by these restric-
provides the primary function in a raw gas burner. The tions comes from the thermo-chemical reactions that form the
proper flame pattern is generated by the combination of regulated emissions. Emission control issues are discussed in
fuel injection pattern provided by the fuel injectors and the Chapter 6.
burner tile and flame holder which controls the air flow.
The fuel injectors are also often called spuds or tips. The
injectors have fuel injection ports that introduce the main 11.4 BURNER TYPES
portion of the fuel into the air stream in a manner that Burners are typically classified based on the type of fuel
generates the desired flame pattern or shape. In conjunction, being burned. A subdivision in burner type often includes the
the air stream must be shaped in an appropriate manner by method of combustion O2 supply. Therefore, there can be as
the air flow passages provided by the shape of the tile and many as eight basic design criteria.
flame holder. In many cases, the flame shape is round or 1. Gas — premix and partial premix, natural draft and/or
brush-shaped and acceptable in length and diameter (see low pressure drop air
Figure 11.11). In this case, the burner tile is typically round 2. Gas — raw gas or nozzle mix, natural draft and/or low
and the fuel is injected symmetrically. Some furnaces pressure drop air

FIGURE 11.11 Round-shaped flame. --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Burner Design 363

FIGURE 11.12 Flat-shaped flame.

3. Gas — raw gas or forced nozzle mix, forced draft high balanced by an increase in mass of the moving stream. This
pressure drop air additional mass is entrained air.
4. Liquid — natural draft and/or low pressure drop air
Burners designed with an inspirating arrangement typically
5. Liquid — forced draft high pressure drop air
6. Solid fuel — forced draft high pressure drop air (typically) require fewer adjustments to the air control. The utilization of
7. Combination gas and liquid — raw gas and oil (typically), the mass and velocity of the fuel results in that portion of the
natural draft and/or low pressure drop air air inspirated being proportional to the gas flow. Therefore,
8. Combination gas and liquid — raw gas and oil (typically), with the reduction in fuel mass flow, there is a resulting reduc-
forced draft high pressure drop air tion in the entrained air. The efficiency of the venturi and the
restriction imposed by the fuel/air distribution nozzle are the
The final two designs listed are simply extensions of other
limits to the capacity (volume) of air that can be inspirated.
designs. The combination of any two or more burners is
simply a matter of basing the design on the most difficult of Another benefit to this design is in the volume of the flame
the fuel types and adapting the other fuel distribution systems generated. Since the majority of the air is initially intimately
to the base design. mixed with the fuel, the resulting flame volume of this pre-
mixed burner will be much smaller than that of any other low
11.4.1 Premix and Partial Premix Gas air pressure drop burner. Conversely, if the efficiency of this
Premix is a term applied to burners that inspirate part or all of premixing is sufficiently low, the flame will actually be larger
the total air required for combustion. This type of burner than a raw gas burner. This is a result of the reduced secondary
provides for intimate mixing of the fuel and combustion O2 air mixing energies. By utilizing the majority of the available
prior to the ignition zone. energy of the fuel stream to achieve the primary premix, the
In inspiration, the motive energy is supplied by the low-mass, remaining energies for any required secondary air mixing are
high-potential-energy fuel. The fuel gas is metered through one reduced to only that available from the air due to the draft loss.
or more orifices at the entrance to a venturi or mixer. The Further benefits provided by this style of burner lie in the
entrained air stream is made available at zero, or virtually zero, fuel metering configuration. Because all of the fuel is metered
velocity at the same location. The conversion of the potential through a single or minimal number of orifices, the size of
energy, pressure of the fuel stream, to kinetic energy, jet veloc- those orifices will be maximized for the conditions. Larger-
ity, is achieved within the zone of air supply. The “free jet” of diameter orifices, as long as they do not jeopardize the func-
the fuel immediately begins to expand and decelerate. Conser- tion of the burner, are a benefit because they minimize the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
vation of momentum requires that the reduction in velocity be chances of plugging from dirty fuels.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 364 Tuesday, March 22, 2005 10:08 AM

364 The John Zink Combustion Handbook

One of the basic limitations to this type of burner is founded gen will typically require multiple points of fuel injection.
in the burning characteristics of fuels. Each fuel chemical This breaking up of the fuel flow into multiple metering
compound has its “rate of reaction” with oxygen. This rate orifices while maintaining high potential energy, pressure,
is dependent on concentration levels of the fuel, the oxygen, can require small orifices. These small orifices are highly
and any inert components. This rate is also dependent on the subject to fouling problems due to fuel quality and foreign
temperature of the mixture. Another way to describe this rate material in the piping.
is in terms of “flame speed,” or the velocity at which the flame Natural draft on the low pressure drop, combustion O2 side
will propagate. The design of the distribution tip/system into often requires a burner design utilizing quiescent zones for
the ignition and combustion zones is dependent on this same stabilization. These low flow, low pressure zones can be gen-
burning characteristic of the design fuel. Changes in the firing erated through the use of flow stream disruption. Cones or
rate or the fuel will result in a change in the volume, velocity, bluff bodies located in the throat of the burner are a common
and burning characteristics at the distribution tip. If the form of flow stream disruption. Flow disrupters or shields
fuel/air mixture is within combustible limits, and the velocity around fuel tips and ledges or sharp changes in tile profile
is not maintained above the flame propagation speed, the are also common. In all cases, these mechanisms are basically
result is a translation of the flame front back along the fuel/air designed to prevent a portion of the combustion O2 from
flow streams. In the worst case, this flash-back condition may leaving the designed ignition zone prior to the introduction
result in flames being translated completely outside of the of fuel. They provide “pockets” of a continuously renewing
designed ignition and flame zone, causing mechanical and combustible mixture at a location in which the velocity is
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

structural thermal damage. lower than the flame propagation speed.

High combustion O2 pressure drop designs have another
11.4.2 Raw Gas or Nozzle Mix available form of flame stabilization — vortex recirculation.
Raw gas (nozzle-mix) burners are designed to introduce the Through the utilization of swirl mechanisms and tile profiles,
fuel and air separately into the combustion zone. They the combustion O2 stream can be forced into a vortex. The
provide all mixing of fuel and combustion O2 at or after the physical properties of a vortex provide a low pressure zone
ignition point of the burner. Typically, these burners provide on the interior of the rotation that actually generates a back-
for the major, or metering, pressure drop for both the fuel and flow to the point of origin. The shape and strength of the
air immediately prior to the ignition zone. vortex is determined by the flow and pressure loss through
the swirler, the flow around the perimeter of the swirler, and
With all of the fuel maintained as segregated from the
the profile of the surrounding air throat. By injecting fuel into
combustion O2 until mixed in the ignition zone, there is no
the swirling O2 stream, a portion of the fuel and oxygen is
possibility of flashback. Therefore, the raw gas style of design
continuously recirculated to the centralized ignition zone.
can effectively handle a wide range of fuels without concern
for equipment damage or personnel safety. Turndown on a
raw gas burner is limited only to the method of stabilization, 11.4.3 Oil or Liquid Firing
the flammability limits of the fuel, and the controllable and Oil firing is more complicated than fuel gas firing because
safety lower pressure limits of the fuel system. the oil must be “atomized” and vaporized before it can be
With all of the air being supplied through and metered by properly burned. Atomization involves producing a relatively
the throat of the burner, it becomes necessary to adjust the fine spray of droplets that will vaporize quickly. Industrial oil
air flow for efficient operation. Independent of whether the burners typically utilize twin fluid atomizers and employ
burner is designed for natural draft, low forced draft, high steam as the atomizing medium. Compressed air, or even
induced draft, or high forced draft, the ability to efficiently high-pressure fuel gas, can be utilized in some applications
combust the fuel and transfer the heat is directly related to rather than steam.
the amount of excess air. Therefore, it is important to control Some internal mix twin fluid atomizers form a gas/liquid
the air with every change in fuel flow or fuel composition or emulsion which then issues from the fuel tip. This type of
there will be a related change in efficiency. atomizer is shown in Figure 11.13. The oil issues from a
Another problem intrinsic to the raw gas design condition single port at the entry to the atomizer section. Steam is
comes from the fuel distribution system, especially in burners injected into the oil stream through multiple ports, forming
designed for natural draft and/or low combustion O2 pressure the emulsion. This oil steam emulsion then travels to the
drop. Patterning of the flame, distribution of the fuel into the chamber between the atomizer section and the exit ports.
combustion air stream, and mixing of the fuel with the oxy- Multiple exit ports are placed on the tip, which allows for
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 365 Tuesday, March 22, 2005 10:08 AM

Burner Design 365

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 11.13 Internal mix twin fluid atomizer.

enhanced mixing of the spray and the combustion air, flame

shaping, and stabilization. Internal mix atomizers often utilize
a steam pressure maintained at 10 to 30 psig (70 to 210 kPa)
above the oil pressure over the operating range of the burner.
The oil pressure may be 100 to 120 psig (700 to 800 kPa) at
maximum firing rate.
Port mix twin fluid atomizers (see Figure 11.14) use the
atomizing media to form a thin liquid annulus on the inner
surface of the port. The high velocity of the atomizing media
pushes the liquid along the length of the port shearing and
stretches the liquid into a thin film. These atomizers generally
operate with a constant steam pressure, typically 120 to
150 psig (800 to 1000 kPa) over the entire operating range
of the burner. The oil pressure may be 100 to 150 psig (700 to
1000 kPa) at maximum firing rate.
In each case, internal mix or port mix, droplets are formed
as the liquid forms a sheet as it exits the port, and this sheet
then breaks up as it expands. The size of the droplets formed FIGURE 11.14 Port mix twin fluid atomizer.
depends on the relative velocity of the liquid sheet and the
surrounding media. It also depends on the viscosity, surface
tension and density of the liquid, size of the port, momentum fuel oil, bunker C oil, residual fuel oil, pitch, tar, and vacuum
of the fuel, and ratio of fuel to atomizing media. tower bottoms. Most heavy fuel oils are the residue from the
oil refining process and are considered a low value by-
11.4.3.1 High Viscosity Liquid Fuels product. This viscous oil must be heated to be efficiently
The bulk of the industrial liquid fuel (see Chapter 5) fired is atomized. Typically, the best atomization requires the
high-viscosity or heavy fuel oil. Heavy oils include No. 6 viscosity of the oil to be in the range of 100 to 250 SSU. This
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 366 Tuesday, March 22, 2005 10:08 AM

366 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 11.16 Typical conventional raw gas burner.

refractory tiles that redirect the intense flame radiation back

to the root of the spray to enhance vaporization rates and help
stabilize the oil flame. Some other heavy oil burners utilize
swirlers to promote internal recirculation of the hot combus-
tion products back into the flame root to enhance heating of
the spray (see Figure 11.15).

11.4.3.2 Low-Viscosity Liquids

Low-viscosity liquid fuels include light oils such as No. 2
FIGURE 11.15 (a) Regen tile and (b) swirler. fuel oil, naphtha, and by-product waste liquids containing a
variety of hydrocarbon by-products such as alcohols.
These liquids are also typically fired using twin fluid
often requires heavy oil to be heated to the range of 200 to
atomizers. However, care must be taken to avoid
250°F (90 to 120°C). However, a heavy pitch may require a
overheating low-boiling-point fuels. Vaporization of the
temperature of 600°F (320°C)to achieve the proper viscosity
liquid within the oil gun can cause disruption in flow
for atomization.
within the atomizer and unsteady operation. Many light
Heavy fuel oils have high boiling points and are therefore liquids require compressed air rather than steam
difficult to vaporize. Many heavy oil burners have special atomization. Because light oils generally have lower
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 367 Tuesday, March 22, 2005 10:08 AM

Burner Design 367

FIGURE 11.17 Typical premix gas burner.

boiling points, they are easier to vaporize and their flames

can often be stabilized using simple bluff-body stabilizers
rather than swirlers or special tiles.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
11.5 CONFIGURATION (MOUNTING
AND DIRECTION OF FIRING)
Burners can be mounted in the furnace or heater floor to fire
vertically upward; in the heater wall to fire horizontally; or
in the roof to fire vertically downward. The major
consideration required for burner design is to ensure proper
support for the burner tile. Burner blocks for floor-mounted
service can typically be simply placed on the furnace steel
or burner mounting plate. Burner blocks for horizontal
mounting must be supported by tile case assemblies and
must be held in place so that they do not move if subjected
to vibration. Roof-mounted tiles must have support surfaces
cast into them so that they can hang from steel supports in
the furnace roof.
FIGURE 11.18 Typical round flame combination burner.
11.5.1 Conventional Burner, Round Flame
The following figures illustrate the different types of burners conventional burner. This burner is used where NOx emissions
that are typically used in refinery and petrochemical furnaces. are not a primary concern and a short flame is desired.
The round flame burner is the most universal design and used Figure 11.17 shows a typical premix gas burner. A premix
in many applications. Figure 11.16 illustrates a typical raw gas round flame burner is useful when a short heater does not
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 368 Tuesday, March 22, 2005 10:08 AM

368 The John Zink Combustion Handbook

The advantage in using this particular burner is that because

of fuel-staging, the concentration of NOx reduces extensively.
Based on fired capacity, one can utilize single or multiple
primary and staged injectors to effect the proper fuel distri-
bution to produce the desired flame pattern. A flat flame
burner can be fired in two basic configurations — wall fired
or freestanding.

11.5.2.1 Wall Fired

A wall-fired flat flame burner is similar to a freestanding flat
flame burner except that it is installed and fired against a
refractory wall. By directing the fuel jets toward the wall, the

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
flame heats the refractory wall, which in turn radiates heat to
the process tubes facing the wall. Wall-fired burners are
typically used in ethylene furnaces.

11.5.2.2 Freestanding
A free-standing flat flame burner is used in applications
where it is necessary to fire a burner between two sets of
process tubes. The tube spacing often requires the use of a
flat flame burner. A staged-fuel, flat flame is shaped by firing
stage fuel jets opposite one another. This makes the flame
FIGURE 11.19 Typical round flame, high-intensity com- shape into a fan. The flame thickness is typically less than or
bination burner. equal to the burner tile width.
have enough draft to supply the required combustion air. The
premix burner uses the fuel jet as a motive force to allow the
burner to pull in part or all of its combustion air.
Figure 11.18 shows a typical round/conical flame combi-
nation oil and gas burner. This burner can be used to burn
gas or liquid fuels. This versatility is desirable to applications
with liquid fuels or where gas fuels may be in short supply
at various times throughout the year and an alternative fuel
must be fired to maintain a process.
Figure 11.19 shows a typical round/conical flame high
intensity combination oil and gas burner. This burner is used
in applications where a high heat release per burner is
required but a short flame length is required.

11.5.2 Flat Flame Burner

Some applications require a flat or fan-shaped flamed due to
the close proximity of process tubes or due to the fact that the
burner is fired along a wall. Figure 11.20 shows a typical
staged-fuel flat flame gas burner. This burner produces a
freestanding flame and is used in applications where the
process tubes are close to the centerline of the burner. FIGURE 11.20 Typical staged-fuel flat flame burner.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 369 Tuesday, March 22, 2005 10:08 AM

Burner Design
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
369

FIGURE 11.21 Typical radiant wall burner.

11.5.3 Radiant Wall 11.6 MATERIALS SELECTION

Burner materials are selected for strength, corrosion resis-
Radiant wall burners are used in cracking furnaces. These
tance, and in many cases temperature resistance. Carbon steel
burners are mounted through the furnace wall and produce a
is used for most metal parts unless temperature or corrosion
thin, flat circular disk of flame adjacent to the wall. There are
considerations require more resistant alloys.
typically several rows of burners, with burners equally
Cast iron or carbon steel can be utilized for fuel gas mani-
spaced in a grid pattern on walls. The burners uniformly heat
folds. The fuel gas riser material between the manifold and the
the walls, which radiate heat to the process tubes on the
fuel injector tip is generally carbon steel for ambient air service
centerline of the furnace. Figure 11.21 shows a typical
and 304 SS for parts in contact with air temperatures higher
radiant wall burner.
than 700°F (370°C). If H2S is present in the fuel gas and pre-
Many applications, such as reformers and cracking fur- heated air over 300°F (150°C) is used, 316L stainless steel may
naces, require high furnace temperatures in the radiant section be required for the gas piping passing through the windbox.
of the heater to produce the required high process tempera- Fuel injector tips for premix burners may be cast iron. Cast
tures. In addition, very uniform heat transfer to the process steel or cast stainless steel may be required for fuel gases con-
tubes is needed. In such operations, a flat flame burner that taining appreciable levels of hydrogen (typically >50 vol%).
fires adjacent to a refractory wall is utilized. Raw gas burners typically use 300 series stainless steel gas
tips. Tips for oil burners are normally 416 SS. For oils con-
taining erosive particles, T-1 tool steel is generally used. Atom-
11.5.4 Downfired izers for oil service are normally brass, or 303 SS for oils
A downfired burner is a burner that is fired vertically into a containing sulfur. Other oil injector parts are carbon steel,
furnace from the ceiling. Typically, these burners produce a although nitride-hardened parts can be used for erosive liquids.
flame that is round/conical in shape. These burners are uti- Flame stabilizer cones and swirlers are normally 300 series
lized in applications such as hydrogen reformers or ammonia stainless steel.
reformers. Downfired burners are often dual-fuel burners Mineral wool is commonly used for noise reduction in
firing a makeup fuel and a waste gas. The makeup fuel can be plenums and as insulation in preheated air service up to
natural gas, naphtha, No. 2 fuel oil, diesel oil, or a 1000°F (540°C). Ceramic fiber insulation is used for higher
propane/butane gas fuel. temperature applications.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch11-Frame Page 370 Tuesday, March 22, 2005 10:08 AM

370 The John Zink Combustion Handbook

High-strength, low-alloy structural steel (ASTM A 242/A REFERENCES

242M) or 304 SS is used for other metal parts subject to
>700°F (370°C) air preheat. 1. C.E. Baukal, Ed., Oxygen-Enhanced Combustion, CRC
Burner block material is typically 55 to 60% alumina Press, Boca Raton, FL, 1998.
refractory with a 3000°F (1650°C) service temperature. Pri- 2. American Petroleum Institute, Burners for Fired
mary oil tiles may require 90% alumina refractory if the Heaters in General Refinery Services, API 535, 1st ed.,
combined vanadium and sodium in the oil is greater than Washington, D.C., July 1995.
50 ppm by weight.
3. American Petroleum Institute, Fired Heaters for
General Refinery Services, API 560, Washington, D.C.,
November 1996.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 12
Combustion Controls
Joe Gifford and Jim Heinlein

TABLE OF CONTENTS

12.1 Fundamentals .......................................................................................................................................... 374

12.1.1 Control Platforms...................................................................................................................... 374
12.1.2 Discrete Control Systems.......................................................................................................... 376
12.1.3 Analog Control Systems ........................................................................................................... 376
12.1.4 Failure Modes ........................................................................................................................... 378
12.1.5 Agency Approvals and Safety................................................................................................... 379
12.1.6 Pipe Racks and Control Panels ................................................................................................. 381
12.2 Primary Measurement ............................................................................................................................. 383
12.2.1 Discrete Devices ....................................................................................................................... 383
12.2.2 Analog Devices ......................................................................................................................... 385
12.3 Control Schemes ..................................................................................................................................... 389

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
12.3.1 Parallel Positioning ................................................................................................................... 389
12.3.2 Fully Metered Cross Limiting................................................................................................... 392
12.4 Controllers............................................................................................................................................... 394
12.5 Tuning ..................................................................................................................................................... 398
References ................................................................................................................................................................ 399

373
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 374 Tuesday, March 22, 2005 11:48 AM

374 The John Zink Combustion Handbook

12.1 FUNDAMENTALS There are also a few disadvantages. Once a certain com-
This chapter discusses the various control system compo- plexity level is reached, relay systems can quickly become
nents, concepts, and philosophies necessary for understanding massive. Although individual relays are very reliable, a large
how control systems work, what the systems are designed to control system with hundreds of relays can be very unreliable.
accomplish, and what criteria the controls engineer uses to Also, relays take up a lot of expensive control panel space.
design and implement a system. The interested reader can find Because relays must be physically rewired to change the
further information on controls in numerous references.1–11 operating sequence, system flexibility is poor.
The purpose of the control system is to start, operate, and
shut down the combustion process and any related auxiliary 12.1.1.2 Burner Controller
processes safely, reliably, and efficiently. The control system A variety of burner controllers is available from several dif-
consists of various physical and logical components chosen ferent vendors. They are prepackaged, hardwired devices in
and assembled according to a control philosophy and arranged different configurations to operate different types of systems.
to provide the user with an informative, consistent, and easy- Burner controllers will execute a defined sequence and moni-
to-use interface. tor defined safety parameters. They are generally located in a
local control panel. Like relays, they generally have no
A combustion system typically includes a fuel supply, a
analog capability.
combustion air supply, and an ignition system that all come
Advantages of burner controllers include the fact that they
together at one or more burners. During system startup and
are generally inexpensive, compact, simple to hook up,
at various times during normal operation, the control system
require no programming, and are fail-safe and very reliable.
needs to verify or change the status of these systems. During
They are often approved for combustion service by various
system operation, the control system needs various items of
safety agencies and insurance companies.
process information to optimize system efficiency. Addition-
There are also some disadvantages. Burner controllers can-
ally, the control system monitors all safety parameters at all
not control combustion systems of much complexity. System
times and will shut down the combustion system if any of the
flexibility is nonexistent. If it becomes necessary to change
safety limits are not satisfied.
the operating sequence, the controller must be rewired or
replaced with a different unit. Controllers also require the use
12.1.1 Control Platforms of attached peripherals from the same vendor, so some design
The control platform is the set of devices that monitors and flexibility is lost.
optimizes the process conditions, executes the control logic,
and controls the status of the combustion system. There are 12.1.1.3 Programmable Logic Controller (PLC)
several different types of platforms and several different ways A programmable logic controller (PLC) is a small modular
that the tasks mentioned above are divided among the types computer system that consists of a processing unit and a num-
of platforms. Following is a list and a brief description of the ber of input and output modules that provide the interface to
most commonly used platforms. the combustion components. PLCs are usually rack-mounted
and modules can be added or changed (see Figure 12.1).
12.1.1.1 Relay System There are many types of modules available. Unlike the relays
A relay consists of an electromagnetic coil and several and burner controllers above, PLCs have analog control capa-
attached switch contacts that open or close when the coil is bility. They are generally located in a local control panel.
energized or de-energized. A relay system consists of a num- PLCs have the advantage of being a mature technology.
ber of relays wired together in such a way that they execute a They have been available for over 20 years. Simple PLCs are
logical sequence. For example, a relay system can define a inexpensive and PLC prices are generally very competitive.
series of steps to start up the combustion process. Relays can They are compact, relatively easy to hook up, and because
tell only if something is on or off and have no analog capabil- they are programmable, they are supremely flexible. They can
ity. They are generally located in a local control panel. operate systems of almost any level of complexity. PLC reli-
Relays have several advantages. Relays are simple, easily ability has improved over the years and is now very good.
tested, reliable, and well-understood devices that can be wired Disadvantages of PLCs include having to write software
together to make surprisingly complex systems. They are for the controller. Coding can be complex and creates the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

modular, easily replaced, and inexpensive. They can be con- possibility of making a programming mistake, which can
figured in fail-safe mode so that if the relay itself fails, com- compromise system safety. The PLC can also freeze up, much
bustion system safety is not compromised. like a desktop computer freezes up, where all inputs and
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 375 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 375

FIGURE 12.1 Programmable logic controller.

outputs are ignored and the system must be reset in order to However, DCSs are often difficult to program. Each DCS
execute logic again. Because of this possibility, standard vendor has a proprietary system architecture, so the hardware
PLCs should never be used as a primary safety device. Special is expensive and the software is often different from any
types of redundant or fault-tolerant PLCs are available that other vendor’s software. Once a commitment is made to a
are more robust and generally accepted for this service, but particular DCS vendor, it is extremely difficult to change to
they are very expensive and generally difficult to implement.
a different one.

12.1.1.4 Distributed Control System (DCS)

12.1.1.5 Hybrid Systems
A distributed control system (DCS) is a larger computer sys-
If one could combine several of the systems listed above and
tem that can consist of a number of processing units and a
wide variety of input and output interface devices. Unlike the build a hybrid control system, the advantages of each system
systems described above, when properly sized, a DCS can could be exploited. In practice, that is what is usually done. A
also control multiple systems and even entire plants. The typical system uses relays to perform the safety monitoring, a
DCS is generally located in a remote control room but PLC to do the sequencing, and either dedicated controllers or
peripheral elements can be located almost anywhere. an existing DCS for the analog systems control. Sometimes,
DCSs have been around long enough to be a mature tech- the DCS does both the sequencing and the analog systems
nology and are generally well-understood. They are highly control, and the safety monitoring is done by a fault-tolerant
flexible and are used for both analog and discrete (on/off) logic system. Most approval agencies and insurers require the
control. They can operate systems of almost any level of safety monitoring function to be separate from either of the
complexity and their reliability is excellent. --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- other functions.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 376 Tuesday, March 22, 2005 11:48 AM

376 The John Zink Combustion Handbook

FIGURE 12.2 Touchscreen.

12.1.1.6 Future Systems The discrete control system does safety monitoring and
Over the next decade or so, it is expected that embedded sequencing. Typically, the system monitors all of the discrete
industrial microprocessors using touchscreen video interfaces inputs, and if they are all satisfactory, allows combustion
(see Figure 12.2) will start to appear in combustion control. system startup. If a monitored parameter is on when it should
These interfaces will communicate with field devices such as be off, or vice versa, the startup process is aborted and the
valves and switches via a single communications cable. They system must be reset before another startup is permitted. The
will use a digital bus protocol such as Profibus or Fieldbus. system also controls such things as which valves are opened
These systems are becoming common on factory floors around in what order, if and when the pilot is ignited, and if and
the world. Because establishment of a single standard has not when main burner operation is allowed. Once the system
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

yet happened and combustion standards are slow to change, starts, the discrete system has little to do other than monitor
these systems have not yet achieved widespread acceptance in safety parameters. If any of the defined safety parameters are
the combustion world. not satisfactory, the system immediately shuts down.
Figure 12.3 is a simplified flow diagram showing a standard
12.1.2 Discrete Control Systems burner light-off sequence.
The world of discrete controls is black and white. Is the valve
open or shut? Is the switch on or off? Is the button pressed or 12.1.3 Analog Control Systems
not? Is the blower running or not? There are two basic types The world of analog controls is not black and white — it is
of discrete devices: (1) input devices (sensors) that have elec- all gray. How far open is that valve? What is the system tem-
trical contacts that open or close depending on the status of perature? How much fuel gas is flowing?
what is being monitored; and (2) output devices, or final There are two categories of analog devices with familiar
elements, that are turned on or off by the control system. names: (1) sensors, which measure some process variable (e.g.,
In a typical control system, sensors such as pressure switches, flow or temperature) and generate a signal proportional to the
valve position switches, flame scanners, and temperature measured value; and (2) final elements (e.g., pumps and valves),
switches do all the safety and sequence monitoring. These which change their status (speed or position, for example) in
devices tell the control system what is happening out in the response to a proportional signal from the control system.
real world. They are described in more detail in Section 12.1.3. In contrast to the discrete control system, the analog control
The final elements carry out the on/off instructions that come system usually has few tasks to perform until the system
from the control system. These are devices such as solenoid completes the startup sequence and is ready to maintain nor-
valves, relays, indicating lights, and motor starters. These mal operation. Most analog devices are part of a control loop.
devices allow the control system to make things happen in the A simple loop consists of a sensor, a final element, and a
real world, and are described in more detail in Section 12.1.3. controller. The controller reads the sensor, compares the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 377 Tuesday, March 22, 2005 11:48 AM

Combustion Controls
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
377

FIGURE 12.3 Simplified flow diagram of a standard burner light-off sequence.

378 The John Zink Combustion Handbook

THERMOCOUPLE
CONTROLLER (SENSOR)
TIC TE
100 100

TV BURNER
TEMPERATURE
100

LEGEND
TV - Temperature Valve
FUEL TO TIC - Temperature Indicator
BURNER Controller
TE - Temperature Element
CONTROL VALVE
(FINAL ELEMENT)

FIGURE 12.4 Simple analog loop.

measured value to its setpoint set by the operator, and then doubles? There will no longer be enough combustion air in
positions the final element to make the measured value equal the system to allow destruction of all the waste. Unburned
the setpoint. Figure 12.4 illustrates a simple analog loop. waste will burn at the tip of the smokestack and clouds of
In this case, the thermocouple transmits the temperature to smoke will billow from the stack. Phone calls from irate
the controller. If the temperature is higher than the setpoint, neighbors will soon begin to accumulate. Using feedforward,
the controller will decrease its signal to the control valve. as shown in Figure 12.5, the waste flow is measured. When
This will decrease the fuel flow to the burner, thus lowering it doubles, the combustion airflow setpoint immediately
the temperature. In this way, the loop works to maintain the increases by a similar amount, avoiding all of the unpleasant
desired temperature — also known as the setpoint. consequences listed above.
The previous illustration is a good example of a simple
feedback system. After the controller adjusts the control 12.1.4 Failure Modes
valve, the resulting change in temperature is fed back to the Almost everything fails eventually. No matter how well the
controller. This way, the controller “knows” the result of the components of a control system are designed and built, some
adjustment and can make a further adjustment if it is of them will fail from time to time. One of the primary tasks
required. Another good example of feedback takes place of the controls engineer is to design the control system so that
whenever one drives a car. If one gets on the expressway and failure of one or even several components will not cause a
decides to drive at 60 mph, one presses the accelerator and safety problem with the combustion system.
watches the speedometer. Near 60 mph, one begins to ease All components used in a control system have one or more
off the accelerator so as not to overshoot. From then on, one defined failure modes. For example, if a discrete sensor fails,
glances at the speedometer every now and then and adjusts it will most likely cause the built-in switch contacts to fail
one’s foot position as necessary. open. To design a safe system, the controls engineer must
Feedback alone is not enough, however. What if there is choose and install the sensor so that when an alarm condition
traffic congestion? One begins to slow down in anticipation. is present, the switch contacts will open. Thus, the alarm
This is called feedforward, which occurs when one changes condition coincides with the failure mode. If it did not, the
the operating point because some future event is about to sensor could fail and the control system would still think
happen and one needs to prepare for it. Feedforward is com- everything was normal and attempt to keep operating as
monly used in combustion control systems. A good example before — a condition that could be catastrophic.
from the world of combustion is waste flow. In a waste In addition to sensors, final control elements also have
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

destruction system, what happens if the waste flow suddenly failure modes. The controls engineer can usually select the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 379 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 379

LEGEND
FIT FIT - Flow Indicating Transmitter
100 FV - Flow Valve
FLOW METER FY - Flow Relay
(SENSOR) FIC - Flow Indicating Controller
WASTE GAS
FLOW TO
BURNER

SUMMER CONTROLLER
FY FIC
100 100

FV
100

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
COMBUSTION
AIR TO
BURNER
CONTROL VALVE
(FINAL ELEMENT)

FIGURE 12.5 Feedforward loop.

desired failure mode. If there is an actuated valve that turns safe operation of various combustion processes. No single
the fuel gas supply on and off, the actuator is installed so that organization does all of the things listed.
the valve will spring closed (fail shut) upon loss of air. In The design of combustion systems in the United States
addition, assume there is a solenoid valve that turns air to the should include specifications that meet National Fire Protec-
actuator on and off. The solenoid valve should be selected tion Association (NFPA) and National Electrical Code (NEC)
and installed to rapidly dump air from the actuator upon loss standards. In accordance with the applicable standards and
of electrical power, thus closing the valve. These designs years of experience in the field, systems should be designed
ensure that the two most likely circumstances of component with some or all of the following safety features.
failure enhance system safety.
Construction of a well-designed system ensures that every 12.1.5.1 Double-Block-and-Bleed for Fuel Supply
component that can fail is installed so that component failures This means that there are two fail-shut safety shutoff block
do not compromise system safety. At its core, that is what valves with a fail-open safety shutoff vent valve located
controls engineering is all about. between them, as shown in Figure 12.6. Each of the three
safety shutoff valves (SSOVs) in the double-block-and-bleed
12.1.5 Agency Approvals and Safety system has a position switch not shown in the figure. For a
Worldwide, there are hundreds of private, governmental, and system purge to be valid, the block valves must be shut and
semi-governmental safety organizations. Each ostensibly has verified. For burner light-off, the vent valve is shut. After the
the proper implementation of safety at the top of its agenda. vent valve position switch confirms that the valve is shut, the
Some agencies are concerned with the electrical safety and two block valves are opened. If there is a system failure, all
reliability of the components used in a control system; others three of the valves de-energize and return to their failure
are concerned about preventing explosions caused by spark- positions. Note that if the upstream block valve ever leaks,
ing equipment in a gaseous atmosphere; and still others are the leakage will preferentially go through the open vent valve
concerned with the proper design of control systems to ensure and vent to a safe location rather than into the burner.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 380 Tuesday, March 22, 2005 11:48 AM

380 The John Zink Combustion Handbook

FIGURE 12.6 Double-block-and-bleed system.

FIGURE 12.7 Failsafe input to programmable logic controller (PLC).

12.1.5.2 Unsatisfactory Parameter System Shutdown the relays will shut down the system. This is an excellent
An unsatisfactory parameter for any critical input immedi- example of redundancy, fail-safe design principles, and effec-
ately shuts down the system. The control system typically tive design philosophy.
receives critical input information as shown in Figure 12.7.
The pressure switch PSLL-03073 is wired so that if it fails, 12.1.5.3 Local Reset Required after System Shutdown
the voltage is interrupted to the relay (CR-xx) and the pro-
After a system shutdown caused by an alarm condition, the
grammable logic controller (PLC). The PLC will then shut
system allows a remote restart only after an operator has
down the system. If either the switch or the relay fails, the
pressed a reset button located at the combustion system. The
system shuts down.
operator should perform a visual inspection of the system and
In addition, the relay has another contact in series with all
verify the correction of the condition that caused the shutdown.
the other critical contacts. If any of these contacts open, the
power cuts off to all ignition sources (all fuel valves, igniters,
etc.), immediately shutting down the combustion system. 12.1.5.4 Watchdog Timer to Verify PLC Operation
In Figure 12.8, if there is a failure anywhere in the circuitry, If the PLC logic freezes, a separate timer fails to receive an
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

the system shuts down. Even if damage to the PLC occurs, expected reset pulse from the PLC and shuts down the system.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 381 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 381

FIGURE 12.8 Shutdown string.

FIGURE 12.9 Typical pipe rack.

12.1.6 Pipe Racks and Control Panels Typical control panels are shown in Figures 12.10a and b.
Figure 12.11 shows the inside of the large control panel. The
For most combustion control systems, two major assemblies
control panel is usually attached to the pipe rack. All of the
comprise the bulk of the system: the pipe rack and the control
devices on the pipe rack, as well as the field devices, are
panel. A pipe rack is shown in Figure 12.9. Sometimes called
electrically connected to the control panel. The control system
a skid, the pipe rack is a steel framework that has a number of
pipes and associated components attached to it. Usually, most usually resides inside the control panel. In addition to the
of the combustion process feeds such as air, fuel, and waste wiring, maintenance, and troubleshooting benefits mentioned
have their shutoff and control elements located on the pipe above, another benefit to packaging the control system is far
rack. This makes maintenance and troubleshooting more con- more important — the people who designed and built the
venient and reduces the amount and complexity of wiring system can test and adjust it at the factory. When the control
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

systems required to connect all of the components. system arrives at the job site, installation consists mostly of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 382 Tuesday, March 22, 2005 11:48 AM

382 The John Zink Combustion Handbook

FIGURE 12.10a Large control panel. FIGURE 12.10b Small control panel.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 12.11 Inside the control panel.

Combustion Controls 383

hooking up utilities and the interconnecting piping, minimizing

the amount of expensive on-site troubleshooting, and tuning.

12.2 PRIMARY MEASUREMENT

This section describes a number of different analog and dis-
crete devices used to provide the interface between combus-
tion systems and control systems. All of these devices are
available from numerous suppliers around the world. There
are vast differences in price, quality, and functionality among
the different devices and suppliers.

12.2.1 Discrete Devices

12.2.1.1 Annunciators
An annunciator is a centralized alarm broadcasting and mem- (a)
ory device. Typically, all of the alarms in a control system are
routed to the annunciator. When any alarm is triggered, the
light associated with that alarm will flash and the annuncia-
tor horn will sound. If the annunciator is a “first-out” type,
any subsequent alarms that occur will trigger their associ-
ated lights to come on solidly — rather than flashing, as the
first alarm does. This is very useful for diagnosing system
problems. When a safety shutdown occurs, other alarms are
usually triggered while the system is shutting down. With a
first-out capability, the original cause of the shutdown can
easily be determined.

12.2.1.2 Pressure Switches

Pressure switches (see Figure 12.12a and b) are sensors that
attach directly to a process being measured. They can be used
to detect absolute, gage, or differential pressure. The
switches generally have a pressure element such as a dia-
phragm, tube, or bellows that expands or contracts against an
adjustable spring as pressure changes. The element attaches
to one or more sets of contacts that open or close upon reach-
ing the setpoint. The devices are used in a number of ways,
but in combustion systems, they are usually used to test for FIGURE 12.12 Pressure switches.
high and low fuel gas pressure. They are normally set so the
contacts are open when in the alarm condition.
with integrated beacons and other visual devices. They are
12.2.1.3 Position Switches normally installed so that their contacts are closed only
Also called limit switches, these sensors attach to or are built when the valve is in the desired safety position.
into valves, insertable igniters, and other devices. Position
switches usually employ a mechanical linkage, but proxim- 12.2.1.4 Temperature Switches
ity sensors are also quite common. They are adjustable, and Temperature switches are usually attached to auxiliary equip-
can tell the control system if a valve is open, closed, or in ment such as tanks or flame arresters. These sensors gener-
some defined intermediate position. Position switches are ally do not have the range to test for combustion system
not usually used for alarms. In combustion systems, they are temperatures, so those applications use other devices. The
generally used to check valve positions during purges and switches usually use a bimetallic element, where the differen-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

burner light-off sequences. Position switches are often used tial expansion of two different metals generates physical
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 384 Tuesday, March 22, 2005 11:48 AM

384 The John Zink Combustion Handbook

movement. The movement opens or closes one or more sets panel, but some newer systems have everything located in
of contacts. The failure mode of temperature switches is not the scanner housing, which mounts on the burner end plate.
always predictable. Generally, installation requires open con- One limitation with flame scanners is the possibility of the
tacts when the switch is in the alarm condition. power wire to the scanner inducing false flame indications in
the signal wire from the scanner. If there are separate wires
12.2.1.5 Flow Switches for scanner power and scanner signal, they must run in sepa-
Flow switches are sensors that generally insert into the pipe rate conduits or shielding to prevent false signals caused by
or duct in which flow is measured. Because of the lack of a induction.
quantitative readout and the improved reliability of analog
transmitters in this service, these devices are becoming less 12.2.1.8 Solenoid Valves
common. Their failure mode is not always predictable. Usual Solenoid valves are turned on or off by the presence or
installation requires open contacts when the switch is in the absence of voltage from the control system. A solenoid valve
alarm condition. has a relay coil that links mechanically to a valve disc mecha-
nism. Energizing the solenoid causes the linkage to push
12.2.1.6 Run Indicators against a spring to reposition the valve disc. De-energizing
A run indication sensor shows whether or not a pump or fan is the solenoid allows the spring to force the valve to the failure
running. It is usually possible to order a motor starter with a position. The most common types of solenoid valves are two-
built-in set of signal contacts that close when the starter motor way and three-way valves. Two-way valves have two posi-
contacts are closed. However, that does not always ensure that tions. They either allow flow or they do not. They are often
the pump is running and pumping fluid. A magnetic shaft used to turn pilot gas on and off. Three-way solenoid valves
encoder rotates a magnetic slug past a pickup sensor every have three ports but still only two positions. If ports are
revolution and provides positive indication of shaft revolution, labeled A, B, and C, energizing the valve may allow flow
but that too does not always ensure that the pump is pumping between ports A and B, while de-energizing the valve may
fluid. It is usually preferable to have a pressure or flow indica- allow flow between ports A and C. It is very important to
tor that shows that the system is functioning normally and carefully select, install, and test three-way solenoid valves.
moving fluid. Three-way solenoid valves typically attach to control valves
and safety shutoff valves (SSOVs). In the case of control
12.2.1.7 Flame Scanners valves, when the solenoid valve is energized, the control
Flame scanners are crucial to the safe operation of a com- valve is enabled for normal use. When the solenoid valve is
bustion system. If the flame is out, the fuel flow into the de-energized, the instrument air is dumped from the control
combustion enclosure is stopped and the area is purged valve actuator diaphragm, causing the control valve to go to
before a re-light can be attempted. Flame scanners come in its spring-loaded failure position. For safety shutoff valve
two main varieties: infrared and ultraviolet. The name tells (SSOV) service, the solenoid valve is hooked up so that when
which section of the electromagnetic spectrum it is designed energized, instrument air is allowed to reposition the SSOV
to see. Generally, ultraviolet scanners are preferred because actuator away from its spring-loaded failure position. When
they are more sensitive and quicker to respond. The detector the solenoid valve is de-energized, the air is dumped from the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

is a gas-filled tube that scintillates in the presence of flame SSOV actuator, causing the control valve to go to its spring-
ultraviolet radiation and emits bursts of current, called an loaded failure position. The failure modes of the solenoid
avalanche, several hundred times per second as long as the valve, control valve, and SSOV are coordinated to maximize
flame continues. When the flame stops, the current stops. system safety no matter which component fails.
There is a 2- to 4-second delay, to minimize spurious shut-
downs, and then the contacts open to designate the alarm 12.2.1.9 Ignition Transformers
condition. Most systems have two flame scanners and both Ignition transformers supply the high voltage necessary to
scanners must fail to achieve system shutdown. Use of infra- generate the spark used to ignite the pilot flame during sys-
red scanners is desirable if there is a waste stream, such as tem light-off. The type of transformer usually used converts
sulfur, that absorbs ultraviolet light and makes operation of standard AC power to a continuous 6000 V DC. This voltage
ultraviolet scanners unreliable. Self-checking scanners are then continuously jumps the spark gap at the igniter, which is
usually used. They have output contacts that open on either located at the head of the pilot burner. High-energy igniters
loss of flame or failure of the self-check. Usually, the con- provide a more intense spark. A high-energy igniter is similar
tacts are part of an amplifier/relay unit located in the control to the transformer mentioned above except that a capacitor is
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 385 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 385

included to store energy and release it in spurts, resulting in a

more intense spark. Both types of transformers are usually
located close to the burner in a separate enclosure and hooked
to the igniter using coaxial cable similar to the spark plug
wire used in cars.

12.2.2 Analog Devices

12.2.2.1 Control Valves
Control valves are among the most complex and expensive
components in any combustion control system. Numerous
books document the nearly infinite variety of valves. Misap-
plication or misuse of valves compromise system efficiency
and safety. Controls engineers cannot simply pick control

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
valves from a catalog because they are the right size for the
line where they will be used. Control valves must be engi-
neered for their specific application. A typical pneumatic
control valve is shown in Figure 12.13.
As shown in Figure 12.14, the type of service and control
desired determines the selection of different flow character-
istics and valve sizes. Controls engineers use a series of cal- FIGURE 12.13 Pneumatic control valve.
culations to help with this selection process. A typical control
valve consists of several components that are mated together 100
before installation in the piping system.

80 Quick Open
12.2.2.1.1 Control Valve Body
Percent Flow

The control valve body can be a globe valve, a butterfly

valve, or any other type of adjustable control valve. Usually, 60
special globe valves of the equal percent type are used for
fuel gas control service or liquid service. Control of combus-
tion air and waste gas flows generally require the use of but- 40 Linear
terfly valves — often the quick-opening type. Because the
combustion air or waste line usually has a large diameter, and Equal %
20
because the cost of globe valves quickly becomes astronomi-
cal after line size exceeds 3 or 4 in., butterfly valves are usu-
0
ally the most economical choice. In Section 12.3, a 0 20 40 60 80 100
discussion of parallel positioning describes how controls
engineers use a globe valve and a butterfly valve together to Percent Open
work smoothly for system control.
FIGURE 12.14 Control valve characteristics.
12.2.2.1.2 Actuator
The actuator supplies the mechanical force to position the
side of the diaphragm and the spring on the other. The air
valve for the desired flow rate. For control applications, a dia-
pressure forces the actuator to move against the spring. If air
phragm actuator is preferred because compared to a piston-
pressure is lost, the valve fails to the spring position, so the
type actuator, it has a relatively large pressure-sensitive area
actuator is chosen carefully to fail to a safe position
and a relatively small frictional area where the stem touches
(i.e., closed for fuel valves, open for combustion air valves).
the packing. This ensures smooth operation, precision, and
good repeatability. Proper selection of the actuator must take
into account valve size, air pressure, desired failure mode, 12.2.2.1.3 Current-to-Pressure Transducer
process pressure, and other factors. Actuators are usually The current-to-pressure transducer, usually called the I/P
spring-loaded and single-acting, with control air used on one converter, takes the 24 V DC (4 to 20 mA) signal from the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 386 Tuesday, March 22, 2005 11:48 AM

386 The John Zink Combustion Handbook

generally used, the spring in the actuator forces the valve

either fully open or fully closed, depending on the engineer’s
choice of failure modes when specifying the valve. Obvi-
ously, a control valve that supplies fuel gas to a combustion
system should fail closed, while the control valve that sup-
plies combustion air to the same system should fail open. In
an application in which the failure mode of the valve is irrele-
vant — and there are some — solenoid valves are not used.

12.2.2.1.6 Mechanical Stops

Mechanical stops are used to limit how far open or shut a con-
trol valve can travel. If it is vital that no more than a certain
amount of fluid ever enters a downstream system, an “up”
stop is set. If it is necessary to ensure a certain minimum flow
(e.g., for cooling purposes), a “down” stop is set. In the case
FIGURE 12.15 Thermocouple. of a fuel supply control valve, the “down” stop is set so that
during system light-off, an amount of fuel ideal for smooth
and reliable burner lighting is supplied. After a defined set-
tling interval, usually 10 seconds, the 3-way solenoid valve is
energized and normal control valve operation is enabled.

12.2.2.2 Thermocouples
Whenever two dissimilar metals come into contact, current
flows between the metals, and the magnitude of that current
flow, and the voltage driving it, vary with temperature. This
phenomenon is called the Seebeck effect. If both metals are
carefully chosen and are of certain known alloy compositions,
FIGURE 12.16 Thermowell and thermocouple. the voltage will vary in a nearly linear manner with tempera-
ture over some known temperature range. Because the tem-
perature and voltage ranges vary depending on the materials
controller and converts it into a pneumatic signal. The signal employed, engineers use different types of thermocouples for
causes the diaphragm of the actuator to move to properly different situations. In combustion applications, the K type
position the control valve. thermocouple (0 to 2400°F, or –18 to 1300°C) is usually used.
When connecting a thermocouple (see Figure 12.15) to a
12.2.2.1.4 Positioner
transmitter, the transmitter should be set up for the type of
The positioner is a mechanical feedback device that senses
thermocouple employed. Installing thermocouples in a protec-
the actual position of the valve as well as the desired position
tive sheath know as a thermowell (see Figure 12.16), prevents
of the valve. It makes small adjustments to the pneumatic
the sensing element from suffering the corrosive or erosive
output to the actuator to ensure that the desired and the actual
effects of the process being measured. However, a thermowell
positions are the same. Current conventional wisdom states
also slows the response of the instrument to changing temper-
that positioners should be used only on “slow” systems and
ature and should be used with care.
not on “fast” systems, where they can actually degrade per-
formance. There is no defined border between “fast” and
“slow,” but virtually all combustion control applications are 12.2.2.3 Velocity Thermocouples
considered to be “slow,” so positioners are almost always Also known as suction pyrometers, the design of velocity
used in these systems. thermocouples attempts to minimize the inaccuracies in tem-
perature measurement caused by radiant heat. Inside a com-
12.2.2.1.5 Three-way Solenoid Valve bustor, the thermocouple measures the gas temperature.
When energized, the three-way (3-way) solenoid valve However, the large amount of heat radiated from the hot sur-
admits air to the actuator. When de-energized, it dumps the roundings significantly affects this measurement. If a thermo-
air from the actuator. Because single-acting actuators are couple is shielded from its surroundings by putting it in a
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Combustion Controls 387

FIGURE 12.17 Velocity thermocouple.

hollow pipe as shown in Figure 12.16, the response time is in which it is calibrated. In recent years, these devices have
slowed because the thermocouple is now located in a shield- become more accurate and sophisticated, with onboard intel-
created low-flow zone. Drawing suction on the shield quickly ligence and self-calibration capabilities. They are available in
pulls gas in from the combustor and the response time a wide variety of configurations and materials, and can be
improves. Using velocity thermocouples (see Figure 12.17) used in almost any service. It is possible to remotely check
provides a high degree of precision in combustion tempera- and reconfigure these “smart” pressure transmitters using a
ture measurement. handheld communicator.

12.2.2.4 Resistance Temperature Detectors (RTDs) 12.2.2.6 Flow Meters

The resistance of any conductor increases with temperature.
There are many different types of flow meters and many
If a specific material and its resistance are known, it is possi-
reasons to use one or another for a given application. The
ble to infer the temperature. Similar to the thermocouples
following is a list of several of the more common types of
described above, the linearity of the result depends on the
flow meters, how they work, and where they are used.
materials chosen for the detector and their alloy composition.
Engineers sometimes use RTDs in place of thermocouples
when higher precision is desired. Platinum is a popular mate- 12.2.2.6.1 Vortex Shedder Flow Meter
rial for RTDs because it has good linearity over a wide tem- A vortex shedder places a bar in the path of the fluid. As the
perature range. Like thermocouples, installation of RTDs in fluid goes by, vortices (whirlpools) form and break off con-
thermowells is common. stantly. An observation of the water swirling on the down-
stream side of bridge pilings in a moving stream reveals this
12.2.2.5 Pressure Transmitters effect. Each time a vortex breaks away from the bar, it causes
A pressure transmitter (see Figure 12.18) is usually used to a small vibration in the bar. The frequency of the vibration is
provide an analog pressure signal. These devices use a dia- proportional to the flow. Vortex shedders have a wide range,
phragm coupled to a variable resistance, which modifies the are highly accurate, reasonably priced, highly reliable, and
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
24 V DC loop current (4 to 20 mA) in proportion to the range useful in liquid, steam, or gas service.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 388 Tuesday, March 22, 2005 11:48 AM

388 The John Zink Combustion Handbook

FIGURE 12.18 Pressure transmitter (left) and pressure gage (right).

12.2.2.6.2 Magnetic Flow Meter wide range. They are generally more expensive and their reli-
A magnetic field, a current-carrying conductor, and relative ability is not as good as some other types.
motion between the two create an electrical generator. In the
case of a magnetic flow meter, the meter generates the mag- 12.2.2.6.5 Ultrasonic Flow Meter
netic field and the flowing liquid supplies the motion and the When waves travel in a medium (fluid), their frequency shifts
conductor. The voltage produced is proportional to the flow. if the medium is in motion relative to the wave source. The
These meters are highly accurate, very reliable, and have a magnitude of the shift, called the Doppler effect, is propor-
wide range, but are somewhat expensive. They are useful tional to the relative velocity of the source and the medium.
with highly corrosive or even gummy fluids, as long as the The ultrasonic meter generates ultrasonic sound waves, sends
fluids are conductive. Only liquid flow is measured. them diagonally across the pipe, and computes the amount of
frequency shift. These meters are reasonably accurate, have a
12.2.2.6.3 Orifice Flow Meter fairly wide range, are reasonably priced, and are highly reli-
able. Ultrasonic meters work best when there are bubbles or
Historically, almost all flows were measured using this
particulates in the fluid.
method and it is still quite popular. Placing the orifice in the
fluid flow causes a pressure drop across the orifice. A pressure 12.2.2.6.6 Turbine Flow Meters
transmitter mounted across the orifice calculates the flow from A turbine meter is a wheel that is spun by the flow of fluid

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
the amount of the pressure drop. Orifice meters are very accu- past the blades. A magnetic pickup senses the speed of the
rate but have a narrow range. They are reasonably priced, rotation, which is proportional to the flow. These meters can
highly reliable, and are useful in liquid, steam, or gas service. be very accurate but have a fairly narrow range. They must be
very carefully selected and sized for specific applications.
12.2.2.6.4 Coriolis Flow Meter They are reasonably priced and fairly reliable. They are used
The Coriolis flow meter is easily the most complex type of in liquid, steam, or gas service.
meter to understand. The fluid runs through a U-shaped tube
that is being vibrated by an attached transducer. The flow of 12.2.2.6.7 Positive Displacement Flow Meters
the fluid will cause the tube to try to twist because of the Positive displacement flow meters generally consist of a set of
Coriolis force. The magnitude of the twisting force is propor- meshed gears or lobes that are closely machined and matched
tional to flow. These meters are highly accurate and have a to each other. When fluid is forced through the gears, a fixed
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 389 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 389

amount of the fluid is allowed past for each revolution. Count- 12.2.2.7.4 Nitrogen Oxides (NOx) Analyzer
ing the revolutions reveals the exact amount of flow. These Nitrogen oxides (NO, NO2, etc.; see Chapter 6) are one of the
meters are extremely accurate and have a wide range. Because main components of smog and are the result of high-temperature
there are moving parts, the meters must be maintained or they combustion. Noxidizers are combustion systems that use an
can break down or jam. They also cause a large pressure drop, extended low-temperature combustion process designed to
which can sometimes be important for certain applications. minimize the formation of nitrous oxide compounds. Noxidiz-
ers use NOx analyzers. To properly control the process, the
12.2.2.7 Analytical Instruments NOx analyzer output goes to a controller that controls airflow
There are many different types of analytical instruments used into the system, minimizing NOx formation.
for very specific applications. Unlike the sensors described
12.2.2.7.5 Carbon Monoxide (CO) Analyzer
previously, these devices are usually systems. They are a com-
CO is also an undesirable pollutant and is a product of
bination of several different sensors linked together by a pro-
incomplete combustion. The output of the CO analyzer (see
cessor of some sort that calculates the quantity in question.
Figure 10.4a) is often used in the analysis of system effi-
Unlike a pressure transmitter, most analytical instruments
ciency or to control airflow to the combustion system.
sample and chemically test the process in question. Because
the process takes time, the engineer, when designing the
system, must plan for a delayed response from the analytical
instrument. A detailed discussion of the design and operation 12.3 CONTROL SCHEMES
of analytical instruments is beyond the scope of this book; Other chapters of this book present the combustion process
however a list of several of the more common types and their and the definition of the terms used to describe it. This sec-
uses is given below. tion describes methods used to control the process. Gener-
ally, controlling the process means controlling the flow of
12.2.2.7.1 pH Analyzer fuels and combustion air.
Almost any combustion system occasionally requires the
scrubbing of effluent or other similar processes. pH monitor- 12.3.1 Parallel Positioning
ing is needed to ensure that the water going into the scrubber Designers use analog control schemes to modulate valve
is the correct pH to neutralize the acidity or alkalinity of the position and control fan and pump speeds to achieve the
effluent. The analyzer sends information to a controller that is required mix of fuel and oxygen in a combustion system.
responsible for opening or closing valves that add alkaline Simple systems often use parallel positioning of fuel and air
chemicals to the water to raise pH. valves from a single analog signal.

12.2.2.7.2 Conductivity Analyzer 12.3.1.1 Mechanical Linkage

Conductivity analyzers are often used in conjunction with pH A common method of parallel positioning is to mechanically
analyzers. Where the pH analyzer system functions to raise link the fuel and air valves to a single actuator. Adjustment of
pH, the output from the conductivity analyzer is usually sent a cam located on the fuel valve supplies the proper amount of
to a controller responsible for opening or closing valves that fuel throughout the air valve operating range. Figure 12.19
dilute the water to lower pH. shows the arrangement.
In the figure, the temperature indicating controller (TIC)
12.2.2.7.3 Oxygen (O2) Analyzer operates an actuator attached to the air valve. A mechanical
Oxygen (O2) or combustibles analyzers monitor the amount of linkage and an adjustable cam operate the fuel valve in par-
oxygen or combustibles in the exhaust of a combustion sys- allel with the air valve. Springs or weights attached to the air
tem. The analyzer sends data back to the control system, valve shaft force a full open position of the air valve if the
which uses it to tightly control the amount of combustion air mechanical connection to the air valve fails. The system uses
coming into the system. This has the dual result of making the a fail-closed actuator to ensure that a low fire failure mode
system more efficient and reducing the amount of pollutants results from loss of signal or loss of actuator power.
that result from the combustion process. Different models Mechanical linkage is simple in operation but requires con-
have varying methodologies, accuracies, and sample times, siderable adjustment at startup to obtain the correct fuel:air
but there are two major types: (1) in situ analyzers carry out ratio over the entire operating range. Predictable flow rates
the analysis at the probe; and (2) extractive analyzers remove of fuel and air throughout valve position require a fixed supply
the sample from the process and cool it before analysis. pressure to the valves and constant load geometry downstream
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

390 The John Zink Combustion Handbook

FIGURE 12.19 Mechanically linked parallel positioning.

of the valves. Analytical feedback to control fuel gas or com- Electronically linked parallel positioning works well if
bustion air supply pressure can make dynamic corrections for properly designed. Good design requires valves with known
fuel variations, temperature changes, and system errors. coefficients throughout valve position and the use of high-
Dynamic adjustments should be small, trimming adjustments, performance positioners. Supply pressure of fuel and air to
rather than primary control parameters. each valve must be constant or repeatable. System load down-
stream of the valves must be of fixed geometry. Section
12.3.1.2 Electronic Linkage 12.3.1.3 shows an example of how to calculate and configure
a characterizer for the air valve.
Electronically linked fuel and combustion air valves for par-
Figure 12.21 shows a variation of parallel positioning that
allel positioning have many advantages over mechanically
permits use of the combustion air valve for the multiple
linked valves. Figure 12.20 illustrates the scheme.
purposes of:
In the example, a TIC (temperature indicating controller)
generates a firing rate demand. The controller applies an 1. supplying combustion air during normal operation
output of 4 to 20 mA to the fuel valve and to a characterizer 2. supplying quench air when burning exothermic waste
in the air valve circuit. Electronic shaping of the characterizer
3. using another heat source requiring quench air
output positions the air valve for correct airflow. Predictable
and repeatable valve positions require the use of positioners When showing a range of milliampere signals, the first value
at each valve. Without positioners, valve hysteresis causes is the minimum valve position and the second value is the
large errors in flow rate. maximum valve position. This convention aids system analy-
Signal inversion (1 minus the value being measured) is sis and is especially useful for complex systems. The TIC
sometimes integral to the characterizer. Signal inversion is output is split-ranged. The top half (12 to 20 mA) is used for
necessary because the air valve fails open and the fuel valve firing fuel gas. When burning exothermic waste requiring
fails closed. Safety concerns dictate failure modes. Fuel should quench air, the temperature controller output decreases, pro-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

always fail to minimum and air should fail to maximum. viding low fire fuel at 12 mA, then quench air below 12 mA.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 391 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 391

FIGURE 12.20 Electronically linked parallel positioning.

FIGURE 12.21 A variation of parallel positioning.

The TIC output is actually 4 to 20 mA. The description of 12.3.1.3 Characterizer Calculations
the action of the receivers uses the term “split-ranged.” For Parallel positioning of a globe-type fuel gas valve and a
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

example, the TIC applies the entire 4 to 20 mA range to the butterfly-type combustion air valve requires characterizer
fuel gas valve, but the valve is configured to respond only to calculations as described below. Figure 12.20 shows the
the partial range of 12 to 20 mA. control scheme.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 392 Tuesday, March 22, 2005 11:48 AM

392 The John Zink Combustion Handbook

Gas Valve %Open vs %Flow the valve is approximately 25% open for many applications.
120 Maximum firing rate occurs between 70 and 80% open, result-
100
ing in a near-linear function of flow rate vs. valve position
80
throughout the firing range. The linear function is not neces-
% Flow

60
40
sary for configuring a combustion air characterizer, but is
20 useful for the application of a dynamic fuel:air ratio correction
0
0 20 40 60 80 100
to the control circuit.
% Open Use of a positioner on the fuel gas valve establishes equal-
ity between percent control signal and percent valve opening.
FIGURE 12.22 Fuel flow rate versus control signal. Columns 1 and 2 of Table 12.1 show gas valve data.

TABLE 12.1 Gas Valve Data 12.3.1.3.2 Step 2: Air Flow Rate vs. Air Valve Position
Fuel Gas Calculation of air flow rate versus vs. position that is predict-
Control Signal TIC Gas Valve Flow Rate able and repeatable requires:
Output % % Open %

10 10 10
1. known and repeatable valve inlet pressure vs. flow rate
20 20 16 2. near-constant temperature
30 30 25 3. high-quality positioner on the valve
40 40 39 4. knowledge of valve coefficient and pressure recovery fac-
50 50 58 tor of the air valve at all valve positions
60 60 83
70 70 100
5. fixed and known flow (pressure drop) geometry down-
80 80 107 stream of the control valve
90 90 109
100 100 110 Figure 12.23 shows the results of a typical butterfly-type
valve calculation for Step 2. The low fire mechanical stop is
normally set at approximately 20% open.

12.3.1.3.3 Step 3: Air Valve Characterizer

Three general steps are required to define the characterizer: Table 12.2 combines data from Figure 12.23 with data from
1. Calculate and graph fuel flow rate vs. control signal. Table 12.1. Air valve graph data are tabulated in columns 3
2. Calculate and graph combustion air flow rate vs. control and 4. Figure 12.24 is a plot of the data from columns 1 and 4
signal. and represents the required shape of the air valve character-
3. Tabulate and graph air valve characterizer. izer. The TIC output signal is the characterizer input and is
plotted on the x-axis. The characterizer output is the percent
12.3.1.3.1 Step 1: Fuel Flow Rate vs. Control Signal open of the air valve and is shown on the y-axis. Many char-
Predictable and repeatable calculations of fuel gas flow rate acterizer instruments are available that will model a curved
vs. control valve position require: response using straight-line segments. This characterizer is
1. pressure regulator upstream of control valve to provide
sufficiently defined using three segments.
constant inlet pressure (varying inlet pressure can be used
only if it is repeatable with flow rate.) 12.3.2 Fully Metered Cross Limiting
2. constant temperature and composition of fuel gas Development of a fully metered control scheme for modulating
3. high-quality positioner on the control valve to eliminate fuel and air to a burner begins with the electronically linked par-
hysteresis and to ensure that valve percent open equals allel positioning scheme as previously shown in Figure 12.20.
percent control signal Figure 12.25 adds flow meters and flow controllers.
4. knowledge of valve coefficient vs. valve position Flow meters are linear with flow rate. Meter output signal
throughout the control valve range, including the pres- scaling provides the firing rate and air:fuel ratio required for
sure recovery factor
the application. The combustion air characterizer used for
5. fixed and known pressure drop geometry downstream of
parallel positioning is not required because the transmitters
the control valve
are linear with flow rate.
6. subsonic regime throughout the flow range
In the illustration, the temperature controller TIC output
Results of Step 1 are shown in Figure 12.22 for a typical
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
sets the firing rate by serving as setpoint to each flow con-
fuel gas valve with equal percent trim. Low fire position of troller. Signal inversion, shown as (1 minus parameter value)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 393 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 393

in the parallel positioning scheme, is not required. Instead, Air Valve %Open vs %Flow
controller output mode is configured to match the valve 160
140
failure mode. 120

% Flow
100
Controller output mode, reverse or direct acting, defines the 60
80
change in output signal direction with respect to process vari- 40
20
able changes. For example, if the controller output increases 0
0 50 100
as the process variable increases, the controller mode is direct % Open
acting. In combustion control schemes, fail-closed fuel valves
require a reverse-acting flow controller, while fail-open com- FIGURE 12.23 Typical butterfly-type valve calculation.
bustion air valves require direct acting flow controllers. From
controller mode definitions, it is clear that the temperature TABLE 12.2 Data for Characterizer
controller (TIC) should be reverse acting. That is, the TIC Fuel Gas Combustion
Control Signal TIC Flow Rate Air Flow Rate Air Valve
output should decrease, reducing the firing rate, in response Output % % % % Open
to an increase in temperature, the process variable.
10 10 10 5
Addition of high and low signal selectors provides cross lim- 20 16 16 13
iting of the fully metered control scheme, as shown in Figure 30 25 25 22
12.26. The low signal selector (<) compares demanded firing rate 40 39 39 29
50 58 58 33
from the TIC to the actual combustion air flow rate and applies
60 83 83 41
the lower of the two signals as the setpoint to the fuel flow 70 100 100 46
controller. The low signal selector ensures that the fuel setpoint 80 107 107 48
cannot exceed the amount of air available for combustion. 90 109 109 49
100 110 110 50
The high signal selector (>) compares demanded firing rate
from the TIC to actual fuel flow rate and applies the higher
of the two signals as the setpoint to the airflow controller. Air Valve Characterizer
60
This ensures that the air setpoint is never lower than required 50
% Output

for combustion of actual fuel flow rate. 40

30
Together, the high and low signal selectors ensure that 20
unburned fuel does not occur in the combustion system. 10
0
Unburned fuel accumulations can cause explosions. Cross 0 50 100
limiting by the signal selectors causes air flow to lead fuel % Input

flow during load increases and for air flow to lag fuel flow
during fuel decreases. This lead/lag action explains why the
FIGURE 12.24 The required shape of the air valve
characterizer.
fully metered cross-limiting control system is often called
“lead-lag” control. Whatever the name, the system performs
the function of maintaining the desired air:fuel mixture during
multiplier gives a fixed trim gain. Substituting a summing
load changes. The system also provides fuel flow rate reduc-
function for the multiplier would result in high trim gain at
tion in the event air flow is lost or decreased.
low flow rates and could produce a combustion air deficiency.
It is possible to trim the control scheme using measurement
of flue gas oxygen content, as illustrated in Figure 12.27. For Oxygen trim may be applied to the combustion airflow
most systems, the oxygen signal should be used to “trim,” and controller setpoint rather than the flow transmitter signal. If
not be a primary control. Many oxygen analyzers are high this technique is used, the airflow signal to the low signal
maintenance and/or too slow in response to be used as a selector must retain trim modification (see Figure 12.28 for
primary control in the combustion process. As shown, the the scheme).
oxygen controller is utilized for setpoint injection and provides
tuning parameters to help process customization. High and Multiple fuels and oxygen sources are accommodated by
low signal limiters restrict the oxygen controller output to a the cross-limiting scheme, as shown in Figure 12.29. When
trimming function, normally 5 to 10% of the normal combus- multiple fuels are used, heating values must be normalized
tion air flow rate. by adjusting flow transmitter spans or by addition of heating
A multiplication function (X) in the combustion airflow value multipliers. Similar methods are used to normalize
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
transmitter signal makes the oxygen trim adjustment. The oxygen content for multiple air sources.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 394 Tuesday, March 22, 2005 11:48 AM

394 The John Zink Combustion Handbook

FIGURE 12.25 Fully metered control scheme.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 12.26 Fully metered control scheme with cross limiting.

12.4 CONTROLLERS other devices by analog signals. The analog signal is usually
4 to 20 mA, DC.
Controllers have historically been called analog controllers In Figure 12.30, the setpoint is a signal representing the
because the process and I/O signals are usually analog. Con- desired value of a process. If the process is flow rate, the
troller internal functions performed within a computer or setpoint is the desired flow rate. Setpoint signals can be gen-
microprocessor by algorithm are sometimes called digital erated internally within the controller, called the local set-
controllers, although the I/O largely remains analog. Some point, or may be an external signal, called the remote setpoint.
digital controllers communicate with other devices via digital Controller output, called the controlled variable (CV) or
communication, but for the most part, controllers connect the manipulated variable (MV), connects to a final element
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 395 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 395

FIGURE 12.27 O2 trim of air flow rate.

in the process. In this example of flow control, the final of error by increasing controller gain is limited by controller
element is probably a control valve. Feedback from the pro- instability at high gain.
cess, called the process variable (PV) in this example, could Offset is the term given to the difference between the set-
be the signal from a flow meter. point and process variable. Correction of offset was the first
The controlled variable (CV) is generated within the con- improvement made to the original controller. Offset correc-
troller by subtraction of feedback (PV) from the setpoint (SP), tion was accomplished by adding bias to the controller output:
generating an error signal e, which is multiplied by a gain K.
Output(CV) = eK + Bias (12.2)
The product eK is the controller output (CV).
Bias adjustment required operator manipulation of a knob or
Output(CV) = (SP − PV)( K ) = eK (12.1) lever on the controller, which added bias until setpoint and
process variables were equal. The operator considered the
This simple controller is an example of the first controller controller “reset” when equality occurred. Each setpoint
built in the early 20th century and is called a proportional change or process gain change required a manual reset of the
controller. The output is proportional to the error signal. The controller. Figure 12.31 illustrates the proportional controller
proportional factor is the gain K. Proportional controllers with manual reset.
require an error (e) to produce an output. If the error is zero, Many operators prefer the term “proportional band” when
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

the controller output is zero. To obtain an output that produces describing controller gain. Proportional band is defined as:
the correct value of the process variable, the operator is
required to adjust the setpoint higher than the desired process 100 100%
Proportional band ( PB) = = (12.3)
variable in order to create the requisite error signal. Reduction Gain K
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 396 Tuesday, March 22, 2005 11:48 AM

396 The John Zink Combustion Handbook

FIGURE 12.28 O2 trim of air setpoint.

Proportional band represents the percent change in the pro- reset) is highest with large errors and continues until the error
cess variable (PV) required to change the controller output is reduced to zero. Figure 12.32 illustrates automatic reset.
100%. For example, a controller gain of 1 (K = 1.0) requires Automatic reset is the “I” component of a PID (three-mode)
a PV change of 100% to obtain a 100% change in controller controller. P is proportional gain and D is derivative (rate)
output. Proportional bands for combustion process variables gain. I is expressed as repeats per minute (RPM), or as min-
are generally in the range of 1000% to 20% (Gain = 0.1 utes per repeat (MPR), depending on the controller manufac-
to 5.0). Flow controller gains are always less than unity. turer’s choice of terms. Some controllers permit user selection
Temperature controller gains vary from 0.1 to 3.0, depend- of the term. Controller output is the same regardless of termin-
ing on the process gain. Pressure controllers generally have ology, but the operator must know and apply proper tuning
gains higher than those of flow or temperature controllers. constants. For example, if tuning requires an I of 2 RPM, the
Controller output becomes unstable (oscillatory) when the operator must enter 0.5 MPR into the controller if MPR is
gain is too high. When instability occurs, the controller oper- the terminology in use. Integral gain of 0.5 MPR means that
ating mode must be changed from automatic to manual to automatic reset equal to the proportional gain will be applied
stabilize the process and prevent equipment damage. A at the controller output each 30 seconds. Integral gain is a
reduction of controller gain must occur before a return to smooth continuous process that contributes phase lag to the
automatic mode. system. Additional phase lag contributes to system instability
Automatic reset was the next improvement to the process (oscillation), which prohibits high values of integral gain.
controller. This was a most welcome addition that eliminated Derivative gain D is a function of how fast the process
the need for manual reset. Automatic reset is a time integral variable is changing. For slow changing processes, derivative
of the error signal, summed with the proportional gain signal gain is of little use. Derivative gain is not used on flow control
to produce the controller output. Integral gain (controller
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
loops with head meters or on other loops with noisy process
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 397 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 397

FIGURE 12.29 Multiple fuels and O2 sources.

Controller Process

Controlled
Setpoint (S.P.) (+) e Variable (CV)
+ K

(-)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Feedback (PV)

FIGURE 12.30 Controller.

variable signals. High noise levels will drive derivative gain direction of controller output when the process variable
to instability. Derivative gain contributes phase lead that can changes. Reverse acting means the controller output decreases
sometimes be beneficial. if the process variable increases. An example illustrates how
Controllers have many modes of operation. P, I, D, auto- to select reverse or direct.
matic, and manual modes have been discussed. Reverse or In this example of flow control, the process variable
direct mode is another choice that must be configured to (flow rate) increases when the final element (control valve)
match the process. Reverse or direct describes the change in opens. In addition, the flow meter output or process variable
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 398 Tuesday, March 22, 2005 11:48 AM

398 The John Zink Combustion Handbook

Controller Process

Controlled
Setpoint (S.P.)
(+) e Variable (CV)
+ K +
(-) Bias

Feedback (PV)

FIGURE 12.31 Analog controller with manual reset.

Controller Process

Controlled
Setpoint (S.P.)
(+) e Variable (CV)
+ K +
(-) Time
Integral

Feedback (PV)

FIGURE 12.32 Analog controller with automatic reset.

(PV) increases with increased flow rate. If the control valve manual tuning. A controller tuned at low flow rates or low
fails closed, and opens on increasing signal (increasing con- temperature could become unstable at high flow rates or high
troller output), the controller mode must be reverse acting. temperatures. Control of most process loops benefits from
That is, if PV increases, the controller output must decrease addition of feedforward components that relieve the feedback
to close the valve and restore flow rate to the correct value. controller of primary control. Operation improves if the feed-
If the control valve fails open (closes on increasing signal), back controller functions as setpoint injection and error trim-
the controller mode must be direct acting. This example illus- ming of the feedforward system.
trates the need to know if each element in a control loop is
Many processes controlled by a current proportional con-
reverse or direct acting, including transmitters, isolators,
troller successfully use the tuning procedure below. The pro-
transducers (such as I/Ps), positioners, and actuators. Proper
cess must be upset to produce oscillations of the process
selection and configuration of loop elements provide not only
variable. A graphic recorder should be used to determine
proper operation, but also proper failure mode. Reversal of
when the oscillations are constant and to ascertain the time
any two elements within a loop will not affect loop response,
for one cycle (oscillation).
but failure modes will change.
1. In manual mode, adjust the output to bring the process
variable (PV) near the desired value.
12.5 TUNING
Many modern controllers have built-in automatic tuning rou- 2. Set the Rate Time to 0 minutes and set Reset Time to the
tines. Tuning parameters calculated automatically require a maximum value (50.00 min), or set repeats per minute
loop upset to enable calculation. Parameters are normally de- (RPM) to the minimum value to minimize reset action.
tuned considerably from optimum because process gains are 3. Increase gain (decrease proportional band PB) signifi-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

often nonlinear. Variable loop gain can also be a problem for cantly. Try a factor of 10.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch12-Frame Page 399 Tuesday, March 22, 2005 11:48 AM

Combustion Controls 399

4. Adjust local setpoint to equal PV and switch to automatic 2. Combustion Control, 9ATM1, Fisher Controls, Mar-
mode. shalltown, IA, 1976.
5. Increase the setpoint by 5 or 10% and observe PV
response. 3. Boiler Control, Application Data Sheet 3028, Rose-
6. If the process variable oscillates, determine the time for mount, Inc., Minneapolis, MN, 1980.
one oscillation. If it does not oscillate, return to the orig-
inal setpoint, increase the gain again by a factor of 2, and 4. Instrumentation Symbols and Identification, ANSI/ISA
repeat Step 5. – S5.1 – 1984, Instrument Society of America, Research
7. If the oscillation of Step 6 dampens before cycle time is Triangle Park, NC, 1984.
measured, increase the gain slightly and try again. If the
oscillation amplitude becomes excessive, decrease gain 5. Temperature Measurement Thermocouples, ANSI –
slightly and try again. MC96.1 – 1984, Instrument Society of America,
8. Record the current value of gain, and record the value for Research Triangle Park, NC, 1984.
one completed oscillation of PV.
9. Calculate gain, reset, and rate: 6. F.G. Shinskey, Process Control Systems, Application,
a. For PI (two-mode controller): Design, and Tuning, 3rd ed., McGraw-Hill, New York,
Gain = Measured gain × 0.5 1988.
Reset time = Measured time/1.2 (MPR)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

b. For PID (three-mode controller): 7. M.J.G. Polonyi, PID controller tuning using standard form
Gain = Measured gain × 0.6 optimization, Control Engineering, March, 102-106,
Reset time = Measured time/2.0 (MPR) 1989.
Rate = Measured time/8.0 (min)
10. Enter the values of gain, reset, and rate into the controller. 8. D.W. St. Clair, Improving control loop performance,
11. Make additional trimming adjustments, if necessary, to Control Engineering, Oct., 141-143, 1991.
fine-tune the controller.
9. Controller Tuning, Section 11, UDC 3300 Digital
12. To reduce overshoot: less gain, perhaps a longer rate time.
13. To increase overshoot or increase speed of response: more Controller Product Manual, Honeywell Industrial
gain, perhaps shorter rate time. Automation, Fort Washington, PA, 1992.

10. F.Y. Thomasson, Five steps to better PID control, CON-

TROL, April, 65-67, 1995.
REFERENCES
11. API Recommended Practice 556: Instrumentation and
1. J.O. Hougen, Measurement and Control Applications Control Systems for Fired Heaters and Steam Generators,
for Practicing Engineers, CAHNERS Books, Barnes & 1st ed., American Petroleum Institute, Washington,
Noble Series for Professional Development, 1972. D.C., May 1997.

Copyright CRC Press

response-surface
contours
p
all direction of
a2l steepest
ascent

old design
new design
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 13
Experimental Design for
Combustion Equipment
Joseph Colannino

TABLE OF CONTENTS

13.1 Introduction to Experimental Design...................................................................................................... 402

13.1.1 The Power of SED: A Burner NOx Example ........................................................................... 402
13.1.2 Statistical Experimental Design Principles............................................................................... 404
13.1.3 The Method of Least Squares ................................................................................................... 405
13.1.4 Matrix Solution ......................................................................................................................... 406
13.1.5 Linear Transformations............................................................................................................. 407
13.2 Important Statistics ................................................................................................................................. 408
13.2.1 The Analysis of Variance (ANOVA) ......................................................................................... 409
13.2.2 The F-distribution ..................................................................................................................... 409
13.2.3 ANOVA with Separate Model Effects ...................................................................................... 409
13.2.4 Pooling Insignificant Effects..................................................................................................... 411
13.2.5 Standard Errors of Effects......................................................................................................... 412
13.3 Two-level Factorial Designs.................................................................................................................... 412
13.3.1 Interactions................................................................................................................................ 412
13.3.2 Pure Error and Bias................................................................................................................... 412
13.3.3 Two-level Fractional Factorials................................................................................................. 414
13.3.4 Screening Designs..................................................................................................................... 415
13.3.5 Method of Steepest Ascent ....................................................................................................... 416
13.3.6 Serial Correlation and Lurking Factors..................................................................................... 416
13.3.7 Foldover .................................................................................................................................... 416
13.3.8 Orthogonal Blocking................................................................................................................. 417
13.3.9 Including Categorical Factors ................................................................................................... 418
13.4 Second-order Designs ............................................................................................................................. 419
13.4.1 Central-composite Designs ....................................................................................................... 419
13.4.2 Practical Considerations ........................................................................................................... 420
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- 401
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 402 Wednesday, April 6, 2005 4:39 PM

402 The John Zink Combustion Handbook

13.5 Accounting for Fuel Mixtures.....................................................................................................................................420

13.5.1 Experimental Designs for Mixtures ..............................................................................................................421
13.5.2 Orthogonal Mixture Designs.........................................................................................................................423
13.5.3 Combining Mixture and Factorial Designs...................................................................................................423
13.5.4 Building 2n-level Factorials from Two-level Factorials ................................................................................424
13.6 Combining Domain Knowledge with SED.................................................................................................................424
13.6.1 Practical Considerations................................................................................................................................425
13.6.2 Semi-empirical Models.................................................................................................................................425
13.6.3 Sequential Experimental Strategies ..............................................................................................................426
13.7 Linear Algebra Primer.................................................................................................................................................427
13.7.1 Taylor and MacLaurin Series Approximations .............................................................................................427
13.7.2 Matrix Multiplication....................................................................................................................................428
13.7.3 Identity, Inverse, and Transpose....................................................................................................................428
13.7.4 Matrix Addition.............................................................................................................................................429

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
References ....................................................................................................................................................................................429

13.1 INTRODUCTION TO Third, statisticians rather than engineers write most SED
EXPERIMENTAL DESIGN texts. Therefore, SED texts usually do not incorporate the
Statistical experimental design (SED) is a method for con- domain knowledge that engineers find so indispensable.
structing experiments that will mute the muddling effect of Domain knowledge is a specialized understanding in a
experimental error and increase experimental efficiency. SED nonstatistical branch of science or engineering — for example,
leads to better and less-expensive data analysis. It is far supe- NOx formation or combustion fundamentals. This chapter
rior to the classical one-factor-at-a-time experimentation often provides a cursory overview of SED related to the performance
taught in school. Classical experimentation cannot account for characterization of burners and combustion equipment.
variable interactions. Classical experimentation often leads to
false conclusions, for example, that one has arrived at an opti- 13.1.1 The Power of SED: A Burner NOx
mal place, when in fact one has not. Figure 13.1 contrasts clas- Example
sical and SED methods. SED is a powerful tool used by too Consider a manufacturer that makes many different burners.
few engineers. This disuse is due to several factors. First, engi- To compete, the company must make certain performance
neers and scientists can successfully (though not as efficiently) guarantees. These guarantees could concern heat release,
experiment without SED. SED is a power tool. In the hands of turndown, flame length, heat flux profile, or combustion-
a skilled practitioner, it reduces the time for experimentation related pollutants like NOx and CO. Uncertainty in any of
and squeezes the most information from the data. these areas may force the company to decline to bid or to
Second, SED vocabulary contains alien terms because stat- increase the bid price to cover the risk of redesign. Neither of
isticians first applied the methods to agricultural problems. these alternatives is attractive. Declining to bid surrenders the
Terms like “treatment” and “block” have obvious meanings in job to a competitor. Inflating the price to cover risk will make
agriculture and obscure meanings to the practicing engineer. the burner less competitive. Even if the company wins the
Engineers prefer terms like variable and experimental series. job, it may spend more than it anticipated achieving the per-
This chapter is written from an engineering perspective, rather formance it guaranteed. Thus, performance uncertainty trans-
than a statistical perspective. However, the bulk of SED knowl- lates directly to lost profit and lost opportunity.
edge is in the statistical literature. Therefore, this chapter con- Suppose a burner generates too much NOx. Many factors
tains a judicious choice of vocabulary to allow the interested influence NOx response; Table 13.1 provides a partial list.
reader to consult statistical treatises for further reference. The terms “response” and “factor” are used in a very specific
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 403 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 403

Classical (one-factor-at-a-time) SED

General Procedure General Procedure
1. Run a series of experimental points in sequential 1. Design a balanced experimental series using SED
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

increments holding all other factors constant. principles.

2. Search for the “optimum” along that factor. 2. Run the experimental points in random order.
3. Hold the factor constant at its “optimum” and 3. Construct an interpolating function and perform
repeat Steps 1 and 2 for the next factor. various statistical tests.
4. Continue in this way for all factors, until arriving 4. To arrive at an optimum, use validated SED tools
at the “optimal” point. (e.g., interpolation, method of steepest ascent,
sequential design, etc.).
Specific Example
NOx (contours shown above) is a function of two Specific Example
factors (x1 and x2) in the above hypothetical example The investigator begins with a factorial design (points
(perhaps firing rate and excess oxygen, or whatever). 1-6). SED uses the centerpoints (2, 4) to test (and
The investigator seeks to minimize NOx, stepping along reveal) curvature. So, the investigator adds
x1. The NOx continues to decrease along x1 until complementary points (7-12) forming a central
reaching point 5. So, the investigator backtracks and composite design. This design gives a full second-
arrives at an “optimum” (~45ppm, point 6). The order interpolating polynomial for the region and
investigator holds x1 at its “optimum” level and properly estimates the optimum.
proceeds to investigate along x2. In like manner, the
investigator arrives at point 12 and declares the point General Considerations
“optimal.” 1. The investigator better estimates the true optimum
in the experimental space.
General Considerations
2. The design gives a second-order polynomial that
1. The investigator is unaware that the “optimum” is
accounts for curvature and interactions between the
specious – more improvement is available.
factors.
2. The design does not account for interaction
3. The balanced design gives good estimates over the
between factors.
entire experimental region, not just near the factor
3. The unbalanced design poorly maps the region:
axes.
classical designs require more points than SED
4. In general, SED gives better information and more
designs to generate the same information. This is
statistically valid models than classical designs.
especially so as the number of factors increase.

FIGURE 13.1 Contrast of classical experimentation and SED methods.

sense. The response is any dependent variable or output of action of the factors must be known. What is the form of the
interest. A factor is any input or independent variable that model? What explicit equation should be used? What are the
affects the response. A given system of interest can have many coefficient values? Proceeding from a purely theoretical basis,
factors and associated responses. it would not be possible to arrive at a reliable explicit formu-
Specifying a response and its factors is a necessary first lation for NOx performance. However, using statistical meth-
step, but it is not sufficient to solve the problem. The inter- ods, the solution is quite tractable. Equation (13.1) gives the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 404 Wednesday, April 6, 2005 4:39 PM

404 The John Zink Combustion Handbook

TABLE 13.1 Some Potential Factors Affecting NOx Response from a Burner
Operating Factors Burner Factors Furnace Factors Ambient Factors
(Constrained by Process) (Controlled by Burner Manufacturer) (Controlled by Heater Design) (Uncontrollable)

Degree of air preheat Burner throat diameter Available air-side pressure drop Ambient humidity
Firing rate Degree of air staging Burner-to-burner spacing Ambient air temperature
Fuel composition Degree of fuel staging Burner-to-furnace wall spacing Barometric pressure
Furnace temperature Fuel port arrangement Heat release/furnace vol. ratio
Fuel pressure Fuel port diameter
Oxygen concentration Multiple combustion zones
Process fluid flow rate Number of fuel ports

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
2200 x1: the reciprocal of the burner spacing
x2: the reciprocal preheat temperature
Furnace Temperature, F

x3: the reciprocal furnace temperature

2100 x4: the oxygen concentration
x5: the heat release density
x6: the hydrogen mole fraction in the fuel
2000 x7: the mole fraction of alkenes (olefins)
x8: the mole fraction of non-methane alkanes (saturates).
1900 Equation (13.1) reveals:
NOx = 10 ppm 20 30 40 50

1800
1. an explicit functional relationship of the response to the
1 2 3 4 given factors
2. the basis for derivative figures and graphs such as Figure
Excess Oxygen, %
13.2
FIGURE 13.2 NOx contours for furnace temperature 3. the statistical significance of each coefficient (using the
and oxygen concentration based on Eq. (13.1). standard errors)
4. safety margin for NOx guarantees
5. an equation for feedforward NOx control1

relation for a family of burners manufactured by John Zink, One can use this knowledge to derive a competitive advan-
and Figure 13.2 shows a graphical representation for two of tage or control process units with simultaneous constraints
the factors, furnace temperature and excess oxygen. for product and performance.

13.1.2 Statistical Experimental Design

y = 3.062[±0.10] + 0.107[±0.011]x1 Principles
−0.096[±0.011]x 2 − 0.142[±0.020]x3 A team of engineers proposes to correlate NOx data collected
from the operation of a fuel-staged burner. Fuel staging seg-
+0.576[±0.018]x 4 + 0.083[±0.014]x 5 regates the combustion into two or more distinct zones and
lowers NOx. Figure 13.3 shows a fuel-staged burner. Based
+0.067[±0.011]x 6 + 0.057[±0.013]x 7 on experience, the team proposes the model of Eq. (13.2).
+0.027[±0.013]x8 (13.1)
y = a0 + a1ξ1 + a2 ξ 2 + a3 ξ 3 + a12 ξ1ξ 2 + ε (13.2)
where y is the natural log of the NOx mole fraction (ppm).
The values in brackets [ ] are the standard errors associated where y is the NOx response (ppm)
with each coefficient, which are described in Section 13.2. a0 – a12 are the coefficients
The factors have the following associations. ξ1 is the heat release (106 Btu/hr)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 405 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 405

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 13.3 Fuel-staged burner.

ξ2 is the oxygen concentration (%) TABLE 13.2 NOx as a Function of Burner Geometry and
ξ3 is the percent of fuel to the primary Operation
combustion zone (%) NOx Firing Rate, Oxygen, Conc., Fuel to Primary,
Run (ppm) [106 Btu/hr (ξ1)] [% (ξ2)] [% (ξ3)]
ξ1ξ2 is the interaction between the heat release
and the oxygen (% × 106 Btu/hr) 1 13, 14 7 1 20
2 18, 19 13 1 20
ε is the experimental error (ppm) 3 27, 24 7 5 20
4 26, 24 13 5 20
Interaction accounts for synergy between factors — some- 5 18, 19 7 1 50
6 21, 22 13 1 50
thing classical experimentation cannot do. In the present 7 28, 29 7 5 50
case, the investigators proposed Eq. (13.2) based on their 8 29, 27 13 5 50
experience. However, with SED, one arrives at the form
directly, as Section 13.2.3 will demonstrate.
squares calculates coefficients such that they minimize the
13.1.3 The Method of Least Squares sum of the squared deviations from the presumed model. The
Table 13.2 gives two replicates of the NOx response for nom- model need not be linear and may contain quadratics, tran-
inal values of each factor. The replicates differ presumably by scendental functions, etc. Table 13.2 comprises eight experi-
experimental error. The value of the coefficients is found mental conditions and 16 NOx values. Equation (13.3)
using the technique of least squares. The technique of least explicitly indexes each response.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 406 Wednesday, April 6, 2005 4:39 PM

406 The John Zink Combustion Handbook

∑ [ y − (a ]
 y1   a0 + a1ξ1,1 + a2 ξ 2,1 + a3 ξ 3,1 + a12 ξ1,1ξ 2,1  ∂
+ a1ξ1 + a2 ξ 2 + a3 ξ 3 + a12 ξ1ξ 2 ) = 0 (13.9)
2

 y2   a0 + a1ξ1,2 + a2 ξ 2,2 + a3 ξ 3,2 + a12 ξ1,1ξ 2,2  ∂a2 0

  = 
 M   M M M M M 
∑ [ y − (a ]
 y  a + a ξ + a ξ + a ξ + a ξ ξ  ∂
+ a1ξ1 + a2 ξ 2 + a3ξ 3 + a12 ξ1ξ 2 ) = 0 (13.10)
2
 16   0 1 1,16 2 2 ,16 3 3,16 12 1,16 2 ,16 
∂a3 0

 ε1 

∑ [ y − (a ]
 ε2  ∂
+ a1ξ1 + a2 ξ 2 + a3 ξ 3 + a12 ξ1ξ 2 ) = 0 (13.11)
2
+  (13.3) ∂a12 0
 M 
ε 
 16  Equations (13.7) through (13.11) reduce to five simulta-
neous equations, one for each coefficient. Equation (13.12)
Although there are 16 observations of the response, only expresses them in a single (symmetrical) matrix form:
five coefficients — a0, a1, a2, a3, and a12 — are calculated.
Equation (13.4) is expressed in matrix form. For a brief review
of matrices, see Section 13.7.

 ∑ y   n ∑ ξ ∑ ξ ∑ ξ ∑ ξ ξ   a  1 2 3 1 2

∑ ξ y   ∑ ξ ∑ ξ ∑ ξ ξ ∑ ξ ξ ∑ ξ ξ   a 
0
 1 1
2
1 1 2 1 3
2
1 2
 1

 a0 
 y1  1 ξ1,1 ξ 2,1 ξ 3,1 ξ1,1ξ 2,1     ε1 


∑ ξ y  =  ∑ ξ ∑ ξ ξ ∑ ξ ∑ ξ ξ ∑ ξ ξ   a 
2 2 1 2
2
2 2 3
2
1 2 2

a
 y2  1 ξ1,2 ξ 2,2 ξ 3,2 ξ1,2 ξ 2,2   1   ε 2  

∑ ξ y   ∑ ξ ∑ ξ ξ ∑ ξ ξ ∑ ξ ∑ ξ ξ ξ   aa 
3 3 1 3 2 3
2
3 1 2 3
3

  =
 M  M M M M
  a2  +  
M    M 
(13.4) 
 ∑ ξ ξ y  ∑ ξ ξ ∑ ξ ξ ∑ ξ ξ ∑ ξ ξ ξ ∑ ξ ξ 
1 2 1 2
2
1 2
2
1 2 1 2 3
2 2
1 2
4

 y  1 ξ  a3  
 16   1,16 ξ 2,16 ξ 3,16 ξ1,16 ξ 2,16     ε16  (13.12)
 a12 
Applying the data of Table 13.2 to Eq. (13.12) gives Eq.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

To obtain a compact notation, denote the response vector (13.13):

by y, the test matrix by ξ, the coefficient vector by a, and the
error vector by ε. The single underline denotes a vector (single  358  16 160 48 560 480  a0 
column), and a double underline denotes a matrix comprising  3622  160 1744 480 5600 5232  a1 
any number of rows and columns.     
 1214 =  48 480 208 1680 2080  a2  (13.13)
12950  560 5600 1680 23200 16800  a 
y = ξa + ε (13.5)      3 
 12158  480 5232 2080 16800 22672  a12 
Equations (13.2 through 13.5) represent identical models.
The various forms are used as needed. With the method of Equation (13.13) yields its solution by the usual rules of
least squares, the intent is to drive the sum of the squared matrix arithmetic (see next section):
errors as close to zero as possible:
 a0   3.19
n =16 n =16
 a1   0.85
∑ ε = ∑ [ y − (a
i =1
2
i
i =1
i 0 + a1ξ1,i + a2 ξ 2,i + a3 ξ 3,i   
 a2  =  4.06

(13.14)
(13.6)  a   0.12
)]  3   
2
+ a12 ξ1,i ξ 2,i →0  a12   −0.19

Equating first derivatives to zero gives the minimum. For

13.1.4 Matrix Solution
simplicity, drop the indices in Eqs. (13.7) through (13.11):
Consider once again Eq. (13.5). Multiply both sides of the
∂
∑ [ y − (a ]
+ a1ξ1 + a2 ξ 2 + a3 ξ 3 + a12 ξ1ξ 2 ) = 0 (13.7)
2 equation by the transpose of the test matrix ξ ( ) (see Section
T

∂a 0 0
13.7.3) to obtain Eq. (13.15):

∑ [ y − (a ]
∂
+ a1ξ1 + a2 ξ 2 + a3 ξ 3 + a12 ξ1ξ 2 ) = 0 (13.8)
( )
2
ξ y = ξ ξa + ξ ε = 0
T T T
∂a1 0 (13.15)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 407 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 407

The reader can verify that Eq. (13.15) and Eq. (13.12) are TABLE 13.3 Transforms for Table 13.4
the same least-squares matrix. Thus, Eq. (13.15) provides a Factor Raw Value Transform Transformed
shortcut for obtaining the least-squares equations without the i Description (ξi) (ξi to xi) (xi)
need for differential calculus. Premultiplying by the inverse 1 Firing rate, 7 –1
ξ 1 − 10 × 10 6 Btu/h

( )
−1 106 Btu/h 13 x1 = +1
3 × 10 6 Btu/h
matrix ξ ξ
T
yields the solution for the coefficient vector:
2 Oxygen 1 ξ 2 − 3% –1
concentration, % 5 x2 = +1
2%

(ξ ξ )
−1 3 Fuel to 20 ξ 3 − 35% –1
ξ y=a
T T
(13.16) primary, % 50 x3 = +1
15%

Equations (13.15) and (13.16) are general solutions. This TABLE 13.4 Transformed Data for Fuel-staged Burner
result is quite important because most computer spreadsheets Run x1 x2 x3 y
can perform the transpose, matrix multiply, and matrix invert
1 – – – 13, 14
operations. 2 + – – 18, 19
But what do the coefficients mean? Because they come 3 – + – 27, 24
from simultaneous equations, it is impossible to gauge the 4 + + – 26, 24
5 – – + 18, 19
effect of a single coefficient viewed alone. Their magnitudes
6 + – + 21, 22
cannot be compared because the associated factors have dif- 7 – + + 28, 29
ferent ranges and units. Moreover, if the model is modified, 8 + + + 29, 27
all coefficients must be recalculated because Eq. (13.12) is a
system of simultaneous equations. Some method of scaling
the factors and transforming the equations to an independent understand they are unit values, only their sign need be
system is desirable. shown. With these transforms, the off-diagonal values in Eq.
(13.12) sum to zero. For the data of Table 13.2, the transfor-
13.1.5 Linear Transformations mation leads to Eq. (13.18), a diagonal matrix. (For conve-
Transforming the factors to a common dimensionless unit nience, the zero elements are omitted.)
scale allows a direct comparison of the coefficients and sim-
plifies the analysis. Linear transforms are preferred because
 358 16   a0 
they do not bend or nonuniformly stretch the data and they  14    a1 
16
are easy to invert. The following transformation satisfies the     
conditions:  70 =  16   a2  (13.18)
 28  16 a 
     3 
 −18  16  a12 
ξ i − ξi
xi = (13.17)
1
(ξ − ξ i − )
2 i+ If xTx generates a diagonal matrix, then x is orthogonal.
Orthogonal matrices generate independent least-squares
where xi is the ith variable, transformed to a dimensionless equations. The solution is as easy as dividing each element
range of ±1 of y by each diagonal element of x:
ξi is the ith untransformed variable in the
conventional metric (e.g., Btu/hr, %, etc.)  a0   22.375
ξ i is the mean value of ξi in the conventional metric  a1   0.875
ξi+ is the maximum value of ξi in the conventional    
metric  a2  =  4.375 (13.19)
 a   1.750
ξi– is the minimum value for ξi in the conventional  3   
metric  a12   −1.125

This transform normalizes ξi to xi , which is dimensionless Note also that removing any terms (e.g., eliminating the term
and spans the range –1 to +1. Table 13.3 gives the transforms, a1x1) does not affect the value of the remaining coefficients.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

and Table 13.4 shows the transformed factors. Since we This is not so for the simultaneous system, Eqs. (13.12) to
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 408 Wednesday, April 6, 2005 4:39 PM

408 The John Zink Combustion Handbook

(13.14). One can directly compare all the coefficients because n

they are dimensionless and have uniform range. The coeffi-

cients provide the following:
case, ∑ )
y−y = ∑ ( y − y) ). The simpler notation is used
i =1
i i

for convenience.)
1. the average NOx is 22.375 ppm for this experimental

∑ ( y − y)) , has the advantage of

series 2
The sum of the squares,
2. the firing rate influences the NOx by ±0.875 ppm over its
range a non-zero sum that increases as the error grows larger. How-

∑ ( y − y))
2
3. the oxygen influences NOx by ±4.375 ppm over its range ever, should grow larger with n and smaller with
4. the percentage of fuel to the primary zone influences NOx
p. If possible, the noise statistic should be independent of the
by ±1.75 ppm over its range
number of observations or the number of parameters in the
5. firing rate and oxygen interact to moderate their effect on
NOx by m1.125 ppm over their joint range model. This reasoning leads to Eq. (13.22), called the mean
square residual (MSR), as a logical measure of random error:
From an examination of the coefficients, oxygen has the
greatest influence on NOx, followed by the percentage of fuel
∑ ( y − y)
) 2
to the primary combustion zone, the interaction between firing
MSR = (13.22)
rate and oxygen, and the firing rate. Actually, the coefficient n− p
for the firing rate seems small. Is it significant? How can one
be sure that the effect is real? How does the coefficient com-
The divisor (n – p) is called the degrees of freedom.
pare to the experimental noise? The next section addresses
)
these questions. In the worst case, y is worthless and none of its factors
actually influence the response. In this case, the data merely
represent n replicates differing only by random error. The
13.2 IMPORTANT STATISTICS model is no better than the mean of all observations, y, defined
It is not enough to know the form of a model, or even calcu- by Eq. (13.23).
late its coefficients explicitly. Some indication of the influ-
ence of background noise and its relationship to the
coefficients is also necessary. The terms background noise,
noise, error, experimental error, pure error, and random error y=
∑y (13.23)
) n
are used interchangeably in this chapter. y is the best approx-
imation of the true but unknown model
Equation (13.23) is the simplest possible model, having
) only a single parameter (p = 1); the parameter is y, or a0 for
y = ξa (13.20)
the orthogonal designs considered. In such a case, Eq. (13.22)
reduces to Eq. (13.24), called the mean square total or MST:
The problem is to recover Eq. (13.20) from Eq. (13.5). The
data contain n total observations and the model contains a
∑ (y − y )
2
coefficient vector, a, comprising p coefficients. Substituting Eq.
(13.20) into Eq. (13.5) and solving for ε yields Eq. (13.22): MST = (13.24)
n −1
)
ε= y−y (13.21)
The squared deviation normalized by its degrees of free-
dom is the variance. The estimated mean and variance (y and
Equation (13.21) represents a vector of n values implicitly s2, respectively) are unbiased estimators for the actual mean
comprising p coefficients. A single number (statistic) to quan-
and variance (µ and σ2, respectively). In an experimental
tify the error vector is desired. Summing the vector
context, the actual mean and variance are usually unknown.
(∑ ) )
y − y is a logical place to start. However, if ε is truly One problem with s2 is that it does not have the same units
as y. The problem is remedied by taking the square root of
random and unbiased, then ∑ y − y) ~ 0. (For the present
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Eq. (13.22) or Eq. (13.24).
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 409 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 409

TABLE 13.5 Generic ANOVA Table

Component SS DF MS F

∑ (y − y)
) 2
Model (M) SSM = DFM = p – 1 MSM = SSM/DFM MSM/MSR

SSR = ∑ ( y − y )
Residual (R) ) 2
DFR = n – p MSR = SSR/DFR
i

SST = ∑ ( y − y )
2
Total (T) i
DFT = n – 1 MST = SST/DFT

1. the desired confidence in rejecting H0


∑ (y − y )
2
 ) 2. the degrees of freedom for the numerator (DFM in the
± MST = ± n −1
, if y = y present case)
s= (13.25)
∑ (y − y))
2 3. the degrees of freedom for the denominator (DFR in the
 ) present case)
± MSR = ± , if y ≠ y
 n− p These quantities have the notation Fconf(DFM, DFR), where
conf is the percent confidence. For example, F95(4,11) = 3.36
In Eq. (13.25), s is the standard deviation, having the same means an F-distribution with DFM = 4 and DFR = 11 will
)
units as the response. The hypothesis that y = y is called the reject the null hypothesis with at least 95% confidence if it
null hypothesis (H0). is greater than or equal to 3.36. The source for the critical
value of 3.36 is Table 13.6.
13.2.1 The Analysis of Variance (ANOVA) With this background, it is possible to test Model 13.2
Table 13.5 arranges the variances in a very convenient form, (Eq. 13.2) for significance. Table 13.7 gives the ANOVA for
referred to as the ANOVA table. The table has the following the data of Table 13.4 and the transformed Eq. (13.2). It appears
column headings: SS is the sum of squares, DF is the degrees that the model is significant, accounting for ~50 times the
of freedom, MS is the mean square, and F is a special ratio variance of the residual error. The calculated F-ratio of 48.7
described below. The row headings are M for model, R for with 6 and 9 degrees of freedom is compared to F95(6, 9) =
residual, and T for total. The column and row headings are 3.37. Because 48.7 is much greater than 3.37, the null hypoth-
combined to obtain the cell heading. For example, the cell esis is rejected. There is greater than 95% confidence that the
located at column SS and row M is SSM, standing for sum- model is significant. In fact, the confidence level is greater than
of-squares, model. 99%, because 48.7 > F99(6, 9) = 5.80. It is also possible to
The mean squares (MS) are calculated by dividing the sum estimate σ 2 ≈ MSR = 1.33 σ ≈ ± MSR = ±1.2 ppm NO x . ( )
of squares (SS) by the appropriate degrees of freedom (DF). However, although the model as a whole is significant, some
One should divide SST by (n – 1), the total number of data of the individual terms of the model may not be significant.
points, less one for the mean. SSM should be divided by DFM To see consider each term separately as presently described.
= (p – 1), the number of coefficients in the model, less one
for the mean. Finally, SSR should be divided by the remaining
13.2.3 ANOVA with Separate Model Effects
(or residual) degrees of freedom, DFR = (n – p). Now the
A single ANOVA table comprising separate model effects is
statistics needed to explore tests for significance are in place.
only possible with orthogonal data. This is another reason to
MSR estimates the noise. The ratio MSM/MSR is called F.
) use SED. To test nonorthogonal data, look at ANOVA tables
If F ~ 1, then y = y and the null hypothesis is accepted. If F
for every possible model combination comprising one to four
>> 1, then H0 is rejected and the model is significant. To find
effects. Test each model in a separate ANOVA table. This
how large F must be to reject H0, the properties of the F
requires 2p (16) different ANOVA tables.
distribution must be known.
For orthogonal data, one can construct a single table, Table
13.8, comprising every possible factor and two-factor interac-
13.2.2 The F-distribution tion. The ANOVA shows that only x1, x2, x3, and x1x2 are
Ratios of variance have an F-distribution. The distribution significant. The significant entries are shown in bold type. In
depends on three quantities: the previous section, s = 1.2 ppm was obtained. But if some
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

410 The John Zink Combustion Handbook

TABLE 13.6 F-Distribution, 99%, 95%, and 90% Confidence

DF ν1
DF ν2 Conf 1 2 3 4 5 6 7 8 9 10 15 20 25 30 50 100 Infinity

99% 4052 4999 5404 5624 5764 5859 5928 5981 6022 6056 6157 6209 6240 6260 6302 6334 6366
1 95% 161 199 216 225 230 234 237 239 241 242 246 248 249 250 252 253 254
90% 39.86 49.50 53.59 55.83 57.24 58.20 58.91 59.44 59.86 60.19 61.22 61.74 62.05 62.26 62.69 63.01 63.33

99% 98.50 99.00 99.16 99.25 99.30 99.33 99.36 99.38 99.39 99.40 99.43 99.45 99.46 99.47 99.48 99.49 99.50
2 95% 18.51 19.00 19.16 19.25 19.30 19.33 19.35 19.37 19.38 19.40 19.43 19.45 19.46 19.46 19.48 19.49 19.50
90% 8.53 9.00 9.16 9.24 9.29 9.33 9.35 9.37 9.38 9.39 9.42 9.44 9.45 9.46 9.47 9.48 9.49

99% 34.12 30.82 29.46 28.71 28.24 27.91 27.67 27.49 27.34 27.23 26.87 26.69 26.58 26.50 26.35 26.24 26.13
3 95% 10.13 9.55 9.28 9.12 9.01 8.94 8.89 8.85 8.81 8.79 8.70 8.66 8.63 8.62 8.58 8.55 8.53
90% 5.54 5.46 5.39 5.34 5.31 5.28 5.27 5.25 5.24 5.23 5.20 5.18 5.17 5.17 5.15 5.14 5.13

99% 21.20 18.00 16.69 15.98 15.52 15.21 14.98 14.80 14.66 14.55 14.20 14.02 13.91 13.84 13.69 13.58 13.46
4 95% 7.71 6.94 6.59 6.39 6.26 6.16 6.09 6.04 6.00 5.96 5.86 5.80 5.77 5.75 5.70 5.66 5.63
90% 4.54 4.32 4.19 4.11 4.05 4.01 3.98 3.95 3.94 3.92 3.87 3.84 3.83 3.82 3.80 3.78 3.76

99% 16.26 13.27 12.06 11.39 10.97 10.67 10.46 10.29 10.16 10.05 9.72 9.55 9.45 9.38 9.24 9.13 9.02
5 95% 6.61 5.79 5.41 5.19 5.05 4.95 4.88 4.82 4.77 4.74 4.62 4.56 4.52 4.50 4.44 4.41 4.36
90% 4.06 3.78 3.62 3.52 3.45 3.40 3.37 3.34 3.32 3.30 3.24 3.21 3.19 3.17 3.15 3.13 3.10

99% 13.75 10.92 9.78 9.15 8.75 8.47 8.26 8.10 7.98 7.87 7.56 7.40 7.30 7.23 7.09 6.99 6.88
6 95% 5.99 5.14 4.76 4.53 4.39 4.28 4.21 4.15 4.10 4.06 3.94 3.87 3.83 3.81 3.75 3.71 3.67
90% 3.78 3.46 3.29 3.18 3.11 3.05 3.01 2.98 2.96 2.94 2.87 2.84 2.81 2.80 2.77 2.75 2.72

99% 12.25 9.55 8.45 7.85 7.46 7.19 6.99 6.84 6.72 6.62 6.31 6.16 6.06 5.99 5.86 5.75 5.65
7 95% 5.59 4.74 4.35 4.12 3.97 3.87 3.79 3.73 3.68 3.64 3.51 3.44 3.40 3.38 3.32 3.27 3.23
90% 3.59 3.26 3.07 2.96 2.88 2.83 2.78 2.75 2.72 2.70 2.63 2.59 2.57 2.56 2.52 2.50 2.47

99% 11.26 8.65 7.59 7.01 6.33 6.37 6.18 6.03 5.91 5.81 5.52 5.36 5.26 5.20 5.07 4.96 4.86
8 95% 5.32 4.46 4.07 3.84 3.69 3.58 3.50 3.44 3.39 3.35 3.22 3.15 3.11 3.08 3.02 2.97 2.93
90% 3.46 3.11 2.92 2.81 2.73 2.67 2.62 2.59 2.56 2.54 2.46 2.42 2.40 2.38 2.35 2.32 2.29

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
99% 10.56 8.02 6.99 6.42 6.06 5.80 5.61 5.47 5.35 5.26 4.96 4.81 4.71 4.65 4.52 4.41 4.31
9 95% 5.12 4.26 3.86 3.63 3.48 3.37 3.29 3.23 3.18 3.14 3.01 2.94 2.89 2.86 2.80 2.76 2.71
90% 3.36 3.01 2.81 2.69 2.61 2.55 2.51 2.47 2.44 2.42 2.34 2.30 2.27 2.25 2.22 2.19 2.16

99% 10.04 7.56 6.55 5.99 5.64 5.39 5.20 5.06 4.94 4.85 4.56 4.41 4.31 4.25 4.12 4.01 3.91
10 95% 4.96 4.10 3.71 3.48 3.33 3.22 3.14 3.07 3.02 2.98 2.85 2.77 2.73 2.70 2.64 2.59 2.54
90% 3.29 2.92 2.73 2.61 2.52 2.46 2.41 2.38 2.35 2.32 2.24 2.20 2.17 2.16 2.12 2.09 2.06

99% 9.65 7.21 6.22 5.67 5.32 5.07 4.89 4.74 4.63 4.54 4.25 4.10 4.01 3.94 3.81 3.71 3.60
11 95% 4.84 3.98 3.59 3.36 3.20 3.09 3.01 2.95 2.90 2.85 2.72 2.65 2.60 2.57 2.51 2.46 2.40
90% 3.23 2.86 2.66 2.54 2.45 2.39 2.34 2.30 2.27 2.25 2.17 2.12 2.10 2.08 2.04 2.01 1.97

99% 9.33 6.93 5.95 5.41 5.06 4.82 4.64 4.50 4.39 4.30 4.01 3.86 3.76 3.70 3.57 3.47 3.36
12 95% 4.75 3.89 3.49 3.26 3.11 3.00 2.91 2.85 2.80 2.75 2.62 2.54 2.50 2.47 2.40 2.35 2.30
90% 3.18 2.81 2.61 2.48 2.39 2.33 2.28 2.24 2.21 2.19 2.10 2.06 2.03 2.01 1.97 1.94 1.90

99% 9.07 6.70 5.74 5.21 4.86 4.62 4.44 4.30 4.19 4.10 3.82 3.66 3.57 3.51 3.38 3.27 3.17
13 95% 4.67 3.81 3.41 3.18 3.03 2.92 2.83 2.77 2.71 2.67 2.53 2.46 2.41 2.38 2.31 2.26 2.21
90% 3.14 2.76 2.56 2.43 2.35 2.28 2.23 2.20 2.16 2.14 2.05 2.01 1.98 1.96 1.92 1.88 1.85

99% 8.86 6.51 5.56 5.04 4.69 4.46 4.28 4.14 4.03 3.94 3.66 3.51 3.41 3.35 3.22 3.11 3.00
14 95% 4.60 3.74 3.34 3.11 2.96 2.85 2.76 2.70 2.65 2.60 2.46 2.39 2.34 2.31 2.24 2.19 2.13
90% 3.10 2.73 2.52 2.39 2.31 2.24 2.19 2.15 2.12 2.10 2.01 1.96 1.93 1.91 1.87 1.83 1.80

99% 8.68 6.36 5.42 4.89 4.56 4.32 4.14 4.00 3.89 3.80 3.52 3.37 3.28 3.21 3.08 2.98 2.87
15 95% 4.54 3.68 3.29 3.06 2.90 2.79 2.71 2.64 2.59 2.54 2.40 2.33 2.28 2.25 2.18 2.12 2.07
90% 3.07 2.70 2.49 2.36 2.27 2.21 2.16 2.12 2.09 2.06 1.97 1.92 1.89 1.87 1.83 1.79 1.76

99% 8.53 6.23 5.29 4.77 4.44 4.20 4.03 3.89 3.78 3.69 3.41 3.26 3.16 3.10 2.97 2.86 2.75
16 95% 4.49 3.63 3.24 3.01 2.85 2.74 2.66 2.59 2.54 2.49 2.35 2.28 2.23 2.19 2.12 2.07 2.01
90% 3.05 2.67 2.46 2.33 2.24 2.18 2.13 2.09 2.06 2.03 1.94 1.89 1.86 1.84 1.79 1.76 1.72

Copyright CRC Press

Experimental Design for Combustion Equipment 411

TABLE 13.6 (continued) F-Distribution, 99%, 95%, and 90% Confidence

DF ν1
DF ν2 Conf 1 2 3 4 5 6 7 8 9 10 15 20 25 30 50 100 Infinity

99% 8.40 6.11 5.19 4.67 4.34 4.10 3.93 3.79 3.68 3.59 3.31 3.16 3.07 3.00 2.87 2.76 2.65
17 95% 4.45 3.59 3.20 2.96 2.81 2.70 2.61 2.55 2.49 2.45 2.31 2.23 2.18 2.15 2.08 2.02 1.96
90% 3.03 2.64 2.44 2.31 2.22 2.15 2.10 2.06 2.03 2.00 1.91 1.86 1.83 1.81 1.76 1.73 1.69

99% 8.29 6.01 5.09 4.58 4.25 4.01 3.84 3.71 3.60 3.51 3.23 3.08 2.98 2.92 2.78 2.68 2.57
18 95% 4.41 3.55 3.16 2.93 2.77 2.66 2.58 2.51 2.46 2.41 2.27 2.19 2.14 2.11 2.04 1.98 1.92
90% 3.01 2.62 2.42 2.29 2.20 2.13 2.08 2.04 2.00 1.98 1.89 1.84 1.80 1.78 1.74 1.70 1.66

99% 8.18 5.93 5.01 4.50 4.17 3.94 3.77 3.63 3.52 3.43 3.15 3.00 2.91 2.84 2.71 2.60 2.49
19 95% 4.38 3.52 3.13 2.90 2.74 2.63 2.54 2.48 2.42 2.38 2.23 2.16 2.11 2.07 2.00 1.94 1.88
90% 2.99 2.61 2.40 2.27 2.18 2.11 2.06 2.02 1.98 1.96 1.86 1.81 1.78 1.76 1.71 1.67 1.63

99% 8.10 5.85 4.94 4.43 4.10 3.87 3.70 3.56 3.46 3.37 3.09 2.94 2.84 2.78 2.64 2.54 2.42
20 95% 4.35 3.49 3.10 2.87 2.71 2.60 2.51 2.45 2.39 2.35 2.20 2.12 2.07 2.04 1.97 1.91 1.84
90% 2.97 2.59 2.38 2.25 2.16 2.09 2.04 2.00 1.96 1.94 1.84 1.79 1.76 1.74 1.69 1.65 1.61

99% 7.77 5.57 4.68 4.18 3.85 3.63 3.46 3.32 3.22 3.13 2.85 2.70 2.60 2.54 2.40 2.29 2.17

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
25 95% 4.24 3.39 2.99 2.76 2.60 2.49 2.40 2.34 2.28 2.24 2.09 2.01 1.96 1.92 1.84 1.78 1.71
90% 2.92 2.53 2.32 2.18 2.09 2.02 1.97 1.93 1.89 1.87 1.77 1.72 1.68 1.66 1.61 1.56 1.52

99% 7.56 5.39 4.51 4.02 3.70 3.47 3.30 3.17 3.07 2.98 2.70 2.55 2.45 2.39 2.25 2.13 2.01
30 95% 4.17 3.32 2.92 2.69 2.53 2.42 2.33 2.27 2.21 2.16 2.01 1.93 1.88 1.84 1.76 1.70 1.62
90% 2.88 2.49 2.28 2.14 2.05 1.98 1.93 1.88 1.85 1.82 1.72 1.67 1.63 1.61 1.55 1.51 1.46

99% 7.31 5.18 4.31 3.83 3.51 3.29 3.12 2.99 2.89 2.80 2.52 2.37 2.27 2.20 2.06 1.94 1.80
40 95% 4.08 3.23 2.84 2.61 2.45 2.34 2.25 2.18 2.12 2.08 1.92 1.84 1.78 1.74 1.66 1.59 1.51
90% 2.84 2.44 2.23 2.09 2.00 1.93 1.87 1.83 1.79 1.76 1.66 1.61 1.57 1.54 1.48 1.43 1.38

99% 7.17 5.06 4.20 3.72 3.41 3.19 3.02 2.89 2.78 2.70 2.42 2.27 2.17 2.10 1.95 1.82 1.68
50 95% 4.03 3.18 2.79 2.56 2.40 2.29 2.20 2.13 2.07 2.03 1.87 1.78 1.73 1.69 1.60 1.52 1.44
90% 2.81 2.41 2.20 2.06 1.97 1.90 1.84 1.80 1.76 1.73 1.63 1.57 1.53 1.50 1.44 1.39 1.33

99% 6.90 4.82 3.98 3.51 3.21 2.99 2.82 2.69 2.59 2.50 2.22 2.07 1.97 1.89 1.74 1.60 1.43
100 95% 3.94 3.09 2.70 2.46 2.31 2.19 2.10 2.03 1.97 1.93 1.77 1.68 1.62 1.57 1.48 1.39 1.28
90% 2.76 2.36 2.14 2.00 1.91 1.83 1.78 1.73 1.69 1.66 1.56 1.49 1.45 1.42 1.35 1.29 1.21

99% 6.63 4.61 3.78 3.32 3.02 2.80 2.64 2.51 2.41 2.32 2.04 1.88 1.77 1.70 1.52 1.36 1.00
Infinity 95% 3.84 3.00 2.60 2.37 2.21 2.10 2.01 1.94 1.88 1.83 1.67 1.57 1.51 1.46 1.35 1.24 1.00
90% 2.71 2.30 2.08 1.94 1.85 1.77 1.72 1.67 1.63 1.60 1.49 1.42 1.38 1.34 1.26 1.18 1.00

TABLE 13.7 ANOVA Table for Equation 13.2 Applied to TABLE 13.8 ANOVA for Table 13.7 with Separate Effects
Data of Table 13.4 Component SS DF MS F F95(1,9)
Component SS DF MS F F95(6,9) Model (M): a1 12.25 1 12.25 9.19 5.12
Model (M) 389.75 6 64.95 48.7 3.37 a2 306.25 1 306.25 229.69 5.12
Residual (R) 12.00 9 1.33 a3 49.00 1 49.00 36.75 5.12
a12 20.25 1 20.25 15.18 5.12
Total (T) 401.75 15 a13 1.00 1 1.00 0.75 5.12
a23 1.00 1 1.00 0.75 5.12

Residual (R) 12.00 9 1.33

model effects are insignificant, it might be best to move them
Total (T) 401.75 15
from the model to the residual and re-estimate σ with a
greater DFR.

true for these factors, then the MSR should not change signif-
13.2.4 Pooling Insignificant Effects icantly. In fact, as the table shows, s2 remains nearly
To pool insignificant effects, their SS and DF are moved to unchanged, decreasing from 1.33 to 1.27. This changes the
the residual, generating Table 13.9. If the null hypothesis is estimate of σ from 1.2 to 1.1 ppm. In the absence of replicate
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 412 Wednesday, April 6, 2005 4:39 PM

412 The John Zink Combustion Handbook

TABLE 13.9 ANOVA for Table 13.7 with Pooled Effects Later, Section 13.3.3 introduces fractional factorials as a way
Component SS DF MS F F95 (1,11) to reduce the required testing. As their name suggests, two-level
Model (M): a1 12.25 1 12.25 9.63 4.84
designs comprise factors at two levels: high (+) and low (–).
a2 306.25 1 306.25 240.63 4.84 Consider the design used tacitly in Table 13.2. Table 13.4
a3 49.00 1 49.00 38.50 4.84
gives the transformed matrix in a standard order comprising
a12 20.25 1 20.25 15.91 4.84
2 3 runs. A general factorial design in f factors will require
Residual (R) 14.00 11 1.27 2 f runs. The sign of the factors will alternate in blocks of 2a–1,
where a is the factor subscript. For example, the sign of the
Total (T) 401.75 15
first factor alternates every run (21–1); the sign of the second
factor alternates in blocks of two (22–1); the sign of the third
factor alternates in blocks of four (23–1); etc. This continues
observations, insignificant effects are the only available esti- for all f factors. After construction of the design, domain
mates for the error. knowledge is used to assign values to the high and low levels.

13.2.5 Standard Errors of Effects 13.3.1 Interactions

An estimate for σ provides the standard errors for each fac-
Factorial designs comprising 2 f unique points can specify
tor’s coefficient. The standard error of each effect (si) is esti-
models with 2 f coefficients, as in Eq. (13.28):
mated by the product of s and the respective diagonal element
of the inverse least-squares matrix: si = s(xTx)–1i,i . For orthogo-
nal data, the diagonal elements of the inverse matrix are sim-
ply the reciprocals of the diagonal elements of xTx. The reader
y = a0 + ∑a x + ∑∑a x x
i i
i< j j
ij i j

may verify that the final model becomes Eq. (13.26) with the
standard errors enclosed in brackets [ ]: + ∑∑∑a
i< j j <k k
x x j xk + L
ijk i (13.27)

y = 22.375[ ±0.28] + 0.875[±0.28]x1 + 4.375[±0.28]x 2 The summations continue until all 2 f terms are specified.
(13.26)
+1.750[±0.28]x3 − 1.125[±0.28]x1 x 2 The additive sequence reaches 2 f coefficients when it
reaches the f-factor interaction. For example, if f = 6, the
last term in the series for Eq. (13.27) will be a six-factor
The ratio of an effect to its standard error is known as the
interaction, a123456 x1x2x3x4x5x6. At that time, the series will
t-ratio. It distributes as the square root of the F(1,n)-distribu-
have grown to comprise (26) 64 terms. Third- or higher-order
tion. For example, for x1, the 95% confidence interval is
interactions are rarely a concern. If the researcher does not
x1 ± t95(11)s1, or 0.875 ± 0.62. This follows from t95(11)s1 =
require higher interactions he can reduce the number of
± F95 (1, 11)s1 = ±2.20(0.28) = ±0.62. Because 0.875 is greater required runs. Section 13.3.3 shows how.
than 0.62, the confidence interval does not include zero, and
one rejects the null hypothesis. The t- and F-ratios provide 13.3.2 Pure Error and Bias
equivalent tests with identical confidence.
One problem not yet discussed is that in addition to noise, the
residual may contain bias (also called lack of fit). Statisti-
cians sometimes use the term “pure error” to distinguish ran-
13.3 TWO-LEVEL FACTORIAL dom error from the entire residual, which may contain bias.
DESIGNS For example, omission of model terms that should have been
The test matrix given in Table 13.2 is a special one known as included will bias the residual. A biased residual inflates the
a two-level factorial design. This type of test matrix offers error estimate. Larger error estimates make the F-test less
special benefits. Two-level factorial designs are powerful sensitive, and increase the probability of falsely accepting the
methods that allow for very efficient experimentation. That null hypothesis. Replicating some of the runs quantifies pure
is, they maximize the information available for a given set of error. Subtracting pure error from the residual gives an esti-
factors. To construct them requires a minimum of 2 f runs mate of the bias. It is not necessary to replicate every design
(experimental tests), where f is the number of factors in the point as was done in Table 13.4. Adding replicate center-
investigation. For f > 5, these designs can involve many runs.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- points allows testing for both bias and pure error. Factorial
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 413 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 413

FLUE GAS

EXHAUST MUNICPAL
STACK SOLID WASTE

NO x
REDUCTION
ZONE
RECIRCULATED
AMMONIA
INJECTION FLUE GAS
(OPTIONAL)

FEED CHUTE

COMBUSTION
ZONE
STOKER
GRATE
UNDERGRATE
ASH COMBUSTION
AIR

Municipal solid waste (MSW) enters the feed chute onto a perforated stoker grate. Combustion air
enters under the grate. The burning the MSW exits the stoker as ash, but creates a hot combustion
zone that generates NOx. A series of nozzles inject ammonia, which chemically reduces NOx.
Varying the injector pressure changes the penetration of ammonia into the hot furnace and affects
NOx reduction. Optionally, flue gas may be added to the combustion air as an additional NOx
reduction strategy.

FIGURE 13.4 Municipal solid waste boiler using ammonia injection to control NOx.

designs remain orthogonal with any number of replicate cen- TABLE 13.10 Factorial Design with Replicate Centerpoints
terpoints. ξ1 ξ2 y
Run Pressure, psig NH3/NOx x1 x2 NOx, lb/h
Consider a municipal solid waste boiler2 that injects NH3
1 20 1.4 – – 31.0
to reduce NOx as in Figure 13.4. It is necessary to correlate 2 40 1.4 + – 25.0
the injection pressure (x 1) and the NH3/NOx ratio (x2) with 3 20 1.8 – + 16.4
4 40 1.8 + + 18.1
the NOx concentration (y): 5 30 1.6 0 0 23.0
6 30 1.6 0 0 25.9
7 30 1.6 0 0 25.2
y = a0 + a1 x1 + a2 x 2 + a12 x1 x 2 (13.28)

Table 13.10 gives the data, comprising m = 5 unique points the following two sums to separate the residual into bias and
with one centerpoint replicated three times. The “0” values pure error components.
represent the centerpoints.
• Calculate SSE as the squared difference between the
To test separately for bias in the ANOVA table, partition
actual response (yi) and the replicate means ( yi ). When
SSR into two parts: one for bias (SSB) and the other for pure there are no replicates, then yi = yi for the purposes of the
error (SSE). The degrees of freedom (DF) for each are DFE summation, adding 0 to SSE. For the case at hand, only
= n – m, and DFB = m – p, where m is the number of unique three centerpoints are replicated. They generate a single
points. Table 13.11 gives the generic ANOVA for replicated
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
replicate mean [(23.0 + 25.9 + 25.2)/3 = 24.7]. Therefore,
data, and Table 13.12 gives ANOVA for the case at hand. Use SSE = (23.0 – 24.7)2 + (25.9 – 24.7)2 + (25.2 – 24.7)2 = 4.58.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 414 Wednesday, April 6, 2005 4:39 PM

414 The John Zink Combustion Handbook

TABLE 13.11 Generic ANOVA for Factorial Design with Replicates

Component SS DF MS F, Case 1 F, Case 2

∑ (y − y)
) 2
Model (M) i
p –1 SSM/DFM MSM/MSE MSM/MSR
i =1

∑ ∑ (y − y )
) 2
Residual Bias (B) k
m–p SSB/DFB MSB/MSE
i =1 k

∑ ∑ (y − y )
2
Pure Error (E) i k
n–m SSE/DFE
i =1 k

∑ (y − y)
2
Total (T) i
n–1
i =1

Case 1: F = MSM/MSE if MSB/MSE ≥ F95(DFB, DFE)

Case 2: F = MSM/MSR if MSB/MSE < F95(DFB, DFE)
Note: for Case 2, MSR = (SSB + SSE)/(DFB + DFE) = SSR/DFR

TABLE 13.12 ANOVA for Table 13.10 and Equation entire residual, as in Table 13.13. The larger number of
Component SS DF MS F F95(1,2) degrees of freedom should better estimate σ2. Table 13.13
Model (M) a1 4.62 1 4.62 2.02 18.51
shows that besides the mean, only a2 is significant. In a like
a2 115.56 1 115.56 50.46 18.51 manner, partition the residual of Table 13.9 as MSE = 11.00/8
a12 14.82 1 14.82 6.47 18.51 = 1.375, and MSB = 3.00/3 = 1.000. An F-test proves the
model contains no significant bias (F = 1.000/1.375 = 0.72,
Residual (R) Bias (B) 7.38 1 7.38 3.22 18.51
F95(3,8) = 4.07, 0.72 < 4.07). This justifies the use of the
Pure error (E) 4.58 2 2.29 entire residual for the earlier F-tests.
Total (T) 146.97 6

Note: MSB/MSE < F95(1,2).

13.3.3 Two-Level Fractional Factorials
Fractional factorial designs always result in fewer runs than
TABLE 13.13 ANOVA for Factorial Design with full factorials. Specifically, fractional factorials comprise nFF
Centerpoint Replicates Case 2 of Table 13.11 runs where:
Component SS DF MS F F95(1,3)

Model (M) a1 4.62 1 4.62 1.16 10.13

a2 115.56 1 115.56 28.96 10.13
nFF = 2 f − d + c ≥ f + 1 + c (13.29)
a12 14.82 1 14.82 3.71 10.13

Residual (R) 11.96 3 3.99 The “fraction” is 2–d. For example, the quarter fraction com-
prises 2 f–2 runs (d = 2, 2–2 = 1/4). The nomenclature FF(f, –d, c)
Total (T) 146.97 6 is used to specify a full or fractional factorial comprising
2 f–d + c runs, where c is the number of centerpoints. Sometimes
it is necessary to study a large group of factors, without know-
• Calculate SSB by finding SSR in the usual way and ing in advance which are likely to be influential. Other times,
subtracting SSE. Alternatively, calculate SSB directly as constraints limit the total number of possible experiments. The
the squared difference between the predicted values ( ŷ i) maximum fractionation (dmax,FF) is determined by solving Eq.
and the replicate means ( yi ). Substitute the value yi = yi (13.30) for d, which must be an integer:
whenever a point comprises only a single value. However,
ŷ i – yi will not necessarily be zero in these cases.

Table 13.12 gives the MSB/MSE ratio (F = 3.22) and

dmax,FF = f − Ceil log 2 ( f + 1) [ ] (13.30)
shows that bias is not significant at the 95% confidence level.
The conclusion of no significant bias allows pooling of the where Ceil is the ceiling operator. It specifies the smallest integer
bias and pure error. Then one performs the F-test on the greater than or equal to [log2(f + 1)]. To calculate log2(x) observe
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Experimental Design for Combustion Equipment 415

that log2(x) = ln(x)/ln(2). For example, if f = 12, log2(12 + 1) = TABLE 13.14 1 2

/ Fractional Factorial [FF(3,–1,0)]
3.7, then Ceil[log2(12 + 1)] = 4, and f-Ceil[log2(f)] = 8; thus, Run x1 x2 x3 x4
the maximum fractionation is d = dmax,FF = 8. 1 – – – –
Consider a full factorial for f = 4. It requires 24 runs. 2 + – – +
Suppose, however, only eight runs are possible — half the 3 – + – +
4 + + – –
number needed. The 1/2 fraction given in Table 13.14 com-
5 – – + +
prises the required number of runs. 6 + – + –
How is it generated? First, generate the full factorial for f-d 7 – + + –
8 + + + +
factors (columns x1 to x3 in Table 13.14). Then, alias the fourth
factor with the highest order interaction (x1x2x3). The fourth factor
is aliased by multiplying the values of x1, x2, and x3 together and
assigning the result to x4. For the first row, this gives (–1)(–1)(–1) is Resolution IV (typically expressed in Roman numerals). In
= (–1). Continue for each of the eight rows to generate the full a Resolution IV design, the main effects (factors to the first
matrix shown. order) alias with three-factor interactions. In a Resolution III
Because x4 is just another name for x1x2x3, x1x2x3 is aliased design, the main effects alias with two-factor interactions. As
with x4. That is, the factor levels for the x1x2x3 interaction are long as the resolution is at least III, no main effect will be
identical to x4. Thus, the response to x1x2x3 is confounded with aliased with any other main effect. Resolution IV designs or
that of x4. So, the one effect is impossible to distinguish from higher have main effects and two-factor interactions clear of
the other. The price for reducing the number of runs is to give each other.
up the ability to separately assess the three-factor interaction A fractional factorial requires specification of d defining
from x4. With careful inspection, it is also possible to determine contrasts. For the quarter fraction, d = 2, and two defining
the following aliases. Simply reporting the subscript of the contrasts, I1 and I2, are needed. Suppose there are five factors
factor sufficiently identifies the alias structure. to study in eight runs. It can be done with a 1/4 fraction. The
contrasts must be chosen, and the choice has consequences.
0 ↔ 1234, 1 ↔ 234, 2 ↔ 134, 3 ↔ 124, 4 ↔ 123, If I1 and I2 are chosen, I3 = I1 × I2 is also tacitly picked. In
12 ↔ 34 13 ↔ 24 14 ↔ 23, general, choose defining contrasts to be as large, as equal,
and with as little overlap as possible. This suggests something
The symbol “↔” means is aliased with. Thus, if the 1234
like I1 = 123 and I2 = 345. Then I3 = I1 × I2 = 123245 = 1245.
interaction is significant it will be completely aliased with
Suppose I1 = 1234 and I2 = 2345 were chosen. They would
the overall mean.
generate I3 = 15, a Resolution II design. This is a very poor
Likewise, if the 234 interaction is significant, its effect will defining contrast, aliasing x1 with x5! With I1 = 123 and I2 =
be completely buried in the a1 coefficient. At most, it is only 345, then I3 = 1245. The design is Resolution III. Therefore,
possible to estimate seven effects plus the mean. This makes no main effect will be aliased with any other. Typically, stud-
sense because the design comprises only eight runs. ies require estimates for main effects and two-factor interac-
The reader may notice a pattern to this confounding. When tions. Three-factor and higher interactions are usually
multiplying the alias groups together, the word 1234 is always negligible. Reporting only main effects and two-factor inter-
obtained. This word is referred to as the defining contrast (I). actions gives the alias structure: 1↔23, 2↔13, 3↔12↔45,
The word length is the number of factors the word contains. 4↔35, 5↔34, 14↔25, and 15↔24 when I1 = 123 and I2 = 345.
In this case, the word length of 1234 is four.
The alias structure can be computed with a simple algo-
rithm: multiply each effect by I — a squared effect cancels. 13.3.4 Screening Designs
For example, 1 × I = 1 × 1234 = 12234 = 234. So, x1 is aliased Screening designs comprise a minimal number of runs to find
with x2x3x4. To figure out the alias structure for x1x2, use 12 × I statistically significant factors among a large body of possible
= 34. Thus, it is understood that x1x2 ↔ x3x4. Because the factors. One kind of screening design is the saturated facto-
defining contrast is 1234, generate the alias structure using rial. When f – d = log2(f + 1), the design becomes saturated.
the relation 123 = 4. Therefore, I = 123 × 4 = 1234; and Saturated designs have exactly f + 1 runs to specify f + 1
conversely, 123 × I = 4. coefficients. Such designs are also called simplex designs.
The shortest word length among the defining contrasts is Actually, one can construct a simplex design3 for any integer
known as the resolution. For the example at hand, the shortest value of f. Saturated designs, on the other hand, are a subset
(and only) defining contrast is I = 1234. Therefore, the design
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
of simplex designs restricted to f = 2J – 1, where J is an inte-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 416 Wednesday, April 6, 2005 4:39 PM

416 The John Zink Combustion Handbook

ger greater than 1 (e.g., f = 3, 7, 15, …). Nonetheless, it is A good rule for step size, in the absence of other criteria, is
recommended that for f ≠ 2J – 1, one should not use the sim- to move the new design center p units from the old one
plex, but the next larger fractional factorial, also known as a along the path of steepest ascent. For the current case, p = 2.
highly fractionated design. For highly fractionated designs, To scale the coefficient vector to 2 length, multiply by λ,
Eq. (13.31) gives the degree, dHF . defined by Eq. (13.35):

[
dHF = f − Ceil log 2 ( f + 1) ] (13.31)
λ=
p
(13.35)

Thus, for f = 12, the FF(12, –8,0) design comprises 16 runs.

∑a 2
i

For highly-fractionated designs, construct the ANOVA as

2
shown in Section 13.3.3. For saturated and other simplex This results in λ = = 0.31. Accordingly, move
designs, there are no degrees of freedom to calculate MSR. −0.9 2 + 4.4 2
the new design center from (0, 0) to 0.31·(–0.9, 4.4) ≈ (–0.3, 1.4).
However, with so many factors, it is likely that several will
Figure 13.5 shows the procedure graphically.
be insignificant. Therefore, to construct the residual, look for
factors with approximately the same low magnitude, presum-
ing that they represent the noise. 13.3.6 Serial Correlation and
Lurking Factors
13.3.5 Method of Steepest Ascent Experimental designs are generated in standard order. How-
Sometimes, the initial screening finds several factors, but the ever, running the designs in standard order can introduce
response does not meet expectations. It may be that the range error caused by serial correlation and lurking variables.
for the initial factors was too conservative. The factor space is Lurking variables are influential but unknown factors. Serial
the region delimited by the joint range of the factors. In this correlation refers to responses that correlate with their run
case, it is necessary to expand the range of factors, or at least order. The subject system may contain hysteresis — a previ-
move to a new location in factor space. The method of steepest ous condition that influences the response. Alternatively, it
ascent is an efficient way to do this. After screening the impor- may be that influential factors correlate with time; for exam-
tant variables, the result is a first-order model: ple, ambient temperature or humidity. Running experiments
in random order virtually eliminates the effects of serial cor-
relation. With randomization, lurking factors may inflate the
y = a0 + ∑a x i i (13.32) background noise, but will not bias the results. Foldover and
blocking techniques, described in the next two sections, are
To find the steepest path, take the first derivatives of Eq. used to reduce the effect of background noise.
(13.32). This generates the coefficient vector of Eq. (13.33).
13.3.7 Foldover
 a1  Sometimes, a complementary experimental series is run after
∂y  a2 
learning something from the initial one. Perhaps it is later
=  (13.33) suspected that a particular interaction is significant, but the
∂x i  M 
a  current design aliases it to another important effect. Conduct-
 n ing a second complement of runs with the signs reversed
from the first comprises a procedure called foldover. This
So then, the path of steepest ascent is along the coeffi- increases the resolution of the design and de-aliases the inter-
cient vector of Eq. (13.33). Consider Eq. (13.34) as an actions. However, because the two groups of designs are run
example: at different times, the researcher must consider the possibility
that some important changes may have occurred. For exam-
y = 5.1 − 0.9 x1 + 4.4 x 2 (13.34) ple, combustion tests with two different batches of fuel oil, or
tests run during different seasons of the year may contain
The direction of steepest ascent is –0.9 units in the x1 systematic differences. To provide for differences between
direction for every 4.4 units in the x2 direction. To minimize one block of experiments and another, use the technique of
the response, step in the opposite direction, for example, +0.9 orthogonal blocking. It generates an experimental series
units in the x1 direction for every –4.4 units in the x2 direction. whose effects are orthogonal to the factors. Therefore, batch
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Experimental Design for Combustion Equipment 417

response-surface
contours
p
all direction of
a2l steepest
ascent

old design
new design
The method of steepest ascent uses the derived coefficients (a1, a2) to find the direction of greatest
increase in the response. l scales the direction vector to move the design p units along this path.
One may reapply the method as many times as desired. However, near the optimum, the first order
design is no longer sufficient and one must augment the design to determine second-order behavior.

FIGURE 13.5 Method of steepest ascent.

effects are prevented from confounding the results. Foldover Batch I, and the higher steam:oil ratios on Heavy Oil
automatically creates an orthogonal block. The next section Batch II, will completely confound the batch and steam:oil
shows how. effects. The preferred test series would ensure that the batch
effects are orthogonal to the steam:oil, oxygen, and firing-
13.3.8 Orthogonal Blocking rate effects. How is this accomplished?
Consider a test conducted with a heavy oil burner using First, construct the experimental design in the usual way
steam to atomize the oil as shown in Figure 13.6. The pur- to evaluate two levels (low and high) of steam:oil ratio (x1),
pose of the test is to correlate flame length with oxygen con- oxygen (x2), and firing rate (x3) (see Table 13.15). Because
centration, steam:oil ratio, and firing rate. If it is feasible to the fuel oil batch may have some effect, study it as an addi-
run 12 tests, simply run a two-level three-factor factorial tional factor. The study of the oil is not interesting in itself.
design of eight runs and add four centerpoint replicates to The oil is more of a nuisance factor, changing unpredictably
give measures of bias and pure error. However, what if the from batch to batch, but the block factor accounts for the
available oil tank will only accommodate six or seven tests? systematic difference between batches. The burner and its
One alternative is to run some tests with the current batch performance at different oxygen levels, steam:oil ratios, and
of oil, then run more tests when a new batch is available. firing rates is the real point of interest. Factors of real interest
Analytical testing could even ensure that the new batch is to the study are called fixed effects. Factors that represent an
similar to the current one. However, it is not possible to know uncontrolled variation are called random effects.
for sure that the new batch will not change in some important Table 13.15 gives the experimental structure. All runs within
but unknown way. Heavy oil is notorious for batch-to-batch Block = I (1, 4, 6, 7, 9, and 10) are randomized and performed.
variation, and subtle changes in the batch could affect the The remaining runs, Block = II (2, 3, 5, 8, 11, and 12), are
response. Running the low steam:oil ratios on Heavy Oil
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
randomized and run for the new batch of oil. To derive the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 418 Wednesday, April 6, 2005 4:39 PM

418 The John Zink Combustion Handbook

REGEN TILE
OIL GUN
GAS RISERS (FOR
COMBINATION FIRING) TERTIARY AIR CONTROL
PRIMARY TILE

GAS PILOT AIR INLET

PLENUM

PRIMARY
AIR CONTROL

SECONDARY AIR CONTROL

GAS RISER MANIFOLD

(FOR COMBINATION FIRING)

FIGURE 13.6 John Zink PLNC combination burner capable of firing either oil or gas, or both simultaneously. Oil flows
through a central oil gun and is stabilized by a regen tile. A surrounding primary tile stabilizes the gas fire. All air flows through
the inlet plenum. However, this version of the PLNC has air control to three zones — primary, secondary, and tertiary — via
respective dampers whose handles are shown. A gas pilot is used for startup but not required for operation.

TABLE 13.15 FF(3,0,4) in Two Blocks the residual variance, so our F-tests will be more sensitive.
Run x1 x2 x3 Block Hunter et al.4 summarize the general philosophy: “Block what
1 – – – I
you can. Randomize what you cannot.” If needed, construct
2 + – – II more orthogonal blocks using additional blocking generators
3 – + – II while taking care to account for their mutual interactions.
4 + + – I
5 – – + II
6 + – + I 13.3.9 Including Categorical Factors
7 – + + I
8 + + + II
The previous discussion presumed continuously distributed
9 0 0 0 I factors. For example, in principle, oxygen can take on any
10 0 0 0 I value over its range. This contrasts with discrete factors that
11 0 0 0 II
can only take specific values. Even some discrete factors can
12 0 0 0 II
be treated as continuous. For example, factors like pipe size
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

or burner throat diameter may only exist in certain discrete

sizes. Nonetheless, for experimental design purposes, they
matrix, put the first three factors in standard order and add the are treated as continuous because intermediate values are the-
centerpoint replicates at the end — half of which are assigned oretically conceivable if not practical. However, some factors
to Block I and the other half to Block II. Generate the blocking such as burner type are clearly discrete and cannot be consid-
variable with the blocking generator B = 123, using the algo- ered continuous.
rithm for defining contrasts in Section 13.3.3. Assign x1x2x3 = –1 Consider again two different burners. The task is to char-
to Block I, and x1x2x3 = +1 to Block II. Now, the block effect acterize flame length (y) as a function of the burner type (ξ I),
is aliased with the x1x2x3 interaction but with no main effects. oil pressure (ξ2), and the differential steam pressure (ξ3). To
Next construct the ANOVA with entries for a1, a2, a3, a12, a13, signify categorical factors, use uppercase Roman numeral sub-
a23, Block, Bias, and Pure Error. The blocking variable reduces scripts to signify a categorical value. A categorical factor, ξ I ,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 419 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 419

distinguishes one category or type of burner from another. TABLE 13.16 Experimental Design with Categorical Factors
Transform ξ I to x I by arbitrarily assigning the values x I = –1 y = Flame
for Burner A and x I = +1 for Burner B. Use capital letters to ξI = Burner ξ2 = Oil ξ3 = Differential Length
distinguish among levels of ξ I. The lack of a continuous scale Run Type Pressure (psig) Pressure (psid) (ft)

for ξ I excludes many of the experimental designs considered 1 A 60 20 8.7

2 A 80 20 10.6
thus far. However, there are some options:
3 A 60 30 8.0
1. Substitute a continuous factor for the categorical one. This 4 A 80 30 9.0
is possible if the burners differ by some influential char- 5 B 60 20 13.3
6 B 80 20 16.3
acteristic, measure, or scale such as throat diameter or
7 B 60 30 10.1
length:diameter ratio. Instead of the categorical factor, use 8 B 80 30 13.5
the continuous characteristic dimension and apply any of
the previous experimental designs.
2. Design the matrix as a full- or fractional-factorial design TABLE 13.17 ANOVA for Table 13.16
in f-k factors, where k is the number of categorical factors. Component SS DF MS F F95(1,4)
Randomize the k 2 f–k–d runs, and conduct the tests. Model (M) xI 35.70 1 35.70 40.98 7.71
x2 10.81 1 10.81 12.41 7.71
Table 13.16 contains the data for a FF(3,0,0) design with x3 8.61 1 8.61 9.88 7.71
a categorical factor. It generates the following equation and
the ANOVA of Table 13.17: Residual (R) 3.49 4 0.87

9.08[±0.33]
Total (T) 58.61 7
if x I ≡ Burner A
y=
14.30[±0.33] if x I ≡ Burner B

+1.16[±0.33]x 2 − 1.04[±0.33]x3 (13.36) 13.4.1 Central-Composite Designs

Table 13.17 shows that burner type is significant, along with Preferably, the second-order design should build upon the two-
steam:oil ratio and oxygen, and Eq. (13.36) quantifies the level designs studied thus far. If the two-level design detects
effects. The analysis can be extended to more than two cate- significant bias, additional runs can be added to create a full
gorical factors. second-order model. Central-composite designs are denoted
with a nomenclature similar to the factorials — CC(f, –d, c).
Central-composite designs require nCC runs as shown in
13.4 SECOND-ORDER DESIGNS Eq. (13.39):
The discussion so far includes first order designs and first-
order designs augmented with second- and higher-order
nCC = 2 f − d + 2 f + c ≥
( f + 1)( f + 2) + c (13.39)
interactions. However, these designs cannot account for cur- 2
vature — pure quadratic factor effects. To account for curva-
(central-composite)
ture, a full second-order model of the form given in Eq. 13.37
is needed: Equation (13.40) gives the maximum fractionation, dmax,CC:

y = a0 + ∑a x + ∑∑a x x + ∑a x
i
i i
i< j j
ij i j
i
2
ii i (13.37)
[
dmax,CC = Ceil f + 1 − log 2 f 2 − f + 2 ( )] (13.40)

(central-composite)
Unfortunately, none of the designs studied thus far can
determine the pure quadratic coefficients, aii , because none A central-composite design uses a two-level design (frac-
of the designs have any factor at three or more levels. A design tional or full) and augments the design with 2f axial runs.
of Resolution V or greater comprising at least three levels per For example, if f = 3, at least ten runs are required for a full
factor is needed. A full second-order model requires nSO runs, second-order model according to Eq. (13.38). Equation (13.40)
as in Eq. (13.38): gives dmax,CC = 1, and Eq. (13.39) shows that a CC(3,–1,0) is
the minimum design needed.
nSO =
( f + 1)( f + 2) (full second-order model) (13.38)
Table 13.18 represents a CC(3,0,6) design for the same
2
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- purpose, comprising 20 runs. Except for the cost of the runs,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 420 Wednesday, April 6, 2005 4:39 PM

420 The John Zink Combustion Handbook

TABLE 13.18 CC(3,0,6) Design 3. Choose α to make the pure quadratics mutually ortho-
Run x1 x2 x3 gonal. This requires a numerical evaluation. (If the exper-
iment is run in a single randomized block, α = 1.525; if
1 – – –
a sequential strategy is used in three randomized blocks,
2 + – –
3 – + – then α = 1.633.)
4 + + – 4. Choose α so that the values are convenient, for example,
5 – – + ±2.
6 + – +
7 – + + Because f, d, and c are restricted to integer values, it is
8 + + + often not possible to satisfy all criteria simultaneously.
9 –α 0 0
10 +α 0 0
11 0 –α 0 13.4.2 Practical Considerations
12 0 +α 0
The general recommendation is to choose values of α that are
13 0 0 –α
14 0 0 +α orthogonal and approximately rotatable. In practical situa-
15 0 0 0 tions, even this can be difficult. Table 13.19 revisits the

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
16 0 0 0 stoker-fired boiler using ammonia injection to reduce NOx
17 0 0 0
18 0 0 0 (Figure 13.4). The mass flow of NOx(y) is a function of the
19 0 0 0 nozzle pressure (x1), the NH3:NOx ratio (x2), and fraction of
20 0 0 0 flue-gas recirculated, FGR (x3). One should choose a single
value for α; say, α = 4 8 for rotatability, or perhaps α = 2 for
TABLE 13.19 CC(3,0,3) Design convenience. However, FGR = on and FGR = off were the
Run x1 x2 x3 y most important states. A central composite gives only one test
1 – – – 31.0 at each of these extreme points. To account for this, the investi-
2 + – – 25.1 gator chose α1 = α2 = 2 and α3 = 1. Consequently, the design
3 – + – 16.4
is not rotatable and without mutually orthogonal quadratic
4 + + – 18.1
5 – – + 31.1 coefficients. The pure quadratics are not orthogonal to one
6 + – + 29.6 another or the mean, and they cannot be considered sepa-
7 – + + 18.0 rately. However, they are orthogonal to the rest of the model.
8 + + + 17.1
9 –2 0 0 28.2 Therefore, it is still possible to use a single ANOVA for the
10 +2 0 0 25.9 model of Eq. (13.37), providing the quadratics are considered
11 0 –2 0 33.1 as a group. From the ANOVA of Table 13.20, the only signif-
12 0 +2 0 11.6
13 0 0 – 25.9
icant term besides a0 is x2. The omitted quadratics are not
14 0 0 + 24.9 orthogonal to a0, therefore, a0 must be recalculated for the
15 0 0 0 23.0 truncated model. The reader may verify that the final model
16 0 0 0 25.9
becomes NOx (ppm) = 24.13 – 5.61[±0.43] x2. The term x2
17 0 0 0 25.2
ranges from –2 to +2. Therefore, in the extreme, NH3 injec-
tion reduces NOx from 35 ppm to 13 ppm.

adding them loses nothing. However, extra degrees of free-

dom to estimate the residual, pure error, and bias. 13.5 ACCOUNTING FOR
A central composite design augments the ±1 structure of FUEL MIXTURES
the factorials with axial points at ±α. So long as α ≠ 1, each Until now, independent or unconstrained factors have been
factor can be tested at five levels with only 2f additional runs. considered. However, the performance of combustion equip-
The choice of α affects the error structure. Consider four ment depends not only on the burner and the furnace, but also
criteria. on the fuel composition. A mixture is a collection of q species
whose properties depend only on their relative proportions.
1. Choose α so that ±α lies on an n-dimensional sphere
(α = ± f ) (e.g., 1.732 for f = 3). zi = q
ξi
(13.41)
2. Choose α so that the error structure is spherical (called
rotatability, α = 4 2 f − d ); for example, 1.682 for CC(3,0,6).
∑i =1
ξi

Copyright CRC Press

Experimental Design for Combustion Equipment 421

TABLE 13.20 ANOVA for Table 13.19 and Equation (13.38)

Component Coef Std err SS DF MS F F95

a0 25.16 (24.13 without pure quadratics) Std err = 0.93

Model (M) x1 –0.71 0.43 7.98 1 7.98 2.72 5.99

SSM = 550.52
x2 –5.61 0.43 502.88 1 502.88 171.62 5.99
x3 0.43 0.54 1.85 1 1.85 0.63 5.99
x1 x 2 1.04 0.61 8.61 1 8.61 2.94 5.99
x1 x 3 0.24 0.61 0.45 1 0.45 0.15 5.99
x2 x 3 –0.51 0.61 2.10 1 2.10 0.72 5.99
x11 0.385 0.36 26.65 3 10.76 3.67 4.35
x22 –0.765 0.36
x33 –1.149 0.87

Residual (R) Bias (Lack of Fit) 15.93 5 3.19 1.39 19.30

SSR = 20.51
Pure Error 4.58 2 2.29
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Upon noting no significant bias, the total residual estimates the error:
(15.93 + 4.58)/(5 + 2) = 2.93. This is the denominator in all the F-tests
except for bias. Bias uses the pure error as the denominator in the F-ratio.
Significance at 95% confidence: no effect exceeds F95 but x2.

Total (T) 571.03 16

In Eq. (13.41), zi is the fraction of the i th component, and ξi is 1. a simulated refinery gas comprising 25% H2, 25% C3H8,
the i th component in the original metric (e.g., %, SCFH, etc). and 50% Tulsa natural gas
The fractions must sum to unity, by definition: 2. a simulated gas low-BTU waste (LBG) comprising 25%
H2, 25% C3H8, 50% CO2
q 3. natural gas comprising mostly CH4 with some higher
∑i =1
zi ≡ 1 (13.42) hydrocarbons

As defined, the system is a three-component system. The

Of the q dependent factors, the constraining relation Eq. relation is illustrated graphically in Figure 13.7. The relation
(13.42) reduces the degrees of freedom by one to give f of Eq. (13.42) removes a degree of freedom and constrains
independent factors: q = 3 components to f = 2 independent factors. A response
will be mapped to the f = 2 system by testing with a sprin-
kling of points (p1, p2 …).
f = q −1 (13.43)
The factor space representing q components is known as
the simplex. Simplex designs are experimental designs with
The fuel components may be pure (e.g., hydrogen) or blends
at least q data points at the extreme vertices of the region.5
(e.g., natural gas, heavy fuel oil, refinery fuel gas, or others).
See p1, p2 , p3 in Figure 13.7. If the centerpoint, p123, is included,
the design becomes a simplex-centroid. If the edge points or
13.5.1 Experimental Designs for Mixtures binary blends, p12, p13, p23, are added, the simplex-centroid
Mixture designs are constructed from SED principles but can design increases in order. If more components are included,
differ in important ways from factorial designs. Simplex then additional points for higher-order interactions such as
designs occupy a central role for mixtures, and third- and ternary, quaternary, quinary, etc. can be added. The design
higher-order interactions are often quite important. Suppose becomes a simplex-axial design if one adds axial points, such
it is desirable to characterize a burner for the three fuel as p1123, p1223, p1233. In general, response functions are mapped
blends below: as a subset of Eq. (13.44):
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 422 Wednesday, April 6, 2005 4:39 PM

422 The John Zink Combustion Handbook

A. FACTOR DISPOSITION B. REGULAR SIMPLEX, q=3

z3
N3
q=3 factor space
p3
p1233
p3 p13 p
p123 23
p1123 p1223

p1 p12 p2
p1233 p23 z1 z2
p13 q=3 projected axes on
p123 f=2 factor space
f=2 surface
p1223
N2
p1133 p2
p12
p1
5zi = 1 constraining plane

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
N1

D. COORDINATES FOR q=3, f=2 C. RIGHT-ANGLE SIMPLEX, q=3

x1
pt z1 z2 z3 x1 x2
p1
p1 1 0 0 2 -1
p2 0 1 0 -1 2 p1123
p3 0 0 1 -1 -1 p13 p12
p123 1/3 1/3 1/3 0 0 p123
p1223 x2
p12 1/2 1/2 0 1/2 1/2 p1233
p13 1/2 0 1/2 1/2 -1
p3 p23 p2
p23 0 1/2 1/2 -1 1/2
p1123 2/3 1/6 1/6 1 -1/2 f=2 orthogonal axes on
p1223 1/6 2/3 1/6 -1/2 1 f=2 surface
p1233 1/6 1/6 2/3 -1/2-1/2

1. A three-component system is constrained by the mixture relation which requires all

components to sum to unity. The intersection of the constraining plane and three-
component space produces an equilateral triangular region (the simplex).
2. The simplex represents all possible mixture fractions: z1 + z2 + z3 =1.
3. We may also transform the region into a right angle triangle by considering q-1
independent factors. The linear transform xi=3zi-1 gives convenient values. Any two
factors may form this region; we have arbitrarily chosen x1 and x2.
4. A table gives the coordinates for the points (solid diamonds) indicated on the diagrams for
comparison.

FIGURE 13.7 Simplex design for q = 3.

q q −1 q q − 2 q −1 q mixture variables. The same subscript notation adopted earlier

y= ∑ i =1
bi zi + ∑∑
i< j j =1
bij zi z j + ∑∑∑
i< j <k j <k k =1
bijk zi z j zk + L (13.44) for factorials is used. Because Eq. (13.44) has q – 1 compo-
nents, the constant coefficient (b0) is eliminated to generate
a solvable matrix.
For clarity, b and z, as opposed to a and x, are used to In general, a q-component design will have n interactions
distinguish the coefficient and component, respectively, for of order k according to Eq. (13.45).
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 423 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 423

 q q! TABLE 13.21 Example of an Orthogonal Subspace for a

n=  = (13.45)
 k  k! (q − k )! q = 3 Simplex
ρ1 = x1 = x2 =
Run z1 z2 z3 z1/(z1+z2) ρ2 = z3 6ρ1 – 3 8ρ2 – 3
Equation (13.46) gives the total number of all mixture terms:
1 0 1 0 0 0 –3 –3
2 1/3 2/3 0 1/3 0 –1 –3
q
 q
∑  k = 2
3 2/3 1/3 0 2/3 0 1 –3
q
−1 (13.46) 4 1 0 0 1 0 3 –3
k =1 5 0 3/4 1/4 0 1/4 –3 –1
6 1/4 1/2 1/4 1/3 1/4 –1 –1
7 1/2 1/4 1/4 2/3 1/4 1 –1
The simplex-axial design of Figure 13.7 generates the spe-
8 3/4 0 1/4 1 1/4 3 –1
cial quartic Eq. 13.47. 9 0 1/2 1/2 0 1/2 –3 1
10 1/6 1/3 1/2 1/3 1/2 –1 1
y = b1 z1 + b2 z2 + b3 z3 + b12 z1 z2 + b13 z1 z3 + b23 z2 z3 11 1/3 1/6 1/2 2/3 1/2 1 1
12 1/2 0 1/2 1 1/2 3 1
(13.47)
13 0 1/4 3/4 0 3/4 –3 3
+ b123 z1 z2 z3 + b1123 z12 z2 z3 + b1223 z1 z22 z3 + b1233 z1 z2 z32 14 1/12 1/6 3/4 1/3 3/4 –1 3
15 1/6 1/12 3/4 2/3 3/4 1 3
This equation represents the response as mapped by the regular 16 1/4 0 3/4 1 3/4 3 3
simplex of Figure 13.7B. If preferred, the equation can be recast Note: See Figure 13.8 for a graphical representation.
in terms of two of the three independent factors, by choosing z1
and z2, and obtaining z3 by the relation z3 =1 – z1 – z2. The
region can also be transformed into a quasi-orthogonal one interest rather than the mixture fraction. Moreover, with judi-
using any convenient linear transform, such as xi = 3zi – 1. cious selection of the ratios one can engulf virtually all of the
With these transforms, Eq. (13.48) depicted in Figure 13.7C simplex. The next section gives an example.
is obtained.
13.5.3 Combining Mixture and
y = a0 + a1 x1 + a2 x 2 + a11 x12 + a12 x1 x 2 + a22 x 22 Factorial Designs
(13.48)
Orthogonal simplex designs may combine mixture and facto-
+ a112 x12 x 2 + a122 x1 x 22 + b1112 x13 x 2 + a1222 x1 x 23
rial points in several ways to form a combined factorial. For
example:
In general, it is not possible to have an orthogonal matrix
over the entire simplex for designs higher than first-order, 1. replicate the full simplex at each factorial point
even with linear transforms. Unlike factorial designs, mix- 2. replicate the full simplex at each fractional factorial point
tures very often involve ternary interactions or higher. How- 3. combine an orthogonal simplex and factorial design and

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
ever, for a region of interest within the simplex, one can have fractionate the entire design as in Section 13.3.3
orthogonal data sets.
To illustrate the last option, the first three columns of Table
13.21 specify a grid of points for q = 3. Figure 13.8A depicts
13.5.2 Orthogonal Mixture Designs the design. The next two columns of Table 13.21 give a ratio
There are two convenient ways to generate orthogonal factor transformation that can be orthogonalized to give x1 and x2.
space. If the region of interest is small, 0 < zi < 1/(q – 1), then The orthogonal coordinates are given in the last two columns
use it directly with the transform xi = 2 f zi – 1 to yield an of Table 13.21 and depicted in Figure 13.8B.
orthogonal factor space in f = q – 1 factors. There are q such Now suppose that in addition to x1 and x2, there are three
overlapping spaces in any simplex. However, any one region non-mixture factors — x3, x4, and x5. The system in Figure
of interest comprises the 2(q – 1)1–q fraction of the factor 13.8C can be depicted as a mixture design replicated at each
space. For example, if q = 3, then 0 < zi < 1/2, xi = 4zi – 1, factorial point. The first option would be to run the complete
comprising 1/2 the factor space. design comprising 16 × 23 = 128 experiments. A second
Another method uses selected ratios. The disadvantage of option would be to use a 1/2-fractional factorial with the full
the method is that ratio transforms are nonlinear so they mixture design comprising 16 × (1/2)23 = 64 points. If even
distort the factor space, especially near a zero denominator. this is too many, the third option is to combine the design and
However, in many cases, the component ratio is the factor of use the quarter fraction. This results in (1/4)·42·23 = 32 tests
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 424 Wednesday, April 6, 2005 4:39 PM

424 The John Zink Combustion Handbook

x1 TABLE 13.22 FF(7,–2,0) Design Generating a Combined

A B
Mixture-Factorial in Five Factors — Two at Four Levels and
z2
Three at Two Levels
x4 = x5 =
x2 Run x1a x1b x2a x2b x1 x2 x3 x1ax1bx3 x2ax2bx3

1 – – – – –3 –3 – – –
2 + – – – –1 –3 – + –
z1 z3 3 – + – – 1 –3 – + –
4 + + – – 3 –3 – – –
C 5 – – + – –3 –1 – – +
x1
6 + – + – –1 –1 – + +
x5
7 – + + – 1 –1 – + +
8 + + + – 3 –1 – – +
x2
9 – – – + –3 1 – – +
10 + – – + –1 1 – + +
x4
11 – + – + 1 1 – + +
12 + + – + 3 1 – – +
x3 13 – – + + –3 3 – – –
14 + – + + –1 3 – + –
15 – + + + 1 3 – + –
16 + + + + 3 3 – – –
17 – – – – –3 –3 + + +
The figure shows a ternary mixture design (A) transformed into an orthogonal one
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

(B). The orthogonal mixture design combines with a factorial design in three factors 18 + – – – –1 –3 + – +
and fractionates to give the combined fractional mixture-factorial design (C). 19 – + – – 1 –3 + – +
20 + + – – 3 –3 + + +
FIGURE 13.8 Mixture factors, a transformation, and a 21 – – + – –3 –1 + + –
combined mixture-factorial. 22 + – + – –1 –1 + – –
23 – + + – 1 –1 + – –
24 + + + – 3 –1 + + –
25 – – – + –3 1 + + –
— a more manageable number. However, the reader will note 26 + – – + –1 1 + – –
27 – + – + 1 1 + – –
that the mixture design is a four-level factorial. Thus far, only
28 + + – + 3 1 + + –
two-level factorials have been studied. The next section 29 – – + + –3 3 + + +
describes how to use the principles of two-level factorials to 30 + – + + –1 3 + – +
generate any 2n-level factorial. 31 – + + + 1 3 + – +
32 + + + + 3 3 + + +

13.5.4 Building 2n-Level Factorials From

Two-level Factorials To add an additional factor, x3, double the number of runs
It is possible construct 2n-level factorials from two-level facto- to 32. The generators I1 = 1a2a34 and I2 = 1b2b35 specify the
rials having n factors. To illustrate, consider the four-level final two columns. Overall, the FF(7,–2,0) design generates a
mixture design of Section 13.5.3. First, split the design into combined mixture-factorial design in five factors. Two of the
four factors labeled x1a , x1b , x2a , and x2b. Then build a 24 design factors (x1, x2) have four levels and three of the factors (x3, x4,
in the usual way, comprising the first 16 runs of Table 13.22. x5) have two levels. Analyze the design in the usual way for
The four possible combinations of x1a and x1b generate the a 27–2 design. The colored points in Figure 13.8C represent the
four values for x1. The four combinations are – –, – +, – –, + +. FF(7,–2,0) design given in Table 13.22.
Use Eq. (13.49) to assign these combinations to the values –3,
–1, 1, and 3 for x1.
13.6 COMBINING DOMAIN
n

xi = ∑2
j = a,b ,Kn
n −i
xia (13.49) KNOWLEDGE WITH SED
Although SED is a powerful tool, there is no substitute for
intelligence. It is important to think carefully about the prob-
In the present case, this equates to x1 = 2x1a + x1b. This keeps lem and the desired outcomes. Combining domain knowl-
the matrix orthogonal and preserves the two-unit difference edge with SED is a very powerful way to tackle experimental
between adjacent levels. problems. Even in complex or poorly understood cases,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 425 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 425

domain knowledge often supplies the form of the model, knowledge. However, it would be misleading to imply that
leaving SED to fit the adjustable parameters. The resulting the form of the model must be exactly correct. All models are
semi-empirical model is often more parsimonious and has wrong, but some are useful. Power transformations, to which
better properties than a purely empirical one. the reciprocal and log functions belong,8 have little effect
As an example, consider NOx formation having the fol- unless the ratio of maximum to minimum values is at least 3
lowing rate law: or greater. Most burners do not have this extreme NOx
response, although some do. Centerpoint replicates and lack-
t b
of-fit estimates usually give good guidance as to the appropri-
[NO x ] = A e ∫ [N ] [O ]dθ
−
T (13.50) ateness of the model. Nonetheless, one should spend quality
2 2
0 time thinking about the factors, and particularly about factor
relationships. Formulation of theoretical models will reduce
where A and b are constants, T is absolute temperature, θ is experimental time, keep the investigator focused on the prob-
time, and the brackets denote volume concentrations of the lem, and partner with SED to generate parsimonious semi-
enclosed species. The rate law shows that NOx is a function of empirical models. As an aid to the reader, some semi-empiri-
temperature, oxygen, and time. Unfortunately, it is not possi- cal forms relating to NOx formation and reduction strategies9
ble to integrate this function over the tortured path of an are itemized in the next section.
industrial burner flame. However, presuming that combustion
takes place at near stoichiometric conditions6 and using a mix- 13.6.2 Semi-empirical Models
ture fraction approach results in the following model:7
Equation (13.53) represents a family of semi-empirical mod-
els that well-represent NOx for a variety of combustion and
y = a0 + a1 x1 + a2 x 2 (13.51)
NOx abatement scenarios:7

where y = Natural log of the NOx fraction, ln[ppm]

a0–a2 = Regressed constants y = a0 + a1 x1 + a2 x 2
x1 = Turndown of the burner
 
x2 = Oxygen concentration in the flue gas  + a3 x3 + a4 x 4 +

∑i ≥5
a1i x1 xi + ∑j ≠i
a2 j x 2 x j 


(13.53)

The turndown is the reciprocal of the heat release of the burner

normalized by the full firing rate. Turndown usually varies
where y is the log of the NOx concentration, ln[NOx]
from 1 to no more than 10 for boilers, and no more than 3 or 4
for process burners. a0, a1, a2, a3, a4, a1i, and a2j are the constant coefficients
The techniques of this chapter are used to fit the coefficients x1 is the unit turndown
of Eq. (13.51). Because the reciprocal and log functions x2 is the oxygen concentration in the flue gas
require an infinite polynomial series, the same model cast in
x3 is the log of the windbox oxygen concentration
terms of heat release and NOx concentration would look like
due to FGR
Eq. (13.52).

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
x4 is the NH3:NOx ratio for ammonia injection (with
or without a catalyst)
β1ψ + β 2 ψ 2 + β 3 ψ 3 + K
(13.52) x1i is any of the following: the degree of water
= α 0 + α1ξ1 + α 2 x 2 + α11ξ12 + K injection, steam injection, other diluent injection
x2j is any of the following: the weight fraction of
Obviously, Eq. (13.52) would require many more coefficients nitrogen in the fuel, degree of air staging, degree
(α0 – α11 …) to adequately represent NOx (ψ) as a function of of fuel staging, or fraction of burners out of
heat release (ξ1). Clearly, Eq. (13.51) is a more parsimonious service
model than Eq. (13.52).
To construct the model, include the first three terms of Eq.
13.6.1 Practical Considerations (13.53) and any other pertinent effects. For example, NOx
The contrast between Eq. (13.51) and Eq. (13.52) under- from combustion of high-nitrogen fuel oil in a burner using
scores the value of applying SED in concert with domain air staging and steam injection generates Eq. 13.54:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 426 Wednesday, April 6, 2005 4:39 PM

426 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 13.9 Flowchart showing a general sequential experimental strategy.

y = a0 + a1 x1 + a2 x 2 + a15 x1 x 5 + a26 x 2 x 6 + a27 x 2 x 7 (13.54) have that luxury. Despite the experimenter’s best efforts, even
the best science contains trial, error, and serendipity. As a
where x5 is the degree of steam injection rule, it is best not to spend more than 25% of the budget in
the first experimental series. That way, subsequent experi-
x6 is the fraction of nitrogen in the fuel
ments are planned in light of knowledge gained. These reflec-
x7 is the degree of air staging
tions point to developing some strategy for learning during
a15, a26, and a27 are the constant coefficients
the experimental process.
Figure 13.9 outlines a general experimental strategy. It
13.6.3 Sequential Experimental Strategies starts by using domain knowledge to identify important can-
Paradoxically, the best time to plan experiments is after one didate factors and conjecture about their relationship. Expe-
has performed them. Only then does one understand which rience shows that while investigators identify significant
are the right experiments. Unfortunately, investigators never factors, they often overlook important candidates. In general,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 427 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 427

A. Fractional Factorial B. Full Factorial Design,

Design, FF(3,-1,2) FF(3,0,4)

Can generate: Can generate:

3 2 3
Sax Si 1aixi + iS< j Sj 1aij xi x j
3
y = ao + i i y = ao +
i =1 = =

C. Central Composite Design,

CC(3,0,6)

Can generate:
3 2 3 3
y = ao + S aixi + iS< j Sj = 1aij x ix j +iS=1a ijx i2
i =1

A sequential design in three orthogonal blocks, starting with a firs

order design and ending with a second-order one.

FIGURE 13.10 Some orthogonal designs for f = 3 arranged in a sequential strategy.

it is best to err on the side of including too many factors rather try again. In this figure, the investigator has found a region
than too few. Screening designs such as those described in of interest that requires better resolution. He augments the
Section 13.3.4 can reduce the number of tests. Foldover and design via foldover in Figure 13.10B and generates several
blocking in, covered in Section 13.3.7, increase the resolution interaction terms. Two additional centerpoint replicates con-
of the factorial design. If diagnostics show a lack of fit as in firm significant lack of fit. Therefore, in Figure 13.10C, the
Section 13.3.2, one may augment the design with axial points investigator augments the design with axial points to form a
to form a second-order design as in Section 13.4. As an central-composite design that accounts for curvature.
optimum is approached, second-order designs often become
necessary. If the resulting models are still insufficient, the 13.7 LINEAR ALGEBRA PRIMER
investigators should carefully examine the way they per- This section reviews Taylor and MacLaurin series approxi-
formed the experiments, review candidate factor lists, and mations and their relevance to SED. It also reviews elemen-
enlist the aid of colleagues. Be especially alert for aberrant tary matrix operations, for example, matrix multiplication,
responses from quality data. Often, the data point the way to identity and inverse, and addition. These linear algebra topics
influential but previously unknown factors and relations. are implicit and fundamental to SED calculations.
Figure 13.10 shows one sequential experimental strategy
for f = 3. The investigator starts with a fractional-factorial 13.7.1 Taylor and MacLaurin Series
design and two centerpoint replicates in Figure 13.10A, gen- Approximations
erating a linear model. If the model is insufficient, the inves- Why do SED methods work? One can use first-, second-, and
tigator can move in the direction of improved response and higher-order curves to approximate engineering functions
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

428 The John Zink Combustion Handbook

because a Taylor series allows it. Consider an arbitrary analytic Thus, a2n refers to the second row, nth column in matrix a. A
function y(ξ). An interval centered at a can be approximated double underline indicates a matrix. A matrix having a single
using a Taylor series: column or row (also known as a vector) uses a single under-
line. In general, a matrix can contain any number of rows and
d 2 y (ξ − a)
2
columns, and the number of rows and columns need not be
y(ξ) = y(a) + ( ξ − a) + 2
dy
+ L (13.55)
dξ dξ 2! equal.

If one lets x = ξ – a, then the series becomes a MacLaurin

series with no loss of generality: 13.7.2 Matrix Multiplication
Matrix multiplication proceeds from left to right with the
d y (x) 2 2
dy result being the sum of the product of rows and columns
y( x ) = y(0) + (x) + 2 +L (13.56)
dx dx 2! taken in order per Eq. (13.59). Therefore, to multiply two
matrices, the number of columns, n, in the first must equal
If one uses two or more variables, then one must account for
the number of rows in the second. The result of multiplying
mixed derivatives. Usually, a second-order function (at most)
an m × n and n × p matrix is an m × p matrix:
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

is sufficient to correlate any response over the interval of inter-

est. Equation (13.57) gives the MacLaurin series for n factors:
 a11 a12 L a1n   b11 b12 L b1 p 
 a21 L a2 n   b21 L b2 p 
y( x1 , x 2 , …, x n ) = y(0, 0, …, 0)
a22 b22
  
 M M O M  M M O M 
∂y ∂2 y a amn   bn1 L bnp 
∑ ∑∑  m1 am 2 L bn 2
+ x + xx (13.57)
i
∂xi i i< j j
∂xi ∂x j i j
 n n n

∂ y  ∑ a1i bi1 ∑ a1i bi 2 L ∑a b 
∑ ∂x
2 1i ip
1
+ 2
xi2 + L  i =1 i =1 i =1 
2!  n n n

∑a b ∑a b ∑
i i

= 2 i i1 2i i 2 L a2 i bip  (13.59)
 i =1 i =1 i =1 
Often, the function y is unknown (or unknowable). This  M M O M 
 n n n 
∑ ∑ ∑
makes it impossible to derive the partial derivatives analytically.
However, Eq. (13.37), developed in Section 13.4, gives the  ami bi1 ami bi 2 L ami bip 
 i =1 i =1 i =1 
general second-order regression model:
In general, matrix multiplication is not commutative (ab ≠ ba,
y = a0 + ∑a x + ∑∑a x x + ∑a x
i
i i
i< j j
ij i j
i
2
ii i (13.37) in general). It is associative, a(ba) = (ab)c. One can multiply
a constant and a matrix, that operation being commutative
ak = ka:
Its coefficients correspond one-for-one with the partial deriv-
atives in Eq. (13.57) up to the second order. Therefore, the
coefficients (ai , aij , and 2!aii ) estimate the partial derivatives  a11 a12 L a1n   ka11 ka12 L ka1n 
in the MacLaurin series. One can estimate the coefficients for  a21 a22 L a2 n   ka ka22 L ka2 n 
Eq. (13.37) using the least-squares technique.   k =  21  (13.60)
 M M O M   M M O M 
Consider a, a general m × n matrix (read m by n): a L amn   ka L kamn 
 m1 am 2  m1 kam 2

 a11 a12 L a1n 

 a21 a22 L a2 n  13.7.3 Identity, Inverse, and Transpose
a=  (13.58)
 M M O M  The identity matrix (I) has the property that Ia = aI = a .
a L amn 
 m1 am 2 The identity matrix comprises a diagonal matrix of unit val-
ues (and zero values elsewhere). For convenience, the zero
Each matrix element is indexed with a double subscript. elements are omitted, but their existence is understood
The first subscript refers to the row, the second to the column. implicitly:
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch13-Frame Page 429 Wednesday, April 6, 2005 4:39 PM

Experimental Design for Combustion Equipment 429

1 0 L 0  1   a11 a12 L a1n   b11 b12 L b1n 

0 1 L 0  1   a21 a22 L a2 n   b21 b22 L b2 n 
  =  (13.61)   + 
M M O M  O   M M O M   M M O M 
    a
0 0 L 1  1  m1 am 2 L amn   bm1 bm 2 L bmn 
(13.65)
 a11 + b11 a12 + b12 L a1n + b1n 
If m = n then a is square. For a square matrix the inverse (a–1),
 a21 + b21 a22 + b22 L a2 n + b2 n 
if it exists, has the property a–1a = aa–1 = I. The transpose = 
operation, denoted aT, switches rows and columns:  M M O M 
a + b am 2 + bm 2 L amn + bmn 
 m1 m1

 a11 a12 L a1n  Matrix addition is both commutative, a + b = b + a, and asso-

 a21 a22 L a2 n  ciative, a + (b + c) = (a + b) + c.
if a =  ,
 M M O M 
a L amn 
 m1 am 2 REFERENCES
 a11 a21 L am1  1. J. Colannino, Control of NOx using MRSM, 1998
 a12 a22 L am 2  American-Japanese Flame Research Committees Inter-
then a T =   (13.62) national Symposium, Maui, October 11-15, 1998.
 M M O M 
a L amn 
2. J. Colannino, Results of a Statistical Test Program to
 1n a2 n
Assess Flue-Gas Recirculation at the Southeast
Resource Recovery Facility (SERRF), Paper 92-22.01,
 b1  presented before the Air & Waste Management Associ-
 b2  ation, 85th Annual Meeting & Exhibition, Kansas City,
if b =   , then b T = (b1 b2 L bn ) (13.63) MO, June 21-26, 1992.
 M
b  3. G.E.P. Box and N.R. Draper, Empirical Model Build-
 n
ing and Response Surfaces, John Wiley & Sons, New
York, 1987.
In general, the inverse is soluble both analytically10 and
4. G.E.P. Box, W.G. Hunter, and J.S. Hunter, Statistics for
numerically.11 However, for diagonal matrices, the inverse is
Experimenters, John Wiley & Sons, New York, 1978.
simply the reciprocal of the diagonal elements.
5. J.A. Cornell, Experiments with Mixtures, Designs
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Models, and the Analysis of Mixture Data, 2nd ed., ISBN

 a11  0-471-52221-X, John Wiley & Sons, New York, 1990.
 a22  6. I. Glassman, Combustion, 3rd ed., Academic Press, San
if a =  ,
 O  Diego, 1996, 273.
 ann 
 7. J. Colannino, NOx and CO: Semi-empirical models for
boilers, Session 49.T28.d, presented at the American
 1  Power Conference, 59th Annual Meeting, Chicago,
 a11  April 1-3, 1997.
 1 
−1   8. G.E.P. Box and D.R. Cox, An analysis of transformations,
then a = a22  (13.64)
J. Roy. Statistical Assoc., B26, 211-252, April 8, 1964.
 O 
 1  9. J. Colannino, Using modified response surface method-
  ology (MRSM) to control NOx, Institute of Clean Air
 ann 
Companies Seminar: Cutting NOx Emissions, Durham,
North Carolina, March 81-20, 1998.
13.7.4 Matrix Addition 10. E. Kreyszig, Advanced Engineering Mathematics, John
One can add two matrices if both their rows and columns are Wiley & Sons, New York, 1979, chap. 19.
equal. The resulting m × n matrix comprises the sum of the 11. A.J. Pettofrezzo, Matrices and Transformations,
individual matrix elements: Dover, New York, 1978.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 430 Tuesday, March 22, 2005 3:27 PM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 14
Burner Testing
Jeffrey Lewallen, Robert Hayes, Prem Singh, and Richard T. Waibel

TABLE OF CONTENTS

14.1 Introduction............................................................................................................................................. 432

14.2 Burner Test Setup.................................................................................................................................... 433
14.2.1 Application................................................................................................................................ 434
14.2.2 Test Furnace Selection Criteria................................................................................................. 434
14.2.3 Selection of Test Fuels .............................................................................................................. 434
14.3 Instrumentation and Measurements ........................................................................................................ 437
14.3.1 Measuring Air-side Pressure and Temperature ......................................................................... 437
14.3.2 Furnace Gas Temperature Measurement................................................................................... 438
14.3.3 Emissions Analysis ................................................................................................................... 438
14.3.4 Fuel Flow Rate Metering .......................................................................................................... 441
14.3.5 Flame Dimensions .................................................................................................................... 441
14.3.6 Heat Flux................................................................................................................................... 442
14.4 Test Matrix (Test Procedure)................................................................................................................... 442
14.4.1 Heater Operation Specifications ............................................................................................... 442
14.4.2 Performance Guarantee Specifications ..................................................................................... 442
14.4.3 Definition of Data to be Collected ............................................................................................ 444
References ................................................................................................................................................................ 444
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

431
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 432 Tuesday, March 22, 2005 3:27 PM

432 The John Zink Combustion Handbook

FIGURE 14.1 John Zink Co., LLC, R&D Test Center, Tulsa, Oklahoma. (With permission.)

14.1 INTRODUCTION conditions can be simulated and the actual operational perfor-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Burner testing provides an opportunity to gather and verify mance of a burner can be measured accurately. Testing allows
valuable information such as operating parameters, pollutant a burner manufacturer to optimize a burner design to closely
emissions, flame dimensions, the heat flux data, safety meet the requirements of a specific application.
limitations, and noise data. Information from test data is often The operating parameters that can be obtained through test-
essential for performance verification of customer applications ing include the heat release envelope of the burner. Burners are
as well as being vital to research and development efforts. sized for maximum heat release with a specified turndown, or
Empirical data collected from burner testing is a valuable a minimum rate at which a burner can be safely operated.
source of information that can be used to improve the Turndown is defined as the ratio of maximum heat release to
predictive capabilities of CFD models, which are becoming minimum heat release. For example, if the maximum heat
more prevalent tools used in the research, development, and release of a burner is 5 MMBtu/hr and the minimum heat
design of combustion equipment at the forefront of technology release of that burner is 1 MMBtu/hr, then the turndown is 5:1.
in the industry. At state-of-the-art test facilities as shown in Another variable operators and engineers may need to know is
Figure 14.1, testing is done year-round to provide furnace what happens to a burner if it is fired beyond its maximum
designers with the data they need to improve heater designs designed heat release. With this performance information, a
and operate their heaters and furnaces more efficiently as well customer can set a target oxygen level in the flue gas to stay
as to develop new technology to meet the ever-increasing above or set an upper pressure limit for a given fuel to stay
demands of customer processes and environmental regulation. below to ensure that the burner does not exceed the designed
While designing a burner appears to involve relatively simple parameters. More importantly, test data can determine the upper
calculations, it is difficult to predict how a burner will operate heat release value at which a burner can be safely operated for
over a broad range of operating conditions. Considering the short durations until a process upset can be corrected.
multitude of heater applications, the wide range of fuels avail- An operator also needs to know the point at which a burner
able to be burned, the required pollutant levels to be met, and will become unstable if fired below the minimum heat release.
the different ways to supply air, the variations between burner The rate at which a burner can be fired below the designed
designs are nearly infinite. Through full-scale testing, specific minimum heat release is defined as the absolute minimum.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 433 Tuesday, March 22, 2005 3:27 PM

Burner Testing 433

A lower pressure limit can be set for the fuel gas to ensure safe decoking procedures can be increased, improving the heater’s
heater operation. This information is especially useful in deter- runtime and thus its efficiency.
mining how many burners should be fired and at what heat Testing can provide a variety of data concerning a burner’s
release for special operations. Decoking and furnace startup to performance. But without a proper setup, the correct instru-
cure refractory are examples of special operations in which a mentation and measurement methods, and a well-defined test
heater is fired at a level different from its usual designed rate. procedure (test matrix), the data collected during a test may
Along with defining the firing envelope of a burner, the be meaningless. This chapter discusses the proper elements
proper air door or air damper settings can be determined required for conducting a test. Items to be covered include
through testing to ensure the efficient operation of a heater identifying the application of the burner, selecting the correct
by controlling the excess oxygen in the flue gas. By running test furnace, and determining the test fuels to be utilized
at lower excess oxygen concentrations, fuel savings can be during testing.
realized, leading to higher heater efficiency. In complex fur-
This chapter also discusses the instrumentation necessary
naces such as ethylene heaters, which may have hundreds of
to record consistent and accurate data — in particular, the
burners in operation at once, advanced knowledge of air door
concentrations of NOx, CO, O2 (wet and dry), unburned
settings for various operating conditions can save time in
hydrocarbons, and particulates in the flue gas, heat flux, and
trimming out the excess air during actual operation.
noise emissions from the burner. Fuel flow metering and flame

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Other information that can be collected during a burner test
measurement are covered as well.
or demonstration includes emissions of pollutants such as
NOx, CO, and unburned hydrocarbons (UHCs). Based on With input from the heater manufacturer and end user, a
theory and field experience, although it is easy to predict meaningful test procedure can be put together that will yield
emissions for a single fuel, modern burners are often expected valuable data in determining a burner’s performance under
to burn a wide range of fuels. As a result, some fuels may different operating conditions and fuels. The test procedure
not be fired at their optimum pressures, and variables such as is designed to answer a specific set of questions regarding the
fuel pressure can significantly affect the emission perfor- performance of the burner. By closely matching the condi-
mance of a burner. By testing a burner in an operating furnace tions of operation expected in the field, data can be collected
prior to final installation, the expected emissions for different that will aid operators in running their furnaces.
operating conditions can be predicted and anticipated with a Finally, this chapter discusses data analysis. Once a test is
greater degree of accuracy. It is likely that the emission tests run, it must be determined if the burner has met the criteria
in a furnace might differ in the field to some extent due to outlined in the test procedure. The criteria include performance
factors such as interaction with other burners, furnace condi- guarantees and operating parameters. With the data collected,
tions, and changes in fuel compositions. the test engineer can optimize the burner to improve emissions,
When firing burners with a wide variety of fuels, flame flame dimensions, stability, and air flow distribution.
dimensions can change, depending on the fuel fired and the Armed with the knowledge described above, customers
operating fuel pressure, as well as the heat release, because will have a greater understanding of what to expect from a
the mixing energy available can significantly affect the vol- burner test, as well as what goes into setting up and conduct-
ume or shape of a flame. By conducting a burner test over ing a test that will provide meaningful data. API 535 gives
the normal operating envelope of a burner, the dimensions of some good guidelines for specifications and data required for
the flame can be determined for all conditions. The flame burners used in fired heaters.1
dimensions are important for ensuring there is no flame
impingement on the process tubes in the furnace.
Another valuable piece of data that can be collected is noise 14.2 BURNER TEST SETUP
data. New plants built today, as well as existing plants, see One of the most important aspects of a burner test is the setup.
stricter requirements for noise levels. Depending on the sever- This includes the selection of a test furnace, which is
ity of the requirement, mufflers can be designed to attenuate determined by the type of burner to be tested and its
the burner noise to acceptable levels. installation configuration. Typically, test furnaces are built with
Some burners are designed to heat a furnace wall. For these one of two methods of cooling: a water-cooled jacket or a
burners, heat flux profiles can be determined through testing series of water-cooled tubes. A water jacket is simply a furnace
to provide heater manufacturers with information about the surrounded by two shells (inner and outer) of carbon steel that
transfer of heat radiated from the wall to the process tubes. contains circulating water between the shells. This keeps
By optimizing the heat flux profile, cycle times between cooling water on the four vertical surfaces to transfer heat. The
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 434 Tuesday, March 22, 2005 3:27 PM

434 The John Zink Combustion Handbook

(adiabatic), specific gravity, and major components of the

fuels to be used in the actual application.

14.2.1 Application
Although any fuel (solid, liquid, or gas) can be used in a
burner designed for a specified fuel type, this chapter is lim-
ited to gas and liquid fuel firing, as they are by far the most
common found in the hydrocarbon and petrochemical indus-
tries. When firing a fuel, the normal products of combustion
are CO2, H2O, N2, O2, and the energy or heat released during
a combustion process. Unfortunately, there are also other less
desirable products that may be released as well. These com-
monly include unburned hydrocarbons (UHCs), particulates,
NOx, SOx, and CO.

14.2.2 Test Furnace Selection Criteria

The selection of a test furnace is important. The furnace
should be large enough to contain the flame without
impingement on the walls or ceiling of the furnace. Also, it
is important to select the proper furnace to keep the furnace
gas temperature close to the customer’s expected furnace
temperature. The following figures show some typical fur-
naces necessary for burner testing.
The furnace shown in Figure 14.2 can be tested in a variety
of configurations. Wall-fired burners can be tested at the floor
level, and radiant wall burners can be tested at higher eleva-
FIGURE 14.2 Test furnace for simulation of ethylene tions in the furnace to match the customer’s configurations.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

furnaces. The number of burners can be changed to simulate the setup

in the field as well as to achieve a certain furnace temperature.
The furnaces shown in Figure 14.3 can be tested in the
other method utilizes cooling tubes that run either horizontally down-fired configuration to simulate certain types of reform-
or vertically along one or two of the furnace walls. ers where the burners are installed on the roof of the furnace
Burners are designed to cover a wide range of applications. and fired down in between the process tubes. Another type
They are vertically up-fired, vertically down-fired, or hori- of furnace is a vertical cylindrical water-jacketed furnace as
zontally fired. They may have round or rectangular flames shown in Figure 14.4.
and may be free-standing or fired along a refractory wall. The Because test facilities are not built for the purpose of heat-
criteria for selecting a burner normally include the fuel to be ing an oil to a desired temperature or creating products, the
fired, air supply method, emission requirements, and heater heat released from the burners must be absorbed by some
configuration. Fuels can be gas, oil, waste gas, or some com- method. The furnace shown in Figure 14.4 is surrounded by
bination of the three. The air supply can be either natural a shell filled with water. Figure 14.5 shows a test furnace used
(induced) draft, forced draft, turbine exhaust gas, or other to demonstrate burners for terrace wall reformers.
sources of oxygen. Emission requirements are primarily
based on NOx, but can include unburned hydrocarbons, car- 14.2.3 Selection of Test Fuels
bon monoxide, SOx, and particulates. The main criteria for fuel selection include:
The test fuel blend is a critical component of a successful • similarity in combustion characteristics with the actual
test. Without the proper blend, a simulated fuel may not fuel specified for the application
provide data that will aid engineers and operators when start- • economics
ing up new units or evaluating the performance of new burners • availability
installed in an existing unit. Test fuels are typically blended • compatibility with the systems, operations, and
to closely simulate the heating value, flame temperature equipment
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 435 Tuesday, March 22, 2005 3:27 PM

Burner Testing 435

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 14.3 Test furnace for simulation of down-fired tests.

Figure 14.6 shows both permanent and portable fuel storage

tanks. Portable tanks can be used when testing is required on
specialty fuels.
Probably the most critical component of successfully test-
ing a burner is the selection of the test fuel. Without matching
key components in the customer’s fuel, the emissions, stabil-
ity, and flame shape shown during a burner test can vary
significantly when compared to the field results. Fuel can be
blended to match the heating value and molecular weight as
specified or as mutually agreed upon with the customer.
Hydrogen and the diluent content of the gas should be similar
in volumetric proportion to the specified actual service fuel
gas if these proportions significantly impact burner perfor-
mance. The Wobbe index is often used as a criteria for
specifying a test blend that will be used to simulate a fuel.
The Wobbe index is the higher heating value (HHV) of a fuel,
divided by the square root of its specific gravity (SG).

HHV
Wobbe index = (14.1)
SG

The specific gravity (SG) for a gas is the ratio of the molecular
weight of a gas to the molecular weight of air. The specific
gravity for a liquid is the ratio of the density of a liquid to the
density of water. It is important to note that the two fluids
should be compared at the same temperature. Two fuels will FIGURE 14.4 Test furnace for simulation of up-fired tests.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 436 Tuesday, March 22, 2005 3:27 PM

436 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 14.5 Test furnace for simulation of terrace wall reformers.

FIGURE 14.6 Test fuel storage tanks.

provide the same heat release from a gas tip at a given supply gaseous fuels available for blending at well-equipped test
pressure if the Wobbe index is the same. facilities include natural gas, propane, propylene, butane,
While the Wobbe index is a good indicator to see if a fuel hydrogen, nitrogen, and carbon dioxide. The composition
is similar, it is important to try and match the lower heating of natural gas varies by geographic location. As an example,
value (LHV), molecular weight, and adiabatic flame tem- Tulsa natural gas has a typical composition as shown in
perature to ensure a good simulation. Commonly available Table 14.1.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 437 Tuesday, March 22, 2005 3:27 PM

Burner Testing 437

TABLE 14.1 Tulsa Natural Gas (TNG) Composition and With the test fuel(s) established, it is time to determine
Properties what measurements will need to be taken and what instru-
CH4 (volume%) 93.4 C4H10 (volume%) 0.20 ments will be required.
C2H4 (volume%) 2.70 CO2 (volume%) 0.70
C3H8 (volume%) 0.60 N2 (volume%) 2.40
LHV (Btu/scf) 913 HHV (Btu/scf) 1012
Molecular weight 17.16 Specific heat ratio @ 60°F 1.3 14.3 INSTRUMENTATION AND
Adiabatic flame temperature (°F) 3452
MEASUREMENTS
Measurements generally required during testing (but not lim-
TABLE 14.2 Example Refinery Gas ited to) are fuel pressure and temperature, air pressure drop
Fuel Component Formula Volume% and temperature, fuel flow measurement, flame dimensions,
Methane CH4 8.13a and emissions measurements.
Ethane C2H6 19.9a
Propane C3H8 0.30
Butane C4H10 0.06 14.3.1 Measuring Air-side Pressure and
Ethylene C2H4 32.0b
Propylene C3H6 0.78
Temperature
Butylene C4H8 0.66 Most applications for the hydrocarbon and petrochemical
1-Pentene C5H10 0.07 industries being tested today are natural-draft applications
Benzene C6H6 0.12
Carbon monoxide CO 0.22
where practical measurement of the air flow cannot be done
Hydrogen H2 37.8c without impacting the quality of the data recorded.
a Balance of fuel is primarily methane and ethane.
b Level of olefins in the fuel. 14.3.1.1 Natural Draft
c Hydrogen content. The combustion air for natural-draft burners is induced
through the burner, either by the negative pressure inside the
TABLE 14.3 Comparison of Refinery Gas to Test Blend furnace or by fuel gas pressure that educts the air through a
Property Refinery Fuel Test Fuel
venturi. Natural-draft burners are the simplest and least
LHV (Btu/scf) 1031 1026 expensive burners, and are most commonly found in the
HHV (Btu/scf) 1124 1121
hydrocarbon and petrochemical industries. Because the
Molecular weight 18.09 18.38
Specific heat ratio @ 60°F 1.27 1.26 energy available to draw air into the burner is relatively low,
Adiabatic flame temperature (°F) 3481 3452 there is no practical way to measure the air flow through the
Wobbe index 1422 1407 burner. As a result, the temperature of the air, the ambient air
pressure, the fuel flow, and the excess air measurements are
critical for accurately calculating the air flow through a
Table 14.2 illustrates an example of a refinery gas and the natural-draft burner.
points of interest in determining a test blend that will effec-
tively simulate the fuel-handling properties, burning charac- 14.3.1.2 Forced Draft
teristics (tendency of a fuel to coke, etc.), and emission levels Forced-draft burners are supplied with combustion air at a
of the customer’s fuel composition. positive pressure. The air is supplied by mechanical means
(air fans/blowers). These burners normally operate at an air-
Based on the available fuels for blending, the hydrogen side delivery pressure that can be in excess of 2 inches of
content is matched, propylene is used to substitute the ethyl- water column (0.5 kPa). They utilize the air pressure to
ene content, and Tulsa natural gas (TNG) is used to simulate provide a superior degree of mixing between fuel and air.
the methane content. By holding the hydrogen content fixed Also, with forced-draft systems, air control can be better
at 38%, TNG and propylene are balanced to obtain a match maintained, thus allowing furnaces to operate at lower excess
air rates over a wide firing range and allowing the operator to
of the lower heat value (LHV) and molecular weight. By
realize economic savings. Figure 14.7 shows an example of a
attempting to balance the LHV, molecular weight, and adia-
mobile air preheater used during forced-draft testing.
batic flame temperature, a test fuel blend of 34% TNG, 28%
With the use of an air delivery system, the air flow can be
C3H6, and 38% H2 would be acceptable to simulate the refin- measured to provide a direct method of measuring the air flow
ery fuel gas illustrated in Table 14.2. Table 14.3 gives a side- to validate the air flow through a burner. Fuel flow metering
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
by-side comparison of the fuel properties. is still used to also determine the air flow. By knowing the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 438 Tuesday, March 22, 2005 3:27 PM

438 The John Zink Combustion Handbook

FIGURE 14.7 Forced-draft air preheater.

amount of fuel burned and the excess air exiting the furnace, comprised of a thermocouple recessed inside a radiation
the amount of air consumed by combustion at the burner can shield. An eductor rapidly aspirates the hot gas across the
be calculated. thermocouple. This configuration maximizes the convective
heat transfer to the thermocouple while minimizing radiation
14.3.1.3 Turbine Exhaust Gas (TEG) exchange between the thermocouple and its surroundings,
Some applications use turbine exhaust gas, often mixed with ensuring that the equilibrium temperature is nearly that of
air, as the source of oxygen for the burners. The turbine the true gas temperature.
exhaust stream or mixture normally contains between 13 and
17 mole % oxygen. These burners are also forced-draft type
burners. When test firing a TEG simulation, it is important to
14.3.3 Emissions Analysis
match the customer’s TEG stream pressure, temperature, and Emission analysis is an important criterion for burner testing.
oxygen content approaching the burner. During this type of The main pollutants in the combustion products are NOx,
test, a second set of probes must be arranged to measure the CO, unburned hydrocarbons, and particulates.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
TEG stream composition.
14.3.3.1 NOx
14.3.2 Furnace Gas Temperature
The chemiluminescent method is most widely used for NOx
Measurement analysis.2 This method is capable of measuring oxides of
A suction pyrometer (also known as a suction thermocouple nitrogen from sub-parts per million to 5000 ppm. Newer
or velocity thermocouple) is widely considered the preferred detector models are free from the disadvantages inherent in
method for obtaining accurate gas temperature measure- analog systems and provide for increased stability, accuracy,
ments in the harsh environment of a furnace. If a bare and flexibility. The principle of operation of these analyzers
thermocouple is introduced into a hot furnace environment is based on the reaction of nitric oxide (NO) with ozone:
for the measurement of gas temperature, measurement errors
can arise due to the radiative exchange between the thermo-
couple and its surroundings. A suction pyrometer is typically NO + O 3 → NO 2 + O 2 + hν (14.2)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 439 Tuesday, March 22, 2005 3:27 PM

Burner Testing 439

The sample, after it is drawn into the reaction chamber, is transparent to the IR radiation and, therefore, produces a
reacts with the ozone generated by the internal ozonator. The measured beam that can be absorbed by CO in the cell. These
above reaction produces a characteristic luminescence with analyzers can measure 0.1 to 1000 ppm CO under well-
an intensity proportional to the concentration of NO. Specif- controlled conditions. The detailed workings of an IR analyzer
ically, light emission results when electronically excited NO2 can be obtained from a standard text on the subject.
molecules decay to lower energy states. The light emission
is detected by a photomultiplier tube, which in turn generates 14.3.3.3 O2 (Wet and Dry)
a proportional electronic signal. The electronic signal is pro- The oxygen concentration is also conveniently measured by
cessed by the microcomputer into an NO concentration read- chromatographic techniques using thermal conductivity
ing. To measure the NOx (NO + NO2) concentration, NO2 is detectors and also by low resolution FTIR spectroscopy.
transformed to NO before reaching the reaction chamber. This Individual measurements of oxygen concentration are most
transformation takes place in a converter heated to about widely done by analyzers based on standard polarographic
625°C (1160°F). Upon reaching the reaction chamber, the techniques. The detailed working of such analyzers can be
converted molecules along with the original NO molecules found in related texts.
react with ozone. The resulting signal represents the NOx.
Further details of the workings of a chemiluminescent gas 14.3.3.4 Unburned Hydrocarbons (UHCs)
analyzer can be found in any standard text. Unburned hydrocarbons (UHCs) increase as the excess air
rate decreases. The combustion of hydrogen and paraffin-
14.3.3.2 Carbon Monoxide rich fuel will produce a minimum of combustibles. The pres-
The carbon monoxide (CO) exiting a burner will initially ence of unsaturated hydrocarbons leads to pyrolysis and
increase slowly as the excess air rate decreases. The increase polymerization reactions, resulting in more combustibles.
will accelerate as excess air levels continue to decline to Unsaturated hydrocarbons, chlorides, amines, and the like
near-zero. Typical control points range between 150 and can plug or damage burner tips, disrupting the desired fuel-
200 ppm CO. This range usually results in the best overall air mixing. This can cause a further increase in the combusti-
heater efficiency. Certain localities may require lower bles level. Heavy oils are more likely to produce greater lev-
emission levels. The presence of unsaturated hydrocarbons els of combustibles than lighter oils. Heavier components
can lead to pyrolysis and polymerization reactions, resulting are not as easily atomized and ignited, and therefore poly-
in a greater possibility that CO will be produced. Burners merization and pyrolysis reactions are more likely to occur.
with greater swirl and/or higher combustion air pressure drop Forced-draft burners provide a better mixing of the fuel-air
(such as forced-draft burners) typically have lower CO mixture and therefore produce reduced combustibles at
emissions at equivalent excess air levels. The reason is that equivalent excess air rates.
these burners provide a superior degree of mixing to allow Chromatographic techniques are the most widely used for
improved combustion at lower excess air levels. VOC determination in refinery off-gases. Their use as a multi-
Although CO can be continuously monitored by chromato- component, completely automated, and continuous emissions
graphic analysis, using thermal conductivity detectors, or by monitor is not documented in the literature.3 Coleman et al.
FTIR spectroscopy methods, individual analysis is best have discussed the use of a gas chromatography-based con-
accomplished using a nondispersive infrared technique. The tinuous emission monitoring system for the measurement of
main advantages of this technique are that it is highly specific VOCs using a dual-column (with DB-5 and PoraPlot U,
to CO and has lower ranges with a wider dynamic range, respectively) gas chromatograph equipped with thermal con-
increased sensitivity and stability, and easy operation because ductivity detectors, in which separation was optimized for
of microcomputer control diagnostics. An added advantage fast chromatography. In this system, nine different VOCs plus
of the technique is that the changes in temperature and pres- methane and CO2 were separated and analyzed every 2 min-
sure of the sample gas are immediately compensated for by utes. Because permits are issued to report emission in pounds
the microcomputer, and the results are thus not affected by or tons of pollutants emitted and not on the basis of parts per
fluctuations in the operating conditions. The basic principle million (ppm), the setup was equipped with a continuous mass
of these analyzers is based on the radiation from an infrared flow measurement device. The data thus collected can be
source passing through a gas filter alternating between CO converted to pounds or mass of VOCs emitted. The DB-5
and N2 due to rotation of the filter wheel. The CO gas filter column separates ethanol, isopropanol, n-propanol, methyl
acts to produce a reference beam that cannot be further atten- ethyl ketone, isopropyl acetate, heptane, n-propyl acetate, and
uated by CO in the sample cell. The N2 side of the filter wheel
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
toluene. The PoraPlot U column separates methane and
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 440 Tuesday, March 22, 2005 3:27 PM

440 The John Zink Combustion Handbook

carbon dioxide. A chromatographic technique using two capability and sensitivity of the method.9–11 A typical low-
fused silica columns — one with Dura-Bond and the other resolution FTIR spectrometer uses spectral resolution, BaF2
with Gas Solid-Q-PLOT — equipped with a flame ionization optics, a Peltier-cooled semi-conductor detector, and a
detector was used by Viswanath to measure VOCs in air.4 temperature-controlled multi-reflection gas cell. The advan-
A technique reported by Pleil et al.5 uses the fact that the tages of low-resolution FTIR over conventional high-resolu-
compounds once identified by retention time in the chromato- tion FTIR include its rugged design, high signal-to-noise
graphic analysis can be confirmed by determining a second ratio without liquid nitrogen-cooled detectors, reduced data
dimension, such as its mass fragmentation pattern or its infra- storage requirements, and increased dynamic range for quanti-
red absorption spectrum from a highly specific detector such tave analysis.9
as a mass selective detector (MSD) or a Fourier transform Jayanti and Jay12 have summarized studies on estimating
infrared system (FTIR). Even with this combination, care VOCs by different techniques employed by various workers.
should be taken to avoid occasional confusion among iso-
meric, co-eluting compounds with similar, strongly absorbing 14.3.3.5 Particulates
functional groups. Using this technique, Pleil et al.5 were able
Proper combustion of gaseous fuels does not generate signifi-
to identify and successfully determine more than 40 com-
cant quantities of combustion-generated particulates. Particu-
pounds in the VOCs. They used a cross-linked methyl silicon
late emissions generally occur with the burner of heavy fuel
megabore capillary column with both flame ionization detec-
oils. Burners with greater swirl and/or higher combustion air
tor and electron capture detector simultaneously. A similar
pressure (such as forced-draft burners) are less likely to pro-
study of VOCs was reported by Siegel et al.6 They used a
duce particulates. They provide a superior degree of mixing
DB-1 column with flame ionization detector and a mass selec-
to reduce the formation of particulates. Greater atomization
tive detector (GC/MSD).
of fuel oil into finer particles will reduce particulate emis-
The U.S. EPA guidelines, as presented in “Compendium of sions. High-intensity burners can considerably reduce partic-
Methods for the Determination of Toxic Organic Compounds ulates. The high degree of swirl, coupled with the high-
in Ambient Air, Method TO-14,” is slowly becoming the cri- temperature reaction zone, induces superior combustion of
terion for VOCs.7 The recommended method uses cryogenic the particulates. However, these burners also emit an
preconcentration of analytes with subsequent gas chromato- increased amount of NOx.
graphic separation and mass spectrometric detection. The
The particulates from hydrocarbon industries are the pollu-
methodology requires detecting nanogram and subnanogram
tants emitted by the effluent gases. The most important crite-
quantities. To obtain this high sensitivity, Method TO-14 rec-
rion for the evaluation of particulates is the particle size. It
ommends the use of a selective ion monitoring (SIM) spec-
has been observed that different results are obtained using
trometric technique. The details of the method are discussed
different techniques of collection and analysis.
by Pleil et al.7 Evans et al.8 have also discussed the use of a
cryogenic GC/MSD system to measure the VOCs in air in The U.S. EPA13 recommended procedures suggest that
different parts of the country. The sample first passes through sampling ports be located at least 8 × duct diameters down-
a fused silica column to resolve the target compounds. The stream and 2 × diameters upstream from any flow distur-
column exit flow splits such that one-third of the flow is bance. Flue gas should be drawn through a U.S. EPA
directed to the chromatographic column (with the flame ion- sampling train.13 It is important to maintain isokinetic sam-
ization detector) and two-thirds of the flow goes to the mass pling conditions.
selective detection system (MSD). The method was found to Particles can be collected by filtration, impaction, and
effectively detect 0.1 ppb by volume of about 25 VOCs. impingement. Glass fiber and membrane filters are efficient
Larjava et al.9 have recently reported a comprehensive tech- for 0.3-µm particles. These filters can be used in an inline filter
nique for the determination of nitric oxide (NO), sulfur dioxide holder. The filter holder can be kept inside the sampling port
(SO2), carbon monoxide (CO), carbon dioxide (CO2), and total such that the filter attains the temperature of the gas stream.
hydrocarbons (CxHy) in the air. The technique used single- For particulate collection by impaction, an Anderson type
component gas analyzers in parallel with a low-resolution in-stack sampler is used. In cases where the sampler cannot
Fourier transform infrared (FTIR) gas analyzer. This technique be accommodated inside the sampling port flange, it can be
successfully demonstrated that the results obtained by single- put outside, with an arrangement to heat it to prevent con-
component analyzers and FTIR were very close. Online analysis densation within the sampler. In this type of sampler, the
of stack gases with FTIR spectrometry has recently received collecting plates are coated by a thin film of silicone grease
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

considerable attention because of the multi-component analysis formed by immersing the plates in a 1% solution of silicone
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 441 Tuesday, March 22, 2005 3:27 PM

Burner Testing 441

grease dissolved in benzene and drying them overnight at and can be changed to match the flow rate and fuel to be
100°C (212°F). metered. Also, there is a significant amount of data concerning
The collection of particulates by impingement consists of measuring fuel flow via an orifice plate.15 Finally, there are
using a series of three or four liquid impingers. These no moving parts to wear out. The drawbacks to orifice plates
impingers each contain 250 ml distilled water. A common are that they are precision instruments and the following must
practice is to use impaction and impingement techniques, be considered: the flatness of the plate, the smoothness of the
followed by glass fiber backup filters. plate surface, the cleanliness of the plate surface, the sharp-
Sampling times also vary according to the technique ness of the upstream orifice edge, the diameter of the orifice
employed; 1 to 5 minute samples are common for filters and bore, and the thickness of the orifice edge.15 The critical
Anderson-type units, while 20 minutes are needed for impingers. inaccuracies due to these items can be alleviated by the pur-
Filter samples are analyzed by light microscopy and scan- chase of ASTM-approved plates, rather than machining the
ning electron microscopy (SEM). The liquid samples from plates. Another drawback is loss of accuracy when measuring
the wet impingement device are filtered onto 0.2-µm mem- flow rates of dirty fuels. While dirty fuels are a way of life
brane filters and examined by SEM. The samples from an for the refining industry, test fuels are clean (no liquid or solid
Anderson sampler can be analyzed by the recommended pro- particles in the gaseous fuels), and this concern is minimized.
cedure or by calculations based on Ranz and Wong equations. Two items that should be verified when testing with orifice
The Anderson plates can also be examined by SEM to deter- plates are that they are installed in the right direction (the
mine the range of particles trapped on the plates. Particle paddle usually indicates the inlet side) and that the correct
counts can be done by light microscopy using an oil-immer- orifice bore is in the correct flow run. While orifice plates can
sion lens system.14 Individual particles are compared on the be used to meter liquids, coriolis meters are often preferred
basis of equal area to previously calibrated circle sizes con- for measuring liquid fuel flows such as No. 6 oil or diesel oil.
tained on a size comparator. In an exhaustive study, Byers14
took data on particulates from a refinery effluent and sug- The coriolis meter operates on the basic principle of motion
gested that membrane filters should be preferred when the mechanics.15 The coriolis meter is able to measure the mass
gas being sampled is at a temperature less than 300°F by measuring the amount of vibration the tube carrying the
(150°C). For higher temperatures, the Anderson sampler is fluid is undergoing. The coriolis meter is a more expensive
suitable provided the plates are suitably coated. Sampling means of measurement, but this is often offset by its degree
techniques causing agglomeration, such as glass fiber filters, of accuracy and its low maintenance requirements.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

wet collection, bulk grab samples, and scrapping of deposits

from collecting surfaces, should be avoided.
14.3.5 Flame Dimensions
14.3.4 Fuel Flow Rate Metering The flame shape and dimensions are determined by the
burner tile, the drilling of the gas tip, the fuel, and the
One of the most important aspects of burner testing is fuel
metering. When firing a natural-draft burner, it is difficult to aerodynamics of the burner. Round burner tiles are used to
measure the flow rate of combustion air. Therefore, accu- produce a conical or cylindrical shape. Flat flame burners are
rately metering the flow rate of each individual component designed with rectangular burner tiles and produce fishtail-
used to make up a fuel blend is necessary to measure the heat shaped flames. Many of the liquid fuel burners are designed
release of the burner and its performance. There are many with round burner tiles and produce a conical flame. The
ways to measure flow: differential pressure, magnetic, mass, drilling of the oil tip determines the shape and length of the
oscillatory, turbine, and insertion flow meters, just to name a flame. The normal included angle of a burner tip is 40 to 70°.
few. For purposes of burner testing, the differential flow With a 50° included angle, the flame length will be
meter is discussed. Even limiting this discussion to differen- approximately 2 ft per MMBtu/hr for natural-draft burners.
tial flow meters, there are still several different methods of Reducing the angle to 40° produces a longer, narrower flame.
measurement available. Measuring the differential pressure Increasing the angle to 70° produces a shorter, bushier flame.
across a known orifice plate is a common method of measur- Forced-draft burners produce a shorter flame because of the
ing the gaseous fuel and steam flow. For liquid fuel firing, a better mixing between the fuel and the air. Firing in
coriolis meter is often used. combination with both liquid and gas fuels will increase the
The orifice plate is the most commonly applied method.15 length and volume of the flame and can cause coking of the
The advantage of using orifice plates is that they are versatile oil and gas tips.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 442 Tuesday, March 22, 2005 3:27 PM

442 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 14.8 Heat flux probe schematic.

14.3.6 Heat Flux utilize a crystal window, gas screen, or a mirrored ellipsoidal
cavity to negate convective heat transfer to the sensor. A radi-
Several techniques have been developed to measure heat flux ometer is also often equipped with a gas purge in an effort to
levels at different locations within a furnace. The instruments keep the crystal window or mirrored ellipsoidal cavity clean
designed to successfully obtain heat flux data in the hostile and free from fouling. Critical parameters to consider when
environment of a full-scale furnace are typically water-cooled using a heat flux meter include the ruggedness, sensitivity,
probes, which are inserted through a furnace port at the loca- calibration method, and view angle of the instrument.
tion of interest. The probes may utilize pyrometers that mea-
sure radiant or total (radiant + convective) heat flux levels.
The sensing element is typically composed of a thermopile-
14.4 TEST MATRIX
type sensor that outputs a voltage proportional to the tempera-
ture difference between the area of the element exposed to
(TEST PROCEDURE)
heat transfer from the furnace and the area that is cooled and
kept at a relatively constant temperature per the element 14.4.1 Heater Operation Specifications
design. Sensor element designs differ chiefly between the Some of the parameters normally measured in burner testing
geometry and configuration of the thermopile-type sensing are fuel pressure, airside pressure drop, noise emissions, NOx
element. Figures 14.8 and 14.9 show schematics of typical emissions, CO emissions, UHC emissions, particulate emis-
heat flux and radiometer designs, respectively. Common sions, heat flux profiles, and flame dimensions.
designs utilize a plug-shaped thermopile element with the
exposed face at one end and the opposite end cooled by con- 14.4.2 Performance Guarantee Specifications
tact with a heat sink. Others use a disk-shaped sensor with the The primary reasons for conducting a burner test is to deter-
temperature gradient existing between the center of the disk mine the operating envelope of the burner as well as the
receiving radiant energy and the radial edge, which is cooled emissions performance. With this data collected, the burner’s
by contact with a heat sink. A sensor designed to measure performance in the field will be more predictable and easier
only the radiant component of heat flux (radiometer) can to operate.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 443 Tuesday, March 22, 2005 3:27 PM

Burner Testing 443

FIGURE 14.9 Ellipsoidal radiometer schematic.

14.4.2.1 Emissions Guarantees The fuel-side pressure drop is displayed during the test. It
It is important to identify which fuels are the operation or is important that the test engineer ensure that the customer’s
online fuels and which fuels are for start-up or emergency use fuel will meet the design pressure requirement based on the
only. By identifying which fuels the emissions guarantees data collected on the test fuels.
apply to, the burner can be better optimized to run on the When verifying the air-side pressure drop, the test engineer
operation fuels. must determine the elevation and the range of the ambient air
temperatures that the burners will be subjected to once
14.4.2.2 Noise installed in the field.
Noise emissions are becoming as important as stack

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
emissions. With some refineries located near populated areas, 14.4.2.4 Flame Dimension Guarantees
it is important to keep noise to a minimum. Burner testing is
Flame dimensions in a full-scale furnace are typically made
usually conducted on a single burner and noise emissions are
by subjective measurement. The flame envelope is most often
usually measured 1 m (3 ft) from the burner. Data collected
determined by visual observation. This operating parameter
during the test include an overall dBA measurement and an
is important to ensure that the flame will not impinge on the
octave band measurement ranging from 31.5 to 8000 Hz.
furnace process tubes or interact with another burner’s flame.
When collecting noise data, it is important to measure it with
Flame impingement on the tubes can damage the process
the burner operating and without it operating, to obtain the
tubes and cause the furnace to prematurely shut down for
background noise, which may or may not be required to
repairs — at great expense to the operator. Flame interaction
determine the noise contribution from the burner.
between two or more burners can result in longer, more
uncontrollable flames and higher emissions. It is important to
14.4.2.3 Fuel and Air-side Pressure Drop identify the burner spacing, the furnace dimensions, and the
The fuel and air-side pressure drop also need to be verified customer’s desired flame dimensions. With this information,
during the test. The test confirms that the burner will have the the test engineer can fine-tune the flame envelope to improve
correct capacity for proper operation. the burner’s performance in the customer’s heater.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 444 Tuesday, March 22, 2005 3:27 PM

444 The John Zink Combustion Handbook

14.4.3 Definition of Data to be Collected 7. J.D. Pleil, T.L. Vossler, W.A. McClenny, and K.D. Oliver,
Prior to installation and testing a burner, a test matrix (test Optimizing sensitivity of SIM mode of GC/MS
procedure) must be developed. A sample fuel gas specifica- analysis for EPA’s TO-14 air toxics method, J. Air
tion is shown in Table 14.4. A typical test procedure might Waste Manag. Assoc., 41, 287, 1991.
resemble that shown in Table 14.5. With a well-developed
test procedure, the data collected from a test will be meaning- 8. G.F. Evans, T.A. Lumpkin, D.L. Smith, and
ful and will assist the operator in running the furnace and pre- M.C. Somerville, Measurement of VOCs from the
dicting the performance from the furnace. TAMS network, J. Air Waste Manag. Assoc., 42, 1319,
1992.

REFERENCES 9. K.T. Larjava, K.E. Tormonen, P.T. Jaakkola, and

A.A. Roos, Field measurements of flue gases from
1. Burners for Fired Heaters in General Refinery
combustion of miscellaneous fuels using a low
Services, API Publication 535, 1st ed., American
resolution FTIR gas analyzer, J. Air Waste Manag.
Petroleum Institute, Washington, D.C., July 1995.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Assoc., 47, 1284, 1997.
2. B.K. Gullett, M.L. Lin, P.W. Groff, and J.M. Chen,
NOx removal with combined selective catalytic 10. K. Wülbern, On line messung von rauchgasen mit
reduction and selective non catalytic reduction: pilot einen FTIR spektrometer, VGB Kraftwerkstechnik, 72,
scale test results, J. Air Waste Manag. Assoc., 44, 1188, 985, 1992.
1994.
11. J.C. Demirgian and M.D. Erickson, The potential of
3. W.M. Coleman, L.M. Dominguez, and B.M. Gordon, A
continuous emission monitoring of hazardous waste
gas chromatographic continuous emission monitoring
incinerators using FTIR spectroscopy, Waste
system for the determination of VOCs and HAPs, J. Air
Waste Manag. Assoc., 46, 30, 1996. Management, 10, 227, 1990.

4. R.S. Viswanath, Characteristics of oil field emissions in 12. R.K.M. Jayanti and B.W. Jay (Jr.), Measurement of
the vicinity of Tulsa, Oklahoma, J. Air Waste Manag. toxic and related air pollutants, J. Air Waste Manag.
Assoc., 44, 989, 1994. Assoc., 40, 1631, 1990.
5. J.D. Pleil, K.D. Oliver, and W.A. McClenny, Ambient
13. Standard Performance for New Stationary Sources,
air analyses using nonspecific flame ionization and
Environmental Protection Agency, Federal Register,
electron capture detection compared to specific
detection by mass spectroscopy, J. Air Waste Manag. 36, No. 247, December 23, 1971.
Assoc., 38, 1006, 1988.
14. R.L. Byers, Evaluation of effluent gas particulate
6. W.O. Siegel, R.W. McCabe, W. Chun, E.W. Kaiser, collection and sizing methods, API Proc. Division
J. Perry, Y.I. Henig, F.H. Trinker, and R.W. Anderson, Refining, 53, 60, 1973.
Speciated hydrocarbon emission from the combustion
of single component fuels. I. Effect of fuel structure, 15. D.W. Spitzer, Practical Guides for Measurement and
J. Air Waste Manag. Assoc., 42, 912, 1992. Control, Instrument Society of America, 1991.

Copyright CRC Press

Burner Testing 445

TABLE 14.4 Test Procedure Gas Specification Sheet

John Zink Company Burner Performance Demonstration
Tulsa, OK (English) Burner Specification (Gas)

Date: 2/4/00 Rev. No: 0 J.Z. Quote No:

Customer: Customer P.O No:
Burner: PSFFG-45M J.Z. Sales Order No:
Drawing: Capacity Curve No:
User: Project Engineer
Jobsite: Test Engineer

Customer Heater Data

Spec Reference: Direction of Firing:

Item No: Setting Thickness:
Quantity of Burners: 60 Burner Spacing: 39 inches (1 meter)
Type of Heater: Ethylene Pilot: ST-1S manual pilot
Firebox Dimensions: Elevation: <1000 feet ASL
Draft Type: induced

Specifications

Fuel Composition LHV M.W. Fuel A Fuel B TNG

Component: Btu/scf #/# mol vol% vol% vol% vol%
Tulsa natural gas 913.0 17.160 34.00 80.00 100.00
Hydrogen 273.8 2.022 38.00 20.00
Propane 2314.9 44.100
Propylene 2181.8 42.080 28.00
Carbon Dioxide 44.010
Butane 3010.8 58.120
Lower Heating Value: Btu/scf
Molecular Weight: (#/# mol) 18.385 14.132 17.160
Isentropic Coefficient: (Cp/Cv) 1.2600
Temperature: (degF) 60 60 60
Pressure Available: psig 20.0 20.0 20.0
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Heat Release per Burner:

Design Maximum: (MMBtu/hr) 6.800 6.800 5.913
Normal: (MMBtu/hr) 5.913 5.913 1.360
Minimum: (MMBtu/hr) 1.360 1.360 6.800
Flame Dimensions: @ ft
Cross Section (Dia.): 4.000 4.000 4.000
Length: 16.000 16.000 16.000
Turndown: 5:1 5:1 5:1
Excess Air @ Design: 10 10 10
Conditions @ Burner:
Heater Draft Available: ("w.c.) 0.80 0.80 0.80
Burner dP @ Design: ("w.c.) 0.80 0.80 0.80
Combustion Air Temp.: (degF) 100 100 100
Guarantees:
NOx: ppm(vd) 100 100 100
CO: ppm(vd) 50 50 50
Note 1 Particulate: #/MMBtu (lhv)
Note 1 UHC: #/MMBtu (lhv)
DSCF/MMBtu @ 3% O2(d) (lhv)
Noise: dBA (spl) @ 3 ft. 85 85 85
Conditions of Guarantees:
% Oxygen Corrected to: 3 3 3
Combustion Air Temp.: (degF) 100 100 100
Heat Release: (MMBtu/hr) 4.500 4.500 4.500
Furnace Temperature: (degF) 2100 2100 2100
Comments: General — Heat Releases are shown as Net or Lower Heating Value
Note 1: Particulate and UHC are NOT measured during Test Demonstration
Courtesy of John Zink Co., Tulsa, OK.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch14-Frame Page 446 Tuesday, March 22, 2005 3:27 PM

446 The John Zink Combustion Handbook

TABLE 14.4 (continued) Test Procedure Gas Specification Sheet

Typical Refinery Fuel Composition

Fuel Component

Name Formula volume %

Methane CH4 8.13*

Ethane C2H6 19.9 *
Propane C3H8 0.30
Butane C4H10 0.06
Ethylene C2H4 32.0**
Propylene C3H6 0.78
Butylene C4H8 0.66
1-Pentene C5H10 0.07
Benzene C6H6 0.12
Carbon Monoxide CO 0.22
Hydrogen H2 37.8***

Tulsa Natural Gas Composition

CH4 (Volume %) 93.4 C4H10 (Volume %) 0.20

C2H4 (Volume %) 2.70 CO2 (Volume %) 0.70
C3H8 (Volume %) 0.60 N2 (Volume %) 2.40
LHV (Btu/scf) 913 HHV (Btu/scf) 1012
Molecular weight 17.16 Specific heat rate 1.3
Adiabatic flame temperature (F) 3452

Refinery fuel: Test fuel:

LHV (Btu/scf) 1031 1026

HHV (Btu/scf) 1124 1121
Molecular weight 18.09 18.38
Specific heat ratio @ 60F 1.27 1.26
Adiabatic flame temperature (F) 3481 3452
Wobbe Index 1422 1407
* Balance of fuel is primarily methane and ethane.
** Note: level of olefins in the fuel.
*** Note: hydrogen content

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Burner Testing 447

TABLE 14.5 Example Test Procedure

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
John Zink Company Burner Performance Demonstration Date: 2/4/00
Tulsa, OK (English) Burner Test Procedure Rev. No:
J.Z. SO No: Customer: User: Burner Drawing No: Furn.
Burner: PSFFG-45M P.O. No: Jobsite: Capacity Curve:

Data Liberation Air Temp. Excess O2 Draft deltaP

Point Fuel (MMBtu/hr) (°F) (%O2 (dry)) (in. w.c.) (in. w.c.) Description/Comments

1 TNG 60 0.80 Burner light off on Tulsa natural gas (TNG).

2 A 6.800 60 2.0 0.80 Maximum heat release, set air to 10% excess air, record noise and heat flux
profile.
3 A 60 0.80 Above maximum heat release, increase fuel flow until CO > 250 ppm.
4 A 5.910 60 2.0 0.80 Reduce heat release to normal heat release, set air damper to maintain 10%
excess air.
5 A 1.360 60 0.80 With damper set for normal heat release, reduce fuel flow to minimum heat
release.
6 A 60 0.80 Determine absolute minimum heat release.
7 B 6.800 60 2.0 0.80 Maximum heat release, set air to 10% excess air, record noise and heat flux
profile.
8 B 60 0.80 Above maximum heat release, increase fuel flow until CO > 250 ppm.
9 B 5.910 60 2.0 0.80 Reduce heat release to normal heat release, set air damper to maintain 10%
excess air.
10 B 1.360 60 0.80 With damper set for normal heat release, reduce fuel flow to minimum heat
release.
11 B 60 0.80 Determine absolute minimum heat release.

General Comments: Data to be Recorded:

1) Refer to the schematic of the equipment set-up for approximate instrument and sight port locations. Fuel Flow × Burner dP ×
2) The firebox temperature will be within the range of 1900°F and 2100°F for Normal and Maximum Heat Release. It will be Fuel Press. × NOx ×
lower at reduced rates.
3) Once the O2 has stabilized, at the target value for a given test point data will be recorded and adjustments for the next point Fuel Temp. × CO ×
will begin.
4) All data points will be run with the furnace draft as specified, controlling excess O2 with the blower (forced draft), Air Temp. × O2 ×
or with the damper (register(s)) (natural draft).
5) The proposed time frame for the duration of the test is approximate and could be longer or shorter depending on equipment Draft × Noise ×
operation and/or weather.
6) Standard tolerances on measurements will be as follows: Air Temp. ±20°F, Fuel Temp. ±20°F, O2 ±0.2%, Draft or dP ±6% Box Temp. ×
of specified.

Test Procedure Acceptance:

_ Approved _ Approved as Company: Signature: Date:
Noted

Test Acceptance:
Name: Company: Signature: Date:

Courtesy of John Zink Co., Tulsa, OK.

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 15
Installation and Maintenance
Roger H. Witte and Eugene A. Barrington

TABLE OF CONTENTS

15.1 Introduction............................................................................................................................................. 450

15.2 Installation............................................................................................................................................... 450
15.2.1 Prepare the Heater..................................................................................................................... 450
15.2.2 Burner Pre-installation Work .................................................................................................... 450
15.2.3 Burner Mounting....................................................................................................................... 451
15.2.4 Tile Installation ......................................................................................................................... 453
15.2.5 Connecting the Burner to the Heater ........................................................................................ 455
15.2.6 Burner Installation Inspection................................................................................................... 456
15.2.7 Air Control ................................................................................................................................ 456
15.2.8 Fuel Piping Design.................................................................................................................... 458
15.3 Maintenance ............................................................................................................................................ 459
15.3.1 Gas Tip and Orifice Cleaning.................................................................................................... 462
15.3.2 Oil Tip and Atomizer Cleaning ................................................................................................. 463
15.3.3 Tile ............................................................................................................................................ 463
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

15.3.4 Flame Stabilizer ........................................................................................................................ 463

15.3.5 Air Registers and Dampers ....................................................................................................... 464
15.3.6 Pilot Burners ............................................................................................................................. 465
References................................................................................................................................................................. 465

449
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 450 Tuesday, March 22, 2005 3:31 PM

450 The John Zink Combustion Handbook

15.1 INTRODUCTION characteristics and make it difficult to operate burner compo-

The correct installation of burners into a heater is essential nents such as air registers.
for good performance. During installation — whether in the The burner steel should be level so that tolerances can be
initial heater construction or at turnaround — the critical maintained when the burner is installed. If the heater steel is
dimensions and orientations of the burners should be more than 0.25 in. (6 mm) out of level, a “donut ring” or
observed. Adequate maintenance will allow the burners to adapter plate should be installed to provide a level burner
deliver the design performance for many years. mount. The adapter plate is a steel ring installed in the open-
ing of the burner and welded to the heater steel to provide a
level mounting surface for the burner (see Figure 15.4).

15.2 INSTALLATION
15.2.2 Burner Pre-installation Work
Improperly installed burners will not operate efficiently and
may damage the heater during service. The burners must be 15.2.2.1 Bill of Materials
installed in accordance with the burner manufacturer’s rec- The burner should be unpacked and inspected to ensure that
ommended procedure. all parts are in accordance with the bill of materials included
with the burner. Missing parts or parts that appear to be
incorrect or damaged should be immediately reported to the
15.2.1 Prepare the Heater burner manufacturer for correction. The bill of materials will
15.2.1.1 Safety list the main burner parts, such as the burner assembly and
A safety checklist for the work to be performed should be pre- burner tile, and other miscellaneous parts. Confirm that any
pared and reviewed by all personnel working on the heater. preassembled parts are as shown on the burner drawing.
This will vary, depending on the type of project and the
requirements of the installing and purchasing companies. A 15.2.2.2 Burner Drawing
common requirement is unimpeded access for the final burner The burner manufacturer provides a drawing of each type of
placement. A permit detailing the immediate safety require- burner supplied to each customer. Do not try to install a
ments, will normally be issued by the user company. burner without the appropriate drawings and instructions.
The burner drawing will show the outside tile dimensions;
15.2.1.2 New Heaters burner orientation; and positions of the gas tips, the pilot tip,
Installing new burners on new heaters is usually easier than the oil tip (if provided), and the burner tile(s). The burner
retrofitting burners to an existing heater. The installer com- mounting bolt pattern will also be shown. All of this informa-
pares the heater manufacturer’s and burner manufacturer’s tion is required to ensure proper installation.
drawings for the required cutout in the heater steel and the
burner mounting bolt pattern, including bolt circle and size 15.2.2.3 Installation and Operating Instructions
(see Figure 15.1). These are compared to the field measure- With each burner project, the manufacturer typically sends a
ments of the same dimensions. If differences are discov- copy of installation and operating instructions for the burner
ered, they must be resolved before proceeding with the and any auxiliary equipment provided. These instructions
installation work. will provide the following basic information:
The heater steel is checked for interferences with the burner 1. safety summary
steel and for flatness. Any necessary work should be com- 2. design specifications
pleted before attempting burner installation. 3. reference drawings
4. receiving and handling of the equipment
15.2.1.3 Existing Heaters 5. installation of the burner and auxiliary equipment
Installing burners on existing heaters is usually more difficult 6. operation of the burner and auxiliary equipment
because some refractory must often be removed, the new 7. troubleshooting recommendations
8. maintenance instructions
burners may be of different sizes than the old burners, and/or
9. recommended spare parts list
the heater steel will often be warped after years of service
10. service available
(see Figure 15.2). The warping may make it difficult to main-
tain the installation tolerances required for proper burner per- The burner installer should be familiar with the installation
formance. Figure 15.3 depicts a burner improperly installed and operating manual as it provides the information necessary
at an angle. Such an installation can cause undesirable flame for a satisfactory installation. If old burners are being rein-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Installation and Maintenance 451

FIGURE 15.1 Heater cutout and burner bolt circle on FIGURE 15.3 Burner improperly installed at an angle
a new heater. due to a warped shell.

FIGURE 15.4 Donut ring for leveling burner mounting

FIGURE 15.2 Warped steel on the shell of a heater. onto the warped shell of a heater.

stalled on a heater, a new set of installation instructions should

be requested from the burner manufacturer.

15.2.3 Burner Mounting

15.2.3.1 Heater
Burners are mounted on the heater steel in a variety of ways.
The design of the heater, the heater manufacturer’s practices,
and the process to which the heater is applied all impact how
the burners are mounted. Some mounting options include:

• on the floor steel, firing vertically upward (see Figure 15.6)

• on the wall steel, firing horizontally (see Figure 15.7)
• on the radiant-zone roof steel, firing vertically down-
ward (see Figure 15.8)
• on a common air or noise reduction plenum, which is in
turn attached to the heater steel (see Figure 15.9)
• on an individual noise reduction plenum, which is
attached to the heater steel (see Figure 15.5)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- FIGURE 15.5 Individual noise reduction plenum.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 452 Tuesday, March 22, 2005 3:31 PM

452 The John Zink Combustion Handbook

FIGURE 15.6 Burner mounted on the floor of a heater.

FIGURE 15.8 Burner mounted on the top of a heater.

FIGURE 15.7 Burners mounted on the side of a heater.

15.2.3.2 Burner
The burner is commonly attached to the heater steel in one of
three ways:

1. The air register mounting flange can be bolted directly to

the steel casing. The burner air register supports the burner
parts in the throat of the burner (see Figure 15.6).
2. The burner front plate is bolted to the air plenum, which
is attached to the steel casing. Figure 15.9 shows a burner
mounted in a plenum in common with other burners. The FIGURE 15.9 Burner mounted in a common plenum.
plenum may be provided for noise abatement or for dis-
tribution of pressurized and/or preheated air from a forced
draft system. sidering the process and heater design characteristics, and
3. The burner with integral individual plenum box for noise local regulations. Complete a dimensional check of the burner
reduction is bolted to the heater steel (see Figure 15.10).
mounting plate and the heater steel mounting location to iden-
The type of mounting is a function of the burner model tify any problems. If the burner is plenum mounted, the air
selected jointly by the heater and burner manufacturers, con- register and plenum depths should be checked prior to
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Installation and Maintenance 453

attempting assembly to identify any possible problems. This

is particularly important if burners are provided for more than
one heater. Ensure that the proper burners are allocated to
each heater.

15.2.3.3 Burner Piping

The fuel and atomizing medium piping to the burners should
have zero loading on the burner connection; the burners are
not designed to carry piping loads. Piping loads can cause
burner misalignment (see Figure 15.11) and failure to main-
tain tolerances or position on the gas and oil tips within the
burner. Unanticipated loads can also cause air registers to
bind and be difficult or impossible to operate. All piping
should be supported independently of the burner.
The piping should be installed with consideration for the
need to remove the complete burner or individual parts with
minimum piping rework. Flanges or unions should be located
so that the piping can be easily removed when necessary. If
the piping is to be insulated, it must be installed with the
necessary clearances to allow the covering to be installed. FIGURE 15.10 Burner in an individual plenum box
The piping should not interfere in any way with the operator’s mounted to a heater.
access to the burner viewports or operating functions (e.g.,
air register adjustment). Inspection should confirm all of these
features.
The burner piping connections may be either threaded or
welded, as the purchasing company’s standards require. There
should be enough flexibility in the piping to easily move the
burner gas and oil tips into and out of the burner (for proper
positioning) when the piping is connected. Sometimes, flex-
ible metal hoses are used for the burner fuel connections to

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
make the positioning easier. These flexible connectors should
be made of steel to meet the required temperature and pres-
sure ratings and provide reasonable durability. The end con-
nections of flexible steel hoses are points of weakness unless
designed and fabricated by reliable suppliers. The hoses
should have a braided armor covering to resist impact and
rough handling. Care should be taken in handling and instal- FIGURE 15.11 Piping improperly loaded on the burner
lation to avoid sharp bends or “kinks” that can be the cause inlet.
of catastrophic hose failure.

15.2.4 Tile Installation 15.2.4.1 Heater Refractory

The burner tile is key to forming the correct flame pattern The refractory on the inside surface of the heater casing is
within the firebox. It also forms the orifice that controls the provided to protect the steel casing and structure from the
flow of air to the combustion reaction. If not installed in the internal temperatures and to reduce heat loss from the
heater as designed by the burner manufacturer and shown on heater. The refractory materials and thickness are com-
the drawings, the flame patterns and heat flux distribution monly selected to limit the casing steel outside temperature
within the heater are likely to be adversely affected. Depend- to a specified value. API Standard 560 provides guidance
ing on the model of the burner, the burner tile can be desig- as to the selection of the casing design temperature and
nated as primary, secondary, and in some cases, tertiary tile. refractory materials.1 The heater designer determines the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 454 Tuesday, March 22, 2005 3:31 PM

454 The John Zink Combustion Handbook

FIGURE 15.12 Cross section of a burner tile.

heater refractory thickness. The burner secondary tile projec- ceramic fiber refractory in the expansion gap between the tile
tion from the casing is at least as great as the heater refrac- and the refractory. The heater refractory should be installed
tory thickness. Some burner tiles may extend an additional prior to burner tile placement.
1.5 in. beyond the heater refractory. An alternative secondary tile installation method involves
the burner manufacturer placing the tile into a (usually metal-
15.2.4.2 Secondary Tile lic) tile case. In that case, the burner manufacturer is respon-
The secondary tile is normally the main burner tile attached sible for maintaining the burner tile dimensions and
to the heater floor, walls, or roof, depending on burner orien- tolerances. The installation contractor attaches the tile case
tation. If the burner is floor-mounted, the secondary tile rests to the heater and installs the ceramic fiber insulation in the
on the steel floor around the burner opening. If wall- expansion gap around the tile case.
mounted, the burner is supported by both the heater steel and
the refractory wall of the heater. The secondary tile of a roof- 15.2.4.2.2 Tile Tolerances
mounted burner is specifically designed to be hung from the Consult the manufacturer’s burner drawing for the burner tile
roof steel of the heater firebox. dimensional tolerances; the tolerances for the secondary tile
The burner tile is commonly designed for a maximum are typically ± 0.5 in. (± 1.3 cm). Each piece of a tile is nor-
service temperature of 2400 to 3000°F (1425 to 1650°C) and mally sealed to its neighbors with relatively thin mortar com-
fabricated in one or more pieces, depending on the size and pletely covering and sealing the joined sides. If too much
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

design of the burner. Figure 15.12 reveals a cross-sectional mortar is applied, the tile tolerances cannot be maintained to
view of a round burner tile with a number of tile sections. specification. After all tile sections are installed and before
The tile is normally supplied by the burner manufacturer. the mortar sets, the burner dimensions are checked. On a
Because the secondary tile forms the air orifice to control round flame burner, the tile outside diameter is measured in
the airflow to the combustion reaction, its dimensions are four directions to ensure that the dimensions are within the
critical. The installation must not alter the outside dimensions tolerances (see Figure 15.13). If the tile is rectangular, check
shown on the burner manufacturer’s drawing. the dimensions at two locations on each of the long and short
sides (see Figure 15.14). After all other interior work on the
15.2.4.2.1 Tile Installation heater is complete, recheck the tile condition and dimensions
Most commonly, the burner manufacturer will specify the and confirm that no loose material is blocking the throat.
outside diameter of the burner tile to the heater manufacturer All dimensions of the tile are on the burner drawing, and
and installer. The manufacturer or heater installer will form the tile installation must be in accordance and within the range
the refractory covering the surface to which the burner of tolerances. Table 15.1 shows the difference in area of a
mounts, leaving an opening for the burner that typically burner that has different tile dimensions. The flow rate and
includes an expansion gap of 0.5 in. (1.3 cm) around the distribution of air through the burner are a function of the tile
periphery. The burner installer will then install the secondary open area and shape. Larger open areas allow a higher flow
tile on the centerlines of the burner opening and place rate of air than smaller areas for the same pressure drop or
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 455 Tuesday, March 22, 2005 3:31 PM

Installation and Maintenance 455

draft loss. Improper tile areas result in maldistribution of air

flow between multiple burners.

15.2.4.2.3 Tile Height

The height of the tile installed in the burner opening is shown
on the manufacturer’s drawing. The tile height is related to the
thickness of the adjacent heater refractory. Some burner mod-
els have the tile height the same as the refractory thickness.
For other models, the tile height is 1 to 1.5 in. (3 to 4 cm)
greater than the adjacent refractory thickness. The installer
and the inspector must confirm that the tile height and the
adjacent heater refractory thickness are in accordance with
the heater drawings. If the burner tile height is not correct,
the burner may exhibit flame instability, resulting in poten-
tially unsafe operating conditions.
It is not good practice to install a burner without a tile, using FIGURE 15.13 Sketch showing a round tile measured
only the heater refractory to form the burner opening. Burner in three different diameters.
tile is much more durable than common heater refractories

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
and less likely to be damaged in service. If the opening is
damaged or worn, the air distribution and flame quality suffer.

15.2.4.2.4 Expansion Joints

The burner tile is surrounded by the heater refractory. When
the burner is fired, the flame burns on or near the tile, and the
tile temperature becomes greater than that of the adjacent
refractory. The burner tile expands into the ceramic fiber-
packed expansion gap mentioned previously. If the expansion
gap is not provided, the burner tile may be crushed by the
adjacent refractory. The tile may crumble in spots and the
dimensions of the air flow passage may change, thus
adversely affecting the burner performance and potentially
causing a poor flame pattern and unsafe conditions. A burner
tile surrounded by ceramic fiber heater refractory does not
require an expansion gap.

15.2.4.2.5 Oil Burner Tile (or Primary Tile)

Fuel oil tiles are an integral part of some combination burn-
ers, so named because they can fire a combination of fuels, FIGURE 15.14 Sketch showing a rectangular tile mea-
liquid or gas. This tile is located in the center of the burner sured at different lengths and widths.
secondary tile and is normally some distance below the top of
the secondary tile. These tiles are often set in or poured in a
steel tile case that is mounted to the burner front plate. On new heaters, bolts are mounted on flat steel surfaces
and the connection of the burner to the heater is easily within
15.2.5 Connecting the Burner to the Heater tolerances. On heaters that have seen service, the heater steel
is often not flat and the bolts and bolt pattern may not fit well
The burner is commonly bolted to the heater casing steel.
to the new burner. The diameter of the selected bolt holes
The number and size of bolts and the bolt pattern are estab-
may have to be enlarged in the field. The modifications must
lished by the burner manufacturer. The bolts support and
not compromise the dimensions and tolerance of the finished
properly locate the burner on the heater steel and are instru-
installation or poor burner performance may result.
mental in holding the proper tolerances of burner installation
in the center of the tile opening.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 456 Tuesday, March 22, 2005 3:31 PM

456 The John Zink Combustion Handbook

TABLE 15.1 Burner Throat Area for Different Tile 15.2.6 Burner Installation Inspection
Dimensions Once the burners are bolted onto the heater or plenum and the
Diameter Area in Diameter in Area in Square surrounding refractory is placed, everything must be
in Inches Square Inches Centimeters Centimeters
inspected to confirm that all dimensions and orientations are
10 78.54 25.40 506.71 correct. In addition to checking the tile(s) as previously dis-
12 113.10 30.48 729.66
14 153.94 35.56 993.15 cussed, the positions and orientations of the gas tips, oil tip,
16 201.06 40.64 1297.17 and pilot tip must be checked. Any protective tape must be
18 254.47 45.72 1641.73 removed from the tips, and it is necessary to confirm that the
20 314.16 50.80 2026.83
24 452.39 60.96 2918.63
ports are clear of foreign material.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

15.2.6.1 Gas Tips

The position of the gas tips in relation to the flame holder must
be correct in order to maintain stable burner operation. The
flame holder may be a conical diffuser, a swirler, or a ledge in
the secondary tile. The jets of fuel from the gas tip ports must
be directed correctly, so the orientation of the tip ports is also
important. The installer must position and orient the gas tips
in accordance with the burner manufacturer’s drawings.

15.2.6.2 Oil Tips

Figure 15.15 shows an oil tip located within an oil tile.
Because there are different designs of oil tiles and tips used
FIGURE 15.15 Oil tip in combination burner showing in burners, the drawings must be reviewed to determine the
oil tip locations. types used and the correct installation parameters. The
dimension shown on the oil gun drawing from the burner’s
front plate to the centerline of the steam and oil connection is
an estimate and must be checked after installation of the gun.
The tip location with respect to tile features is also shown on
the drawings. Another way to check the tip position within
the tile is to insert welding rods into the oil tip exit ports (see
Figure 15.16). These rods indicate the oil spray pattern that
will be discharged from the exit ports. The rods should be
about 1 to 1.5 in. (3 to 4 cm) from the edge of the oil tile
when the oil gun is correctly positioned. This is a technique
sometimes used by the operator to verify correct positioning.

15.2.6.3 Pilot Tip

The pilot tip is typically located close to the main gas tips so
that the pilot flame will contact and ignite the gas discharged
from the main tips. The burner drawing will show the correct
FIGURE 15.16 Welding rod in an oil tip. pilot tip location. In Figure 15.17, the pilot is not properly
located and should be repositioned correctly per the drawing
in Figure 15.18. The pilot tip is usually at the plane of the dif-
On forced-draft, air-preheated fired heaters, the connection fuser flame holder. On spider gas tips, the pilot may be
between the burner and the heater steel should be installed located 1 to 1.5 in. below the plane of the spider.
with a gasket to ensure that there is no leakage of hot air from
the air plenum. Failure to seal the air may cause a safety
15.2.7 Air Control
hazard to operating personnel.
The maximum air flow across the burner is primarily con-
trolled by the pressure drop across the burner tile throat. The
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 457 Tuesday, March 22, 2005 3:31 PM

Installation and Maintenance 457

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
total air flow is adjusted with the burner air registers or
dampers, the heater stack damper(s), or with dampers or
speed control on fans. Air flow is usually adjusted to meet a
target level of excess oxygen in the firebox.

15.2.7.1 Burner Air Registers

The common air register consists of two steel plates, each
rolled and welded into a cylinder, with similar openings in
each cylinder’s sides. One cylinder is mounted so that it is
stationary; the other cylinder is free to rotate (see Figure
15.19). When the free cylinder is rotated such that the open-
ings match or are aligned, the maximum amount of air can
flow through the register. As the free cylinder is rotated, the
area for air flow through the openings is varied, as is the
amount of air flowing to the burner. Air registers should be
exercised after installation and periodically thereafter to ver- FIGURE 15.17 VYD burner gas tip in a diffuser with
ify functionality. If binding is noted, it should be corrected; a pilot.
try lubrication with graphite or removal of foreign material
between the steel cylinders.
Older burners may have cast iron registers that operate as
above. Other registers may consist of vanes whose positions
are fixed or variable, depending on burner design. Changing
the position of the vanes will vary the air flow rate to the
burner.
After all interior work on the heater has been completed,
the registers should be inspected to ensure operability and
that loose debris has been cleared.

15.2.7.2 Burner Air Dampers

On some newer burners, the air registers just described have
been replaced by a damper on each burner’s air inlet box. The
damper can be a single- or multiple-bladed design to control
the incoming air flow. Seals can be specified and installed on
the blade(s) in an attempt to minimize air flow when the
damper is closed. FIGURE 15.18 VYD sketch showing the diffuser cone
The dampers should be exercised after initial installation and and the pilot tip.
periodically thereafter to verify functionality. Also check that
any loose debris has been cleared from the air passages. The
Exercise and check these dampers for proper function
seals should be inspected after a period in service and adjusted
after installation and after any repair of the damper or
so that they seal properly when the damper is in the closed
actuator. If these dampers fail to operate correctly, the
position.
burners may operate with either far too much air or insuf-
ficient air for complete combustion.
15.2.7.3 Fan Dampers
Fan dampers are used to control fan volumetric capacity and
should be installed near the inlets of forced-draft or induced- 15.2.7.4 Stack Dampers
draft fans. These are usually multiple blade units, with either Stack dampers are installed in heater stacks to control the
parallel or opposed blade operation, that mount in the fan pressure within the heater fireboxes. The dampers may be
inlet box. An alternate design is the inlet vane assembly that single- or multi-blade units, depending on the diameter of the
mounts directly on the fan air inlet. stack. The negative pressure within the firebox at the burner
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 458 Tuesday, March 22, 2005 3:31 PM

458 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 15.19 Example air register.

level(s) defines the driving force to push air through the burn- and viscosity to be more reliably maintained at the levels
ers and supply the air for combustion. These dampers should required for adequate atomization. Lighter fuel oils with less
be exercised for proper function after installation, after any concern for viscosity control may or may not need to be
repair of the damper or actuator, and after any combustion heated (see Figure 15.22).
upset event that could overheat the damper. Refractory has
been known to fall from stacks, even during installation, and 15.2.8.1 Line Sizing
cause the stack damper to not function correctly. Inspect to Good engineering design attempts to equalize the flows in the
ensure that no debris affects the operation of the damper or fuel lines leading from a common manifold to individual
restricts the flow of flue gas from the heater. burners. Both momentum and friction effects on flow distri-
bution from the manifold should be considered. Equalizing
the burner flows promotes uniform flame patterns and even
15.2.8 Fuel Piping Design
heat distribution within the firebox.
The fuel piping from the header to the burner should leave In heated fuel oil systems, it is common to circulate more
the header vertically upward and then run to the burner (see oil than is burned to minimize the lowering of oil temperature
Figure 15.20). This reduces the possibility that liquids in a at the burners due to heat losses from the piping. The pro-
gas fuel system or solids in a liquid fuel system will leave the portion of oil returned unburned to the fuel preparation and
header and enter burners. Liquids in a gas burner can cause heating facilities decreases as the amount of oil burned
fouling and plugging of the tip ports or orifice and can, in increases. This is due primarily to the reduction in the ratio
large amounts, extinguish the flame. Solids in a liquid fuel of insulation surface area to oil volume circulated as the oil
system will plug the passages in oil atomizers and tips. flow and piping size increase. The return flow will vary from
The fuel oil piping for a system handling no. 6 or heavier about three-quarters of the total flow in small systems to one-
fuel should be an insulated circulating system, as shown in third of total flow in very large systems or, for every barrel
Figure 15.21. This type of system allows the oil temperature burned, 0.5 to 3 barrels will be returned, depending on the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 459 Tuesday, March 22, 2005 3:31 PM

Installation and Maintenance 459

FIGURE 15.20 Typical fuel gas piping system.

specific gravity of the oil. If time allows, a heat trans- fuel oil tips and atomizers, flame stabilizer cone or tile ledge,
ferred/heat balance comparison calculation provides the pilot tip, and burner tile. The air registers and dampers should
required flows. The excess or return flow provides heat to require less frequent maintenance. Operators and inspection
ensure against solidification of heavy residual oils, possibly personnel should inspect all of these when developing a
requiring piping disassembly and cleaning or total replace- maintenance worklist.
ment. Lighter oil fuels with less concern for viscosity control Many burner maintenance activities can be completed
and no need to heat the fuel do not require a return system. while the heater continues to operate. Some are simple and
may occur fairly frequently. Among these are removal of oil
guns, gas tips, and pilot burners for cleaning, the removal of
15.3 MAINTENANCE center-fired raw gas guns, and maintenance on the regis-
Burner performance typically deteriorates with operating ters/dampers of burners that are not in plenums. Complete
time due to fouling, plugging, and wear on the components. removal of an inactive burner on a natural-draft heater can be
The fouling, plugging, and wear reduce the effectiveness of done while the heater operates, but safety issues must be
the fuel and air mixing and can affect the flame and heat flux addressed. The heater must be kept operating steadily so that
patterns in ways that reduce the heater efficiency and heating the firebox pressure stays negative. Procedures to minimize
capability. The burner parts that usually require maintenance the exposure of personnel, proper protective clothing, tools
to avoid serious performance loss or safety issues include the to handle the burner (particularly if it is a heavy floor-mounted
fuel gas tips, the fuel gas orifice and mixer (premix designs),
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- unit), and temporary sealing of the burner opening must all
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 460 Tuesday, March 22, 2005 3:31 PM

460 The John Zink Combustion Handbook

FIGURE 15.21 Typical heavy fuel oil piping system.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Installation and Maintenance 461

FIGURE 15.22 Typical light fuel oil piping system.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

462 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
• amine compounds from the fuel gas hydrogen sulfide
removal process
• coking of condensed heavy or unsaturated hydrocarbons
in the fuel gas
• polymers that form inside the burner heater risers or tips
• hydrocarbon mists that vaporize or react in hot risers or
fuel tips

Foreign materials must be carefully removed from the fuel

orifices so as not to alter the orifice dimensions. Replace fuel
tips if any orifices exceed the specified diameter by more than
one to two twist drill sizes. The type of material plugging the
ports determines how the burner part should be cleaned.
If the material is pipe scale or gums, a twist drill of the
same size as the orifice or tip ports is used for cleaning. The
FIGURE 15.23 Typical fuel gas tips.
drill bit is gently twisted manually and pushed through the
plugged port to remove the deposit. If the material is not
be addressed. Depending on burner construction, all parts can easily removed, soak the burner part in a solvent to loosen
be removed and repaired, or all but the secondary and tertiary the deposit, and try again with the twist drill. Never use a
tiles are removable during service. power tool with the twist drill because it is likely to enlarge
the port(s); this will lead to increased fuel gas flow and flame
Burner maintenance for planned shutdowns can usually be
or combustion problems. An alternative that is often success-
identified by problems observed during operation or by con-
ful is to use a welder’s file to gently remove the scale from
ditions visually observed within the firebox. Many problems
the orifices. If the scale cannot be removed, the fuel tip or
with burners develop because of broken or poorly maintained
orifice must be replaced.
parts. See Chapter 17 for discussion of several such problems.
Repairs of damaged tile are a common shutdown item that If the source of the foreign material is an amine compound,
may be identified by observed damage or by flame problems. the orifices can be cleaned with wet steam or hot water,
Other readily identified work items include inoperable regis- because amines are water soluble. Amine plugging is fairly
ters and dampers, damaged tips and risers on combination or common in the tips of premix-type burners (i.e., spiders or
raw gas burners, damaged diffusers (flame holders), and radiant wall tips). If the amine plugging occurs frequently,
whole burners and/or plenums damaged by fire or oil spill. the ports can be cleaned without removing the burner by
Replacement parts should be ordered from the original shutting off the fuel gas valve and injecting steam into the
manufacturer so as to maintain the appropriate quality and burner. The length of the steam injection period depends on
durability. Copied parts normally do not have the same the amount of deposit in the burner manifold, risers, gas tips
tolerances as the original parts and often cause inferior or orifices, and associated piping. Note that injection cleaning
operation. may not clear all tip ports equally.
Coking of hydrocarbons in the fuel gas tips may be so
15.3.1 Gas Tip and Orifice Cleaning severe that the tips must be replaced. Light coking can be
The fuel gas tips (see Figure 15.23) and the fuel gas orifice removed with a twist drill as described earlier.
have drilled ports that direct the flow of fuel into the air Plugging with polymers and as a result of heavy hydrocar-
stream and combustion zone. These ports must be kept free bon vapors and high temperatures in the fuel line and tip can
of foreign material that would reduce the effective port size. sometimes be removed by soaking the tips in a hydrocarbon
If the ports become partially or completely plugged, the solvent, followed by cleaning with a twist drill. Severe fouling
amount and distribution of fuel entering the combustion zone may require replacement of the tip or orifice.
vary from the design intent, and combustion problems will
Coking, polymer, and solid deposits can be removed or
likely occur. Sources of foreign materials that plug tips and
converted to easily removed compounds by oxidization in a
orifices include:
small, high-temperature furnace. Many tips can be cleaned at
• pipe scale and gums from the fuel gas piping one time with this technique.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 463 Tuesday, March 22, 2005 3:31 PM

Installation and Maintenance 463

15.3.2 Oil Tip and Atomizer Cleaning

Fuel oil tips are more difficult to clean than gas tips and often
require more frequent cleaning. Tips in light fuel oil service
may have only a slight carbon-like deposit on the surface.
This can be easily removed with a wire brush. Tips in heavy
fuel oil service will usually have a tenacious hard deposit on
the surface and in the ports. These tips should be removed
from the burner when fouling is observed and placed in a
naphtha or diesel oil bath to soften the deposits. Wire brush-
ing, steam cleaning, or the use of a twist drill as described
earlier, individually or in combination, may be effective in FIGURE 15.24 Example oil gun atomizer.
removing the deposits. Do not use a power drill to clean the
ports or a power tool to clean the oil tip surface. Any nicks or
indentations on the surface of the oil tip act as a site to collect
oil and accelerate the tip coking problem.
Coking and deposits on the surface of the tip can be
removed without taking the oil burner out of service in nat-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

ural-draft and in some forced-draft applications, depending

on the design of the burner. A spoon-like implement on a long
handle can be used to reach through the air register and scrape
off the deposit while oil and atomizing medium continue to
flow. The deposit must be identified early, before the fouling
becomes extreme, for this to be successful.
An oil gun atomizer is shown in Figure 15.24. There are
two orifices in the atomizer: one is for steam and can supply
steam to three or four ports, and the other is for oil supply to
one to four ports. The steam orifice can become plugged with
pipe scale that is easily removed. If heavy oil has entered the FIGURE 15.25 Catalyst deposit within an oil burner
steam orifice, due to a (probably momentary) pressure imbal- tile.
ance on either the steam or oil, the atomizer should be soaked
in a solvent followed by blowing steam through the orifice
to remove remaining oil or solid deposit.
of the tile condition and replacement when the tile is identi-
When removing the atomizer from the oil gun, the atomizer
fied as damaged. Any major cracks, particularly in wall- or
must be handled carefully to protect the labyrinth seal. This
roof-mounted burner tile where the tile pieces are likely to
seal separates the oil from the steam by a series of rings on
fall to the floor, are reason to replace the tile. Multiple cracks
the atomizer with a tight tolerance between the atomizer and
in a section provide evidence of crushing due to restrained
the oil gun body. Sometimes, the atomizer is galled and seizes
expansion and are reason for replacement.
and cannot be removed from the gun without breaking the
atomizer. If this is a continuing problem, the atomizer met- The primary or oil tile should be checked. If it is badly
allurgy should be changed to a different metallurgy than that pitted or cracked, the recirculation of gases within the tile is
of the oil tip. uneven and coking can occur on the tile. This coking can lead
to oil dripping and spillage from the burner. Figure 15.25
depicts a catalyst deposit within an oil tile; the catalyst entered
15.3.3 Tile as a component of the blended fuel oil. If this is observed,
Burner tiles are difficult to repair and return to service suc- the tile must be removed and cleaned.
cessfully because the unique shape of the tile surface is diffi-
cult to reproduce. Also, the refractory of the tile has
undergone phase transformation in service, and refractory 15.3.4 Flame Stabilizer
repairs usually do not adhere to the surface for very long. Several flame stabilizer designs are used to hold a stable flame
Maintenance of burner tiles is generally limited to inspection in the combustion zone in the burner. These designs include
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 464 Tuesday, March 22, 2005 3:31 PM

464 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 15.26 Typical diffuser cone. FIGURE 15.27 Typical spin diffuser.

the diffuser cones (see Figure 15.26), spin diffusers (see

Figure 15.27), tile ledges, and tapered tiles in natural-draft
burner designs. All of these must be inspected, condition and
dimensions compared to the appropriate drawings, and the
device replaced or repaired if it will no longer act as a reliable
stabilizer. A stabilizer in poor condition, or missing, can result
in unsatisfactory flame shape or an unsafe flame than can lift
off the burner and leave the zone where combustion is initi-
ated. Figure 15.28 depicts a damaged stabilizer that should be
replaced. Operation with this stabilizer would result in a lop-
sided flame pattern with part of the flame lifting off the burner.
Replacement of the diffuser cone can usually be done with the
heater in operation.

15.3.5 Air Registers and Dampers

Air registers and dampers are used to vary and control the
amount of air flowing through the burner. If the registers or
dampers are not adjustable, the targeted level of excess oxy-
gen at the burner cannot be maintained over the desired range
FIGURE 15.28 Example of a damaged stabilizer. of operation. This will result in either too much or too little
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch15-Frame Page 465 Tuesday, March 22, 2005 3:31 PM

Installation and Maintenance 465

oxygen, both resulting in inefficient combustion and the latter

resulting in a possibly unsafe or damaging operation.
All registers and dampers should be exercised periodically.
If dampers are inoperable, check the actuator and repair if
necessary. Dampers are often inaccessible during operation
and repairs must wait until the heater is shut down. If an air
register is not operable, it can often be removed while the
heater remains in service. Clean the register of foreign mate-
rial such as refractory pieces wedged in the register, oil that
has spilled, insulation that blocks the air flow, and rust or
sand. If the register is deformed, possibly due to efforts to
operate it, or the operating handle is broken, repair or replace
the part.

15.3.6 Pilot Burners

There are several types of pilot burners used in the process
industry, but the most common is the small heat release pre-
mix burner type. The basic parts include the pilot tip (possi-
bly with a wind-resistant shield), the gas mixer and mixing FIGURE 15.29 Example of a damaged pilot tip.
tube, and the gas orifice. The gas orifice is commonly
1/16 in. (1.6 mm) in diameter and is easily plugged by pipe
scale. This is cleaned by hand with a twist drill as previ- REFERENCES
ously described.
1. API Standard 560, Fired Heaters for General Refinery
If inspection reveals a damaged pilot tip (see Figure 15.29),
Service, American Petroleum Institute, Washington,
the tip should be replaced (see Figure 15.30).
D.C., 1996.
If the pilot is equipped with electronic ignition (see Figure
15.31), the ignition rod should be inspected to ensure that 2. John Zink Burner School Course Notes, October 2000,
the spark arc is properly located for ignition of the pilot gas. copyrighted 2000.
Check for unwanted electrical grounds that will inhibit 3. E.A. Barrington, Fired Process Heaters, Course Notes,
proper arcing. copyright 1999.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

466 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 15.30 New ST-1S pilot tip without an electronic ignitor.

FIGURE 15.31 New ST-1SE pilot with electronic ignitor.

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 16
Burner/Heater Operations
Roger H. Witte and Eugene A. Barrington

TABLE OF CONTENTS
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

16.1 Burner/Heater Operation......................................................................................................................... 470

16.2 Measurements ......................................................................................................................................... 470
16.2.1 Draft .......................................................................................................................................... 470
16.2.2 Excess Air or Excess Oxygen ................................................................................................... 471
16.2.3 Fuel Flow .................................................................................................................................. 473
16.2.4 Fuel Pressure............................................................................................................................. 473
16.2.5 Fuel Temperature ...................................................................................................................... 476
16.2.6 Combustion Air Temperature.................................................................................................... 476
16.2.7 Flue Gas Temperatures.............................................................................................................. 476
16.2.8 Process Tube Temperature ........................................................................................................ 477
16.2.9 Process Fluid Parameters .......................................................................................................... 479
16.3 The Heater and Appurtenances ............................................................................................................... 479
16.3.1 Burner ....................................................................................................................................... 479
16.4 Further Operational Considerations ........................................................................................................ 488
16.4.1 Target Draft Level ..................................................................................................................... 488
16.4.2 Target Excess Air Level ............................................................................................................ 489
16.4.3 Heater Turndown Operation...................................................................................................... 492
16.4.4 Inspection and Observations Inside the Heater......................................................................... 493
16.4.5 Inspection and Observations Outside the Heater ...................................................................... 495
16.4.6 Heater Performance Data.......................................................................................................... 496
16.4.7 Developing Startup and Shutdown Procedures for Fired Heaters ............................................ 497
16.4.8 Developing Emergency Procedures for Fired Heaters.............................................................. 499
References ................................................................................................................................................................ 499

469
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 470 Tuesday, March 15, 2005 7:35 AM

470 The John Zink Combustion Handbook

16.1 BURNER/HEATER OPERATION heater, the burner draft loss is the difference between the
The governing principles for fired process heater and burner pressure in the firebox at the burner elevation and the atmo-
operation are: spheric pressure at that elevation. In a forced-draft heater, the
burner draft loss is the difference between the pressure in the
1. operate safely
windbox or plenum (often positive) and the pressure in the
2. protect the environment
firebox, both at the burner elevation.
3. avoid damage to the fired equipment
4. satisfy the processing heat requirements Most of the process heaters within HPCI operate with a
5. maximize heater efficiency negative pressure in the firebox. Because the firebox is not
To accomplish these goals, procedures are established to completely sealed, if there is any air leakage it will be outside
guide heater startup, continuing operation, efficiency air leaking into the firebox rather than combustion gases
improvement, handling of emergencies, and operation of the leaking out of the firebox. Positive pressure inside the heater
fuel and air supply systems. In each case, the operator must can cause flue gas leakage and damage to the furnace casing
refer to measurements to properly control the combustion and structure. A positive pressure can even create a safety
reaction and monitor the performance of the heater. hazard to operating personnel.
The most important measurements for safely controlling Almost all fired process heaters should operate with a neg-
combustion are draft (or pressure within the casing), excess ative static pressure, or draft, throughout the flue gas path.
air (measured as excess oxygen), fuel flow and pressure, and This draft should be measured at specific points (see Figure
liquid fuel and atomization medium pressure and temperature. 16.1). Most important is to measure and control the draft at
The operator should also be aware of the combustion air the location of highest pressure within the heater; this typi-
temperature. For proper heater operation, the operator must cally occurs at the roof of the radiant section (or firebox).
also monitor the temperature of the hot gases at the exit from The draft is the lowest at this point, and maintaining a slight
the firebox (bridgewall temperature), the temperature of the negative pressure at this point normally ensures a negative
flue gases entering the stack, the visual appearance of the pressure throughout the heater. Another location for draft
flames and tubes, the temperature of the process tubes, the measurement is at the elevation(s) of the burners. This is
appearance and condition of the refractory, the process fluid checked to ensure that all burners have an adequate draft loss
flow rate in each pass, and the process fluid pressure drop available to supply the necessary combustion air flow. The
and outlet temperature for each pass. third important location for draft measurement is at the flue
gas outlet from the convection section, often located below
the stack damper. By combining this measurement with the
16.2 MEASUREMENTS draft value at the roof of the firebox, in many common heater
16.2.1 Draft designs, one can determine the draft loss across the convec-
Draft is defined in API Standard 560 as the negative pressure tion tube bank. This can help identify the occurrence of dam-
of the flue gas measured at any point within the heater.1 Draft age or excessive fouling in the convection section.
can be expressed as inches (in.) of water, millimeters (mm)
Draft (static pressure) can be measured with an inclined
of water, or kiloPascals (kPa). Negative pressure or draft
manometer (see Figure 16.2) or with a dial gauge manometer.
occurs because the hot flue gas within the confined volume of
Draft transmitters can be mounted externally at the firebox
the heater and its appurtenances (e.g., ducts, stacks, air pre-
roof level to provide remote draft indication or recording.
heater, etc.) is less dense than the surrounding atmospheric
Once the draft target at the firebox roof elevation has been
air. All other factors being equal, the hotter the flue gas
properly determined and made known to the operator, this is
and/or the colder the surrounding air, the greater the differ-
the only draft that requires frequent monitoring. The static
ence in densities and the greater the draft or negative pressure
pressures within the firebox and convection section will
within the heater. The difference in densities causes air to
always be less than this value in a properly designed heater.
flow into the heater, through the burners or through other
The draft at the firebox roof is controlled by adjusting the
openings, and the hot flue gases to flow out of the heater.
damper in the stack or, if an induced-draft fan is provided,
Draft loss is the pressure drop of air or flue gas as it flows
by adjusting the fan damper or speed.
through ducts, burners, firebox volume, or air preheaters, and
across tube banks. In burner terminology, the draft loss across Erratic draft readings can be caused by pulsating flames or
the burner is the pressure drop of the combustion air as it by sample lines that leak, are plugged, or contain water from
flows through the throat of the burner tile. In a natural-draft the products of combustion.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Burner/Heater Operations 471

FIGURE 16.1 Typical draft measurement points.

FIGURE 16.3 Excess air indication by oxygen content.

Because excess oxygen is monitored to ensure complete

combustion, it is best to sample the flue gas at a location that
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

is most representative of the combustion at the burners. The

sampling must also be at a location where the flue gas is
representative of the combustion as a whole in a multi-burner
FIGURE 16.2 Inclined manometer. heater (i.e., where the flue gas is well mixed). The location
that best satisfies these criteria and the correct sample point
for controlling the combustion reaction is at the flue gas outlet
16.2.2 Excess Air or Excess Oxygen from the radiant section. The most common point is at the
top of the radiant section (see Figure 16.4).
Excess air is defined in API Standard 560 as the amount of air
above the stoichiometric requirement for complete combus- Because the heater operates under negative pressure, any
tion, expressed as a percentage. The excess oxygen is the openings will allow air to leak into the heater. The air leaking
amount of oxygen in the incoming air not used during com- into the heater that does not pass through the burners cannot
bustion and is related to percentage excess air as shown in participate in the combustion process. The oxygen analysis
Figure 16.3.2 Excess oxygen is easy to measure and is used as used to determine the excess air cannot differentiate between
a proxy for excess air. If there is an excess of oxygen in the air that enters via the burners and air leaking into the heater.
flue gas, good fuel/air mixing at the burner, and a stable The amount of air leaking into the heater is typically low in
flame observed in the firebox, the operator can be reasonably a well-maintained, well-operated radiant section, but convec-
assured that combustion is complete at the point in the heater tion sections usually have many more sources of air infiltra-
where excess oxygen is measured. tion. Therefore, sampling for excess oxygen at the flue gas
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 472 Tuesday, March 15, 2005 7:35 AM

472 The John Zink Combustion Handbook

The preceding paragraphs discuss the determination of the

combustion efficiency, that is, the efficiency of converting the
hydrocarbons into carbon dioxide. The operator can directly
affect the combustion efficiency by managing the excess oxy-
gen. At times, it is desired to calculate the overall heater
efficiency, or the amount of heat transferred into the process
fluid from the amount of heat released by the fuel. For this
purpose, and for some emissions reporting, the oxygen and
excess air in the flue gas are determined by sampling down-
stream of the last heat transfer surface or in the stack before
the flue gases are released to the atmosphere. See API Stan-
dard 560 for the calculation procedure.
Today, oxygen analysis is done with electronic instruments.
Portable analyzers (see Figure 16.5) that will measure
oxygen, carbon monoxide, carbon dioxide, NOx, nitrogen
dioxide, hydrocarbons, and smoke in the flue gas are avail-
able. When purchasing a portable analyzer, it is important to
realize that sample lines will be long, from the sampling point
FIGURE 16.4 Location for measuring excess oxygen. to grade, and a sturdy built-in pump with high head capability
is required to ensure rapid, accurate readings. Portable ana-
lyzers typically have a dessicant chamber at the flue gas inlet
exit from the firebox, rather than at the stack inlet, gives values port. This removes water vapor and protects the analyzer cells,

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
most representative of the combustion process. The combus- increasing cell life and reliability. The instrument therefore
tion process is what the operator can affect by register and provides a “dry” analysis of the flue gas, without the moisture
stack damper adjustments. If the operator depends on an contributed by combustion and incoming air humidity. Por-
excess oxygen reading taken at the stack, that reading may table analyzers are used for “spot” analyses and monitoring,
indicate an excess oxygen level that does not exist at the but are not suitable for continuous flue gas analysis.
burner level. Thus, the adjustments to reduce excess oxygen In situ extractive analyzers are mounted on the heater and
that the operator would typically make, such as throttling air continuously measure oxygen in the flue gas and, with added
registers, could lead to insufficient combustion oxygen at the features, can also measure combustibles. These can be man-
burners as well as unburned hydrocarbons and CO (carbon ufactured to withstand flue gas sampling temperatures up to
monoxide) in the flue gases. 3000°F, although the typical standard construction is limited
The unburned hydrocarbons and CO may lead to a condi- to about 1800°F. The oxygen measurement response time is
tion known as “afterburn.” Afterburn is the term given to a few seconds, a speed that allows the output signal to be
combustion that occurs near or within the convection tube used for automatic control of the combustion air supply. The
bank. If hot, unburned combustibles leave the burner area due sensitivity is such that fuel composition changes can be
to inadequate air flow through the burners, they can burn noticed almost immediately. The modern in situ oxygen ana-
wherever they come in contact with oxygen within the heater. lyzer is so reliable that it typically requires instrument tech-
nician attention and a calibration check about once a month.
Because the amount of air leaking into the heater is typically
The combustibles analyzer feature requires more frequent
large at the convection section, this is where the afterburn is
attention and is not commonly added unless very low excess
most likely to occur.
air operation is planned. The in situ continuous analyzer is
One additional consideration further improves the accuracy mounted so as to sample the flue gas leaving the firebox, thus
of oxygen analysis. It has been found that air that enters providing information on the quality of combustion. The oxy-
through casing openings tends to stay near the firebox walls gen and combustibles contents, in percent, are based on the
as it flows to the exit. If the flue gas is sampled further into actual flue gas composition and therefore the instrument pro-
the flowing stream, the gas is more representative of the vides a “wet” analysis.
combustion at the burners. Therefore, it is recommended to The justification for controlling the excess oxygen in the
sample through a probe that extends typically 18 in. (46 cm) heater is shown in Figure 16.6 for fuel gas and Figure 16.7
or more from the wall into the flue gas. for liquid fuel oil. The information required to determine the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 473 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 473

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 16.5 Oxygen analyzer.

annual savings achievable includes the heat release per burner fully sized so that it reacts quickly without imposing an
or the total heat release for the heater; the operating excess excessive pressure drop on the fuel supply to the burners. A
oxygen; the stack temperature; and the cost of the fuel. poorly sized valve may encourage operation with the valve
Oscillating or erroneous oxygen analyzer readings can be bypass open, thereby reducing the effectiveness of any burner
caused by leaking or plugged sample lines. Liquid water management system.
condensed from the products of combustion of gases entering
the analyzer may result in instrument damage. 16.2.4 Fuel Pressure
The pressure of the fuel, whether gas or liquid, is the major
16.2.3 Fuel Flow energy source used within the burner to effect the required
The fuel flow — or rate of heat release — is one of the most mixing of the fuel and the air. The design burner pressure
important controlled variables in a process heater. Each oper- for gas fuel will typically be 15 to 30 psig (1 to 2 barg) at
ator should be aware of the maximum design heat release or the maximum design heat release with the design fuel com-
the maximum heat release that has been proven by successful position as measured at the burner (see Figure 16.8). For
operation. As operation approaches this maximum, it is liquid fuels, the design burner pressure may be 100 to 150
important to intently monitor critical variables and be alert to psig (6.8 to 10 barg) at design conditions. Higher fuel pres-
the effects of small variations. sures allow a greater range of heat release, known as turn-
The fuel flow control valve most often acts to control the down. Turndown is the ratio of maximum heat release to the
process bulk outlet temperature, modulating the rate of fuel minimum heat release. The atomization medium pressure
input to maintain or change the desired temperature. Com- will depend on the type of oil gun used. Either a constant
mon valve control loops use cascade control techniques to atomization medium pressure of about 70 to 250 psig (5 to
minimize the effect of pressure and composition fluctuations 17 barg) or a pressure controlled to 15 to 30 psi (1 to 2 bar)
in the fuel supply system. The control valve should be care- above the fuel pressure is typical. The pressures of both the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 474 Tuesday, March 15, 2005 7:35 AM

474 The John Zink Combustion Handbook

FIGURE 16.6 Cost of operating with higher excess oxygen levels (natural gas).
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 16.7 Cost of operating with higher excess oxygen levels (No. 6 oil).
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 475 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 475

FIGURE 16.8 Fuel gas pressure measurement.

fuel and any atomization medium should be monitored at a

point downstream of any control valve and close to the
burner block valves. The pressure of the fuel and atomiza-
tion medium must be checked with the individual burner
block valves fully opened. The fuel pressure at pilot burners
will typically be regulated at a constant pressure, usually 5
to 15 psig (0.35 to 1.05 barg), depending on the pilot model
employed.
The burner manufacturer normally provides a curve of the
fuel pressure versus heat release (or fuel flow) for each burner
and for each specified fuel composition. This curve (see Figure
16.9) will indicate the maximum and minimum recommended FIGURE 16.9 Graph of fuel pressure vs. heat release.
fuel pressures. These pressures should not be exceeded unless
adequate burner performance has been proven beyond these
extremes by field observations and testing. determined during a performance test of the burner at the
Upper and lower fuel pressure limits are set at levels that burner manufacturer’s facility. After the burners are installed
prevent the heater from operating outside the range desired in the heater, the limits should be rechecked. Usually, it is
by the customer. The burner’s upper and lower pressure limits only practical to check the low pressure limit, which may
at which instability of the combustion may occur can be
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- anyway be the most important. The pressure limit is used to
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 476 Tuesday, March 15, 2005 7:35 AM

476 The John Zink Combustion Handbook

select the pressure at which low fuel pressure causes the to maintain the oil temperature between the fuel heater and
burner management system to interrupt the fuel supply to the the burners. Damaged or missing insulation should be
burners in order to maintain safe heater operations. repaired or replaced. Figure 16.10 is a typical plot of fuel oil
To check the low fuel pressure limit, install a pressure gage viscosity versus temperature for many types of liquid fuels.
at one burner on the pipe between the burner and its block The viscosity measured at two temperatures can be plotted
valve. With the heater firing steadily under normal conditions and the line extended to the desired viscosity to determine
of draft and heat release, slowly throttle the valve while the required fuel oil temperature.
watching the flame. The pressure at which flame instability
is initially observed is the low pressure limit. Adjust the low 16.2.6 Combustion Air Temperature
fuel pressure alarm 1 to 1.5 psig above the low pressure limit The burner designer sets the tile air flow area in the burner on
and the fuel trip 0.5 to 1.0 psig above the low pressure limit. the basis of a design air density, determined by the air supply
Fuel composition changes can affect the measured fuel pressure and temperature. The higher the air temperature, the
pressure. With gas fuels, if the density and heating value lower the weight (moles) of oxygen in a given volume of air.
decrease, the controls will call for a higher burner pressure If the air temperature is significantly greater than design, the
to pass more fuel gas and provide the required heat release. amount of oxygen flowing through the tile decreases, and the
If the gas fuel density and heating value increase, perhaps desired amount of fuel cannot be burned without experienc-
due to adding propane or butane, the controls will reduce the ing incomplete combustion and the possibility of some oper-
burner pressure. The energy for mixing in the burner will be ating problems. A significantly lower than design air
decreased and the flame appearance will likely change. A temperature can lead to an increase in air flow and excess
good operator will visually check the flame quality after a oxygen, leading to a reduction in efficiency. In either case, an
significant fuel composition or pressure change and make the adjustment of the stack or fan dampers or the burner air regis-
necessary adjustments that maintain adequate flame quality, ters must be made.
good efficiency, and good combustion. Typically, the natural draft air temperature is the ambient
If fuel or atomization medium pressures change substan- air temperature that varies from –30°F to 150°F (–35°C to
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

tially, the controls and the supply source should be checked, 66°C). This temperature range does not normally cause any
along with a visual check of the flames. Atomization pressure problem in the operation of the burners. Today many heaters
will be controlled from the fuel pressure in those oil guns have air preheat systems installed to reduce the amount of
where a pressure difference is maintained. fuel required. The air preheat system will normally require a
forced-draft blower and an induced-draft blower to control
16.2.5 Fuel Temperature the combustion air and flue gas flows and pressures within
Fuel gas temperature can be at any value below the autoigni- the heater. The preheated air temperatures range from 300 to
tion temperature. Typically, the fuel gas temperature is 80 to 850°F (150 to 450°C). Some processes can provide combus-
180°F (25 to 82°C). The temperature is specified for burner tion air temperatures as high as 1200°F (650°C).
orifice and tip sizing, and any change that varies the density
of the gas will affect the pressure and volume of fuel flowing 16.2.7 Flue Gas Temperatures
at a given pressure. See Chapter 4 for correcting the fuel flow The flue gas temperatures of primary interest are the bridge-
for the fuel temperature changes. Normally, any changes will wall temperature (i.e., the temperature of the flue gas leaving
cause corrective action by the fuel control valve and a burner the radiant section and entering the convection section) and
pressure increase or decrease (see Section 16.2.3). the temperature of flue gases leaving the convection section
If low ambient and fuel temperatures occur together with and entering the stack. The bridgewall temperature is indica-
supplementing of higher molecular weight gas components, tive of the radiant section performance and the degree of
expect liquid condensation in the fuel line. Measurements fouling of the radiant section tubes. The bridgewall tempera-
have shown that as much as one-half of the heavy gases ture will rise as fouling occurs on the radiant section tubes or
condense under extreme conditions. If proper knockout facil- as excess oxygen in the firebox increases. The excess oxygen
ities are not included, the liquid can extinguish burners. can be adjusted with the burner registers and dampers, but
The variation of temperature of heavier (low API gravity) removing fouling deposits in the radiant section tubes usually
liquid fuels is used to control the oil viscosity. Most manu- requires heater shutdown. The stack gas temperature is a
facturers require a viscosity of about 200 SSU (43 centi- rough measure of the overall heater efficiency; as it rises, the
stokes) at the atomizer. The fuel lines must be well insulated heater efficiency decreases. A rising stack temperature is
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 477 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 477

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 16.10 Viscosity vs. temperature for a range of hydrocarbons.

most commonly caused by high excess air, afterburning due 16.2.8 Process Tube Temperature
to high levels of combustibles entering the convection sec- The process heater tube materials become weaker and less
tion, or convection tube fouling, which lowers heat transfer able to withstand internal pressure as the temperature of the
from the stack flue gases. tubes increases. To provide safe operation and satisfactory
tube life, avoiding process tube ruptures, the tube-wall or
These temperatures are commonly measured using thermo-
tube-skin temperatures must be limited. The basic reference
couples inserted in fixed thermowells. Great accuracy is not
for defining the allowable tube-skin temperatures is API
important, but observation of trends will give the best infor-
Standard 530.3
mation. This is good because the radiant, convection, and
Tube metal temperatures can be measured in several ways.
conduction heat transfer to and from a thermowell in these
Most general-service process heaters with alloy tubes will
services, combined with the typical arrangements in a heater, have tube-skin thermocouples installed at selected locations.
cause readings that are many degrees lower than the actual These will measure and usually provide a record of temper-
flue gas temperatures. If good accuracy is required for a test atures of the tube external surface at the selected installation
run, for debottlenecking studies or heater efficiency studies, points. The shielded style of tube-skin thermocouple has been
these temperatures should be measured with a suction pyro- found to give the greatest accuracy with acceptable life.
meter, also known as a velocity thermocouple (see Reed,2 Infrared thermography gives a photograph of the heater fire-
p. 46; see also Figure 16.11). box interior where the temperature of the tubes and refractory
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 478 Tuesday, March 15, 2005 7:35 AM

478 The John Zink Combustion Handbook

FIGURE 16.11 Velocity thermocouple.

can be accurately inferred from color or contrast. This tech- filters, the variety of contributions means that, mathemati-
nique requires special equipment, appropriate training, firing cally, the errors become of the same magnitude as the indi-
with gas fuel, and fairly generous sight-port sizing. This is cated temperature in “cooler” fireboxes, and the results are
not practical as a continuous, or even frequent, monitoring unreliable.
technique. It is limited to identifying hot tube areas caused High tube temperatures due to internal fouling can often
by flame impingement, internal process tube fouling, or be visually observed. A reddish or silvery spot on the tube is
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

erratic hot flue gas flow patterns. It can be used for monitoring an indication of localized overheating.
the progress of fouling, confirming the accuracy of tube-skin When hot spots are identified, operation of the heater
thermocouples, and identifying problem burners. should be modified to keep the tube from becoming hotter.
Tube temperatures in high-intensity or high-temperature Also, tube cleanout should be considered. There are three
heaters are generally monitored with infrared pyrometers. generally applicable techniques for avoiding further increases
These hand-held optical instruments are aimed at the selected in the tube temperature. One option is to increase the process
tube and measure the infrared energy entering the instrument. flow in the pass containing the hot tube. The increased flow
Using an assumed tube metal emissivity, the instrument cir- rate increases the convection heat transfer inside the tube and
cuits electronically convert the incident energy to an indicated removes heat from the tube wall more rapidly. This should
metal temperature. The heaters commonly monitored in this be undertaken with knowledge of the impact on the other
way are those in high-temperature pyrolysis (e.g., ethylene passes, where flow is reduced. The temperatures of tubes in
production) and steam-hydrocarbon reforming (e.g., ammonia these passes can be expected to rise and must therefore be
and hydrogen production) services. carefully observed.
The infrared pyrometer can accurately measure, within The second option is to reduce the radiating effectiveness
50°F (28°C), high tube temperatures over the entire visible of the hot gases in the radiant zone by increasing the excess
length of the tube. This is typically done twice a shift, or six air. The radiating capability of the hot gases is inversely pro-
times a day. The infrared pyrometer is not restricted to mon- portional to the concentration of symmetrical molecules, such
itoring point locations as are tube-skin thermocouples, and as oxygen and nitrogen. As the concentration of these
measurements can be easily taken by a trained operator. The increases, the energy radiated from the gases to the radiant
use of the infrared pyrometer is limited to high-temperature tubes decreases. Because hot spots usually occur in the radiant
services because of instrument construction. The infrared zone or firebox, this can be an effective action, particularly if
energy entering the instrument comes not only from the tube, firing gas fuel. The impact of this action on the remainder of
but also from the refractory, adjacent tubes, and hot gases. the heater is an increase in the amount and temperature of the
While most of the hot gas contribution can be removed by flue gases entering the convection section. The convection
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 479 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 479

section temperatures and heat absorption will rise. Duty has in the pass heat absorption, the outlet temperature will rise.
been shifted from the radiant section to the convection section. Conversely, if the flow is increased, the outlet temperature
The second option will be of limited effectiveness if the will drop. If a change in pass outlet temperature occurs with-
heat transfer to the hot spot is largely radiation from a flame. out a corresponding and appropriate change in the indicated
In such a case, the most effective action is to reduce or pass flow, the operator should immediately investigate. An
eliminate the heat input from that flame by throttling or clos- instrument problem is likely, and a flow interruption or reduc-
ing the burner valve. This action should be reported to man- tion, with potential tube overheating, is possible.
agement because it has the effect of possibly shifting the hot
spot to another location. Section 16.3.1.7 discusses the prob-
lems and limitations involved in throttling burner valves. 16.3 THE HEATER AND
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

APPURTENANCES
16.2.9 Process Fluid Parameters The typical heater consists of two major process sections: the
The process fluid flow rate, pressure drop, and outlet temper- radiant section and the convection section. The radiant section,
ature are measured with conventional instruments. The flow also called the firebox, includes the burners and air plenum(s),
rate through each individual pass is monitored in the majority tubes containing the process fluid, supports for the tubes, and
of heaters. In some cases, the overall flow is measured and refractory. The major mode of heat transfer to the tubes is radi-
the flow to individual passes is governed by restriction ori- ation, although some convection heat transfer also occurs. The
fices. In other designs, where the addition of measuring ele- convection section contains the shock (or shield) tubes, con-
ments and control valves would impose an uneconomical vection tubes, refractory, and tube supports and hangars. The
pressure drop, careful distribution manifold design equalizes major mode of heat transfer is convection, although radiation
pass flows as much as possible. also occurs. The radiant and convection sections are enclosed
The flow of the process fluid acts to cool the tubes and by steel casing and structural members that support the refrac-
maintain an acceptable metal temperature. A reduction in tory, anchor the tube supports, and structurally support the
flow will lead to a rise in the tube-skin temperature and, tubes and stack(s). The casing, the structure steel supports,
often, a phase change or chemical degradation in the fluid. refractory, along with header boxes, ducts, burner tile, and
An unexpected formation of vapor in a normally liquid-filled stack(s), are collectively known as the heater setting. Other
tube pass will increase the pressure drop and cause further appurtenances may include an air preheater, a forced- and/or
reduction in flow and cooling capability with further increase induced-draft fan, instruments, and necessary fuel and process
in tube temperature. piping and valves, along with safety features such as smother-
Chemical degradation of hydrocarbons typically generates ing or purge steam connections.
some solids and some vapor. This may devalue a product, as This section describes the components, reviews their
when the overheating discolors a lubricating oil component. proper operation or condition, discusses any visual clues,
More frequently, it will cause formation of a solid layer of and suggests corrective actions to identified problems.
material, referred to as coke, on the tube internal surface. This Burner problems and corrective actions are discussed in
layer impedes the conduction of heat from the tube to the detail in Chapter 17.
fluid, and the tube temperature must increase to maintain the
amount of heat absorbed by the fluid. The pressure drop in 16.3.1 Burner
the pass may increase measurably. Ultimately, the tube metal The burner is the mechanical device that mixes the fuel and
temperature rises to a level where corrective action is needed, air, initiates combustion, shapes the flame, and releases the
usually a cleaning of some type, before the tube fails. heat required by the process. There are many different fuel
Both of the above effects can be largely avoided if the flow gas burners installed in heaters. These burners are all of two
rate of the process fluid is kept near the design rate. The pass basic types: the premix burner and the raw gas or diffusion
flows should be nearly equal and not vary by more than 10% burner, both described in Chapter 11. The common oil burner
when the heater is firing at greater than 75% of the rated heat is a variant of the diffusion burner type, in which the combus-
release. At lower heat releases, the heat flux is unlikely to tion is heterogeneous (fuel and air phases differ) rather than
cause the problems mentioned above. homogeneous (fuel and air phases are the same). The Lo
Thermocouples are often installed in the outlets of each NOx burner can be either a premix or a raw gas (diffusion)
pass. These can be used to check the pass flow controls and burner. Each type of burner may differ in the way it adjusts
indicators. If the flow in a pass is reduced, without a change combustion air flow to the burner and in the flame pattern
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 480 Tuesday, March 15, 2005 7:35 AM

480 The John Zink Combustion Handbook

Mixing energy is significant in both the fuel gas pressure and

the air stream pressure drop. Because of the high energy of
mixing for the fuel and air, a forced-draft burner can release
several times the heat release available from a natural-draft
burner of the same footprint.

16.3.1.1 Air Flow Control

Most burners are equipped with some type of device to con-
trol the amount of combustion air flowing into the burner.
This device is commonly known as an air register. In a simple

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
configuration, it may consist of two metal cylinders with
equal openings cut in each. As in Figure 16.12, one cylinder
rotates and the other cylinder is stationary. As the free cylin-
der is rotated, the area for air to flow into the burner changes,
FIGURE 16.12 Air control device schematic. causing variation in the air flow across the burner. The air
control device can also be a damper located in the burner air
inlet path (see Figure 16.13). Either the register or damper,
whichever is applied, is also used to block air flow to the
burner, for example, when the fuel flow is stopped and the
operator wants to stop the air flow. Good operating practice
calls for closing the air register or damper of any burner that
is out of service. This forces all air entering the heater at the
burner level to contact fuel in the active burners and improves
the mixing energy, thereby improving the heater efficiency or
fuel utilization.
Other types and combinations of registers are also applied.
Some forced-draft burners will have a fixed air register whose
function is to swirl the incoming air to the burner. In this
application, the air flow will be varied by modulation of a fan
damper or the fan speed.

FIGURE 16.13 Picture of air control device. 16.3.1.2 Premix Burner

The combustion air flow on a natural-draft premix burner is
controlled with a primary air door and a secondary air regis-
produced. An operator must know what to expect regarding ter. A 100% premix burner has no secondary air supply and
flame patterns so that he or she can visually identify abnor- uses only a primary air door. The primary air door is located
mal situations and make the appropriate adjustments. on the mixer, or eductor, outside of the heater (see Figure
16.14). This door is adjusted to vary the air flow area so that
Natural-draft burners are designed to provide the required
the primary or premixed air flow is maximized and the flame
combustion air flow across the burner with very low pressure
on the burner tip is stable. The gas tip is in the heater, cen-
drop, usually less than one in. of water column. The primary
tered in the burner tile. The ideal flame will initiate and
source for mixing the fuel and air is the kinetic energy in the
appear to stabilize in a position within 0.5 to 1.5 in. (1.3 to
fuel stream (fuel pressure is measured in psig, kg/nm3, or
3.8 cm) of the tip. If the flame appears to be less than 0.5 in.
kilopascals [kPa]). These burners can also be applied to
(1.3 cm) from the tip or even resting on the tip, the flame
forced-draft or balanced-draft systems with air blowers. The
speed (or ignition velocity) is close to the flowing velocity of
pressure drop in the forced-draft system can achieve 3 to 4 in.
the air/gas mixture leaving the tip. If the flame speed exceeds
of water column.
the flowing velocity, a flashback to the orifice at the entrance
Forced-draft burners utilize the high pressure drop of com- of the mixer will occur, with a flame outside the heater and in
bustion air across the burner from forced-draft blowers. The the operating area. This may damage the burner and restrict
air-side pressure drop may be up to 20 in. of water column. its heat release (see Chapter 17.4 for corrective action). The
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 481 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 481

FIGURE 16.14 Premix burner.

appropriate primary air door adjustment will position the the burner, separated by a restriction orifice. The velocity
flame on the burner tip to the ideal position. If the flame is through the orifice must be maintained above a minimum value
more than 1.5 in. (3.9 cm) from the burner tip, the gas/air or combustion can travel into the premix chamber, potentially
mixture velocity leaving the tip holes is too high, and the causing damage. The operating parameters of a forced-draft
flame is at risk of lifting off the burner and extinguishing (see premix burner are specific to the burner, and the burner man-
Section 17.8 for corrective action). The flame liftoff from the ufacturer should provide complete operating guidance.
gas tip and the flame extinguishing indicate flame instability.
It is evident that the gas/air mixture volume and the flame 16.3.1.3 Diffusion or Raw Gas Burner
speed are balanced in a properly operating burner, and that The conventional natural-draft diffusion or raw gas burner has
this balance is affected by the primary air door adjustment. one air register or damper controlling the air flow across the
The secondary air register is positioned open or closed to burner. Because there is no primary register or air door, air flow
adjust the excess air, as monitored by an oxygen analyzer, control on the diffusion gas burner is with a secondary air regis-
to meet the targeted excess oxygen. If the secondary air ter or damper which is opened or closed to meet the excess oxy-
register is fully closed, any control of excess air occurs via gen target, depending on the oxygen analyzer reading.
the primary air door. Both the primary air door and secondary Raw gas burners are preferred over premix burners for wide
air registers must be in good working order to operate the variations in fuel gas compositions and for high burner turn-
heater efficiently. down requirements in forced/balanced-draft systems. Typi-
A good natural-draft premix burner produces a very compact cally in these systems, the forced-draft fan damper or motor
and short flame as compared to a raw gas or diffusion flame speed control is used to adjust the heater excess oxygen
burner. Natural-draft premix burners operate best with a fuel toward the target. The individual burner registers/dampers are
gas of constant composition or specific gravity. These burners adjusted to balance the air flow equally to all active burners.
can be designed to handle gases of up to 90% (by volume) The goal is to equalize the air flow within 10 to 15% between
hydrogen content. The normal turndown of a premix burner is active burners to ensure equal heat release from each burner
3 or 4:1. and even heat distribution within the firebox and between tube
Forced-draft premix burners also mix the fuel and air before passes. The forced-draft fan dampers and the burner regis-
the mixture enters the combustion chamber. In this case, the ters/dampers must be in good working order and easily oper-
premix chamber and the combustion chamber are both part of able over their full range to operate the heater efficiently.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

482 The John Zink Combustion Handbook

tion air supply to the burner. Hence, the flow measurement of

each fuel being fired is very important to the operation of a
combination burner. When gas and liquid fuels are fired in
combination, it is recommended that the minor fuel flow be
base loaded and that the outlet process temperature controller
control the major fuel flow. The result is that each burner has
the same flow of fuel or combination of fuels. The flame pat-
terns should be the same on all the burners if the fuel flow and
air are the same.
When firing in combination, the resulting flame length is
normally 20% larger than if firing a single fuel. This is a result
of the gas fuel mixing faster with the combustion air than the
liquid fuel. The mixing of gas and air is rapid because both
FIGURE 16.15 Burner ignition ledge. are in the gas phase. The oil flame, with atomized oil droplets
spraying into the air flow, is a heterogeneous process. The
fuel phase (liquid) is not the same as the air phase (gas); the
A good raw gas burner flame will be attached to a diffuser combustion reaction is slower than when fuel gas is burned.
or tile ledge to maintain a stable burning flame (see Figure The liquid fuel must receive heat from the combustion prod-
16.15). The flame will be larger than a premix flame of equal ucts to vaporize components that then burn homogeneously.
heat release because the mixing of the fuel and air is not as The remaining material burns as a liquid on the droplet sur-
rapid as the fuel/air mixing in a premix burner. The tip drilling face and as a solid when all liquids are consumed.
and the tile design will govern the flame shape. Narrow tip
The majority of natural-draft combination burners will have
drilling promotes longer flames with a narrow shape; wider
a separate air register providing air that flows through the
drilling favors bushier and wider flames.
primary or oil tile. This is termed the primary register and is
Because the fuel ports in the tips of raw gas burners are
sized to provide 20 to 30% of the air required by the oil for
small, as compared to premix burner tip ports, they are sus-
complete combustion. The remainder of the air for liquid fuel
ceptible to plugging. Scale from the fuel piping, carryover
combustion is provided through a conventional secondary air
from treating processes, and reactive components in the fuel
can all contribute to plugging. Plugging is a common cause register. The secondary register also passes the air for fuel
of irregular flames, flame liftoff, flame instabilities, high fuel gas combustion. Inlet dampers can be substituted for registers
pressure, and insufficient heat release. Refer to Chapter 17 in individual and common plenum designs.
for corrective actions. The combination burner for low air pressure drop has a tile
The raw gas burner can satisfactorily handle a wide range shape with a tile ledge (see Figure 16.15). The tile ledge is
of fuel gas compositions, including unlimited hydrogen con- used to stabilize and hold the flame on the burner from the
centrations and inert gas components (e.g., water vapor, car- minimum to the maximum firing rate. On some high air
bon dioxide, nitrogen). A raw gas burner may have turndowns pressure drop burners, the flame is stabilized with a spin
of up to 10:1, while premix burners are limited to 3 or 4:1. diffuser or diffuser cone.
Any liquids should be removed from the fuel gas before the
fuel gas reaches the burner. Fuel gases containing mists and The flame envelope of an oil flame is normally the same
significant concentrations of liquid hydrocarbons will likely as a gas flame of similar heat release. The oil flame will be
cause tip fouling or coking problems and high gas pressures. bright yellow and opaque, a good radiator for heat transfer,
The liquid hydrocarbons may cause coking within the gas tip, and will cool relatively rapidly due to the radiation heat loss.
resulting in tip cracking. The challenge with an oil flame is to mix the liquid droplets
with the combustion air and complete combustion before the
16.3.1.4 Combination Burners flame cools to below the ignition temperature. When the
Combination burners are diffusion flame burners designed liquid droplets fall below the ignition temperature, soot and
with the ability to fire either gas or liquid fuels, or a combina- smoke are generated. If smoky flames are observed and excess
tion of the two. When firing gas and liquid fuel in combination, oxygen is adequate, poor atomization or wet atomization
the total amount of fuel must be limited to match the combus- steam may be causing the problem (see Chapter 17.11).
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Burner/Heater Operations 483

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 16.16 Gas tips.

16.3.1.5 Pilot Burners burner manufacturer can obtain a predictable flame pattern
Pilot burners are premix gas burners. Pilots may or may not from the burner. The fuel tip drill patterns can be varied
be installed on a burner, depending on the requirements of the based on testing and experience to obtain short, bushy flames
operator. When pilots are required, they are part of the burner or long, narrow flames depending on the requirements of the
assembly. The pilot may use the same air supply as the main process heater (see Table 16.2 for typical flame patterns
burner, or it may be located outside the main burner air expected).
source to ensure stable operations. The burner tile shape and condition are critical to obtaining
The primary function of the pilot is to provide a small source the desired flame shape. Round tile shapes provide round
of heat input for the ignition of the main burner fuel. The pilot flame patterns; rectangular tile shapes provide flat flame pat-
may be shut off after lighting the main burner, or may continue terns. Missing tile, or poorly maintained tile with holes and
to operate after the main burner is lit. The pilot must have a cracks, can cause a poor flame pattern. Substitution of tiles
stable flame and be located in the correct position near the with ones of a different design can restrict air flow, result in
main burner fuel discharge for ignition of the main fuel. poor fuel/air mixing, or cause flame instability because a tile
ledge or other feature is missing or incorrectly located.
16.3.1.6 Flame Patterns The flame patterns and dimensions and the impact of dif-
The flame pattern from a burner is developed jointly by the ferent tip drillings are determined by the burner manufacturer
heater designer and the burner manufacturer. The burner and can be confirmed by performance testing at the manu-
manufacturer selects the gas or oil tip drilling pattern, the facturer’s test facilities. Today, multi-burner testing is avail-
type of diffuser (if applied), and the tile shape to achieve the able at some burner manufacturers to determine the effect of
desired flame. See Figures 16.16 and 16.17 for different gas the interaction between burners and the flue gas circulation
and oil tip drillings. Figure 16.18 shows how variation in gas within the firebox on the flame patterns. It should be noted
tip drilling pattern can create short and bushy or long and that the flame dimensions observed in a single burner test,
narrow flames. Oil tip drillings can be tailored to produce particularly a diffusion flame burner, will rarely be observed
similar flames. There are many different types and shapes of in a multi-burner firebox. Variation of the fuel and air flow
fuel tips. Each fuel tip is drilled with a given pattern to meter between burners, flame interactions, flue gas circulation cur-
the fuel and to inject the fuel into the combustion air. With rents, and air leakage all act to vary the flame pattern between
the tile shape, diffuser, and the fuel tip drilling pattern, the burners.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 484 Tuesday, March 15, 2005 7:35 AM

484 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 16.17 Oil tips.

FIGURE 16.18 Long, narrow and short, bushy flame.

The flame patterns within the heater should be visually Flame patterns are designed to stay within an envelope that
inspected as often as necessary, and any change in shape or is a safe distance from the process tubes and the refractory.
dimensions of an individual flame should be noted. The Figure 16.19 shows a typical flame envelope, X-Y-Z, inside
flames from all active burners of the same size should be a cabin heater. API 560 cites some standards and recommen-
uniform because they all have the same heat release, the dations as to burner-to-tube and burner-to-refractory. Some
same air flow across the burner, and the same fuel flow at users apply greater clearances between burners and between
each burner. Table 16.1 can be used to estimate flame burners and tubes than the API 560 recommendations. The
dimensions in order to evaluate observed flame patterns to burners installed on existing heaters many years ago may not
determine if the flame patterns are typical and what is comply with current API 560 recommendations. If installation
expected by the burner manufacturer. of burners into an old heater are to comply with API 560, the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 485 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 485

TABLE 16.1 Static Draft Effect

per Foot of Height
Temperature (°F) Inches

300° 0.0044
400° 0.0056
500° 0.0066
600° 0.0074
700° 0.0080
800° 0.0086
900° 0.0090
1000° 0.0094
1100° 0.0097
1200° 0.0100

TABLE 16.2 Typical Flame Dimensions for Different

Burner Types
Natural-Draft Forced-Draft LO NOx
Burners Burners Burners

Height (ft/106 Btu/hr) 0.8–1.2 0.2–1.0 0.8–2.5

Diameter (ft) Tile Size × 1.5 Tile Size × 1.5 Tile Size × 1.5

number of burners installed may need to be evaluated. More

burners at a lower heat release will provide smaller flames.
Any flames that are visually different or unusual, either in
dimension or stability, should be investigated and any prob- FIGURE 16.19 Typical flame envelope with x-y-z axes.
lem corrected. Chapter 17 discusses several visual indications
of flame problems, along with appropriate corrective actions.
In the combustion reaction, there are many different types
of ions formed, depending on the fuel components being
burned. The ions are excited by the high flame temperatures
and radiate at wavelengths that are visible to the human eye.
These visible wavelengths are the definition of the flame
patterns on the burners. In Figure 16.20, sodium ions in the
flame radiate in bright yellows and oranges. The color of the
flame is affected by the flame temperature, the intermediate
ions formed during the combustion reaction, and the amount
of carbon particles in the flame. Regardless of the flame color,
if combustion is completed, the amounts of heat released and
heat available for heat transfer to the process fluid in the tubes
are the same. FIGURE 16.20 Sodium ions in the flame.
Figure 16.21 shows an example of a good flame within the
firebox, while Figure 16.22 provides an example of a very
bad flame pattern. Good flame patterns alone do not protect 16.3.1.7 Burner Block Valves
against localized tube overheating. Flue gas circulation flows All burners are installed with manual block valves on the
within the firebox and uneven firing of burners will affect the individual fuel line (and atomization medium line, if
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
distribution of heat to tubes. provided) to each burner. Good operating practice on most
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 486 Tuesday, March 15, 2005 7:35 AM

486 The John Zink Combustion Handbook

FIGURE 16.21 Example of a good flame within the firebox.

maintain adequate fuel pressure on active burners; to balance

the heat release in the firebox; or to repair or maintain the
burner taken out of service.
If the operator uses individual burner block valves to control
flame patterns at selected burners, an uneven distribution of
heat within the firebox may result. One possible consequence
is overheating of tubes at some locations, causing coking of
the tube contents and potential tube failure. It is also common
for an operator to react to a high tube temperature alarm by
throttling the block valve on the burner closest to the alarming
tube-skin thermocouple. While this stops the alarm and
reduces the temperature at the alarming thermocouple loca-

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
tion, the total fuel input does not change. The fuel pressure
and input on the other active burners increase and tube-wall
temperatures rise in response. The new locations of high tube-
wall temperature likely are not at one of the few thermo-
couples, and tube overheating goes undetected until failure or
FIGURE 16.22 Example of a very bad flame pattern in severe fouling occurs. The operator should adhere to the best
a firebox. practice of leaving the burner valves fully open or fully closed,
and should react to the alarm in other ways to cool the tube.
If block valves on burners are partially closed, the fuel
heaters is to keep these valves fully open or fully closed and pressure on the burners varies, depending on valve positions.
never to throttle the amount of fuel to the burner. The fuel to If the heater controls reduce the fuel flow, one or more burners
the burners should be regulated with the main fuel control may experience a fuel pressure below the stable limit and
valve. The fuel (and atomizing medium) block valve is used may extinguish. The low pressure alarm and trip instruments
to stop fuel flow to a burner taken out of service in order to on the burner fuel manifold will not sense the same low
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 487 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 487

pressure as at the burner(s) and will not protect the heater as to sulfur trioxide, which then combines with water vapor and

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
expected with the burner management control system. condenses, at normal casing temperatures, to form dilute sul-
The burner block valves on high-temperature heaters, with furic acid. This sulfuric acid condensation and potential cor-
1800 to 2200°F fireboxes, such as in pyrolysis and reforming rosion can best be reduced by coating the steel casing internal
services, are often operated differently. These heaters may surfaces with an acid-resistant barrier suitable for the expo-
have many small wall-mounted burners heating vertical tubes sure temperature, by raising the casing temperature with
located between the burner walls and operate with tube metal external insulation, or by using a monolithic castable refrac-
temperatures near the metal strength limit. When a high tube tory lining. Operation to minimize excess oxygen will reduce
temperature is identified, it is common and accepted practice sulfur trioxide formation and acid corrosion.
to reduce the fuel to the burner opposite the hot spot by Sight doors, mounted on the casing, provide visual access
throttling the burner valve. Tube temperatures are monitored to the firebox interior. The operator can monitor the burner
over the full length with a pyrometer; thus, the effect of flames, the temperature and condition of tubes, and the tem-
throttling a burner can be identified quickly and reliably. Also, perature and condition of the refractory. The sight port doors
the burners are numerous and small; throttling a few has little should ideally be sized and located so that all burners and the
impact on the heat distribution. If the heater controls modulate full length of all radiant section tubes can be observed. If the
the fuel such that fuel pressure at a throttled burner drops sight doors are not adequate to fully observe the inside of the
below the stability limit of a single burner, flameout may firebox, additional sight doors should be added to ensure good
occur. The amount of unburned fuel entering the firebox is visual observations inside the heater.
small and ignites and burns because of the elevated firebox
Before opening a sight door, the draft at the door elevation
temperature. Thus, the combination of small heat release,
should be checked to ensure a negative pressure. A rag should
many burners, tube temperature monitoring over the full
be held at the door to confirm the flow of air into the heater.
length of all radiant tubes, and high firebox temperature
Hot gas exiting through a sight door, under positive firebox
(well above the fuel gas autoignition temperature) leads to a
pressure, will blow the rag outward and can injure the unpro-
safe practice and no equipment endangerment.
tected observer. Proper protective gear should be worn when
opening sight doors for inspecting heaters.
16.3.1.8 Casing and Refractory
Modern heaters have a steel plate casing supported by struc- The refractory protects the steel casing from the heat of the
tural steel columns, often called buckstays. The steel, in turn, combustion process and provides insulation to reduce the heat
supports the internal refractory, the tubes, the access plat- loss to the ambient air. The types of refractory used to line the
forms, and in many designs the convection section and stack. casing steel include ceramic fiber, monolithic castable, brick,
Some heaters may have a separate structure for the convec- and plastic (a moldable form of brick). All have service tem-
tion section and stack. perature limits that must be observed in the selection of the
The casing is normally not designed for airtight construc- proper lining for the heater. The refractory manufacturers pro-
tion. There are typically open seams, sight doors, openings vide recommended maximum exposure temperatures for each
for tubes and manifolds, and doors for maintenance access in material; if this temperature is exceeded, the refractory under-
addition to the burner openings. Because the heater operates goes a phase change and weakens. API Standard 560 provides
under a slight negative pressure, air will leak into the heater guidance in refractory design and selection, including service
through all openings. It is desirable, for best operating effi- temperature recommendations.
ciency and proper burner operation, to have all air enter the The refractory is held in place by (usually metallic) refrac-
heater through the burners. Therefore, all doors should be tory anchors of various designs. The anchor design, spacing,
kept tightly closed during operation and other openings and attachment are critical to obtaining a satisfactory refrac-
should be minimized, or sealed, as much as possible. tory installation and service life. Most refractory failures can
Unwanted openings capable of leaking air into the heater ultimately be traced to anchor failure. These failures are due
can be located by smoke testing using smoke generators to anchors overheating, corrosion, or inadequate anchor spac-
(or smoke bombs), usually placed inside the idle firebox. ing (too far apart) to adequately hold the refractory.
Cracks and seams thus identified are most successfully sealed Refractory failure may first be evident by debris on the
with commonly available silicone caulk. One should repair heater floor. With severe refractory failure, hot areas and
openings due to warped or displaced plates. discolored paint on the outside of the casing may be observed.
If the fuel being burned in the heater contains sulfur, the If this occurs, the casing can be cooled with a low-pressure
casing is subject to acid corrosion. The sulfur partially converts steam spray or water flood until repairs can be completed.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 488 Tuesday, March 15, 2005 7:35 AM

488 The John Zink Combustion Handbook

When viewing the firebox refractory, color can be an leaking into a furnace significantly reduces the operating cost
indication of even heat distribution within the firebox. of a plant.
Refractory color can also be a very rough measure of the Remember that the burners require a draft loss or pressure
refractory temperature. Dark streaks on the refractory pro- drop to force combustion air through the tile and registers.
vide evidence of air leaking into the firebox and cooling the The available draft decreases with height — that available at
refractory. the floor of the firebox will be greater than that at the roof.
This is shown in the typical static draft profile (Figure 16.23).
Thus, burners at or near the floor will usually have ample
16.4 FURTHER OPERATIONAL draft available to exceed the required draft loss; but for burn-
CONSIDERATIONS ers located high on the walls or on the roof, there may not
be enough draft. In such a case, the draft target should be set

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
16.4.1 Target Draft Level
to satisfy the highest level burners at the design heat release,
A target draft is established at the point of highest flue gas
plus a small margin for ambient temperature, fuel composi-
pressure within the heater. The target value is selected to min-
tion, and operational changes.
imize air leaking into the heater and to provide adequate dif-
ferential pressure or draft loss at the burner level for The consequence of too much draft in the firebox is exces-
necessary air flow across all burners. The typical fired heater sive air leaking into the heater and a lower heater efficiency.
data sheet defines the draft at the arch, or radiant zone roof, The additional air adds to the flue gases and increases the
of 0.1 in. (2.5 mm) of water. This is not a design limit, it is draft loss across the convection section and stack. Too little
the reference value from which draft losses are calculated and draft can restrict the air flow through burners, sometimes
flue gas passages sized by the heater designer. This is also the enough to cause flames to exceed the designed flame envelope
location at which the operator normally monitors and con- (flame impingement), flame instability, and the formation of
trols the heater draft performance. Each heater should have carbon monoxide. If the firebox pressure becomes positive at
facilities to monitor the draft available at the arch or at the the arch, hot gases will flow out through openings, potentially
location where the draft (or negative pressure) within the damaging the casing, weakening refractory anchors, creating
heater is at a minimum. unsafe conditions, and restricting heat release.
Good efficient operation minimizes air leaking into the The draft is controlled by modulating the flow of flue gas
heater through paths other than the burners. Minimizing the out of the firebox, usually via the stack and stack damper.
arch level draft minimizes the differential pressure between the The mechanism of control is stack damper positioning or
outside air and the flue gases within the heater. This minimizes variation of an induced-draft fan speed (or fan inlet damper
the driving force pushing air through the openings that may position). The dampers are typically single- or multiple par-
allow air to enter the heater. Air leaking into the heater is an allel blade designs. The draft is measured with a liquid
important consideration when setting a draft target. The higher inclined manometer (see Figure 16.2), a dial gauge, or a
the draft, the more air leaks into the heater. The lower the draft, transmitter instrument.
the less air leaks into the heater. The cost for air leaking into The draft target selected will be a function of how the draft
the heater can be seen by the following example: is controlled. If the draft is controlled automatically, using a
draft transmitter and a pneumatic or electric operator attached
16.4.1.1 Example to the damper shaft, a lower draft target (e.g., 0.05 in.
What does air leakage cost at a draft of 0.5 in. (1.3 cm) of [1.25 mm] or less of water column) is practical. If the draft is
water and a stack temperature of 750°F (400°C)? Assume controlled manually using a manometer, a higher draft target
that one 4-in. (10 cm) by 6-in. (15 cm) peep door is left open. (e.g., 0.1 in. [2.3 mm] of water column) may be required. Even
How much air would leak into the heater? Figure 16.16 indi- further reduction in draft target can be achieved if additional
cates that at 0.5 in. (1.3 cm) of water, the airflow is 170,000 information is provided to the automatic controller. Such infor-
ft3/hr-ft2. If the fuel being burned is methane, the LHV is 910 mation may include oxygen trim, setting a minimum value of
Btu/scf. Using the following formula: excess air in the firebox, process inlet flow and temperature
to anticipate changes in firing rate, and even ambient air tem-
Q = wc p ∆t (16.1) perature if large and relatively sudden swings are expected.
Inputs such as these allow the damper to anticipate the move-
and a fuel cost of $3.50/106 Btu, the cost of additional fuel is ments that will be required to hold the target draft and will
approximately $12,000 per year. Hence, reducing the air stabilize the burner operation.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 489 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 489

FIGURE 16.23 Typical draft profile in a natural draft heater.

16.4.2 Target Excess Air Level on all out-of-service burners. Equalize the air flow to all active
It is not possible to recommend a single target excess air level burners by adjusting all registers to the same opening. Check
for all fired heaters. The condition of the heater, the type and the oxygen and carbon monoxide levels of the flue gas and
composition of the fuel fired, the level and variability of pro- ensure that the draft available at the firebox roof is on target.
cess operation, and the ambient conditions all affect the Next, slightly close the air registers equally on the active
achievable target. burners or the common plenum air supply damper. The draft
The target excess air level can be established by following will probably increase above the target because the incoming
a structured procedure. The recommended procedure for air and the flue gas amounts are lowered, reducing the friction
establishing excess air level targets for a natural-draft heater losses in flow through the heater. Measure the excess oxygen
is given below. The procedures for forced- or balanced-draft and carbon monoxide in the flue gas. Observe the flame
heaters are similar. Essentially, the procedure is that which is condition and monitor the other instruments for satisfactory
used to adjust a heater to the maximum possible combustion operation with no approaching limits. Adjust the draft to the
efficiency, or to optimize the heater performance. The impor- target value with the stack damper.
tant instruments required to be in good working order and Continue to close the burner air registers or plenum damp-
proper calibration include the fuel flow meter, the process ers in slight increments while holding the targeted draft level
outlet temperature indicator or recorder, the draft gauge or with stack damper adjustments. Measure the excess oxygen
controller, the flue gas oxygen analyzer, and the carbon mon- and carbon monoxide levels in the flue gases leaving the
oxide analyzer. The analyzers must be located so as to sample firebox after each burner air register or plenum damper
the flue gases leaving the radiant section. adjustment. The minimum practical excess oxygen level for
Begin the procedure with all possible burners in service and the heater at the current firing conditions and with the cur-
firing equally on one fuel, satisfactory flame appearance, cor- rent fuel composition is reached when the draft reading is
rect fuel temperature and pressure, steady process operation, on target and any stack damper adjustment or burner air
and all potential limits (such as tube metal temperature, etc.) register adjustment (or plenum damper) closure causes the
monitored and recorded. Close the air registers (or dampers)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
carbon monoxide in the firebox to exceed 100 ppm. At this
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 490 Tuesday, March 15, 2005 7:35 AM

490 The John Zink Combustion Handbook

point, there is insufficient air available to burn any additional With the target draft and the target excess oxygen estab-
input of the current fuel. Any further increase in fuel will lished, the operator is now ready to make any adjustments
increase the carbon monoxide level and cause the firebox needed to keep the heater operating as efficiently as possible.
to become “flooded,” or “fuel rich.” The minimum excess The flowcharts in Figures 16.24 (natural draft) and 16.25
oxygen level is achieved. (forced draft) will guide the operator through the necessary
If the operator or engineer has many years of experience adjustments on the heater to achieve set targets on a continual
in combustion and there is no carbon monoxide analyzer operating basis. For example, the operator can control the
available, the minimum practical excess oxygen level can be preset targets as follows.
determined by closing the burner air registers (or plenum The target draft has been determined to be –0.05 in. of
dampers) and keeping the draft on target with the stack water from the previous discussion. The target excess oxygen
damper. When the firebox becomes flooded or fuel rich, the has been determined to be 3.0%. In a natural-draft heater, the
experienced engineer or operator will observe no increase and stack damper and the burner air register are adjusted to control
probably a decrease in process outlet temperature when the the oxygen and the draft. In a forced-induced draft system,
fuel flow increases. the forced-draft damper on the inlet to the forced-draft blower
The targeted excess oxygen level for continuous operations and the induced-draft damper on the induced-draft blower are
should be set at 1 to 2% greater than the minimum practical used to adjust the excess oxygen and draft. See Figures 16.24
level determined above for natural-draft heaters, and 0.5 to 1% and 16.25 for the logic diagram for tuning natural-draft and
greater for forced-draft operations. The higher targeted excess a balanced-draft heaters, respectively.
oxygen level allows for variations in the fuel composition, For this example, the operator begins in the “START ”
variation in ambient air conditions, and variations in firing box of Figure 16.24. The pressure (draft) is measured on
rates for the heater. Highly sophisticated heater instrumenta- the heater at –0.14 in. (–3.6 mm) of water. The pressure is
tion and control systems can safely allow lower excess oxygen
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

below the target of –0.05 in. of water, and the logic box
levels than would be suitable for simple automatic or manual indicates “HIGH.” The excess oxygen is measured on the
systems. The target excess oxygen level should provide safe heater at 5%. The chart indicates that if the excess oxygen
and steady heater operations with good flame patterns (no is also above the target of 3%, then the excess oxygen is in
flame impingement), good flame stability, and tube-skin and the “HIGH” box. The flowchart indicates that the corrective
firebox temperatures within the limits set by the operator for action required is to close the stack damper on the heater.
the heater. If heat input requirements or fuel composition When the stack damper is closed, the pressure within the
changes are significant, then the carbon monoxide should be heater goes from –0.14 to –0.05 in. (–3.6 to –1.3 mm) of
monitored. If the carbon monoxide is observed to exceed the water. The logic chart indicates a return to “START.” The
100-ppm limit, then the minimum practical excess oxygen draft (pressure) is now on target, so go to box “ON TAR-
level should be reestablished for the current conditions. Tables GET.” The excess oxygen measured in the field is 3.9%.
16.3 and 16.4 indicate typical excess air volumes that should The excess oxygen is still above the target of 3%. The
be achieved with this procedure. operator goes to the “HIGH” box.The corrective action indi-
cated is to close the air register or damper on the burner
TABLE 16.3 Typical Excess Air Values for and return to “START.”
Gas Burners
Type of Furnace Burner System
The draft is measured again and determined to be
–0.07 in. (–1.8 mm) of water, above the target again. The
Natural draft 10–15%
Forced draft 5–10%
logic chart indicates to check the excess oxygen. The new
excess oxygen reading is 3.2%. The logic chart indicates to
close the damper again and return to “START.” Return to
TABLE 16.4 Typical Excess Air Values for
“START ” and measure the draft again. The new draft read-
Liquid Fuel Firing
ing is 0.5 in. (13 mm) of water. The draft is on target; thus,
Operation Fuel Excess Air
the logic chart indicates to measure the excess oxygen again.
Natural draft Naphtha 10–15% The excess oxygen reads 3.0% and is therefore on target.
Heavy fuel oil 15–20%
Residual fuel oil 15–20% The logic chart indicates “Good Operations.” The tuning of
Forced draft Naphtha 10–15% the natural-draft heater has been completed. How many of
Heavy fuel oil 10–15% dollars were saved? The charts in Figures 16.6 or 16.7 can
Residual fuel oil 10–20%
be used to determine the savings. The forced-draft heater
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 491 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 491

FIGURE 16.24 Logic diagram for tuning natural-draft heater.

diagram shown in Figure 16.25 is used to make adjustments temperature of the tubes. The effect is to increase the amount
on a balanced-draft heater. of flue gas and the bridgewall temperature, increasing the
While the operator should be encouraged to operate to amount and temperature of gas flowing to the convection
the excess oxygen target as determined above, there may be section. Duty is shifted from the radiant to the convection
conditions where it is desirable to operate with excess oxy- section. Similarly, for heaters where waste heat steam gen-
gen well above the target. One is mentioned in Chapter eration occurs in the convection section, trouble with the
16.2.8 in conjunction with mitigating high tube-skin tem- plant steam boilers may make it desirable to generate more
perature. Here, increasing the excess oxygen reduced the steam in the waste heat generation coils. Increasing the
radiating effectiveness of the firebox gases, lowered the excess oxygen (excess air) above the target value will
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

radiation to the radiant tubes, and lowered the heat flux and increase convection section duty and generate more steam.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 492 Tuesday, March 15, 2005 7:35 AM

492 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 16.25 Logic diagram for tuning balanced-draft heater.

16.4.3 Heater Turndown Operation sure well above any low-pressure trip limits. Flames will be
kept stable and within the normal flame shape and envelope.
During normal or design-level heater operation, the operator Burners are turned off in a pattern that maintains an even
strives for even heat distribution in the firebox by equalizing heat release throughout the firebox. The heater designer
the air and fuel to all burners while maintaining good, stable arranges burners so that each radiant tube pass “faces” the
flames. The same goals apply to turndown operation. When same number of burners. It is wise to maintain this practice
operating at reduced heat-release levels, the operator will when reducing the number of active burners. This helps to
remove burners from service in a selected pattern, closing air ensure that each pass receives the same amount of heat and
registers and fuel valves, so as to maintain adequate fuel pres- avoids having to overly bias pass flow rates to equalize outlet
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 493 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 493

temperatures. If the heater configuration allows (usually the convection section should be increasing, depending on
vertical tubes) and if tube metal temperatures are nearing how much nesting of tubes is taking place. The afterburning

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
upper limits, the best burners to deactivate are usually those may severely oxidize the extended surface (fins or studs) on
opposite the hottest outlet tubes in each pass. the convection tubes such that heat transfer capability is lost,
In natural-draft heaters with diffusion flame burners, the or may cause tube overheating leading to fouling or failure.
excess oxygen can be held nearly constant with air register The stack temperature with afterburning in the convection
or damper adjustments from 100% capacity to approximately section should be increasing significantly above the design
75% capacity. Then the excess oxygen level begins to increase temperature. Check the excess oxygen, the carbon monoxide
as the heater (burner) heat release is further reduced. Forced- level, the process outlet temperature, and the fuel flow.
draft burners, with high kinetic energy in the air stream, are The operator should reduce the fuel input when high carbon
usually capable of maintaining the excess air level down to monoxide levels are indicated or afterburning is known to occur
50%, or even lower, of the design heat release. in the convection section until an excess of oxygen is attained.
In natural-draft heaters with premix burners, the excess Then, opening the stack damper or burner air registers (or damp-
oxygen will follow the fuel gas pressure as the heater duty is ers) should further increase the excess oxygen. Now increase
reduced. The excess oxygen will remain the same from 100% the fuel flow rate to satisfy the process heat input requirements.
duty to 25% of duty on a 100% premix burner, and 100% to Do not increase excess air without first reducing the fuel
50% of duty on a 50% premix and 50% secondary air burner. flow to the point where an excess of oxygen is observed.
The mixing of the fuel and air is affected mainly by the
primary air mixer efficiency. 16.4.4.2 Flame Stability
In observing the flame patterns inside the firebox, the flame
stability should be noted. An unstable flame front will be oscil-
16.4.4 Inspection and Observations
lating on the burner tip as the fuel is mixed with the air. The
Inside the Heater flame front should always be very near the ignition ledge, dif-
Much of successful heater operation depends on frequent and fuser, or fuel tip. If the flame front is detached from the igni-
knowledgeable visual checks of the equipment. In most tion ledge, diffuser, or fuel tip, as in Figure 16.26, the flame is
applications, visual surveillance of the heater as often as nec- unstable and may be on the verge of being extinguished. When
essary to ensure safe and optimum operations is considered the flame is extinguished and fuel is injected into the heater
good practice. Visual observations should also be made after and mixes with the air with no flame, an unsafe and potentially
significant load changes, atmospheric disturbances, fuel gas explosive condition exists within the heater.
composition changes, and utility system upsets. A checklist
of what to observe should be developed for each heater. The 16.4.4.3 Process Tubes
visual inspection of a heater should include the inside of the The process tubes should be periodically checked visually for
heater, the outside of the heater, and observation of the per- evidence of localized hot spots, tube displacement, and pro-
formance data of the heater. cess leaks. Tube hot spots may appear as red or silver spots
and indicate temperatures approaching or exceeding the
16.4.4.1 Flame Pattern mechanical limit of the tube material. The immediate cause
A check into the firebox should show even flame patterns and of the hot spot is usually a fouling deposit on the internal tube
good stability on all active burners. The size, color, and shape wall. This deposit may be the result of flame impingement,
of the flames should be the same because all active burners over-firing, uneven distribution of active burners, or concen-
have the same fuel pressure (heat release) and draft (combus- trated heat input from flue gas circulation currents. Hot spots
tion air flow). Any uneven flame patterns, any flames that are need to be continually monitored; if allowed to continue, they
unstable, or any flames impinging on tubes indicate a prob- will ultimately cause tube failure. The process tube maxi-
lem that needs to be corrected. See Chapter 17 for a full dis- mum temperature limits should be known to the operator and
cussion on troubleshooting that problem. monitored to protect against tube overheating and failure.
A hazy appearance in the firebox may be an indication of Tubes that are out of position may bow due to overheating
high carbon monoxide levels and may cause wildly swirling or because of loss of a tube support or guide. Overheating
flames extending to the arch. If the carbon monoxide levels may be caused by over-firing, by concentration of active
are high enough, afterburning may occur in the convection burners in the firebox, or by flame impingement. The bowing
section, possibly resulting in tube support failure and nesting may be accompanied by internal tube fouling. Look for metal
of the tubes in the convection section. The draft loss across parts from a broken support or guide on the heater floor.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 494 Tuesday, March 15, 2005 7:35 AM

494 The John Zink Combustion Handbook

A process tube leak may initially show as a small flame or

wisp of smoke at the tube surface. Flames or smoke may be
traveling through the tube-sheet into the firebox from a header
box. The process tube leak acts as fuel and consumes oxygen.
This may result in long flames from the burners, due to the
reduction in excess oxygen, and afterburning in the convec-
tion section. The oxygen analyzer will often show a steep
change to lower oxygen levels and the measured carbon mon-
oxide levels will increase.
Black smoke from the stack indicates a major process tube
failure or incomplete combustion of the fuel. The latter occurs
most frequently when firing liquid fuels. The operator must
check the flame appearance to determine whether the problem
is a tube failure or poor combustion at one or more of the
burners. Also check whether high molecular weight compo-
nents have been added to the fuel gas system.

16.4.4.4 Refractory and Tube Support Color

The color of the refractory and tube supports can be an indi-
cation of high temperatures in the firebox. Tube supports
should appear uniform in color if the heat distribution is uni-
form within the heater. If some tube supports appear red or
visibly hotter than other tube supports, then there is uneven
heat distribution within the heater. If the refractory is not uni-
form in color, then there may also be uneven heat distribution
within the heater. Uneven firing on active burners or poor
burner distribution, flame impingement, or flue gas circula-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

tion patterns are all suspects and should be checked.

FIGURE 16.26 Unstable flame.
16.4.4.5 Burner Refractory and Diffuser Condition
Broken burner tiles, burner tiles that have deteriorated from
chemical attack, improper installation of the tile, and burner
diffuser conditions should be noted (see Figure 16.27). Tile
or diffuser in poor condition can lead to poor flames and pos-
sible impingement. The diffuser can normally be replaced
while the heater is in service, but tile replacement may
require heater shutdown.

16.4.4.6 Air Leaks

When observing inside the heater, look for air leaks around
sight doors and other areas where air leakage into the heater
is possible. In hot fireboxes, air leaks may emerge as a dark
line or streak on the refractory surface. See Figure 16.28.
The higher the air leakage into the heater, the higher the
operating cost and likelihood of poor flame patterns. Idled
burners should have air registers completely closed. Air
FIGURE 16.27 Broken burner tile. leaks should be sealed when identified.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 495 Tuesday, March 15, 2005 7:35 AM

Burner/Heater Operations 495

16.4.5 Inspection and Observations

Outside the Heater
16.4.5.1 Stack Damper
The stack damper position should be noted to determine if
there is sufficient control of the draft within the firebox. If the
stack damper is a single-blade damper and is over 75% open,
then there is probably very little capacity left to increase the
draft (negative pressure) within the firebox. However, if this
is a multi-blade damper, there is probably sufficient control
remaining to control the targeted draft and get more capacity
out of the heater.
Check the arrow indicating the stack damper position. Does
it match the position shown in the control room? Does the
damper move when the actuator is adjusted, or is it stuck?
Most dampers control adequately over a range of positions

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
between 20 to 80% open. Compare the present position to this
range to get an idea of the amount of draft control available.
If a black plume is observed being emitted from the stack,
look for a tube rupture, a burner with poor atomization, or
insufficient excess air, possibly due to a fuel composition
change. A blue plume from the stack indicates the burning of
sulfur in the fuel. Operation changes will not affect this plume.

16.4.5.2 Burner Block Valves

Check the positions of the individual burner block valves.
Observe those not fully open or closed. Most heaters should
FIGURE 16.28 Dark line or streak on hot refractory
be operated with these valves either fully open if the burner surface indicating air leaks.
is in operation, or fully closed if the burner is out of service.
Some special heater designs allow throttling of individual
valves to control the tube temperatures. However, in most Hot spots on the casing are the result of internal refractory
heaters, the throttling of these valves can result in poor failure, possibly due to positive pressure in the firebox. Check
flames as well as compromised safety systems and tube the draft in the heater, and, if necessary, make damper adjust-
temperature monitoring. ments to regain a positive draft (negative pressure) at the arch.
The hot spot can be cooled with steam or water spray.
16.4.5.3 Pressure Gauges
The pressure gauges should be located on the fuel line to pro- 16.4.5.5 Burner Damper Position
vide a pressure reading to indicate the amount of fuel flowing The burner damper position or air register position should be
to each burner. The fuel pressure should be recorded and at the same opening for all burners in operation. They should
monitored in the control room and, if possible, it should be be equally open on the active burners and closed on all burn-
trended. The local pressure gauges should be in good workers out of service. Check natural draft premix burners for
ing condition to give an accurate pressure reading. backfiring and correct immediately by adjusting the air door.

16.4.5.4 Heater Shell or Casing Condition 16.4.5.6 Burner Condition

The heater shell or casing should be inspected periodically to The general condition of the burner should be observed. A
determine if there are any hot spots developing on the shell or burner in good condition can be operated at very low excess
casing. Hot spots on the heater shell are an indication of oxygen levels if the heater is in good condition and sealed
refractory failure within the firebox. If left uncorrected, the properly. Burners in poor condition will require higher
heater efficiency will decrease because of the higher shell heat excess oxygen levels to operate safely and should be
loss, and casing failure can lead to the heater being shut down. removed and repaired.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch16-Frame Page 496 Tuesday, March 15, 2005 7:35 AM

496 The John Zink Combustion Handbook

16.4.5.7 Piping 16.2.4.4 Stack Temperature

Inspect the fuel and atomizing medium piping and the pro- Check stack temperature. Stack gas temperature low limits
cess piping for leaks. Also inspect piping for stresses due to are intended to protect against condensation and corrosion;
expansion. Check the insulation and tracing condition on all high limits may help protect air preheater and stack materials
fuel piping to ensure that it is adequate. Listen for any gas from failure. This temperature can also be used to trend the
leaks. Gas leaks may also be detected by the odor of the gas. heater to determine efficiency changes and identify convec-
Check the air lines on all pneumatic instruments and control tion section heat transfer problems.
valves to ensure no leaking or loss of functionality.
16.2.4.5 Bridgewall Temperature
16.4.5.8 Controllers Bridgewall or radiant section inlet flue gas temperature limits
may protect against overfiring and overheating of convection
Check the control air and power air lines at all valves for
section tubes and supports.
leaks, and note the pressures and the positions of the actuator
stems. This can provide an indication of the range of opera-
16.2.4.6 Process Inlet and Outlet Temperature
tion available on process flow, draft regulation, fuel input,
and Pressure
and air supply. Check and monitor the process stream inlet and outlet tem-
perature limits to help prevent overfiring, protect process
16.4.5.9 Air Leaks stream quality and avoid unwanted reactions, and identify
Look for open sight port doors, warped explosion doors, idle upstream problems, such as heat exchanger fouling. Check
burners with open registers, and other locations of obvious the process inlet and outlet pressure to maintain the equip-
air leakage into the heater. Air leaking into the heater ment integrity and to help identify internal tube deposits.
increases the operating cost and may affect flame patterns.
16.2.4.7 Draft Targets
16.4.6 Heater Operation Data Draft targets are set to maintain negative pressures within the
heater and avoid safety and structural problems while ensuring
Every heater should have established control points on cer-
conditions that allow adequate air flow through the burners.
tain operating variables that are selected to ensure safe opera-
tion and maximum heater utilization. Operators must ensure
16.2.4.8 Excess Oxygen Targets
that these control limits are not exceeded. Items that may be
Excess oxygen, measured in the flue gas leaving the firebox,
considered limits will include some of the following and
will provide guidance on adjusting burner registers (or damp-
should be monitored frequently.
ers) to reach the excess oxygen target level. These adjust-
ments must also consider the draft target. Trends in excess
16.4.6.1 Process Flow oxygen can indicate changes in fuel composition, tube leaks,
The process flow rate should be monitored and trended. The and deterioration of heater operation.
process flow is an indication of the amount of work the
heater is to perform. If the process flow is at design condi- 16.2.4.9 Fuel Pressure, Temperature, and Flow Rate
tions, then the burners can be expected to be operating at Check the fuel pressure at the heater. Compare the fuel flow
design conditions. If the process flow is above design con- from the burner capacity curve to the fuel flow measured by
ditions, then the burners can be expected to be operating the heater instruments. Fuel flow should be recorded and
above design conditions. trended. When adjusting excess oxygen, the fuel flow should
be observed to identify when the heater is “fuel rich” and the
16.4.6.2 Pass Flow excess oxygen is low. An increase in fuel flow together with
Check each pass flow. Pass minimum flow limits are estab- no increase or even a decrease in process outlet temperature
lished to avoid maldistribution of process flow through the identifies a “fuel rich” condition. Tube leaks may be identi-
tubes in each pass. Flows that are below the limit can cause fied with careful attention to trends of the fuel flow.
coking of interior tube surfaces or overheating and tube failure. The fuel temperature should be monitored. The liquid fuel
temperature at the burner is most critical to good combustion
16.2.4.3 Tube Metal Temperature of the liquid fuel. If the temperatures are outside the design
Check tube metal temperatures. Tube maximum temperature limits, then the temperature should be corrected to that
limits are established to protect against tube overheating designed by the burner manufacturer for the burner. The fuel
and failure. gas temperatures are normally not monitored.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Burner/Heater Operations 497

16.2.4.10 Atomization Medium, Pressure, systems are functional. Confirm adequate pressure in the
and Temperature fuel and atomizing medium supply systems.
The atomization medium pressure and temperature should be c. Adjust the burner registers to confirm operability over
noted to ensure agreement with the liquid fuel gun being the full range. Leave them set at the 100% open posi-
used. If the atomization medium is steam, the steam lines tion(s) for purging the heater.
should be insulated and have adequate steam traps to ensure 2. Check the heater for readiness:
dry steam to the oil gun. The atomization pressure should be a. Visually check that there is no debris from maintenance
as shown on the capacity curve of the equipment. Any pres- or construction left inside the heater. All doors (main-
tenance, header box, sight, explosion) are to be closed
sures outside the norm should be investigated and corrected.
and protective refractory placed where applicable.
b. Adjust the stack damper over the full range to confirm
16.2.411 Fuel Composition
operability and leave it fully open.
Fuel composition (sometimes indicated by specific gravity)
3. Establish flow through the process tubes:
monitoring can be a “leading indicator” for adjustments of
a. Multi-pass heaters with vertical radiant tubes and liquid
excess oxygen in the heater by adjusting burner registers or
feed require a special procedure that involves filling
plenum damper. one pass at a time. This is to ensure that full and stable
flow is reached in each pass during operation. Heater
16.2.4.12 Stack Emissions manufacturer guidelines and instructions should be fol-
Stack emissions, such as NOx and combustibles (including lowed to ensure that each pass is flowing properly.
CO and VOC), are monitored to help ensure compliance with b. During the filling process, check the operation of the
regulations and operating permits. A change may indicate an pass flow indicators, valves, and controllers. If any
improperly adjusted burner, inadequate air flow to the burn- pass does not have the correct flow indicated or if flow
ers, unstable flames, or a change in fuel composition. fluctuates, analyze and correct the problem before
beginning the purge and burner lighting steps. It is
often helpful to place the pass flow control valves on
16.4.7 Developing Startup and Shutdown
manual operation until the flow reaches about 75% of
Procedures for Fired Heaters the normal flow value. This will overcome the reset
Each plant will need to develop startup and operating proce- windup problems that occur with some control sys-
dures for its particular heater and operations. These proce- tems and valves with minimum flow stops.
dures should be followed for startup and shutdown to ensure 4. Purge any accumulated combustibles from the firebox:
a safe and efficient operation. The following procedures are a. Use steam or fan-supplied air to purge the firebox for
only intended as a guide to help develop startup and operat- at least five volume changes. Normally, this takes
ing procedures for a plant. approximately 15 to 20 minutes, or until a steam
The startup and operating procedures are intended to ensure plume appears at the top of the stack. Avoid excessive
steaming; a long exposure to steam and condensate
avoidance of an explosive mixture occurring in the firebox
can damage the refractory.
and to establish stable flames and process flows. Startup pro-
b. Check the operation of the draft gauge during the
cedures will vary with the degree of automation and instru-
purge period to ensure that the draft gauge is operable
mentation on the heater. The procedures should address the
and reading the draft within the heater.
issues listed below.
5. Light the pilot burners:
1. Fuel system and burner preparation: a. First check that all pilot and main burner individual
a. Confirm that the burners are properly installed (Chap- block valves are closed.
ter 15), with the tips properly positioned and oriented b. Reduce the purge flow 75% and sample the firebox in
and no blockage in the fuel ports or the air flow pas- several locations with an explosimeter. If safe, remove
sages. Isolation valves and line blinds in the fuel sys- the pilot gas line blinds, and confirm that the pilot gas
tems must be closed, including those at each burner. pressure regulator is set accurately and in accordance
b. The fuel lines and valves may be pressure-tested to iden- with the required pressure for the pilot to work properly.
tify leaks or may be blown out to remove scale and debris. Close the burner air register or damper to a position of
Fuel gas lines should then be confirmed as full of hydro- 15 to 25% open. Close the stack damper to 25% open.
carbon fuel, not noncombustible gas. The fuel gas knock- Light the pilots individually; if the first pilot does not
out vessel should be drained of all liquids. The circulation light, briefly repurge and try to light another. Light all
in the fuel oil system must be established and confirma- pilots and keep them burning steadily for 15 minutes to
tion obtained that the heat tracing and atomizing medium confirm their flame stability. Relight any that extinguish.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

498 The John Zink Combustion Handbook

6. Light the main burners: ceramic fiber has been installed since previous opera-
a. Visually recheck that all main burner individual fuel tion, the manufacturer or installer of the material will
valves are closed. Remove any blinds in the main fuel recommend a firing procedure to carefully dry and
supply line. Check the fuel pressure controller setting cure the material to achieve the optimum properties.
for proper operation of the temperature control valve If only Thermbond refractory patches are installed, the
(TCV). Place the TCV in the “manual” mode on initial warm-up rate should not exceed 500°F per hour.
startup because operations will be below the auto- 7. Periodic checks:
matic controller range. Set the TCV to 15% open,
open the main fuel supply block valves, and open one a. Periodically during the startup, check the fuel and pass
individual burner block valve on a burner with a stable flows, the individual pass outlet temperatures, fuel
lighted pilot. If lightoff does not occur within 5 to 9 pressure, draft at the highest pressure point in the fire-
seconds, close the block valve and repurge the heater box, flue gas oxygen content (to ensure correct instru-
(see step 4). If lightoff does not occur within 5 to ment operation and good combustion) at the radiant
9 seconds, check the positioning of the pilot with section outlet, and firebox and tube-skin temperatures.
respect to the main burner tip. b. Draft should be monitored frequently by the operator
b. After the first burner is confirmed to be burning with when heat input is increasing. As the heat input is
a good stable flame, discontinue the firebox purge increased, the air registers (or burner dampers) on a
flow. natural-draft heater and the stack damper will need to
c. Go to another burner with a stable operating pilot and be opened.
repeat the lightoff procedure. A repurge is not neces- c. Place the pass flow controllers on automatic operation
sary if the first burner is lit and this one fails to light. at about 75% of normal flow (see step 3). Watch the
Wait 5 minutes and try a different burner. The heat pass outlet temperatures closely. A high or uneven pass
from the first burner will cause enough draft to safely outlet temperature can be due to a low pass flow, flame
dilute and remove the fuel from the failed burner. impingement, or uneven firing of the burners. This
Continue lighting main burners that have stable pilot must be analyzed and corrected quickly to avoid dam-
flames, manually opening the temperature control age or curtailment of the run length due to internal
valve after each to maintain adequate fuel pressure tube fouling. The pass flow indicators may not be
(minimum firing pressure shown on the burner capac- reliable; the outlet temperature may be the best infor-
ity curve) at the active burners. Failure to do this will mation. If flow stoppage or reduction is suspected or
cause flame instability and likely failure, requiring a a major pass outlet temperature discrepancy cannot be
repurge of the heater if all the pilots and burners are corrected, quickly extinguish the main burners until
extinguished. Light burners in a pattern that distributes the problem is resolved. The pilots can usually be kept
active burners evenly throughout the firebox. lit under these circumstances, thereby avoiding having
d. If operating in the natural-draft mode, increase the fuel to purge the heater again.
rate slowly to warm the heater and establish the draft
that allows more air to enter the heater. Failure to d. Be aware that without process flow, the tube-wall
establish adequate draft can cause a fuel-rich, poten- temperature will rise to within about 200°F (110°C)
tially explosive mixture to develop in the firebox. of the firebox operating temperature, after the main
Check the draft at the top of the firebox radiant section burners are extinguished. Opening the stack damper(s)
to ensure that air is flowing through the burners. fully will increase the flow of cooling air. If the process
e. Visually do a frequent check for stable flames on both flow is interrupted, do not reintroduce fluid into the
main and pilot burners. tubes until they have cooled to below 900°F (485°C).
f. When the process outlet temperature warms to within e. Shutdown of the heater is far simpler. Gradually reduce
the range of the temperature controller, place the fuel the heat input, taking burners out of service in order
control valve on automatic operation. to hold adequate fuel pressure on the active burners.
g. Typical allowable warm-up rates for heaters vary from Do not close the registers on burners taken out of
100 to 200°F (55 to 110°C) per hour for heaters with service. The rate of cooling should be similar to the
plug headers, to up to 350°F per hour for heaters with heating rate in step 6 above. Purge the contents from
fully welded coils. The lower rates for plug headers will the tubes when they have cooled and fires are extin-
avoid excessive thermal stresses that can cause leaks at guished. Close all fuel and atomizing medium valves.
plugs or header attachments. The above temperatures Open the stack and air plenum dampers and the air
are flue gas temperatures measured at the bridgewall. registers to increase the flow of air and the rate of
h. Warm-up rates may be limited by the curing require- cooling of the heater. Install line blinds in the fuel and
ments for new refractory. If new refractory other than process lines as required by safe practice.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Burner/Heater Operations 499

16.4.8 Developing Emergency Procedures pass into the firebox without losing containment. The tubes
for Fired Heaters in the other passes should be purged of contents in the normal
manner to avoid a rupture in another pass adding fuel to the
An emergency can be defined as an off-design condition that,
conflagration. If the failed pass contained hydrogen, allow
if not properly handled, will result in major damage to or
the contents to dissipate into the firebox without purging. A
destruction of equipment and personnel injury or death.
hydrogen fire may become so hot that tubes melt or vaporize,
Emergency procedures most commonly address the problem
and it is important to give the smothering steam a chance to
of rupture of a tube containing a flammable material while the
cool the firebox.
heater is operating. A more common and more easily handled
event is an incident in which unburned combustibles collect in It is likely that there will be unburned hydrocarbon gases
an active firebox due to errors in fuel and air handling. In the or vapors in the heater firebox. Beware of the instinctive
latter case, the appropriate procedure is to gradually reduce reaction to cut off the fuel to the burners. The combustion air
the firing rate, without increasing the air supplied to the fire- freed by such action could result in an explosive mixture in
box, until complete combustion is reestablished. (Adding air the heater, resulting in violent destruction. Avoid creating
first to a combustible mixture at ignition temperature could explosive mixtures in the affected heater and in any ducting
cause a detonation and heater destruction.) Then, the fuel and or stack common with other heaters. These other heaters, if
air can be increased to satisfy the process requirements. any, should be kept firing at low rates and with the absolute
Some events that can result from a tube rupture and for minimum excess air until they can be shut off and their air
which the proper mitigating actions must be developed registers or plenum dampers closed. Then introduce smoth-
include the following: ering steam to them as well.
If valves are available, isolate the heater to minimize the
• melting or vaporization of the tubes and supports amount of flammables that can flow from other equipment.
• detonation in the firebox
Begin depressurizing the plant as soon as possible to mini-
• convection section collapse into the radiant section
mize the amount of fuel that is available.
• heater collapse due to support structure failure
• flaming oil pool spreading to other areas, putting Keep the smothering steam flow on and the combustion air
additional equipment at risk flow blocked until the heater cools to below 600°F. Test the
• explosive vapor cloud forming around the heater and firebox for combustibles with an explosimeter. If no combus-
possibly igniting tibles are found, the plenum damper or air registers can be
• damage to a stack used by several heaters opened to increase the rate of cooling.
• rapid shutdown of heater and unit causing leaking
flanges due to thermal shock
• rapid depressuring from high pressures causing upset of
catalyst beds and distillation column trays REFERENCES
Some actions are almost always appropriate and should be 1. API Standard 560, Fired Heaters for General Refinery
considered when developing emergency procedures. These Service, American Petroleum Institute, Washington,
include the following: D.C., 1996.
1. Always leave the stack damper in position or try to open
2. R.D. Reed, Furnace Operations, 3rd ed., Gulf Publish-
it fully, if possible from a remote location.
ing, Houston, TX, 1981.
2. Attempt to minimize air entering the heater; close air
plenum dampers if possible. 3. API Standard 530, “Calculation of Heater-Tube Thick-
3. Turn on firebox smothering (snuffing) steam to cool the fire. ness in Petroleum Refineries,” American Petroleum
4. Activate firewater monitors and hoses to quench any Institute, Washington, D.C., 1996
spilled oil, to protect adjacent equipment with fogging
sprays, and to cool the structure and stack to avoid pos- 4. John Zink Burner School Course Notes, Sept. 2000,
sible collapse. copyrighted 2000.
Aim at containing the fire inside the heater. Slowly, using 5. E.A. Barrington, Fired Process Heaters, Course Notes,
steam or nitrogen, purge the contents of tubes in the failed copyrighted 1999.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 17
Troubleshooting
Roger H. Witte and Eugene A. Barrington

TABLE OF CONTENTS
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

17.1 Introduction............................................................................................................................................. 502

17.2 Pulsating Flame....................................................................................................................................... 502
17.3 Flame Impingement on Tubes................................................................................................................. 503
17.4 Flashback ................................................................................................................................................ 505
17.5 Irregular Flame Patterns.......................................................................................................................... 505
17.6 Oil Spillage ............................................................................................................................................. 507
17.7 Long Smoky Flames ............................................................................................................................... 507
17.8 Main Burner Fails to Light-Off or Extinguishes While in Service ......................................................... 508
17.9 Leaning Flames ....................................................................................................................................... 509
17.10 High Fuel Pressure .................................................................................................................................. 510
17.11 High Stack Temperature.......................................................................................................................... 511
17.12 Overheating of Convection Section ........................................................................................................ 511
17.13 “Motorboating” When Firing Oil............................................................................................................ 512
17.14 Flame Lift-Off......................................................................................................................................... 513
17.15 Pilot Burner Fails to Ignite or Extinguishes While in Service ................................................................ 514
17.16 Smoke Emission from the Stack ............................................................................................................. 514
17.17 High NOx Emissions .............................................................................................................................. 515
17.18 Summary ................................................................................................................................................. 519
References ................................................................................................................................................................ 519

501
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 502 Tuesday, March 15, 2005 7:48 AM

502 The John Zink Combustion Handbook

17.1 INTRODUCTION to personnel, then a solution must be employed immediately

Diagnosing and solving problems with burners on heaters in — whatever the cost.
the hydrocarbon and petrochemical industries often seem to If a solution is going to be employed, a careful study of
be as much an art as a science. It is a basic scientific assump- symptoms being exhibited by the furnace should be conducted
tion that the principles of physics, chemistry, fluidics, and a probable cause for the symptoms should be identified.
hydraulics, and combustion do not change. Yet the myriad It is important that the possible causes be carefully identified
variables in a typical petrochemical operation sometimes lest an attempted solution escalate the seriousness of the
make it appear that the equipment has a personality. The problem rather than solve it.
complexity of the sciences, multiplied by the many-staged Once a cause has been determined, standard procedures
processes in a typical plant, causes problems to occur that should be followed to solve the problem. All personnel
were not, and could not have been, anticipated by the design involved should be aware of the problem, the planned correc-
engineers. Moreover, although scientific principles remain tive actions, the ways that safety is addressed, the expected
the same, equipment changes with use. Parameters that may results, and the proper action to take should the problem
have been designed correctly change with time. worsen or not be solved. During normal operations, care
With all of the complexity of conditions that may occur, it should be taken to make incremental changes and adjustments
is still the operator’s job to “keep it running.” When produc- to parameters controlling the combustion process. However,
tion suffers because of the inability of equipment to operate under some conditions, a change must be made quickly and
at required capacity, costs go up and profit or product margins with confidence to prevent additional operating problems.
decrease. Frequently, the operator of the heater must be The following chapter sections describe typical problems,
trained and use knowledge of the equipment and process unit their indicators, their likely effects on operations, and their
to make adjustments that bring operations back to the required causes and standard solutions. Additional references are avail-
capacity desired by plant management. able on troubleshooting.1–5
It is essential that troubleshooting be done in a systematic,
well-organized fashion. Effective and safe troubleshooting
involves four basic steps:
17.2 PULSATING FLAME
1. Recognizing the problem 17.2.1 Indications of the Problem
2. Observing indications of the problem The pulsating flame phenomenon is sometimes called
3. Identifying the effects and the cost of the problem on the “woofing” or “breathing.” Rather than a steady, stable flame
operation of constant volume, the flame is pulsing and changes in size
4. Identifying solutions for the problem and taking correc- with the woofing or breathing. During the change in the
tive action flame pattern, the operator can hear low-frequency noise
The initial indication that a problem exists may come from being emitted from the furnace. The flame oscillates along
controls and instruments on the furnace, or from direct obser- its axis and may periodically extend through the burner to
vation of conditions within or outside the furnace. Changes the outside of the heater.
in process temperature, process pressure drop, excess oxygen, A woofing noise may be heard in the area of the heater.
draft, fuel pressure, stack temperature, and emission levels The woofing or breathing noise is a very low-frequency noise
are typically first observed with instruments. Changes in the that is different from normal operational noise around the
noise being emitted from the heater or flame patterns, flame heater. The operator who has been around combustion equip-
impingement, oil spillage, flame instability, and flashback are ment can immediately identify the change in sound. The noise
usually observed directly during inspection of the heater. will continue to increase to such a violent condition that even
When a problem is noted, it is necessary to evaluate its the most inexperienced operator can identify the noise. Visual
likely effect on the process or product being produced. Some inspection of the flame pattern when the woofing first starts
solutions may require the heater to be shut down for the reveals an unsteady and erratic flame — varying in length
problem to be resolved. Then, plant management must deter- and volume. Under extreme woofing conditions, the sides of
mine the economic value of meeting contracts for products the furnace may be observed flexing.
vs. operating the equipment until it fails and has to be shut
down for repairs at a higher cost. On the other hand, if the 17.2.2 The Cause and Effect on Operation
problem might result in large fines from controlling govern- The effect on operation during the woofing condition is
mental agencies, significant damage to equipment, or danger
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- normally just a passing condition if corrected immediately.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 503 Tuesday, March 15, 2005 7:48 AM

Troubleshooting 503

However, if the woofing condition is not resolved and is with a large combustible mixture of fuel and air, which might
allowed to continue, the intensity of the pressure surges within result in a detonation and damage to the equipment.
the furnace may cause the refractory insulation to begin to If the air registers or the stack damper are fully opened,
break up and fall onto the heater floor. The burner tile may then the heater has reached its maximum capacity. Any wind
even begin to deteriorate and fall apart. The heater vibration blowing under the heater or across the top of the stack may
can break piping, tubing, and instruments. The incomplete interrupt the air flow to the burners and start the pulsating
combustion that occurs causes a drop in heat release, and the flame problem. Sometimes, a windscreen or fence around the
heater cannot fulfill its required duty. Hence, the heater may furnace is necessary to prevent the wind effects on heaters
have to be shut down for major repairs. operating at maximum capacity in high-wind locations.
The usual cause of the pulsating flame is lack of oxygen in Another solution would be to consider changing the raw gas
the combustion reaction. When the oxygen or air flow is burners to premix burners, the latter not being affected by the
inadequate, the flame will search for oxygen alternately inside wind conditions but by the fuel gas pressure. If the flue gas
the heater as the air flows across the burner and outside the analyzer is in the stack, consider relocating it to the firebox
heater as it becomes starved for air. As the flame moves into the flue gas exit so that air leaking into the heater does not give
heater, a pressure front is generated, again causing the air flow a false indication of the oxygen available at the burner.
to cease, and the flame moves back to the burner for oxygen.
The movement of the flame and pressure continue to increase 17.3 FLAME IMPINGEMENT
in intensity and with such force as to cause damage to the heater. ON TUBES
In extreme cases the flame may oscillate so far from the burner
that combustion is extinguished and the flame is lost. 17.3.1 Indications of the Problem
The condition of insufficient oxygen in the combustion The most direct indication of flame impingement is visual
zone of the furnace may exist even when the oxygen levels observation by the operator of the flames contacting the
measured in the stack flue gas indicate that sufficient oxygen external tube surface inside the firebox. The operator may
is available. Air can leak into the heater through the flanges also observe tubes with a cherry-red color or bulges in the
between the convection section and the radiant section, open- tube walls. Indirect indications that flame impingement has
ings in the heater shell, sight ports left open, manway flanges, occurred include higher pressure drop on the process side
and corrosion holes in the heater steel shell. Air leakage into because of coke deposition on the tube walls, higher firing
the heater will add oxygen to the flue gases, causing inaccurate due to the loss of heat transfer because of coke deposi-
rate indications of oxygen available for combustion in the tion on the tube walls, and an increase in the bridgewall and
throat of the burner. stack temperatures.
If flame impingement is suspected but cannot be directly
observed, several infrared photographs of the tubes should be
17.2.3 Corrective Action
taken to determine if there are any high tube skin temperatures
As soon as a pulsating flame is observed, the firing rate should as a result of direct or indirect flame impingement. Also,

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
be immediately reduced to establish sufficient oxygen for the components of the flame radiating at frequencies not visible
combustion reaction to go to completion. The operator must to the naked eye may be contacting the tubes. To establish if
reduce the firing rate until the pulsating flame has stopped and the flame is contacting the tubes, the operator can inject some
no woofing noise is heard. When no woofing noise is heard baking soda (Na+) or activated carbon particles into the com-
and the flames are not pulsating, the operator should observe bustion air. The flame temperature causes the Na+ ion to get
good stable combustion. There should be a measurable excess excited and glow a bright yellow or the carbon to burn with
of oxygen in the firebox gases. Then the operator can open the a yellow flame. The glowing Na+ ion or the burning carbon
air registers on the burner and increase the stack damper’s will show the flame pattern being emitted from the burner for
opening to adjust the excess oxygen and draft to the correct a short time span and will also indicate if there is any flame
levels required for the firing rate desired. The firing rate can impinging on the tubes.
then be increased to the burner capacity or to the required heat
release requested by the heater control system. 17.3.2 The Cause and Effect on Operation
The combustion air should not be increased before cutting During normal operation, the process fluid flowing through
back on the fuel and establishing a stable flame. Increasing the tubes will provide sufficient cooling of the tube surface to
the air before cutting back on the fuel may fill the furnace cause the tube color to be essentially black in contrast to the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 504 Tuesday, March 15, 2005 7:48 AM

504 The John Zink Combustion Handbook

of the burner. Hence, the flame is searching for additional

oxygen from the leaking air. Overfiring may result from
instrument failure, installation of wrong burner tips, or a
change in fuel gas composition, affecting the flame pattern.
Another possible cause may be that the burner firing ports
are eroded (see Figure 17.1) such that the fuel is being injected
through the combustion air flow directly toward the tube.
Another possible cause may be that flue gas recirculation
within the firebox is preventing the flame from forming an
acceptable pattern in its allotted space. Flue gas circulation
may be pushing some burner flames onto the tube surface.

17.3.3 Corrective Action

The heater operator must check the excess oxygen level
within the firebox at the top of the radiant section to be sure
that the instrument is operating correctly, that air is not leak-

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
ing into sample lines, and that all air leaks on the radiant sec-
tion are sealed. Then the operator can increase the amount of
combustion air by opening the air registers or stack dampers
FIGURE 17.1 Corroded gas tip. to allow more air to flow across the burner. If fuel oil is being
fired, the operator may need to check the atomizing steam
pressure to ensure complete atomization and efficient com-
tube hangers or brackets. The tube hangers or brackets that
bustion of the oil. The furnace should be checked regularly
support the tubes will normally be black or slightly red. Flame
for air leaks by inspecting the heater shell, the areas around
impingement on the tubes can create hot spots, causing the
the tubes entering and leaving the firebox, the convection sec-
tubes to appear red or orange in color. The tube color indicates
tion flanges, the sight ports, and flanges around manways.
excessive tube wall temperatures that may result in localized
coke formation. The layer of coke insulates the tube wall from If the fuel is oil, the oil tips should be cleaned regularly to
the cooling effects of the process fluid and allows the metal prevent coke buildup and clogging. The exit ports on the oil
temperature to rise. The insulating effect creates two undesir- tip should be checked for erosion. Both oil and gas-fired
able conditions: (1) heat transfer to the process fluid is burners should be checked for proper alignment and position
impeded, thereby reducing efficiency; and (2) the tube is inad- of the fuel tips within the throat.
equately cooled by the process fluid, resulting in hot spots, If the flue gas circulation patterns within the firebox are
more coke deposition, and eventually tube rupture within the pushing the flames into the tubes, then a Reed wall6 or division
heater. wall 7 may need to be installed in the heater firebox. The Reed
Some process liquids do not coke when overheated, but wall redirects the flue gas circulation pattern within the heater
form vapor. If not considered in the heater design, the vapor while the division wall interrupts the circulation. Each allows
may significantly increase the resistance to flow and the pres- the burner flame pattern to develop in the space designed for
sure drop. The lowered flow rate combined with film boiling the flame pattern.
at the location of impingement will reduce the heat transfer Damaged or missing burner tile sections can cause unequal
coefficient and raise the local tube temperature. distribution of air within the burner. This will lead to fuel-
A possible cause of flame impingement on the tubes may rich zones, locally longer flame segments, and the potential
be a deficiency of combustion air in the combustion reaction, to lean toward and into the tubes. Check the tile condition
causing the flame to search for additional combustion air and repair if necessary.
within the firebox. The deficiency of combustion air for the Impingement may be overcome by changing the flame
combustion reaction may be a result of overfiring or of air shape. For example, if the firebox dimensions allow a longer
leaking into the firebox and not flowing across the burner air flame, the burner tip port included angle can be reduced to
orifice. Air leakage through other openings on the heater does obtain a more slender flame and move the flame envelope
not mix well with the combustion air moving across the throat further from the tubes.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 505 Tuesday, March 15, 2005 7:48 AM

Troubleshooting 505

17.4 FLASHBACK TABLE 17.1 Ratio of the Upper and Lower Explosive
Limits and Flashback Probability in Premix Burners for
17.4.1 Indications of the Problem Various Fuels
Flashback is the phenomenon that occurs in premix burners Fuel Ratio1 Probability2
when the flame velocity is greater than the velocity of the flow- Acetone 5.01 1.67
ing mixture and the flame front propagates back through the Acetylene 32.00 Infinite
Acrylonitrile 5.57 1.85
mixer or venturi to the orifice or area where the fuel and pri-
Ammonia 1.71 0.57
mary air are being mixed. Then the fuel and primary air con- Aromatics (Mean) 5.00 1.66
tinue to burn in the venturi or mixer. Flashback is most likely Butadiene 5.75 1.92
Butane 4.52 1.51
to occur in burners using fuels having a high ratio between the
Butylene 4.88 1.63
upper and lower explosive limits of the fuel. Table 17.1 reveals Carbon Disulfide 40.00 Infinite
that the gases most susceptible to flashback include carbon Carbon Monoxide 5.93 1.97
Cyanogen 6.45 2.15
disulfide, acetylene, ethylene oxide, hydrogen, hydrogen
Ethane 4.02 1.34
sulfide, and ethylene. Another reason for flashback may be that Ethyl Alcohol 5.77 1.92
the gas tip design is not optimized for the fuel that is being Ethyl Chloride 3.70 1.23
Ethylene 10.04 3.33
burned. If the velocity of the gas and air exiting the gas tip is
Ethylene Oxide 26.66 Infinite
very low, because of a large-diameter firing port, then the Gasoline (Mean) 5.06 1.68
velocity of the flame front may be greater than the velocity of Hydrocyanic Acid 7.14 2.38
Hydrogen 18.55 Infinite
the fuel/air mixture exiting the tip and flashback will occur.
Hydrogen Sulphide 10.60 3.52
When flashback occurs within the mixer or venturi, a flame Methane 3.00 1.00
will be observed burning in the venturi or mixer. If flashback Methyl Alcohol 5.43 1.81
Methyl Chloride 2.26 0.75
has occurred sometime in the past, the mixer or venturi will Naphtha 5.45 1.81
show signs of oxidation on the outside of the cast iron venturi. Oil Gas 6.84 2.28
When flashback occurs in a premix burner, there is little doubt Propane 5.25 1.75
Propylene 5.55 1.85
in the operator’s mind that flashback is occurring within the Vinyl Chloride 5.42 1.80
burner. A sharp barking noise in the mixer is continually 1 Ratio of the upper and lower explosive limits.
emitted until corrected. 2 Probability of flash-back as compared to methane.
Source: R.D. Reed, A new approach to design for radiant heat transfer in
17.4.2 Effect on Operation process work, Petroleum Engineer, August, C7-C10, 1950.
When flashback occurs and remains uncorrected, the burner
mixer or venturi is damaged from the high temperatures gen-
erated within the mixer from the combustion reaction that is that no flashback will occur. If the fuel composition or process
occurring. The damage to the burner parts will result in heat requirement changes, resulting in lower operating pres-
higher maintenance costs. When flashback occurs, the capac- sures, burners must be shut off to raise the fuel gas pressure or
ity of the burner is restricted; if flashback occurs on many new orifices may be required in the mixers or venturi to keep
burners within the heater, the outlet temperature of the pro- the fuel gas pressure at a level sufficient to prevent flashback.
cess cannot be obtained. The burning inside the venturi tube If raising the fuel pressure does not resolve the flashback
and intermittent open flame constitute a safety hazard. problem, one can look to the flame velocity of the fuel/air
mixture for another solution. The flame velocity is related to
17.4.3 The Cause and Corrective Action the percentage of air in the fuel gas. A change in this percent-
The solution to the problem will vary depending on what is age, by adjusting the burner air door, can raise or lower the
available to the operator. The operator should immediately flame velocity. Try reducing the primary air flow to lower the
check the gas tip discharge port and the main gas orifice to mixture flame velocity. Adjust the secondary air register to
ensure that both are clean. If the gas orifice is dirty, the fuel maintain the target excess oxygen level.
flow may be reduced to the point of creating a flashback condi-
tion. If the gas tip discharge port is dirty, the flow of the fuel/air 17.5 IRREGULAR FLAME PATTERNS
is reduced on the exit discharge port to the point of creating the
flashback condition. Hence, the operator must clean both the 17.5.1 Indications of the Problem
gas metering orifice and the gas tip discharge port to allow the On a single burner, the flame pattern is nonsymmetrical. An
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
fuel/air velocity to remain higher than the flame velocity so irregular flame pattern implies that the flame varies in length
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 506 Tuesday, March 15, 2005 7:48 AM

506 The John Zink Combustion Handbook

refractory or plenum insulation, in the burner throat or by a

damaged tile. Either will affect the local mixing of fuel and
air, delaying mixing in part of the throat and lengthening the
flame on that burner.
In the case of several burners in the same firebox, the prob-
lem may be more complicated. Some of the burners may be
plugged while others are operating normally. The air registers
may be at different settings, causing uneven air flow across
each of the burners. A poorly designed plenum may be supply-
ing more combustion air to some of the burners and less to
others or more to one side of the burners. Maintenance or
installation personnel may have placed the burner tips in an
incorrect position with respect to the refractory ledge. Burner
tips designed to be oriented in a specific direction with respect
FIGURE 17.2 Irregular flame pattern.
to other tips or the refractory ledge may be turned to the wrong
position, causing flame-to-flame interference. The refractory
may be damaged, causing poor or uneven air circulation around
or shape across the width of a burner. For multiple burners in
the burner tips.
a heater, the flame pattern from one burner to another burner
is not of the same shape or volume for the same conditions. Obviously, it is good operational practice to maintain the
burner tips and ensure that they do not become dirty or
Irregular flame patterns can appear in two ways. A single
plugged. Many irregular flame patterns are corrected by sim-
burner may have one side of the flame pattern longer than
ply keeping the burner tips in good operating condition. While
the other side (see Figure 17.2), or part of the flame may be
cleaning the tips, the orifices should be inspected to ensure
emerging at a different angle from the main flame. In a multi-
that no erosion has occurred that would change the burning
burner furnace, some burners may have a longer flame pattern
characteristics of the tip. If the fuel is oil, some of the orifices
than other burners in the furnace when the same fuel pressure
on the tip may become completely plugged with coke. Man-
and same air register opening is being provided at each of
ually inserting a twist drill bit into the orifice and blowing
the burners.
steam through the orifice should clean the orifice. Do not use
powered drills to clean orifices. The use of a power drill may
17.5.2 Effect on Operation change the size or shape of the orifice. A larger or different-
In either the single burner or multi-burner case, the irregular sized orifice then flows fuel at a different rate thereby causing
pattern may cause the furnace to have hot spots on the tubes. the tip to produce an irregular-shaped flame.
In the case of multi-burner furnaces, the burners with the In multi-burner furnaces, all of the burner tips should be
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

longer flame lengths may create flame impingement on the inspected, cleaned, and of the same orifice size. All air reg-
process tubes while the other burners have short, compact isters should be set to the same approximate opening. All fuel
flames and cause no hot spots on the tubes. Irregular flame pressures to the burners should be at the same level. If the
patterns can result in higher operating cost or afterburn due to flames are still irregular, check the manufacturer’s drawings
incomplete combustion of the fuel being burned. for the burner fuel tip position and orientation. Then check
each tip on each burner and ensure that the tips are at the
17.5.3 The Cause and Corrective Action correct height and orientation with respect to the ignition
In the case of an irregular flame from a single burner (see ledge. Inspect the burner refractory tile for damage, especially
Figure 17.2), the problem may be a dirty burner tip, an in the immediate area of the tips. Some tips are designed to
eroded tip, or a tip improperly oriented with respect to the operate in close proximity to the refractory, and a difference
refractory. In a dirty burner gas tip, some orifices may be par- of ±0.25 in. (6.4 mm) can cause irregular flames.
tially plugged while others are operating normally. If part of Air plenum distribution problems are difficult to diagnose.
the flame is emerging at a different angle from the main Such problems usually occur when the plenum is not correctly
flame, tip erosion may have occurred. Improperly placed tips sized. As a result, some burners get more combustion air and
may be too close, too far, or turned at the wrong orientation other burners are starved for combustion air. The operator
with respect to the refractory tile ignition ledge. The irregular needs to ensure that all burner air registers have the same size
flame may also be caused by a foreign material, such as opening. The designer should ensure that the air plenums are
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 507 Tuesday, March 15, 2005 7:48 AM

Troubleshooting 507

properly sized for good air distribution to all burners. In some spray will contact the tile and cause coking on the tile as well
cases, it may be beneficial to have the air plenums modeled as oil drips or spills. This sometimes happens when the oil
so that they are correctly sized. In large plenums, benefit may gun from one heater is mistakenly installed in another. The
be obtained by adding air inlets to improve the distribution operator should have manufacturer drawings available to
of air. check the oil tip location and should have the oil guns for
each heater readily identified.

17.6 OIL SPILLAGE 17.6.3 Corrective Action

17.6.1 Indications of the Problem Operators should be aware of indications of poor atomization
On natural draft heaters, the operator observes fuel oil drip- so that action can be taken before oil spills occur. Showers of
ping and forming pools of oil under the heater. Spilling in an “sparklers” or “fireflies” leaving the flame indicate poor atomi-
air plenum can cause a plenum fire, damaging burners and zation or water droplets in the oil spray. Smoky flames, fouling
the heater structure. The steel air plenum cannot be touched of the burner tip, and uncontrollable flames reaching the con-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

by the operator’s hand because of the excessively high skin vection section may also be indicative of atomization problems.
temperatures. The paint on the air plenum is observed to be The operator should check the fuel temperature at the
blistering and peeling. In extreme cases, the operator may burner and control the temperature to a viscosity of 200 SSU
observe oil burning under the heater where the oil has (45 Centistokes) or less. The operator should inspect all parts
dripped and formed pools of oil at grade. of the fuel oil system, including the fuel oil gun and atomizer.
The fuel oil tip and atomizer should be cleaned, and no foreign
17.6.2 The Cause and Effect on Operation material should be on the tip or inside the fuel oil tip or
An open pool of oil beneath an operating furnace is clearly a atomizer. The fuel oil tip should be inspected for erosion and
fire hazard that must be immediately corrected to ensure replaced if necessary. If the fuel oil tip orifice is +2 drill sizes
operator and equipment safety. If the pool of oil under the larger than the design, then the oil tip should be replaced with
heater is allowed to ignite and burn, damage may occur to the a new case-hardened oil tip to provide longer life.
heater. Second, the oil is not all being consumed in the com-
bustion reaction, causing higher fuel operating costs. The operator should check the burner manufacturer’s draw-
ings to ensure the oil tip is located in accordance with man-
The main cause of oil spillage or dripping from the burner
ufacturer instructions. The steam-to-oil differential pressure
is normally the poor atomization of the fuel oil when the
should also be checked and set in accordance with the burner
temperature of the oil is lower than specified or the atomizer
manufacturer’s instructions. The steam traps on the atomizing
is damaged or plugged. At low fuel oil temperatures with
steam line should be operating correctly to ensure that no
higher oil viscosity, the oil droplets leaving the oil gun are
condensate or water is present in the atomization steam; steam
larger. The large oil droplets may contact the burner tile and
line insulation should be in good repair.
begin flowing down the burner onto the grade.
The temperature of the fuel oil must be at a viscosity of
200 SSU (45 cs) or less at the burner, not at the fuel oil heater
outlet. If the fuel oil lines from the heated fuel oil supply tank 17.7 LONG SMOKY FLAMES
to the burner oil gun connection are not insulated properly,
the fuel oil temperature can be too low at the burner, making 17.7.1 Indications of the Problem
the viscosity too high. Even with higher steam atomization
pressure, the cold fuel oil may not be broken into small Visual observation inside the firebox reveals long, dirty,
enough droplets to burn effectively. smoky flames, possibly reaching into the convection section
The burner fuel oil tip and atomizer may have suffered of the heater. Smoke may be observed exiting the stack. The
erosion from particles present in the fuel oil. The eroded tip flame, rather than being confined in the flame pattern space
will allow the oil to be injected into the tile at a larger angle within the firebox, may be impinging on the process tubes in
than design. The oil then contacts the sides of the refractory the convection section. The combustion zone may appear
tile and begins leaking fuel oil down onto the burner and hazy, rather than bright and clear.
dripping to the grade under the burner. The operator of the heater may notice that the process outlet
Not all spills are due to atomization or fuel tip problems. temperature cannot be achieved. The stack temperature may be
If the oil gun is not inserted far enough into the burner, the above the design specifications.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 508 Tuesday, March 15, 2005 7:48 AM

508 The John Zink Combustion Handbook

17.7.2 Effect on Operation with closely sized burners or with some burners out of service,
Long, smoky flames are the result of incomplete combustion will cause a shortage of oxygen at the burners and long flames
and can consume far more fuel than necessary to achieve the that smoke. Place more burners in service or reduce the fuel
desired heat transfer and outlet temperature. Soot (unburned input. Contact the burner manufacturer if larger burners are
carbon) will be deposited on tube surfaces in the convection required to increase the heat release.
section, reducing heat transfer. Coke formation may be Also check the size of the ports in the burner tips. The ports
observed in the burner tile. The required process outlet tem- may be enlarged or the wrong tips may have been installed
peratures may not be achieved. during maintenance. In the latter case, the tips may even vary
from burner-to-burner. Enlarged ports allow uneven fuel input
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

17.7.3 Corrective Action for the available air. Replace any damaged or improper tips.
The corrective actions are aimed at ensuring that adequate air If changing the operating variables and cleaning the tips and
is provided to each burner and that this air is mixed quickly atomizers, together with replacing damaged parts, do not
and completely with the fuel so as to achieve rapid combus- improve the flame, contact the burner manufacturer. He may
tion and smaller flame volume. Air that is not mixed with fuel recommend a different design for the atomizer and tip, partic-
at the burner passes into the less turbulent firebox where mix- ularly if fuel characteristics are not close to the original design.
ing of the fuel and air is less likely and ignition temperatures When firing gas at low excess air, if flames become long
may be too low for combustion. and smoky, suspect a change in the fuel gas composition. The
The operator should check the draft and excess oxygen substitution of higher heating value, heavier components
level at the top of the radiant section of the heater to determine along with a lowered burner pressure will increase the oxygen
if there is sufficient oxygen to burn the fuel. If the excess requirement. The air registers and possibly the stack damper
oxygen is lower than the target, the draft and excess oxygen must be opened to provide more air.
should be adjusted to the correct levels.
The next item to be checked by the operator is the temper-
ature of the oil at the burner. The higher the specific gravity 17.8 MAIN BURNER FAILS TO
of the fuel oil, the more critical the fuel oil temperature is to LIGHT-OFF OR EXTINGUISHES
atomization in the oil burner. If the viscosity is above 200 SSU
WHILE IN SERVICE
(45 cs), atomization will suffer.
The operator should determine that the oil gun and atomizer
are clean and no foreign material is plugging either the oil or 17.8.1 Indications of the Problem
steam orifices within the oil gun. The fuel oil and steam must The operator follows the usual purge and ignition procedures
be delivered to the oil gun at the burner manufacturer’s design for lighting a burner, but there is no indication of main burner
pressure and temperature. There are many different types of ignition. After the heater is in operation, one or more burners
oil guns used in the burning of fuel oil, and each has different flame out.
pressures for fuel and steam to make it work successfully. For
no. 6 fuel oil and heavier, all fuel oil and all steam lines must 17.8.2 Effect on Operation
be insulated to reduce loss of heat and temperature. The steam
If a burner fails to light, the process unit outlet temperature
traps on the steam line should be checked to ensure that the
may not be achieved or the heater startup may be delayed.
condensate is being removed from the steam and the steam
Additionally, if the burner fails to ignite or flames out, the
remains dry.
furnace may fill with a dangerous mixture of gas and air that
The oil tip should be positioned in the oil tile or diffuser
can result in an explosion in the firebox.
cone in accordance with the burner manufacturer’s drawings
and instructions. Improper location may cause the oil–steam
spray discharging from the oil gun to collapse and produce a 17.8.3 Corrective Action
dense spray. This spray cannot mix with the combustion air The most common cause of ignition failure is improper posi-
quickly so as to burn with a short flame. The flame becomes tioning of the pilot burner in relation to the main burner. If
long and smoky. the pilot is not located so that its flame is directed into the
If the burners still show long flames and most of the burners fuel/air mixture leaving the main burner, the ignition temper-
have the same flame length, the firing rate needs to be ature is not achieved and the main flame is not initiated. The
checked. Firing at above the rated heat release, particularly operator should check all components of the pilot and burner
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 509 Tuesday, March 15, 2005 7:48 AM

Troubleshooting 509

to ensure that they are positioned according to the specifica- quality, as well as the fuel oil temperature, pressure, and flow,
tions in the manufacturer’s design drawings. and correct if necessary. Monitor these conditions regularly.
Check that fuel gas is being supplied to the burner. Before
startup, the fuel supply line may be purged or pressure tested
with an inert gas. If this gas is not completely displaced with 17.9 LEANING FLAMES
fuel, the operator will be attempting ignition of an inert mate-
rial. Procedures should be in place for removing the inert gas 17.9.1 Indications of the Problem
and checking for the presence of a flammable material. In floor or ceiling-fired furnaces, the flames may lean to one
Check for closed or blocked valves, fuel gas line blinds, side rather than burning in a vertical line. In wall-fired fur-
plugged strainers or filters, and plugged burner ports. Plug- naces, the flames may lean to the side rather than firing hori-
ging is a frequent problem during startup if the fuel piping is zontally and curling upward.
new or just revised due to foreign debris left in the lines. Old Observation of the flame pattern inside the firebox reveals
piping may have internal scale that dislodges and enters the that the centerline of the flame does not follow the designed
burners, thereby plugging the ports. The burner ports and fuel path as specified by the burner manufacturer. The flame is
lines may need to be cleaned. A strainer should be installed commonly expected to propagate in the general direction of
to minimize plugging if this problem occurs frequently. the centerline of the air orifice or refractory tile. However, in
When burners flame out in service, the problem is almost some burners, the flame may be designed to propagate in the
always an interruption of either fuel or air being supplied to general direction of the gas orifice.
the burner. Fuel interruptions can be caused by instrument
failures such as the closing and reopening of a fuel control 17.9.2 Effect on Operation
valve. If the burner pilot is lit, the fuel should re-ignite with Flames that do not have the designed pattern and direction
only a “puff” or minor detonation. If there is no pilot, the fuel can create problems such as impingement on the process
may not ignite and fuel gas will continue to flow; a flammable tubes which results in hot spots and may eventually cause
mixture may accumulate and explosion may result. tube rupture.
Gas burners are likely to be extinguished if a significant
amount of liquid enters the burner as anything other than a 17.9.3 Causes and Corrective Action
mist. Gas burners have no ability to atomize liquid into finely
Refer to Sections 17.3, 17.5, and 17.7 for some thoughts on
divided droplets that will burn. Hence, the flame is lost if the
corrective action.
liquid is discharged through the burner. Liquid may be amine
Incorrect positioning and orientation of the burner tip(s)
carryover from treating facilities that remove hydrogen sulfide;
with respect to the refractory walls or floor of the firebox can
from (during cold weather) condensed heavy gases such as
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

cause leaning flames. This may be due to incorrect burner

propane or butane added to supplement the normal fuel supply;
installation or to deformation of the heater steel. Both should
or condensed water vapor. Facilities to remove liquids, such
be checked and corrected. The wrong burner tip, an improp-
as a knockout drum, should be considered.
erly drilled tip, or a tip turned in the wrong direction can also
All burners may flame out if the air supply is interrupted cause flames to lean. All should be checked and the manu-
either by dampers opening and closing or by pressure surges facturer’s drawings referenced to ensure that orientation and
within the firebox. Ensure that conditions allow an appropri- locations are correct.
ate air supply to flow across the burners at all times. Improperly installed or damaged refractory can affect the
Too much air flow can also cause loss of ignition or burner direction of flame propagation. When observed, such sections
flameout problems. Too much air may result in a nonflamma- should be repaired. Damaged diffuser cones can cause leaning
ble mixture at the burner. The large excess of air may also and uneven flame patterns (see Figure 17.3) and should be
quench the fuel–gas mixture below the ignition temperature. replaced. Damaged burner tile should be replaced if it is
If this is suspected, close the air registers and/or the stack contributing to flame leaning.
damper to reduce the excess air in the firebox to the targeted Circulating currents of the flue gas within the furnace can
levels or to a minimum level if during startup. be a cause of leaning flames. The circulation can be observed
When firing oil, the atomizing medium can extinguish the by injecting baking soda or activated carbon into the burner.
flame if the fuel pressure is too low, if fuel temperature is too The glowing particles trace the flue gas circulation currents
low, or if atomizing medium operating conditions are not cor- within the heater. The most effective correction is the instal-
rect. Check the atomizing medium pressure, temperature, and lation of division or Reed walls, as discussed earlier.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 510 Tuesday, March 15, 2005 7:48 AM

510 The John Zink Combustion Handbook

during operation with typical plant fuels and provide reason-

able pressures for fuel/air mixing at turndown.
Under normal circumstances, increasing the fuel pressure
above these levels will increase the fuel flow to the combus-
tion process without problems as long as there is an excess
of oxygen in the flue gases. However, if the oxygen analyzer
does not indicate there is a sufficient excess of oxygen for
additional fuel, then the fuel pressure should not be increased.

17.10.3 Solution and Corrective Action

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

First, look to the pressure gage at the burner fuel manifold.

Experience tells us that pressure gage accuracy deteriorates
with time. Replace the pressure gage with a new or recondi-
FIGURE 17.3 Damaged diffuser cone. tioned and calibrated gage. Also confirm the pressure gage
location. It is not unusual to be using a pressure reading from
a gage near or even upstream of the fuel control valve as rep-
Air infiltration when operating with minimum excess oxy- resentative of the pressure at the burner. This can lead to seri-
gen can cause flames to lean toward the source of the leaking ous error in the “measured” pressure value. A gage at this
air. Air leakage points should be identified with smoke tests location is measuring the pressure at the burner plus the pres-
and sealed. sure drop in the piping plus the pressure drop in any valves,
which may be throttled — not the pressure at the burner.
Make sure that the gage is properly located and, if in heavy
17.10 HIGH FUEL PRESSURE oil service, properly sealed, insulated, and steam-traced.
Next, look to the flow measuring instrument from which
17.10.1 Indications of the Problem the heat release is calculated. If the current fuel gas compo-
The pressure gage on the fuel line at the burner measures a sition and specific gravity do not match the properties used
pressure higher than the burner capacity curve indicates for to determine the meter factor, the flow indicated may be
the required heat release. Typical process burners are inaccurate, and the calculated heat release is wrong. If this is
designed for maximum heat release at 15 to 25 psig (1.0 to the case, the heating value of the fuel has also probably
1.7 barg). Hence, if the pressure observed is 30 to 40 psig changed. A new burner curve using the new fuel properties
(2.0 to 2.7 barg), the fuel pressure may be too high for the should be obtained from the burner manufacturer.
heat release required if the fuel composition is per design. If a problem is still evident, some plugging in the fuel
Typical fuel oil pressure for maximum heat release is around delivery system, strainers, the burner orifice, or tip ports, the
100 psig (6.8 barg). Hence, if the operating fuel oil pressure
wrong-sized fuel lines, and lack of adequate insulation and
is 150 psig (10.2 barg), the operator is experiencing high fuel
tracing on heavy oil piping are all possible causes of the
oil pressures and needs to determine the problem.
restriction of fuel flow. All should be checked and cleaned
or corrected. The burner parts should also be checked for
17.10.2 Effect on Operation damage or incorrect drilling.
Higher than normal fuel gas or fuel oil pressure can cause
The operator may be attempting to fire too much fuel on
irregular flame patterns and hot spots on the process tubes
too few burners. Place more burners in service to reduce the
within the furnace. Again, the flame impingement and hot
fuel pressure. Fully open the fuel valves on all active burners.
spots may lead to process tube rupture and unit shutdown.
With higher than designed fuel pressures observed, the fur- In some circumstances, the higher than designed fuel pres-
nace might not achieve the desired outlet temperature. sure is acceptable as long as there is sufficient excess oxygen
The fuel gas pressure is normally designed at 15 to 25 psig to burn the fuel being injected into the heater, the flame
(1.0 to 1.7 barg) for most process heaters. These pressures patterns are acceptable, there is no flame impingement, and
ensure reasonable orifice diameters to reduce fouling problems the CO level is below 100 ppm.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 511 Tuesday, March 15, 2005 7:48 AM

Troubleshooting 511

17.11 HIGH STACK TEMPERATURE section of the heater, and if the air leaks are minimal, then
the operator must check for foreign deposits on the convection
17.11.1 Indications of the Problem section tubes. If fuel oil is being fired, the operator may need
Under normal furnace operation, the flue gas temperature to blow the soot from the convection section tubes. At times,
measured in the stack will be near the operating temperature a rather simple technique will clean up an externally fouled
predicted by the furnace designer at the design duty. High convection over time, depending on what has fouled the tubes
stack temperatures indicate decreased heater efficiency, and the flue gas temperature. Changing from oil firing to gas
higher fuel consumption, and increased operating cost. High firing, together with a modest increase in excess air, has been
stack temperatures can indicate excessive heat in the convec- observed to remove fouling deposits within a few weeks.
tive section of the heater. Long-term operation at higher than Apparently, the deposits oxidize slowly and disappear up the
normal stack temperatures can damage the furnace, espe- stack as gases. The operator must also determine if there is
cially in the roof area of the convective section. foreign material deposited between the fins on the convection
section tubes or fin damage from previous afterburns. The
17.11.2 Effect on Operation operator will also need to check the fuel oil temperature to
High stack temperatures indicate reduced heater efficiency. ensure proper atomization of the liquid fuel within the burner
The higher the stack temperature, the lower the efficiency of flame envelope, as well as check to ensure there is no water
the heat recovery to the process and the higher the operating in the atomizing steam. Poor atomization will result in large
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

cost to produce a product. Long-term operation at higher than fuel droplets that will carry into the convection section,
normal stack temperatures may result in damage to the fur- deposit on the convection section tubes, and cause high stack
nace stack and convection section. temperatures to occur.
High stack temperatures may indicate afterburning in the Sometimes the high stack gas temperature is caused by the
convection section. The afterburning may be caused by lack
heater construction. Hot flue gas will follow the path of least
of excess oxygen within the burner throat and firebox. Air
resistance to the stack entrance. This is usually at the ends of
leaks into the convection section and the temperature of the
the convection section tubes, where the surface to absorb heat
flue gases is sufficient to cause the completion of the com-
is minimal. Unless steps are taken to minimize this bypassing,
bustion reaction to occur (afterburning) in the convection
which may occur in header boxes or at tube ends in the main
section of the heater. The afterburning will result in the
flue gas passage, a significant amount of hot gas reaches the
destruction of the extended surface on the tubes in the con-
stack without transferring much heat to the convection tubes.
vection section of the heater. The loss of the extended surfaces
results in a loss of heat transfer in the convection section. The Also, if enough stacks or flue gas offtakes from the convec-
afterburning within the convection section may ultimately tion are not provided, the profile of flow across the tubes is not
result in tube failures in the convection section of the heater. even.7 Flue gas in the zones of high velocity flow escapes from
the convection section at greater than design temperatures.
17.11.3 Corrective Action
The stack temperature is controlled by the amount of excess
oxygen in the flue gases. Hence, the operator can reduce the
17.12 OVERHEATING OF THE
stack temperature and lower the excess oxygen in the flue
gases by controlling the stack draft and burner air registers or CONVECTION SECTION
dampers. This should be the first action by the operator.
The operator must then determine if there are any air leaks 17.12.1 Indications of the Problem
into the furnace and plug all possible cracks and openings
leaking air. The excess oxygen at the top of the radiant Upon visually inspecting the inside of the firebox, there is a
section must then be checked to ensure that there is sufficient lot of refractory lying on the floor and in the burners. The
excess oxygen at the burner to burn all the hydrocarbon fuel heater shell and structure on the convection section show
being injected. If there is insufficient excess oxygen, then signs of overheating. The shock tubes and shock tube hang-
the afterburning may take place in the convection section, ers are failing. The draft at the top of the radiant section is
resulting in higher than normal stack temperatures. at a positive pressure. When the sight ports are opened, hot
If there is sufficient excess oxygen within the heater to flue gases are forced out, thereby causing a safety hazard to
ensure all hydrocarbons are burning completely in the radiant the operator.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 512 Tuesday, March 15, 2005 7:48 AM

512 The John Zink Combustion Handbook

17.12.2 Effect on Operation 17.13 “MOTORBOATING”

The convection section is designed to remove heat from the
WHEN FIRING OIL
hot flue gases exiting the radiant section at the bridgewall
temperature of 1200 to 2000°F (650 to 1100°C) and transfer
17.13.1 Indications of the Problem
When firing light fuel oil such as naphtha or propane, a
the heat from the flue gases to the process liquid by convec-
motorboating sound may be emitted from the heater. The
tion. If the hot flue gases leak out of the cracks in the con-
motorboating sound is caused by the alternate flowing of
vection section, then the structural steel and heater
vapor and liquid from the fuel oil gun in the burner. The light
convection section shell are overheated, resulting in damage fuel oil flashes and causes the flow of liquid to cease for a
to the heater. moment until the flashed vapor is emitted, goes back to liquid
The hot flue gases leaking out of the cracks in the heater flow for a moment, and then back to vapor. The flames will
shell result in overheating the carbon steel heater shell and have a rapidly pulsating appearance. The alternate flowing of
vapor and liquid causes a fluttering sound similar to that of a
structural steel supporting the convection section of the heater
motor boat. The motorboating sound is more pronounced by
above the radiant section. With the overheating of the heater
low fuel flow rate, low fuel pressure, and low initial boiling
shell, the refractory anchors are damaged and leave nothing
point oil.
to support the refractory in the arch section of the heater, and The motorboating sound being emitted is normally expe-
the refractory falls to the heater floor. As a result, more heat rienced when light fuel oil with an IBP of 150 to 250°F
reaches the structural steel, and finally the convection section (66 to 120°C) is being burned. The burning of heavy fuel oil
falls into the radiant section of the heater and the unit is shut does not normally produce the motorboating sound because
down for repair. the IBP of heavy fuel oil is usually greater than 400°F (200°C)
and the heavy fuel oil will not vaporize at conditions within
the oil gun.
17.12.3 Corrective Action
The operator must first obtain the draft reading at the top of 17.13.2 Effect on Operation
the radiant section and compare this to the targeted draft read- The noise is more of a nuisance than a problem, possibly
ing. If the draft reading indicates positive pressure, then the requiring the use of hearing protection by operators. The
hot flue gases exiting all cracks must be stopped before the flame position is more of a concern since any flame instabil-
carbon steel shell is overheated and the anchors holding the ity has the potential to further deteriorate flame quality and
refractory are damaged. When the anchors holding the refrac- possibly cause the flame to extinguish.
tory are damaged and can no longer hold the refractory, the
refractory falls onto the floor of the heater. Upon observing
17.13.3 Solution and Corrective Action
The operator must first check the fuel oil pressure and atomi-
the positive pressure in the heater, the operator must imme-
zation steam pressure to ensure they are within the operating
diately open the stack damper on a natural-draft heater, or
parameters for the heat release being demanded by the fur-
the induced-draft fan damper on a balanced-draft system if
nace. If the pressures are within the operating parameters,
there is a measured excess of oxygen at the firebox exit. If then the operator will need to check the fuel oil bypass valve
there is no excess of oxygen, reduce the firing rate until an to ensure it is completely closed and not leaking any steam
excess is attained, then increase fuel input while maintaining through the valve. If steam is leaking through the bypass
the draft and excess oxygen at the targets. A negative pres- valve, the steam will cause the fuel oil to be heated above the
sure must be established to eliminate the overheating of the flashpoint by direct contact.
convection section. If there is no steam leaking across the bypass valve, then
the operator needs to ensure that the fuel oil does not have
If the stack damper or induced-draft fan damper is com-
any water being injected with the oil. Water will have a lower
pletely opened, then the operator must reduce the firing rate boiling point than the fuel oil and will vaporize and may cause
on the heater such that there is a slightly negative pressure at the flame to be extinguished. The loss of flame may result in
the top of the radiant section of the heater. Also check the very unsafe operating conditions.
excess oxygen and reduce it to the target value. High excess The operator can eliminate the motorboating noise from the
air can lead to positive firebox pressure. burner by turning off a burner in a multi-burner installation.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Troubleshooting 513

The fuel oil pressure should increase to keep the same heat 17.14.3 Corrective Action
release in the furnace. The increase in pressure may eliminate
the vaporization within the oil gun and stop the motorboating Flame speed and air/fuel delivery speed must be balanced to
ensure that the flame is attached to the tip rather than rising
sound. The best solution for correcting the problem is to con-
above it.
tact the burner manufacturer and change the oil gun design
from a concentric tube design to a dual tube design. The dual When flame lift-off from a burner is first observed, imme-
tube design separates the oil and the steam until they reach diately reduce the fuel pressure on that burner by partially
the oil tip. Hence, the residence time for transferring heat from closing the specific burner block valve. If the flame continues
the steam to the light fuel oil is minimal, and vaporization of to lift-off, the operator should completely shut off the burner.
the light fuel oil is eliminated. In either case, take the following corrective actions. Flame
lift-off produces unsafe conditions and should be corrected
immediately before a more serious condition occurs.

17.14 FLAME LIFT-OFF The alignment and positioning of the gas burner tip and
the oil burner tip should be checked to ensure they are cor-
rectly installed and positioned in accordance with the burner
17.14.1 Indications of the Problem manufacturer’s drawings and specifications. The gas tips
should be located in relation to the tile ledge or flame holder
The first indication of a flame lift-off problem is when the
as shown on the drawings, and the oil tip should be located
operator observes that the flame is detached from the burner
in relation to the inlet tile throat or diffuser as per the draw-
when inspecting the flame patterns inside the firebox. Nor-
ings. If the oil tip is too high in relation to the tile or flame
mally, flame lift-off occurs at one or two burners in a multi-
holder, the oil flame may lift-off.
burner installation and not at all the burners at the same time.
The process outlet temperature, the excess oxygen in the fur- The fuel gas firing ports and ignition ports should be
nace, the draft in the furnace, the fuel pressures, and the noise checked to ensure they are not plugged. If they are plugged,
from the operation of the burner will give no indication of the ports should be cleaned by manually inserting a twist drill
any flame lift-off problem. the same size as the port and twisting the drill to remove all
foreign material in the port.

17.14.2 Effect on Operation If the burner is a premix burner design, then the primary air
door should be adjusted to the correct position. On a premix
Flame lift-off from the burner is a very significant safety haz- burner, the gas/air mixture exits the gas tip at a given velocity.
ard. If the lift-off is extreme, there may be a total loss of flame The flame burning above the firing port has a flame velocity
at the burner and unburned fuel will be injected into the fire- that is traveling in the opposite direction, that is, it is trying
box. If the refractory remains at a sufficiently high tempera- to get back to the source of the fuel and air mixture. If the
ture or if the pilot remains lit, then re-ignition may occur. Re- gas/air velocity is too high and the flame speed is too low,
ignition may also be initiated by adjacent burner flames. The then the flame begins lift-off from the burner gas tip. To correct
re-ignition may cause a minor or major explosion, depending the exit velocity from the firing port, the primary air door is
on the amount of fuel injected into the firebox, with the extent closed to reduce the amount of primary combustion air enter-
of damage dependent on the heater design and configuration. ing with the gas, hence, a reduction in the gas/air velocity.
If the explosion within the firebox is minor, the explosion With the reduced gas/air velocity, the flame reattaches to the
doors will open and relieve the internal pressure built up burner gas tip or the flame holder.
within the firebox. If there are no explosion doors on the If an oil-fired burner is experiencing flame lift-off, the oper-
heater to relieve the internal pressure buildup, the heater may ator will need to adjust the steam atomization pressure per the
be damaged by an explosion. If the explosion within the fire- burner manufacturer’s instructions. If the atomizing steam
box is a major explosion, the complete heater may be torn pressure is too high, the oil flame will tend to lift-off from the
apart, resulting in the heater and process being shut down. The burner. Raw gas burners that utilize a stabilizing cone can
loss of the heater will cause a loss of product and hence a loss experience lift-off if the cone is missing or damaged. Shut off
of profits being generated from the process unit. In the most any such burner where the cone is partially or completely
severe explosions within the firebox, there may be loss of life missing or is improperly installed, and replace the diffuser
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

or injury to operating personnel. cone or flame holder.

514 The John Zink Combustion Handbook

17.15 PILOT BURNER FAILS TO Sometimes the pilot gas–air mixture leaving the tip is not
IGNITE OR EXTINGUISHES in the flammable region, usually too lean to support combus-
tion. Close the primary air door to enrich the mixture. The
WHILE IN SERVICE
pilot gas may be of the wrong composition or pressure to
induce adequate primary air flow. Check both the gas com-
17.15.1 Indications of the Problem position and the pilot fuel pressure against the manufacturer’s
Pilot burners are usually small premix gas burners designed to specifications and correct either if necessary.
ignite the main burner while releasing only a small amount of A natural draft premix pilot will not stay lit if the firebox
heat. Because pilot burners have small ports, they are suscepti- pressure at the pilot location is positive. Adjust dampers to
ble to plugging. Many users use a very clean, dedicated, and obtain a negative pressure at the burners. With forced draft
reliable pilot gas fuel to eliminate this problem. The pilot flame burners where the pressure at the pilot may be positive, it is
should remain stable and burn throughout the range of main good practice to provide a reliable pressurized source of com-
burner operation. It is intended to re-ignite the main burner bustion air or a special pilot design. Consult with the burner
when the main fuel flow is reestablished, before a potentially manufacturer for the correct pilot in this service.
damaging accumulation of combustible gases can occur.
The pilot may be wet, from condensed steam after firebox
In a problem situation, although the operator follows the purging, for example, and refuse to light. Allow the pilot to
appropriate purge and ignition procedures, the pilot burner may dry or dry it with the handheld torch before introducing fuel.
fail to light or may not continue to burn once lit. There may Wind may also make pilot lighting and continuous operation
be the loss of the pilot burner flame during operation. difficult. Consult with the burner manufacturer for a wind
resistant pilot tip.
17.15.2 Effect on Operation Manually light the pilot and observe the pilot flame for
If the pilot does not light upon heater startup, the correspond- stability. If the pilot flame appears to be too high above the
ing main burner is not placed in operation, or the main burner pilot tip, close the primary air door on the pilot mixer. If the
must be lit using a handheld torch. The latter is a less satis- flame appears to be burning with a yellow flame, open the
factory method because a large gas supply valve must be primary air door on the mixer. If a thermocouple or a flame
opened rather than the small pilot gas supply valve. If a mis- rectification rod is being used to monitor the pilot, the oper-
take is made, a relatively large amount of unburned gas enters ator needs to ensure these devices are providing correct infor-
the firebox, compared to the small amount from a pilot. This mation and are functioning properly.
large amount of unburned gas is an unsafe condition and has
the potential to form an explosive mixture within the firebox.
17.16 SMOKE EMISSION FROM
17.15.2 Corrective Action THE STACK
Ensure that the fuel gas is flowing to the pilot. If the fuel pip-
ing has been pressure tested or purged with an inert gas, this 17.16.1 Indication of the Problem
gas must be completely displaced with the fuel before a suc-
Smoke appears at the top of the stack.
cessful light-off can occur. Also check for closed valves or
blinds, plugging of the small pilot gas lines, or a plugged
strainer or filter. Since the ports in the pilot are small, they are
17.16.2 Effect on Operation and Equipment
also susceptible to plugging. Check for a plugged orifice and Smoke indicates either incomplete combustion or a process
clean it if it appears plugged. tube rupture if the process fluid is a hydrocarbon. As a regu-
Pilots are lit using a handheld torch or a spark ignitor. If lated emission, continued smoking could lead to sanctions,
the lighting torch flame is unstable or is not properly posi- fines, and termination of operations.
tioned, the torch flame might not contact the flammable mix-
ture leaving the pilot tip. Adjust the torch flame to ensure 17.16.3 Corrective Action
stability, and take care in the positioning of the flame relative If smoke appears while burning oil, suspect poor atomization.
to the pilot tip. If spark ignition is used, clean the spark plug Check the fuel oil temperature, the atomizing medium condi-
whenever possible and replace insulators on the ignition rod. tions, oil and atomizing medium pressures, and the condition
Before startup, check that the ignitor sparks and that unwanted of the oil gun. Refer to Sections 17.6 and 17.10. Consider a
electrical grounding is avoided. --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
switch to gas fuel until the problem is resolved.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 515 Tuesday, March 15, 2005 7:48 AM

Troubleshooting 515

Check the excess oxygen in the firebox to identify any

deficiency. Determine if there has been any change in fuel
composition. When operating at low excess air conditions, a
change in fuel gas composition — adding some higher
molecular weight gases — can initiate smoking because of
a lack of increase in air flow.
A process tube leak, spilling hydrocarbon contents into
the firebox, will create an incomplete combustion condition
and smoke at the stack. This situation requires a heater shut-
down and a steam purge into the firebox to cool any confla-
gration and to reduce the amount of air and the intensity of
combustion.

FIGURE 17.4 Effect of excess oxygen on NOx in raw

17.17 HIGH NOx EMISSIONS gas burners.

17.17.1 Indication of the Problem

The measured NOx concentration in the flue gas leaving the
stack exceeds the permitted operating levels.

17.17.2 Effect on Operation

Continued operation above permitted emission levels can
result in sanctions or loss of the operating permit.

17.17.3 Corrective Action

Thermal NOx formation can be significantly reduced by
burner technology, so reduced emissions may be obtained by
a change of burners. See Figure 17.8 for the effect of burner
model on NOx emissions. Since high flame temperatures
favor increased NOx formation, reducing the peak flame tem-
perature will reduce the rate of NOx production. The rate of
FIGURE 17.5 Effect of combustion air temperature on
NOx production is also a function of fuel composition. If the NOx.
fuel contains chemically-bound nitrogen, NOx emissions will
increase in proportion to nitrogen concentration. Figure 17.7
shows the effect of bound nitrogen, usually found in liquid
fuels, on NOx production. Figures 17.4 through 17.6 show
the effect of other variables — excess oxygen, combustion air
temperature, firebox temperature — on NOx formation.
To control the NOx properly the operator must first under-
stand the type and model of the burner installed in the heater
and the design features provided. Types of burners designed
NOx

to reduce NOx emissions include staged air burners, staged

fuel burners, Ultra LoNOX burners, and COOL technology
burners. Since heater operation may depend on low NOx emis-
sions, it is important that the burner manufacturer have exten-
sively tested the selected burner, have a reputation for 2400 2400 2400 2400 2400 2400
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

providing a durable product, and offer reliable field service. Firebox Temperature, ºF
The first thing to do if NOx emissions appear high is
check the analyzer. Like all instruments of this type, it can FIGURE 17.6 Effect of firebox temperature on NOx.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 516 Tuesday, March 15, 2005 7:48 AM

516 The John Zink Combustion Handbook

As firing rates increase, larger flames may result. The larger

flame may have a core that does not lose heat and cool as
rapidly compared to smaller flames typical of lower firing
rates. Some new technologies overcome this disadvantage of
larger flames.
Burners designed for lowered NOx production typically
require that the fuel, air, and any firebox gases or flue gases
injected into the flame envelope be well-mixed and held close
to the design proportions. Anything that will cause the burner
to operate outside of the design parameters will tend to
increase the production of NOx. Potential problems of this
sort and respective solutions are listed below.

• Partially plugged or cracked fuel tip — clean or replace

FIGURE 17.7 Effect of bound nitrogen in the liquid the tip.
fuel on NOx.
• Fuel tip oriented or positioned improperly — refer to the
manufacturer’s drawing and correct the positioning.
• Improper tips installed (usually during maintenance) —
refer to the manufacturer’s drawing for the proper tip
description and replace in accordance with the
recommended positioning.
• Damaged primary flame holder or stabilizer — replace
the damaged part.
• Damaged or improperly installed burner tile — replace
or properly reinstall the tile.
• Unstable flame — see information in this chapter for
corrective actions.
• Air passages partially plugged disrupting uniform air or
recirculation flows — remove the material plugging the
FIGURE 17.8 Effect of burner model on NOx. passages.
• Partially plugged or damaged flue gas orifices — clean
or replace as appropriate.
be out of calibration or the sampling system may not be • Temperature or pressure of the fuel or any injected
operating correctly. material not in accordance with the design — adjust the
operating parameters to the recommended values.
Next check the level of excess oxygen in the flue gas at
the firebox exit. If this level is above the target, follow the The pilot burner is not a low-NOx type of burner. If the
procedure to reduce excess oxygen. Seal all possible air leaks pilot is too large or unstable, it may make a relatively large
to make controlling the excess oxygen easier. Adjust the draft contribution to the measured NOx. Look at the possibility of
to the target value to minimize the amount of air leaking into safely reducing the pilot heat release.
the heater and the excess oxygen at the burners. Adjust reg- The adjustment of the registers to control and proportion
isters in accordance with the burner manufacturer’s recom- the combustion air varies with the type or model of low-NOx
mendations for the specific burner model. burner. Descriptions of register adjustment to reduce NOx
Check the fuel composition and the firing rate to identify follow. The fuel injection nozzle parameters — sizing, loca-
changes that may raise flame temperature and NOx emission. tion, orientation — are determined by the burner manufac-
Fuel gas components such as hydrogen burn with a high flame turer, and no adjustments should be made by the operator
temperature, potentially forming more NOx than a natural gas unless needed to conform to the manufacturer’s drawings.
flame. The concentration of chemically-bound nitrogen in fuel If the NOx emissions remain above the expected level after
oil typically increases with the oil density, i.e., heavy oils will following the directions in this section, consult with the
typically have more nitrogen than light oils. burner manufacturer for other possible solutions.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Troubleshooting 517

17.17.3.1 Staged Air Burner

The staged air burner is utilized for either gas or liquid fuel
firing. This type of burner normally has three air registers to
control the flow rate and distribution of combustion air
through the burner, and only one fuel injection nozzle. How-
ever, some designs have only two air registers.
If the burner has three air registers or dampers they are
termed the primary, secondary, and tertiary air registers
(dampers); each must be correctly adjusted to successfully
minimize NOx production. The primary air register is nor-
mally on burners that have liquid fuel firing capability, and it
controls the air flow to the oil tile. The secondary air register
controls the combustion air flow through the burner throat.
The tertiary air register controls the flow of combustion air
that bypasses the burner throat and is then mixed into the flame
in the secondary combustion zone of the burner. See the sketch
in Figure 17.9b for this type of burner.
To begin adjustment, the primary air register should be as
open as possible and still provide a stable flame. The second-
ary and tertiary air registers should each be approximately
50% open. Further adjustments of the secondary air register
position are made while referring to the oxygen analyzer
output. The NOx analyzer is used to adjust the tertiary air (a)
register to the optimum opening. If the NOx target cannot be
reduced with secondary and tertiary register adjustments, then
the primary air register must be closed and the secondary and
tertiary register adjustments repeated.
In staged air burner models with only two air registers, the
registers are labeled the secondary and tertiary air registers.
The oxygen analyzer will be used to guide adjustment of the
secondary air register, and the NOx analyzer output will guide
tertiary register positioning.

17.17.3.2 Staged Fuel Burner

The staged fuel burner is normally applied only with gas fuels
to reduce NOx formation. However, it can be designed to stage
gaseous fuels while having the independent capability of firing
liquid fuels with standard, not low-NOx, burner technology.
The staged gaseous fuel burner utilizes one combustion air (b)
register or damper to control the air flow across the burner and FIGURE 17.9 (a) Air staging. (b) Staged air burner.
two sets of fuel injection nozzles to stage the fuel flow into the
combustion reaction (see Figure 17.10a). The burner air regis-
ter or damper position is set with reference to the output of the The introduction of the largely inert flue gases into the
oxygen analyzer at the flue gas exit from the radiant section. combustion zones of the burner reduces the peak flame tem-
perature and NOx formation. The combination of staging the
17.17.3.3 Ultra Low-NOx Burner gaseous fuel and the injection of flue gases results in a burner
The Ultra Low-NOx burner utilizes the staged fuel concepts
with the currently lowest possible NOx.
and also inspirates flue gases from the radiant section into
the primary and secondary combustion reaction zones (see This burner type has one air register or damper to adjust
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Figure 17.11). the airflow across the burner. The burner has both primary
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 518 Tuesday, March 15, 2005 7:48 AM

518 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

(a) (a)

(b)
(b) FIGURE 17.11 InfurNOx burner technology.
FIGURE 17.10 Staged fuel burner.
17.17.3.4 COOL TECHNOLOGY Burner
and secondary fuel injection nozzles that act to inject fuel gas When inert components are mixed with the fuel gas, the flame
into the combustion zone together with flue gas from the temperature is lowered and the rate of NOx formation is
firebox. The burner air register or damper position is set with reduced. COOL TECHNOLOGY burners take advantage of this
reference to the output of the oxygen analyzer located at the and require that the amount of inert gas injected with the fuel
flue gas exit from the radiant section. gas be controlled within limits set by the burner manufacturer.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch17-Frame Page 519 Tuesday, March 15, 2005 7:48 AM

Troubleshooting 519

TABLE 17.2 Troubleshooting for Gas Burners

Trouble Causes Solutions
Burners go out Gas/air mixture too lean, or too much draft Reduce total air. Reduce primary air.
Flame flashback Low gas pressure Shut off burners to raise the fuel gas pressure to the operating burners;
reducing the burner orifice size can also be helpful.
High hydrogen concentration in fuel gas Reduce primary air; tape the primary air shut if flashback continues; a
new burner or tip drilling may be required.
Insufficient heat release Low gas flow Increase gas flow; increase burner tip orifice size; make sure that sufficient
air will be available for the increased fuel rate.
Desired heat release exceeds design capacity Larger burner tips or new burners may be required.
Pulsating fire or breathing (flame Lack of oxygen/draft Reduce firing rate immediately; establish complete combustion at lower
alternately ignites and goes out) rate; open stack damper and/or air registers to increase air and draft;
reduce fuel before increasing air.
Erratic flame Lack of combustion air Adjust air register and/or stack damper.
Incorrect position of burner tip Locate tips per manufacturer’s drawings.
Damaged burner block Repair burner block to manufacturer’s tolerances.
Gas flame too long Excessive firing Reduce firing rates.
Too little primary air Increase primary air; decrease secondary air.
Worn burner tip Replace tip.
Tip drilling angle too narrow Change to wide drilling angle tip.
Gas flame too short Too much primary air Increase secondary air; decrease primary air.
Tip drilling angle too wide Change to narrow drilling angle tip.

Check the ports on the injection device to ensure that they Tables 17.2 and 17.3 give summaries of the most frequently
are clear so that the inert gas is being added in the quantities occurring troubleshooting operations for gas- and oil-fired
designed by the burner manufacturer. Ensure that the damper burners, respectively.
on the inert gas supply plenum is open. If flue gas is the inert
material being injected, be sure that the flue gas headers are
properly insulated and that the fuel gas mixture is kept at the REFERENCES
proper operating conditions.
1. Burners for Fired Heaters in General Refinery Services,
COOL TECHNOLOGY burners have one air register or
API Publication 535, 1st ed., July 1995.
damper to adjust the airflow across the burner. The burner has
both primary and secondary fuel injection nozzles that act to 2. N.P. Lieberman, Troubleshooting Process Operations,
inject the fuel gas and inspirate firebox flue gases into the 3rd ed., Penn Well Publishing, Tulsa, OK, 1991.
combustion zone. The burner air register is set with reference 3. R.A. Meyers, Handbook of Petroleum Refining Pro-
to the output of the oxygen analyzer located at the flue gas cesses, 2nd ed., McGraw-Hill, New York, 1997.
exit from the radiant section. 4. W. Bartok, and A.F. Sarofim, Eds., Fossil Fuel Com-
bustion: A Source Book, John Wiley & Sons, 1991.
5. John Zink Co. LLC, John Zink Burner School Notes,
17.18 SUMMARY Tulsa, OK, Sept. 16-18, 1998.

In summary, troubleshooting burners in a furnace involves: 6. R.D. Reed, A New Approach to Design for Radiant
Heat Transfer in Process Work, Petroleum Engineer,
(1) observing the problem; (2) identifying the problem;
August, C-7–C-10, 1950.
(3) determining the effect on the operation of the furnace; and
(4) determining the solution and the corrective action that 7. Fired Heaters for General Refinery Services, API Stan-
should be taken to correct the problem. If there is a problem dard 560, second ed., Sept., 1995.
that cannot be identified and resolved, one should consult with 8. E.A. Barrington, Fired Process Heaters, course notes,
the burner manufacturer to obtain advice on how to proceed. 1999.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

520 The John Zink Combustion Handbook

TABLE 17.3 Troubleshooting for Oil Burners

Trouble Causes Solutions
Burners dripping; coke deposits Improper atomization due to high oil Check fuel oil type; increase fuel temperature to lower viscosity to proper
on burner blocks; coking of viscosity, clogging of burner tip, level; check composition for heavier fractions; clean or replace burner tip;
burner tip when firing fuel oil insufficient atomizing steam, improper confirm burner tip is in proper location; increase atomizing steam; place
only location of burner tip tip in location as per burner drawings.
Failure to maintain ignition Too much atomizing steam Reduce atomizing steam until ignition is stabilized; during start-up, have
atomizing steam on low side until ignition is well established.
Too much primary air at firing rates Reduce primary air.
Too much moisture in atomizing steam Ensure appropriate insulation is on steam lines; confirm steam traps are
functioning; adjust quality of atomizing steam to appropriate levels.
Coking of oil tip when firing oil High rate of gas with a low rate of oil; high Increase atomizing steam to produce sufficient cooling effect to avoid
in combination heat radiation to the fuel oil tip coking; reduce gas firing rate; dedicate individual burners to either fuel.
Incorrect oil gun position Place tip in locations as per burner drawings. If this fails, readjust burner
tip ±0.5 in. until coking ceases.
Erratic flame Lack of combustion air Adjust air damper or register.
Plugged burner gun Clean burner gun.
Worn burner gun Replace burner gun.
High rate of gas firing while firing a low Reduce gas rate; dedicate burners to either fuel.
rate of oil
Damaged burner block Repair burner block.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Excess smoke at stack Insufficient atomizing steam Increase atomizing steam.

Moisture in atomizing steam Requires knockout drum or increase in superheat, check steam line insulation.
Low excess air Increase excess air.

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 18
Duct Burners
Peter F. Barry and Stephen L. Somers

TABLE OF CONTENTS

18.1 Introduction............................................................................................................................................. 524

18.2 Applications ............................................................................................................................................ 524
18.2.1 Cogeneration ............................................................................................................................. 524
18.2.2 Air Heating ............................................................................................................................... 524
18.2.3 Fume Incineration ..................................................................................................................... 526
18.2.4 Stack Gas Reheat ...................................................................................................................... 526
18.3 Burner Technology.................................................................................................................................. 526
18.3.1 In-duct or Inline Configuration ................................................................................................. 526
18.3.2 Grid Configuration (Gas Firing) ............................................................................................... 527
18.3.3 Grid Configuration (Liquid Firing)........................................................................................... 527
18.3.4 Design Considerations .............................................................................................................. 528
18.3.5 Maintenance.............................................................................................................................. 539
18.3.6 Accessories ............................................................................................................................... 540
18.3.7 Design Guidelines and Codes ................................................................................................... 540
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

523
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 524 Wednesday, March 23, 2005 7:27 AM

524 The John Zink Combustion Handbook

18.1 INTRODUCTION all fuels suitable for the engine/turbine, as well as many that
Linear and in-duct burners were used for many years to heat are not, including heavy oils and waste gases.
air in drying operations before their general use in co- Heat recovery for large systems is usually accomplished
generation systems. Some of the earliest systems premixed by convective heat transfer in a boiler (commonly referred to
fuel and air in an often complicated configuration that fired as a heat recovery steam generator, also known by the acro-
into a recirculating process air stream. The first uses in high nym HRSG). Smaller systems utilize either a steam or hot
temperature, depleted oxygen streams downstream of gas water boiler, or, alternatively, some type of air-to-air heat
turbines in the early 1960s provided additional steam for exchanger or direct transfer to a process.
process use in industrial applications and for electrical peak- Supplementary firing is often incorporated into the
ing plants operating steam turbines. As gas turbines have boiler/HRSG design as it allows increased production of
become larger and more efficient, duct burner supplemental steam as demanded by the process. The device that provides
heat input has increased correspondingly. the supplementary firing is a duct burner, so called because
it is installed in the duct connecting the engine/turbine exhaust
Linear burners are applied where it is desired to spread heat
to the heat recovery device, or just downstream of a section
uniformly across a duct, whether in ambient air or oxygen-
of the HRSG superheater (see Figures 18.4 and 18.5). Oxygen
depleted streams. In-duct designs are more commonly used
required for the combustion process is provided by the turbine
in fluidized bed boilers and small cogeneration systems.
exhaust gas (TEG).

18.2.1.2 Combined Cycle

18.2 APPLICATIONS Combined cycle systems incorporate all components of the
simple cycle configuration with the addition of a steam
18.2.1 Cogeneration turbine/generator set powered by the HRSG. This
18.2.1.1 Introduction arrangement is attractive when the plant cannot be located
near an economically viable steam user. Also, when used in
Cogeneration implies simultaneous production of two or
conjunction with a duct burner, the steam turbine/generator
more forms of energy, most commonly electrical (electric
can provide additional power during periods of high or
power), thermal (steam, heat transfer fluid, or hot water), and
“peak” demand.
pressure (compressor). The basic process involves combus-
tion of fossil fuel in an engine (reciprocating or turbine) that
drives an electric generator, coupled with a recovery device 18.2.2 Air Heating
that converts heat from the engine exhaust into a usable Duct burners are suitable for a wide variety of direct-fired air
energy form. Production of recovered energy can be heating applications where the physical arrangement requires
increased independently of the engine through supplemen- mounting inside a duct, and particularly for processes where
tary firing provided by a special burner type known as a duct the combustion air is at an elevated temperature and/or
burner. Most modern systems will also include flue gas emis- contains less than 21% oxygen. Examples include:
sion control devices. A typical plant schematic is shown in
• Fluidized bed boilers (see Figure 18.6): where burners
Figure 18.1. Aerial views of typical combined cycle electric
are installed in combustion air ducts and used only to
power plants are shown in Figure 18.2 and Figure 18.3. provide heat to the bed during startup. At cold
Reciprocating engines (typically diesel cycle) are used in conditions, the burner is fired at maximum capacity with
smaller systems (10 MW and lower) and offer the advantage fresh ambient air; but as combustion develops in the bed,
of lower capital and maintenance costs but produce relatively cross-exchange with hot stack gas increases the air
high levels of pollutants. Turbine engines are used in both temperature and velocity. Burners are shut off when the
small and large systems (3MW and above) and, although desired air preheat is reached and the bed can sustain
more expensive, generally emit lower levels of air pollutants. combustion unaided.
• Combustion air blower inlet preheat: where burners are
Fossil fuels used in cogeneration systems can consist of
mounted upstream of a blower inlet to protect against
almost any liquid or gaseous hydrocarbon, although natural thermal shock caused by ambient air in extremely cold
gas and various commercial-grade fuel oils are most com- climates (–40°F/°C and below). This arrangement is
monly used. Mixtures of hydrocarbon gases and hydrogen only suitable when the air will be used in a combustion
found in plant fuel systems are often used in refining and process as it will contain combustion products from the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

petrochemical applications. Duct burners are capable of firing duct burner.

Duct Burners
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- 525

FIGURE 18.1 Typical plant schematic.

526 The John Zink Combustion Handbook

FIGURE 18.2 Cogeneration at Teesside, England. (Courtesy of Nooter/Eriksen.)

• Drying applications: where isolation of combustion reduce or eliminate potentially corrosive condensation inside
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

products from the work material is not required, such as the stack. A source of ambient augmenting combustion air is
certain paper and wallboard manufacturing operations. often added if the stack gas oxygen concentration is low. This
arrangement may also provide a corollary emissions reduction
18.2.3 Fume Incineration benefit (see Section 18.2.3).
Burners are mounted inside ducts or stacks carrying exhaust
streams primarily composed of air with varying concentrations of
organic contaminants. Undesirable components are destroyed, 18.3 BURNER TECHNOLOGY
both by an increase in the gas stream bulk temperature and
through contact with localized high temperatures created in 18.3.1 In-duct or Inline Configuration
the flame envelope. Particular advantages of the duct burner Register or axial flow burner designs are adapted for
include higher thermal efficiency as no outside air is used, installation inside a duct. The burner head is oriented such that
lower operating cost as no blower is required, and improved the flame will be parallel to and co-flow with the air or TEG
destruction efficiency resulting from distribution of the flame stream, and the fuel supply piping is fed through the duct side
across the duct section with grid-type design. wall, turning 90° as it enters the burner (see Figure 18.7).
Depending on the total firing rate and duct size, one burner
18.2.4 Stack Gas Reheat may be sufficient, or several may be arrayed across the duct
Mounted at or near the base of a stack, heat added by a duct cross-section. Inline burners typically require more air/TEG
burner will increase natural draft, possibly eliminating a need pressure drop, produce longer flames, and offer a less uniform
for induced draft or eductor fans. In streams containing a large heat distribution than grid-type. On the other hand, they are
concentration of water vapor, the additional heat can also more flexible in burning liquid fuels, can be more easily
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 527 Wednesday, March 23, 2005 7:27 AM

Duct Burners 527

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 18.3 Combination (oil and gas) fired duct burners at Dahbol, India. (Courtesy of Enron.)

modified to incorporate augmenting air, and sometimes 18.3.3 Grid Configuration (Liquid Firing)
represent a less expensive option for high firing rates in small
ducts without sufficient room for grid elements. As with the gas-fired arrangement, a series of linear burner
elements comprised of a pipe and flame holders (wings)
span the duct width. However, instead of multiple discharge
18.3.2 Grid Configuration (Gas Firing)
points along the pipe length, liquid fuel is injected down-
A series of linear burner elements that span the duct width are stream of the element through the duct sidewall, and directed
spaced at vertical intervals to form a grid. Each element is parallel to the flame holders (cross-flow to the TEG). This
comprised of a fuel manifold pipe fitted with a series of flame
configuration utilizes the duct cross-section for containment
holders (or wings) along its length. Fuel is fed into one end of
of the flame length, thus allowing a shorter distance between
the manifold pipe and discharged through discrete multi-port
the burner and downstream boiler tubes (see Figure 18.10).
tips attached at intervals along its length, or through holes
The injection device, referred to as a side-fired oil gun, uti-
drilled directly into the pipe. Gas ports are positioned such
that fuel is injected in co-flow with the TEG. The wings meter lizes a mechanical nozzle supplemented by low-pressure air
the TEG or air flow into the flame zone, thus developing eddy (2 to 8 psi) to break the liquid fuel into small droplets (atom-
currents that anchor ignition. They also shield the flame in ization) that will vaporize and readily burn. Although most
order to maintain suitably high flame temperatures, thereby commonly used for light fuels, this arrangement is also suit-
preventing excessive flame cooling that might cause high able for some heavier fuels where the viscosity can be low-
emissions. Parts exposed to TEG and the flame zone are typi- ered by heating. In some cases, high pressure steam may be
cally of high-temperature alloy construction (see Figures 18.8 required, instead of low-pressure air, for adequate atomiza-
and 18.9). tion of heavy fuels.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 528 Wednesday, March 23, 2005 7:27 AM

528 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 18.4 Typical location of duct burners in an HRSG. (Courtesy of Deltak.)

18.3.4 Design Considerations contain high concentrations of unsaturated hydrocarbons with

a tendency to condense and/or decompose inside burner
18.3.4.1 Fuels
piping. The location of burner elements inside the TEG duct,
18.3.4.1.1 Natural Gas
surrounded by high-temperature gases, exacerbates the
Natural gas is by far the most commonly used fuel because it problem. Plugging and failure of injection nozzles can occur,
is readily available in large volumes throughout much of the with a corresponding decrease in online availability and an
industrialized world. Because of its ubiquity, its combustion increase in maintenance costs.
characteristics are well-understood, and most burner designs With appropriate modifications, however, duct burners can
are developed for this fuel. function reliably with most hydrocarbon-based gaseous fuels.
Design techniques include insulation of burner element mani-
18.3.4.1.2 Refinery/Chemical Plant Fuels
folds, insulation and heat tracing of external headers and pipe
Refineries and chemical plants are large consumers of both trains, and fuel/steam blending. Steam can also be used to
electrical and steam power, which makes them ideal candidates periodically purge the burner elements of solid deposits
for cogeneration. In addition, these plants maintain extensive before plugging occurs.
fuel systems to supply the various direct- and indirect-fired
processes, as well as to make the most economical use of 18.3.4.1.3 Low Heating Value
residual products. This latter purpose presents special By-product gases produced in various industrial processes
challenges for duct burners because the available fuels often such as blast furnaces, coke ovens, and flexicokers, or from
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 529 Wednesday, March 23, 2005 7:27 AM

Duct Burners 529

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 18.5 Schematic of HRSG at Teesside, England. (Courtesy of Nooter/Eriksen.)

mature landfills, contain combustible compounds along with

significant concentrations of inert components, thus resulting
in relatively low heating values (range of 50 to 500 Btu/scf).
These fuels burn more slowly and at lower temperatures than
conventional fuels, and thus require special design
considerations. Fuel pressure is reduced to match its velocity
to flame speed, and some form of shield or “canister” is
employed to provide a protected flame zone with sufficient
residence time to promote complete combustion before the
flame is exposed to the quenching effects of TEG.
Other considerations that must be taken into account are
moisture content and particulate loading. High moisture
concentration results in condensation within the fuel supply
system, which in turn produces corrosion and plugging. Pilots
and ignitors are particularly susceptible to the effects of mois- FIGURE 18.6 Fluidized bed startup duct burner.
ture because of small fuel port sizes, small igniter gap tolerance,
and the insulation integrity required to prevent “shorting” of
Solid particulates can cause plugging in gas tip ports or other
electrical components. A well-designed system might include
a knock-out drum to remove liquids and solids, insulation and fuel system components and should therefore be removed to
heat-tracing of piping to prevent or minimize condensation, and the maximum practical extent. In general, particle size should
low-point drains to remove condensed liquids. Problems are be no greater than 25% of the smallest port, and overall loading
usually most evident after a prolonged period of shutdown. should be no greater than 5 ppm by volume (weight).
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 530 Wednesday, March 23, 2005 7:27 AM

530 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 18.7 An inline burner.

FIGURE 18.8 Linear burner elements.

Duct Burners 531

FIGURE 18.9 Gas flame from a grid burner.

FIGURE 18.10 Oil flame from a side-fired oil gun.

18.3.4.1.4 Liquid Fuels is provided by the residual in the turbine exhaust gas instead of
In cogeneration applications, duct burners are commonly from a new, external source of air. Because this gas is already
fired with the same fuel as the turbine, which is typically at an elevated temperature, duct burner thermal efficiency can
limited to light oils such as No. 2 or naphtha. For other exceed 90% as very little heat is required to raise the
applications, specially modified side-fired guns or an inline combustion products temperature to the final fired temperature.
design can be employed to burn heavier oils such as No. 6 TEG contains less oxygen than fresh air, typically between 11
and some waste fuels. and 16% by volume, which, in conjunction with the TEG
temperature, will have a significant effect on the combustion
18.3.4.2 Combustion Air and Turbine Exhaust Gas process. As the oxygen concentration and TEG temperature
18.3.4.2.1 Temperature and Composition become lower, emissions of CO and unburned hydrocarbons
When used for supplementary firing in HRSG cogeneration occur more readily, eventually progressing to combustion
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

applications, the oxygen required for the combustion reaction instability. The effect of low oxygen concentration can be
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 532 Wednesday, March 23, 2005 7:27 AM

532 The John Zink Combustion Handbook

FIGURE 18.11 Approximate requirement for augmenting air.

partially offset by higher temperatures; conversely, higher Grid burners are designed to distribute heat uniformly
oxygen concentrations will partially offset the detrimental across the HRSG or boiler tube bank, and thus require a
effects of low TEG temperatures. This relationship is depicted reasonably uniform distribution of TEG or air to supply the
graphically in Figure 18.11. Duct burner emissions are fuel with oxygen. Inadequate distribution causes localized
discussed in more detail elsewhere in this chapter. areas of low velocity, resulting in poor flame definition along
with high emissions of CO and unburned hydrocarbons. Tur-
18.3.4.2.2 Turbine Power Augmentation bine exhaust flow patterns, combined with rapidly diverging
During periods of high electrical demand, various tech- downstream duct geometry, will almost always produce an
niques are employed to increase power output, and most will unsatisfactory result that must be corrected by means of a
increase the concentration of water vapor in TEG. The corre- straightening device. Likewise, the manner in which ambient
sponding effect is a reduction in TEG oxygen concentration air is introduced into a duct can also result in flow mal-
and temperature with consequent effects on duct burner distribution, requiring some level of correction. Selection and
combustion. Depending on the amount of water vapor used, design of flow-straightening devices are discussed elsewhere
CO emissions may simply rise, or in extreme cases the flame in this chapter (see Figure 18.12).
may become unstable. The former effect can be addressed In instances where bulk TEG or air velocity is lower than
with an allowance in the facility operating permit or by required for proper burner operation, flow straightening alone
increasing the amount of CO catalyst in systems so is not sufficient and it becomes necessary to restrict a portion
equipped. The latter requires air augmentation, a process of the duct cross-section at or near the plane of the burner
whereby fresh air is injected at a rate sufficient to raise the elements, thereby increasing the “local” velocity across
TEG oxygen concentration to a suitable level. flame holders. This restriction, also referred to as blockage,
commonly consists of unfired runners or similar shapes uni-
18.3.4.2.3 Velocity and Distribution formly distributed between the firing runners to reduce the
Regardless of whether TEG or fresh air is used, velocity across open flow area.
flame stabilizers must be sufficient to promote mixing of the Inline or register burners inject fuel in only a few (or
fuel and oxygen, but not so great as to prevent the flame from possibly only one) positions inside the duct, and can therefore
anchoring to the burner. Grid-type configurations can generally be positioned in an area of favorable flow conditions, assum-
operate at velocities ranging from 20 to 90 fps (feet per ing the flow profile is known. On the other hand, downstream
second) and pressure drops of less than 0.5 in. water column. heat distribution is less uniform than with grid designs, and
Inline or register burners typically require velocities of 100 to flames may be longer. As with grid-type burners, in some
150 fps with a pressure drop of 2 to 6 in. water column.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
cases it may be necessary to block portions of the duct at or
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 533 Wednesday, March 23, 2005 7:27 AM

Duct Burners 533

just upstream of the burners to force a sufficient quantity of

TEG or air through the burner.

18.3.4.2.4 Ambient Air Firing (Air-only Systems and

HRSG Backup)
Velocity and distribution requirements for air-systems are
similar to those for TEG, although inlet temperature is not a
concern because of the relatively higher oxygen concentration.
As with TEG applications, the burner elements are exposed to
the products of combustion, so material selection must take
into account the maximum expected fired temperature.
Ambient (or fresh) air backup for HRSGs presents special
design challenges. Because of the temperature difference
between ambient air and TEG, designing for the same mass
flow and fired temperature will result in velocity across the
burner approximately one third that of the TEG case. If the FIGURE 18.12 Duct burner arrangement.
cold condition velocity is outside the acceptable range, it
will be necessary to add blockage, as described elsewhere
in this chapter. Fuel input capacity must also be increased
With their research and development facilities, manufactur-
to provide heat required to raise the air from ambient to the
ers have defined the oxygen requirement with respect to TEG
design firing temperature. By far the most difficult challenge
temperature and fuel composition, and are able to quantify the
is related to flow distribution. Regardless of the manner in
amount of augmenting air required under most conditions
which backup air is fed into the duct, a flow profile different
likely to be encountered. It is usually not practical to add
from that produced by the TEG is virtually certain. Flow-
enough air to the turbine exhaust to increase the oxygen con-
straightening devices can therefore not be optimized for
tent to an adequate level. Specially designed runners are there-
either case, but instead require a compromise design that
fore used to increase the local oxygen concentration. In cases
provides acceptable results for both. If the two flow patterns
where augmenting air is required, the flow may be substantial:
are radically different, it may ultimately be necessary to alter
from 30 to 100% of the theoretical air required for the sup-
the air injection arrangement independently of the TEG duct-
plemental fuel.
straightening device.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

The augmenting air runner of one manufacturer consists of a

18.3.4.3 Augmenting Air graduated air delivery tube designed to ensure a constant velocity
As turbines have become more efficient and more work is across the length of the tube. Equal distribution of augmenting
extracted in the form of, for example, electricity, the oxygen air across the face of the tube is imperative. The augmenting air
level available in the TEG continues to get lower. To some is discharged from the tube into a plenum and passes through a
extent, a correspondingly higher TEG temperature provides second distribution grid to further equalize flow. The air passes
some relief for duct burner operation. through perforations in the flame holder, where it is intimately
mixed with the fuel in the primary combustion zone. This inti-
In some applications, however, an additional oxygen source
mate mixing ensures corresponding low CO and UHC emissions
may be required to augment that available in the TEG when
under most conditions likely to be encountered. Once the deci-
the oxygen content in the TEG is not sufficient for combustion
sion has been made to supply augmenting air to a burner, it is
at the available TEG temperature. If the mixture adiabatic
an inevitable result of the design that the augmenting air will be
flame temperature is not high enough to sustain a robust flame
part of the normal operating regime of the combustion runner.
in the highly turbulent stream, the flame may become unstable.
The problem can be exacerbated when the turbine manu-
facturer adds large quantities of steam or water for NOx 18.3.4.4 Equipment Configuration and
control and power augmentation. A corresponding drop in the TEG/Combustion Air Flow Straightening
TEG temperature and oxygen concentration occurs because The turbine exhaust gas/combustion air velocity profile at the
of dilution. The TEG temperature is also reduced in installa- duct burner plane must be within certain limits to ensure good
tions where the HRSG manufacturer splits the steam super- combustion efficiency; in cogeneration applications, this is
heater and places tubes upstream of the duct burner. rarely achieved without flow-straightening devices. Even in
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 534 Wednesday, March 23, 2005 7:27 AM

534 The John Zink Combustion Handbook

either upstream of the first tube bank or between the first tube
bank and (upstream of) the burner. Although not very com-
mon, some HRSG design configurations utilize two stages of
duct burners with heat transfer tube banks in between, and a
flow-straightening device upstream of the first burner. Such
an arrangement is, however, problematic because the TEG
downstream of the first-stage burner may not have the
required combination of oxygen and temperature properties
required for proper operation of the second-stage burner.
Perforated plates that extend across the entire duct cross-
section are most commonly used for flow straightening
because experience has shown they are less prone to mechan-
ical failure than vane-type devices, even though they require
a relatively high pressure drop. The pattern and size of
perforations can be varied to achieve the desired distribution.
Vanes can produce comparable results with significantly less
pressure loss but require substantial structural reinforcement
to withstand the flow-induced vibration inherent in HRSG
systems. Regardless of the method used, flow pattern com-
FIGURE 18.13 Comparison of flow variation with and plexity — particularly in TEG applications — usually dictates
without straightening device.
the use of either physical or computational fluid dynamic
(CFD) modeling for design optimization.

18.3.4.4.1 Physical Modeling

TEG/air flow patterns are determined by inlet flow character-
istics and duct geometry, and are subject to both position and
time variation. Design of an efficient (low pressure loss) flow-
straightening device is therefore not a trivial exercise, and
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

manual computational methods are impractical. For this rea-

son, physical models, commonly 1:6 or 1:10 scale, are con-
structed, and flow characteristics are analyzed by flowing air
with smoke tracers or water with polymer beads through the
model (see Figure 18.14). Although this method produces
reliable results, tests conducted at ambient conditions (known
as “cold flow”) are not capable of simulating the buoyant
effects that may occur at elevated temperatures.

18.3.4.4.2 Computational Fluid Dynamic (CFD) Modeling

FIGURE 18.14 Physical model of burner.
Flow modeling with CFD, using a computer-generated draw-
ing of the inlet duct geometry, is capable of predicting flow
non-fired configurations, it may be necessary to alter the pattern and pressure drop in the turbine exhaust flow path. The
velocity distribution to make efficient use of boiler heat model can account for swirl flow in three dimensions, accu-
transfer surface. Figure 18.13 shows a comparison of flow rately predict pressure drop, and subsequently help design a
suitable device to provide uniform flow. The CFD model must
variation with and without flow straightening.
be quite detailed to calculate flow patterns incident and
Duct burners are commonly mounted in the TEG duct through a perforated grid or tube bank while also keeping the
upstream of the first bank of heat transfer tubes, or they may overall model solution within reasonable computation time.
be nested in the boiler superheater between banks of tubes. Combustion effects can be included in the calculations at the
In the former case, a straightening device would be mounted cost of increased computation time. Figure 18.15 shows a sam-
just upstream of the burner, while in the latter it is mounted ple result of CFD modeling performed on a HRSG inlet duct.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 535 Wednesday, March 23, 2005 7:27 AM

Duct Burners 535

FIGURE 18.15 Sample result of CFD modeling performed on an HRSG inlet duct.

18.3.4.5 Wing Geometry: Variations from the turbulence in the exhaust gas duct. The low-pressure
18.3.4.5.1 Flameholders zone pulls the flame back onto the manifold. This low-cost
Design of the flame stabilizer, or flameholder, is critical to runner may overheat the manifold, causing distortion of the
the success of supplementary firing. Effective emission metallic parts. Emissions are unpredictable with changing
control requires that the TEG be metered into the flame geometry and CO is usually much higher than the current
zone in the required ratio to create a combustible mixture typically permitted levels of under 0.1 lb/MMBtu.
and ensure that the combustion products do not escape
before the reactions are completed. In response to new 18.3.4.5.3 Low Emissions Design
turbine and HRSG design requirements, each duct burner Modifications to the design for lower emission performance
manufacturer has proprietary designs developed to provide generally have a larger cross-section in the plane normal to
the desired results. the exhaust flow. The increased blocked area protects the fuel
injection zone and increases residence time. The NOx is
18.3.4.5.2 Basic Flameholder reduced by the oxygen-depleted TEG and the CO/UHC is
In its basic form, a fuel injection system and a zone for reduced by the delayed quenching. The correct flow rate of
mixing with oxidant are all that is required for combustion. TEG is metered through the orifices in the flameholder, and
For application to supplemental firing, the simple design the fuel injection velocity and direction are designed to
shown in Figure 18.16 consists of an internal manifold or enhance combustion efficiency. The flame zone is pushed
“runner,” usually an alloy pipe with fuel injection orifices away from the internal manifold (“runner” pipe), creating
spaced along the length. A bluff body plate, with or without space for cooling TEG to bathe the runner and flameholder
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
perforations, is attached to the pipe to protect the flame zone and enhance equipment life.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 536 Wednesday, March 23, 2005 7:27 AM

536 The John Zink Combustion Handbook

FIGURE 18.16 Drilled pipe duct burner.

hydrogen, may not receive the cooling needed to protect the

metal from oxidation. Alternately, fuels subject to cracking,
such as propylene, may not have the oxygen needed to min-
imize coke buildup.
Another manufacturer supplies custom designs to accom-
modate velocity extremes, while maintaining low emissions.
In the design shown in Figure 18.17, the flameholder is
optimized with CFD and research experimentation to
enhance mixing and recirculation rate. Special construction
materials are easily accommodated. This supplier also uses
removable fuel tips with multiple orifices, which can be
customized to counteract any unexpected TEG flow distri-
bution discovered after commercial operation. Figure 18.18
depicts the flow patterns of air/TEG and fuel in relation to
FIGURE 18.17 Low emission duct burner.
the duct burner flameholder.

Each manufacturer approaches the geometry somewhat dif- 18.3.4.6 Emissions

ferently. One manufacturer uses cast alloy pieces welded 18.3.4.6.1 NOx and NO vs. NO2
together to provide the required blockage. These standard Formation of NO and NO2 is the subject of ongoing research
pieces often add significant weight and are difficult to cus-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
to understand the complex reactions (see Chapter 6).
tomize to specific applications. Hot burning fuels, such as Potentially, several oxides of nitrogen (NOx) can be formed
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 537 Wednesday, March 23, 2005 7:27 AM

Duct Burners 537

NO + HO 2 → NO 2 + OH

if the flame is rapidly quenched. This quenching can occur

because of the large quantity of excess TEG commonly
present in duct burner applications. Conversion to NO2 may
be even higher at fuel turndown conditions where the flame is
smaller and colder. NO2 formed in this manner can contribute
to “brown plume” problems and may even convert some of

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
the turbine exhaust NO to NO2.
There are two principle mechanisms in which nitrogen
oxides are formed.
1. Thermal NOx: The primary method is thermal oxidation
of atmospheric nitrogen in the TEG. NOx formed in this
way is called thermal NOx. As the temperature increases
in the combustion zone and surrounding environment,
increased amounts of N2 from the TEG are converted to
NO. Thermal NOx formation is most predominant in the
peak temperature zones of the flame.
1. Fuel-bound nitrogen NOx: The secondary method utilized
to form NOx is the reaction of oxygen with chemically
bound nitrogen compounds contained in the fuel. NOx
formed in this manner is called fuel NOx. Large amounts
of NOx can be formed by fuels that contain molecularly
bound nitrogen (e.g., amines and mercaptans). If a gas-
eous fuel such as natural gas contains diluent N2, it simply
behaves as atmospheric nitrogen and will form NOx only
if it disassociates in the high-temperature areas. However,
FIGURE 18.18 Flow patterns around flame stabilizer.
if the gaseous fuel contains, for example, ammonia (NH3),
this nitrogen is considered bound. In the low concentra-
tions typically found in gaseous fuels, the conversion to
during the combustion process, but only nitric oxide (NO)
NOx is close to 100% and can have a major impact on
and nitrogen dioxide (NO2) occur in significant quantities. NOx emissions.
NO is colorless and NO2 has a reddish-brown color.
Bound nitrogen in liquid fuel is contained in the long
In the elevated temperatures found in the flame zone in a carbon chain molecules. Distillate oil is the most common oil
typical HRSG turbine exhaust duct, NO formation is favored fired in duct burners as a liquid fuel. The fuel-bound nitrogen
almost exclusively over NO2 formation. Turbine exhaust NOx content is usually low, in the range of 0.05 weight percent.
is typically 95% NO and 5% NO2. In the high-temperature Conversion to NOx is believed to be 80 to 90%. For No. 6
zone, NO2 dissociates to NO by the mechanism of: oil, containing 0.30 weight percent nitrogen, the conversion
rate to NOx would be about 50%. Other heavy waste oils or
waste gases with high concentrations of various nitrogen com-
NO 2 + O + Heat → NO + O 2
pounds may add relatively high emissions. Consequently, fuel
NOx can be a major source of nitrogen oxides and may
However, after the TEG exits the hot zone and enters the predominate over thermal NOx.
cooling zone at the boiler tubes, reaction slows and the NO2 is The impact of temperature on NOx production in duct
essentially fixed. At the stack outlet, the entrained NO is slowly burners is not as pronounced as in, for example, fired heaters
oxidized to NO2 through a complex photochemical reaction or package boilers. One reason is that both the bulk fired
with atmospheric oxygen. The plume will be colorless unless temperature and the adiabatic flame temperature are lower
the NO2 increases to about 15 ppm, at which time a yellowish than in fired process equipment.
tint is visible. Care must be taken in duct burner design because When used to provide supplementary firing of turbine
NO can also be oxidized to NO2 in the immediate post-flame exhaust, duct burners are generally considered to be “low NOx”
region by reactions with hydroperoxyl radicals: burners. Because the turbine exhaust contains reduced oxygen,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 538 Wednesday, March 23, 2005 7:27 AM

538 The John Zink Combustion Handbook

FIGURE 18.19 Effect of conditions on CO formation.

the peak flame temperature is reduced and the reaction speed control or operation at partial load, but the primary concern is
for O2 and N+ to form NOx is thus lowered. The burners also the sometimes large CO contribution from supplementary fir-
fire into much lower average bulk temperatures — usually ing. The same low-temperature combustion environment that
less than 1600°F (870°C) — than process burners or fired suppresses NOx formation is obviously unfavorable for com-
boilers. The high-temperature zones in the duct burner flames plete oxidation of CO to CO2. Increased CO is produced
are smaller due to large amounts of flame quenching by the when fuels are combusted under fuel-rich conditions or when
excess TEG. Finally, mixing is rapid and therefore retention a flame is quenched before complete burnout. These condi-
time in the high-temperature zone is very brief. tions (see Figure 18.19) can occur if there is poor distribution
The same duct burner, when used to heat atmospheric air, of TEG to the duct burner, which causes some burner ele-
is no longer considered “low NOx,” because the peak flame ments to fire fuel-rich and others to fire fuel-lean, depending
temperature approaches the adiabatic flame temperature in air. on the efficiency of the TEG distribution device. The factors
Clearly, operating conditions have a major impact on NO affecting CO emissions include:
formation during combustion. To properly assess NOx pro-
duction levels, the overall operating regime must be consid- • turbine exhaust gas distribution
ered, including TEG composition, fuel composition, duct • low TEG approach temperature
firing temperature, and TEG flow distribution. • low TEG oxygen content
• flame quench on “cold” screen tubes
18.3.4.6.2 Visible Plumes • improperly designed flame holders that allow flame
Stack plumes are caused by moisture and impurities in the quench by relatively cold TEG
exhaust. Emitted NO is colorless and odorless, and NO2 is • steam or water injection
brownish in color. If the NO2 level in the flue gas exceeds
about 15 to 20 ppm, the plume will take on a brownish haze. Unburned hydrocarbons (UHCs): In the same fashion
NOx also reacts with water vapor to form nitrous and nitric as carbon monoxide generation, unburned hydrocarbons
acids. Sulfur in the fuel may oxidize to SO3 and condense in (UHCs) are formed in the exhaust gas when fuel is burned
the stack effluent, causing a more persistent white plume without sufficient oxygen, or if the flame is quenched before
combustion is complete. UHCs can consist of hydrocarbons
18.3.4.6.3 CO, VOC, SOx, and Particulates
(defined as any carbon-hydrogen molecule) of one carbon or
Carbon monoxide: Carbon monoxide (CO), a product of
multiple carbon atoms. The multiple carbon molecules are
incomplete combustion, has become a major permitting concern
often referred to as long-chain hydrocarbons. Unburned
in gas turbine-based cogeneration plants. Generally, CO emis-
hydrocarbons are generally classified in two groups:
sions from modern industrial and aero-derivative gas turbines
are very low, in the range of a few parts per million (ppm). There 1. unburned hydrocarbons as methane
are occasional situations in which CO emissions from the tur- 2. non-methane hydrocarbons or volatile organic compounds
bine increase due to high rates of water injection for NOx
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
(VOCs)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 539 Wednesday, March 23, 2005 7:27 AM

Duct Burners 539

The reason for the distinction and greater concern for VOCs is TABLE 18.1 Typical NOx and CO Emissions From Duct
that longer chain hydrocarbons play a greater role in the forma- Burners
tion of photochemical smog. VOCs are usually defined as mole- NOx CO
cules of two carbons or greater, and are sometimes considered Gas (lb/106 Btu fired) (lb/106 Btu fired)
to be three carbons or greater. These definitions are set by local Natural gas 0.1 0.08
air quality control boards and vary across the United States. Hydrogen gas 0.15 0.00
Refinery gas 0.1–0.15 0.03–0.08
UHCs can only be eliminated by correct combustion of the Plant gas 0.11 0.04–0.01
fuel. However, hydrocarbon compounds will always be Flexicoker gas 0.08 0.01
present in trace quantities, regardless of how the HRSG sys- Blast furnace gas 0.03–0.05 0.12
Producer gas 0.05–0.1 0.08
tem is operated. Syn fuels 0.08–0.12 0.08
Sulfur dioxide: Sulfur dioxide (SO2) is a colorless gas Propane 0.14 0.14
Butane 0.14 0.14
that has a characteristic smell in concentrations as low as
1 ppm. SO2 is formed when sulfur (S) in the fuel combines Note: NOx emissions from butane and propane can be modified by direct
steam injection into a gas or burner flame. CO emissions are highly depen-
with oxygen (O2) in the TEG. If oxygen is present (from
dent on TEG approach temperature and HRSG fired temperature.
excess of combustion) and the temperature is correct, the
sulfur will further combine and be converted to sulfur trioxide
(SO3). These oxides of sulfur are collectively known as SOx. sometimes expose parts to excessively high temperatures,
Except for sulfur compounds present in the incoming par- which results in wing warpage and oxidation failure.
ticulate matter, all of the sulfur contained in the fuel is con- 3. Fuel quality/composition: Some refinery fuels or waste
verted to SO2 or SO3. Sulfur dioxide will pass through the fuels contain unsaturated components and/or liquid carry-
boiler system to eventually form the familiar “acid rain” over. Eventually, these compounds will form solids in the
unless a gas-side scrubbing plant is installed. Sulfur trioxide runner pipes or directly in tips, which results in plugging.
can, in the cooler stages of the gas path, combine with mois-
ture in the exhaust gas to form sulfuric acid (H2SO4), which The following are some items to look for when operational
is highly corrosive and will be deposited in ducts and the problems are encountered:
economizer if the exhaust gas is below condensing tempera-
tures. Natural gas fuels are fortunately very low in sulfur and • Plugged gas ports, which are evidenced by gaps in the
flame or high fuel pressure: Gas ports may simply
do not usually cause a problem. However, some oil fuels and
consist of holes drilled into the element manifold pipe,
plant gases can be troublesome in this respect.
or they may be located in individual removable tips.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Particulate matter (PM): Particulate emissions are Designs of the former type may be re-drilled or else the
formed from three main sources: ash contained in liquid fuels, entire manifold pipe must be replaced. Discrete tips can
unburned carbon in gas or oil, and SO3. The total amount of be replaced individually as required.
particulate is often called TSP (total suspended particulate). • Warped flame holders (wings): Some warping is normal
There is concern for the smaller sized portion of the TSP, as and will not affect flame quality, but excessive
this stays suspended in air for a longer period of time. The deformation such as “curling” around the gas ports will
PM-10 is the portion of the total particulate matter that is less degrade the combustion and emissions performance.
than 10 microns (1 × 10–6 m) in size. Particles smaller than Most grid-type burner designs permit replacement of
PM-10 are on the order of smoke. individual flameholder segments.
Typical NOx and CO emissions for various fuels are shown • Oxidation of flame holders (wings) or portions of flame
in Table 18.1. holders: If more than one-third of the flameholder is
missing, it is a good candidate for replacement. Fabricated
and cast designs are equally prone to oxidation over time.
18.3.5 Maintenance
Most grid-type burner designs permit replacement of
1. Normal wear and tear: If nothing has been replaced in individual flameholder segments.
the past 5 years and the burner (or turbine/HRSG set) is • Severe sagging of runner pipes (grid design only): If the
operated fairly continuously, it is likely that some tips and manifold pipe is no longer supported at both ends, it
wings may require replacement. should be replaced. Beyond that relatively extreme condi-
2. Damage due to misuse, system upsets or poor mainte- tion, sagging at midspan in excess of approximately 2 to
nance practices: Older systems designed without suffi- 3 in. (5 to 7 cm) should be corrected by runner replace-
cient safety interlocks (TEG trip, high temperature) ment and/or installation of an auxiliary support.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch18-Frame Page 540 Wednesday, March 23, 2005 7:27 AM

540 The John Zink Combustion Handbook

18.3.6 Accessories own national standards. Specific requirements for burner

18.3.6.1 Burner Management System safety systems are included, but as stated in the foreword,
All fuel-burning systems should incorporate controls that NFPA 8506 does not encompass specific hardware applica-
provide for safe manual light-off and shutdown, as well as tions, nor should it be considered a “cookbook” for the
automatic emergency shutdown upon detection of critical design of a safe system. Prior to the issuance of NFPA 8506,
failures. Control logic may reside in a packaged flame designers often adapted NFPA boiler standards to HRSGs,
safeguard module, a series of electromechanical relays, a which resulted in design inconsistencies.
programmable logic controller (PLC), or a distributed control
system (DCS). At a minimum, the duct burner management 18.3.7.2 FM (Factory Mutual)
system should include the following: An insurance underwriter that publishes guidelines on
combustion system design, Factory Mutual also “approves”
• flame supervision for each burner element specific components such as valves, pressure switches, and
• proof of completed purge and TEG/combustion air flow flame safeguard equipment that meet specific design and
before ignition can be initiated
performance standards. Manufacturers are given permission
• proof of pilot flame before main fuel can be activated
to display the FM symbol on approved devices. Although FM
• automatic fuel cutoff upon detection of flame failure, loss
approval may be required for an entire combustion control
of TEG/combustion air, and high or low fuel pressure
system, it is more common for designers to simply specify
Other interlocks designed to protect downstream equipment the use of FM-approved components.
can also be included, such as high boiler tube temperature or
loss of feedwater. 18.3.7.3 UL (Underwriters Laboratories)
Well-known in the United States for its certification of a
18.3.6.2 Fuel Train broad range of consumer and industrial electrical devices, UL
Fuel flow to the burners is controlled by a series of valves, authorizes manufacturers to display their label on specific

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
safety devices, and interconnecting piping mounted on a items that have demonstrated compliance with UL standards.
structural steel rack or skid. A properly designed fuel train Combustion system designers will frequently require the use
will include, at a minimum, the following: of UL-approved components in burner management systems
and fuel trains. Approval can also be obtained for custom-
• at least one manual block valve
designed control systems, although this requirement gener-
• two automatic block valves in series
• one vent valve between the automatic block valves (gas ally applies only to a few large cities and a few regions in the
firing only) United States.
• flow control valve
• high and low fuel pressure switches 18.3.7.4 ANSI B31.1 and B31.3 (American National
• two pressure gages, one each at the fuel inlet and outlet Standards Institute)
These codes address piping design and construction. B31.1 is
Depending on the custom and operating requirements at a
incorporated in the NFPA 8506 guideline, while B31.3 is
particular plant, pressure regulation, flow measurement
generally used only for refining/petrochemical applications.
devices, and pressure transmitters can also be incorporated.
See Figures 18.20 through 18.27 for typical duct burner fuel
18.3.7.5 Others
system piping arrangements.
The following may also apply to duct burner system designs,
depending on the country where equipment will be operated.
18.3.7 Design Guidelines and Codes
18.3.7.1 NFPA 8506 (National Fire Protection • National Electrical Code (NEC)
Association) • Canadian Standards Association (CSA)
First issued in 1995, this standard has become the de facto • International Electrotechnical Commission (IEC)
guideline for heat recovery steam generators in the United • European Committee for Electrotechnical Standardization
States and many other countries that have not developed their (CENELEC)

Copyright CRC Press

Duct Burners 541

FIGURE 18.20 Typical main gas fuel train: single element or multiple elements firing simultaneously.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 18.21 Typical main gas fuel train: multiple elements with individual firing capability.

Copyright CRC Press

542 The John Zink Combustion Handbook

FIGURE 18.22 Typical pilot gas train: single element or multiple elements firing simultaneously.

FIGURE 18.23 Typical pilot gas train: multiple elements with individual firing capability.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Duct Burners 543

FIGURE 18.24 Typical main oil fuel train: single element.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 18.25 Typical main oil fuel train: multiple elements.

Copyright CRC Press

544 The John Zink Combustion Handbook

FIGURE 18.26 Typical pilot oil train: single element.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 18.27 Typical pilot oil train: multiple elements.

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 19

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Boiler Burners
Lev Tsirulnikov, John Guarco, and Timothy Webster

TABLE OF CONTENTS

19.1 Boiler-Specific Burner Requirements ..................................................................................................... 548

19.1.1 Conventional Burner Technology for Boilers ........................................................................... 548
19.1.2 Low-NOx Burner Technology for Boilers ................................................................................ 550
19.1.3 Staged Burner Design Philosophy ............................................................................................ 551
19.1.4 Design Features of Low NOx Burners...................................................................................... 552
19.1.5 Effects of Burner Retrofits on Boiler Performance................................................................... 553
19.2 Boiler Design Impacts on NOx Emissions Correlations......................................................................... 553
19.2.1 Boiler Design ............................................................................................................................ 553
19.2.2 Excess Air ................................................................................................................................. 555
19.2.3 Boiler Load Influence on NOx.................................................................................................. 558
19.2.4 Boiler/System Condition Impacts on Combustion and NOx Formation .................................. 561
19.3 Current State-of-the-Art Concepts for Multi-Burner Boilers ................................................................. 563
19.3.1 Combustion Optimization......................................................................................................... 563
19.3.2 Methods to Reduce NOx Emissions ......................................................................................... 568
19.4 Low-NOx burners for Packaged Industrial Boilers................................................................................. 575
19.5 Ultra-low Emission Gas Burners ............................................................................................................ 580
19.5.1 Background ............................................................................................................................... 580
19.5.2 Implementation of the NOx Formation Theory for Ultra-low-NOx Emissions ....................... 580
19.5.3 Ultra-low Emissions Burner Design ......................................................................................... 581
19.5.4 Ultra-low Emissions Burner Operation .................................................................................... 582
19.6 Atomizers for Boiler Burners.................................................................................................................. 584

547
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 548 Wednesday, March 23, 2005 7:40 AM

548 The John Zink Combustion Handbook

19.1 BOILER-SPECIFIC BURNER to be CO < 1000 ppm and opacity < 6%. The opacity numbers
REQUIREMENTS were measured using the Ringelman or Baharac methods.
The requirements for furnaces and burners of gas/oil-fired In some countries, the burner performance was evaluated
utilities and industrial boilers are based on long-term opera- specific to each boiler design. For example, in Russia, all indus-
tional experience and comprehensive testing of different new trial and relatively small utility boilers with steam flow capac-
designs and retrofits of boiler equipment. These requirements ities of up to 220,000 lb/hr (100,000 kg/hr) with preheated air
are changing with the times. Figure 19.1 shows typical utility were designed with up to six burners, each with rated heat
boilers, and a typical single-burner industrial boiler is shown inputs of up to 170 × 106 Btu/hr (50 MWt). Larger utility boilers
in Figure 19.2. having capacities from 440,000 to 1,800,000 lb/hr (200,000 to
Prior to the establishment of NOx emission rate control 820,000 kg/hr) superheated steam flow were designed with up
requirements in the United States and other countries, perfor- to eight burners, each with rated heat inputs of up to 400 × 106
mance data for furnace and burner design of utility and indus- Btu/hr (120 MWt). All ICP concentrations in the flue gas,
trial boilers usually had to match the following common including both CO and H2, were required to be less than
accepted requirements: 1000 ppm at O2 < 1%. To obtain lower ICP values, burner air
velocities were increased to at least 160 ft/s (50 m/s), with
1. high reliability during long-term operation 300 ft/s (90 m/s) preferred.
2. simplicity and reliability of gas and oil fuel ignition Until recently, there were two operational parameters used
3. high flame stability while firing either gas or oil fuels, to regulate the combustion processes in boilers to achieve the
even at full turndown required operational characteristics: the variation of fuel and
4. provision for designed superheated and reheated steam air input, and the regulation of reheated/superheated steam
temperatures while firing either gas or oil fuels at full
temperature. The many years of operating this way resulted
turndown during long-term operation
in significant difficulties, including:
5. high thermal efficiency and low concentrations of incom-
plete combustion products (ICPs) while firing either gas 1. reduction in reliability due to frequent failures of high-
or oil fuels using comparatively low excess air, even at temperature heat exchange surfaces
full turndown 2. loss of reheated/superheated steam temperature during
6. low combustion air system pressure drop, especially the full turndown operation
burner register draft loss 3. overfiring of the unit
7. simplicity of burner/windbox maintenance and adjustment
Experience has shown that controlling the combustion pro-
8. ease of automatic mode operation and fuel changeover
cesses to allow maintenance of key operational parameters
between gas and oil
9. provision of restricted flame dimensions to match the
within a given range, including superheated and reheated
dimensions of an existing or newly designed furnace steam temperatures, best solved these problems. Combustion
10. allowing operation with no flame impingement on either control yielded the most efficient and reliable boiler opera-
the target or side walls of the furnace that promoted reli- tion, independent of load, fuel type, or other conditions.
ability of all high-temperature heat exchange surfaces —
especially water-wall, superheater, and reheater tubes 19.1.1 Conventional Burner Technology
To match this requirement for any boiler application, it was for Boilers
necessary to carefully review how a retrofit burner design Older generations of burners, such as the register, or swirl
should be implemented for each particular retrofit furnace. burner shown in Figure 19.3, were considered very reliable
Prior to the establishment of NOx emission rate control and were the staple of the industry for many years. They con-
requirements, burner performance was typically evaluated as sisted of several main parts, including a diffuser, air doors, a
a function of ICP concentration levels at low excess air con- throat ring, and a fuel supply system. They operated on a
ditions. While firing either oil or gas fuels in preheated air, principle of precisely controlling the swirl of the burner by
optimal low excess air burner designs provided carbon mon- adjusting the air doors, either open or closed. In an ideal situ-
oxide (CO), hydrogen (H2), and unburned hydrocarbon ation, each burner in a multi-burner windbox should receive
(UHC) concentrations of less than 200 ppm at excess oxygen an equal amount of air mass flow. However, in the real world,
(O2) levels ranging between 0.2 and 0.6% at full load, as well the burner-to-burner mass flow distribution varied up to
as O2 < 1% over a 3:1 turndown range. In many European ±30% of the average mass flow. The typical compensation for
countries (e.g., Germany, France, Italy, and Belgium), the low deviations in mass flow was to close down on the air doors of
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
excess air gas/oil burners acceptable condition was considered the burners receiving too much air. This in turn affected the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 549 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 549

FIGURE 19.1 Typical utility boilers. (Courtesy of Florida Power & Light)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 19.2 Typical single-burner industrial boiler. (Courtesy of North Carolina Baptist Hospital)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 550 Wednesday, March 23, 2005 7:40 AM

550 The John Zink Combustion Handbook

structure and emission properties of a flame can be varied

using controlled fuel input.

19.1.2 Low-NOx Burner Technology

for Boilers
The objective of the modern typical boiler burner retrofit is to
reduce NOx. Two NOx reduction techniques for boilers are
combustion modification and back-end (post-treatment)
cleanup (see Chapter 6). Combustion modifications include
low-NOx burners, low excess air, over-fired air, flue gas recir-
culation (FGR), fuel-induced recirculation, reburn, and water
tempering. Back-end cleanup techniques include selective cat-
alytic reduction (SCR) and selective non-catalytic reduction
FIGURE 19.3 Swirl burner. (SNCR). This chapter focuses on the combustion modification
techniques of low-NOx burners, low excess air, over-fired air,
flue gas recirculation, and fuel-induced recirculation.
swirl of the burner, taking it away from the optimal single-
When the United States and other countries mandated NOx
burner performance setting.
control, in addition to the requirements listed in the previous
Flame length, which depended largely on heat input, was section, this new requirement became the dictating feature
the critical burner design parameter during this time period. that determined the quality of burner/furnace performance.
Flame impingement on either the target or side walls was not As a result, many low-NOx burners, designed for a plethora
allowed. Figure 19.4 shows the average flame length for nat- of different boiler design and heat input capacities, firing gas
ural gas and No. 6 oil as a function of burner heat input in and/or oil with preheated and ambient air, were developed in
an experimental furnace. Other parameters include preheated the United States and many other countries.
air of ~570°F (300°C) and O2 < 1%. The oil was mechanically The situation is complicated by numerous contradictions
atomized at ~280 to 300 psig (19 to 20 barg) and at 248 to between the operational baseline conditions for the existing
266°F (120 to 130°C). boiler on the one hand, and the new NOx requirements on
Conventional gas and gas/oil burners usually produce the other hand. For example, all NOx reduction combustion
flames with fixed parameters. However, the combustion pro- methods will increase flame length and redistribute the heat
cesses can by modified by varying the aerodynamic and chem- flux between radiant and convective heating surfaces. How-
ical structure of flows fed into the furnace via burners. In this ever, the flame length is restricted by the current furnace
case, in addition to the typical requirements listed above, the dimensions. Also, increasing heat flux in the super-
burner design should provide the following possibilities. heater/reheater surfaces is not allowed because it will degrade
the lifetime of these surfaces. Decreasing the total area of
1. the ability to change the direction of or reverse the fuel/air these surfaces increases their reliability, but reduces the boiler
mixture swirl to cause interaction of the flame vortices;
thermal efficiency.
this simplifies the problem of increasing the furnace wall
heat absorption rate when switching from oil to gas while Another contradiction is related to the desire to get simul-
retaining the same O2 in the furnace taneous NOx reduction and reduction of incomplete combus-
tion products (ICPs). Lowering the O2 will decrease NOx but
2. the ability to change the swirl intensity, flame length, flow
increase ICPs. To reduce ICP concentrations, O2 must be
density, and static pressure across the furnace cross-section,
as well as along the length of the combustion zone increased and fuel/air mixing must be improved. However,
these actions will tend to increase NOx. Solving these prob-
3. the ability to vary the flue gas recirculation (FGR) rates
lems simultaneously is very difficult for both existing and
to control the primary firing zone to enable both stable
new design furnaces, especially on utility boilers.
combustion and low emission formation
There are other complications on dual-fuel boilers. It is nec-
Regulating not only the aerodynamic properties of the essary to provide the designed reheated and superheated steam
flame, but also the chemical structure and emission properties temperatures over long-term operation, while firing fuels gen-
of the gas and oil combustion products, can further increase erating different flame properties over the full turndown load
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

the effect of burners on combustion processes. The chemical range. The luminescence of an oil flame is significantly higher
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 551 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 551

50
45 Natural Gas
#6 Oil
40

35
Flame Length, ft.
30
25
20

15
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

10
5

0
0 50 100 150 200 250 300 350
Burner Heat Input, M M Btu/hr

FIGURE 19.4 Average flame length as a function of burner heat input.

than that of a gas flame (see Chapter 3). Assuming all other a primary air swirl defined by the fixed blade axial swirler.
conditions are the same, furnace heat absorption is less and The swirler determines the size and strength of the recircu-
the furnace exit gas temperature is higher while firing gas. lation zone. A tertiary air stream flows between the venturi
Higher exit gas temperatures create conditions for more base and the burner throat quarl. Tertiary air separates some
intense thermal NOx generation in all industrial boilers and of the combustion air from the main flame, effectively staging
most utility applications for 50 to 150 MWe power units. On combustion and reducing NOx. For natural gas firing, fuel
larger capacity utility boilers, the NOx is higher while firing can be introduced through internal or external pokers or gas
gas than while firing oil. A more detailed explanation will be rings, and can also be injected through a central gas pipe with
given in Section 19.3.2.3. Typically, the NOx while firing oil multiple orifices at the furnace end. A single conventional
is higher due to fuel-bound nitrogen (Nf). Fuel-bound nitrogen atomized burner (oil gun), located along the burner centerline,
usually ranges from 0.2 to 0.45% wt, but can be as high as typically supplies the oil. The oil gun may use dual fluid,
0.7 to 1% in No. 6 oil. Because the effectiveness of various mechanical, or rotary cup atomization.
NOx reduction methods is different with gas and oil firing, it
makes sense — in principle — to implement different NOx 19.1.3 Staged Burner Design Philosophy
reduction methods for each fuel. However, doing so would Most of today’s low-NOx burner designs implement staged
complicate boiler operations so much that such designs would combustion principles as an effective way to reduce NOx.
not be acceptable to the customer. Attempts to implement The staged burner should also be designed to provide the
NOx reduction methods for both fuels create situations where maximum degree of flexibility in achieving high burner turn-
different methods are selected for the same boiler design down, low NOx, and improved flame shaping capability. The
installed at different power plants. The method chosen design basis of a staged burner is to develop a stratified flame
depends on many operational parameters, including annual structure with specific sections of the flame operating fuel-
consumption and seasonal distribution of each fuel, local rich and other sections operating fuel-lean. The burner design
climate, annual average load, levels and frequency of load thus provides for the internal staging of the flame to achieve
peaks, and stack height. NOx reductions while maintaining a stable flame.
Today’s low-NOx burner relies on control of the combustion Controlling combustion stoichiometry to fuel-rich condi-
air in several component streams, as well as the controlled tions inhibits NOx production, especially in the burner’s flame
injection of fuel into the air streams at selected points for front. Operating the flame fuel-rich also reduces the burner
maintaining stable, attached flames with low NOx generation. NOx dependence on the burner zone heat release (BZHR)
Typical venturi-style, low-NOx burners are shown in Figures rate, which is discussed in Section 19.2.1. This is especially
19.5 and 19.6. Primary and secondary air enters the burner important in applications where very high BZHR near full
radially through the venturi and exits the burner axially with load can result in an exponential increase in NOx.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 552 Wednesday, March 23, 2005 7:40 AM

552 The John Zink Combustion Handbook

FURNACE FRONT WALL

PRIMARY / SECONDARY TERTIARY

AIR INLET AIR INLET POKER GAS INJECTOR

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
COOLER OXYGEN RICH
ZONE REDUCES
FUEL SPRAY THERMAL NOx
OIL INLET

LOW EXCESS AIR

ZONE REDUCES
OIL ATOMIZER
FUEL NOx

GAS INLET AIR SWIRLER

FIGURE 19.5 A typical low-NOx burner, venturi-style.

burner design allows control of the stoichiometry of the oxi-

dizing zone to range from being pure air to having varying
degrees of excess oxygen. This controls NOx formation by
causing the fuel burnout to occur in the form of a premixed
flame rather than a diffusion flame. The operational flexibility
is provided in the burner design to control stoichiometry to
achieve minimum NOx.
Flame stability should also be a design focus for a staged
burner. The stability criterion is to maintain a minimum turn-
down capability on a gas-firing ratio of 8:1 over a broad range
of excess O2. To achieve this objective, it is necessary to care-
fully select the burner throat velocity and the velocity profile.
FIGURE 19.6 A typical-low NOx burner, venturi-style
The burner flame front should be a fuel-rich premixed flame
(second example).
designed to increase flame speed and, in turn, flame stability.
Combustion stoichiometry, approaching a perfect premixed
In addition to controlling NOx formation, operating under flame, is established at the burner flame front. This condition
fuel-rich conditions aids in the production of combustion broadens the flammability range of the fuel and ensures the
intermediates that can result in the destruction of previously maintenance of a flammable mixture over a broad range of
formed NOx. In a reducing environment, NO can act as an burner throat velocities and fuel injection rates.
oxidizer to react with these combustion intermediates, result- Controlling air distribution within the burner throat improves
ing in the reduction of NO to N2. Therefore, the NO neces- flame-shaping capability. Provisions are made in the burner
sarily formed to satisfy the requirements of establishing a design to allow the burner to operate as a low excess air burner
strong flame front can be scavenged by the process and as well as a low-NOx burner with an extended flame envelope.
reduced to N2.
To achieve complete fuel burnout at minimum O2 , the 19.1.4 Design Features of Low-NOx Burners
burner design must provide direct interaction of fuel-lean Some of the standard design features of low-NOx burners are:
zones with the center fuel-rich sections. This ensures that the 1. flame stability at low excess air rates for reliable, energy
“rich” products of combustion from the center flame pass efficient boiler operation
through the oxidizing zone for complete fuel burnout. The 2. high turndown ratios for a wide range of boiler operations
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 553 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 553

3. well-distributed air flow to control the flame envelope and

provide even heat flux
4. known flame length and diameter to suit furnace firing
lane without impinging on boiler tubes or furnace walls
5. elimination of combustion-induced vibrations of the
boiler, due to the aerodynamics of the register design and
turbine blade diffuser
6. a strong flame front established within a maximum of 0.5
diffuser diameters of the face of the diffuser, as depicted
in Figure 19.7 (the flame front should not move during
changes in the firing rate, thus providing a stable flame
for scanning, and resulting in reliable operation)
7. for gaseous fuel firing, a multiple-poker gas burner assem-
bly with “poker shoes” (gas nozzles) oriented to provide
internal fuel staging FIGURE 19.7 A strong flame front established within a
maximum of 0.5 diffuser diameters of the face of the diffuser.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
19.1.5 Effects of Burner Retrofits on Boiler
Performance
Low-NOx burners have different flame characteristics from 19.2 BOILER DESIGN IMPACTS ON
their predecessors, the low excess air burners. Therefore, NOx EMISSIONS CORRELATIONS
there exists the probability that the differences will affect The two major mechanisms for the formation of NOx are:
boiler performance. Performance changes would only occur (1) NOx formed by the oxidation of Nf (fuel NOx) and
by an alteration of the heat absorption pattern in the furnace, (2) the thermal fixation of atmospheric nitrogen (thermal
thus affecting the amount of heat going to the boiler back- NOx). Separate NOx correlations have been developed for
pass. The overall furnace heat absorption could either each. These correlations can be combined to predict the total
increase or decrease, due to factors such as flame distance to NOx emissions for a selected burner design based on the fuel
walls, flame temperature, flame emissivity, and characteris- nitrogen content and boiler design.
tics of ash deposits on furnace walls. The impact on boiler The conversion of Nf to NOx is dependent on oxygen
performance will vary by unit and by fuel. In some cases, availability and Nf content. Thus, for a given burner system
there will be no impact; but in most cases, there will be either at constant excess air, the key variable controlling fuel NOx
a positive or negative impact experienced. Historically, how- formation is Nf content. Previous studies have shown that Nf
ever, there has been no net effect on boiler performance that conversion efficiency is inversely proportional to the nitrogen
would be considered extreme. Atomizer tip design can drasti- content of the fuel. High conversion efficiencies are observed
cally impact boiler performance, but total boiler performance with low Nf, and low efficiencies are seen with high Nf.
is always a balancing act between key variables such as NOx
and water wall temperatures.
19.2.1 Boiler Design
To assess the impact on boiler performance, certain infor- Based on the strong dependency of thermal NOx on flame
mation is typically monitored during the testing of low-NOx zone temperature, thermal NOx formation for wall-fired boil-
burners. Important data includes superheated and reheated ers has been correlated with the ratio of heat input to furnace
steam temperatures, superheater and reheater tube surface size and the number of firing walls. For this correlation, the
temperatures, and any operational parameters that affect or ratio of heat input to the furnace burner zone area is defined as
control steam temperatures. The boiler performance variation the burner zone heat release (BZHR) rate. The BZHR repre-
resulting from a change in the overall O2 level is not consid- sents the boiler heat release rate divided by the water-cooled
ered in this analysis. The actual impacts on boiler efficiencies surface area in the burner zone (106 Btu/hr-ft2), and is a mea-
are relatively small, and there are many complicating factors sure of the “bulk furnace temperature.” The NOx created
such as furnace and heat recovery area cleanliness, fuel com- within this zone is dependent on this bulk temperature by an
position, sootblower availability, operational variability, etc. exponential relationship, as discussed in previous chapters.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 554 Wednesday, March 23, 2005 7:40 AM

554 The John Zink Combustion Handbook

FURNACE SIZE EFFECTS ON NOx EMMISSION RATES

UNIT DESIGN HIGH LOW / MEDIUM
PARAMETER HEAT RELEASE RATE HEAT RELEASE RATE

FURNACE
ELEVATION
(SAME MW SIZE UNIT)

BURNER ZONE (3X BURNER SPACING) 1.66X (HIGH HRR VOLUME)

VOLUME X WIDTH X DEPTH

FURNACE SPACE 86,000 BTU / Hr / Ft3 53,000 BTU / Hr / Ft3

HEAT RELEASE

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
LOW EXCESS 0.55 Lb / Million BTU 0.22 Lb / Million BTU
AIR BURNER

LOW NOx 0.39 Lb / Million BTU 0.16 Lb / Million BTU

BURNER

FIGURE 19.8 The effects of boiler design on NOx.

The burner zone is defined as the six-sided surface bounded by The correlations of Nf conversion and thermal NOx can be
the furnace walls and imaginary horizontal planes located one used to extrapolate burner NOx from one boiler to another. The
burner row spacing above the top row and below the bottom increase in NOx observed with fuel oil compared to natural
row of burners. The area of any division walls within this vol- gas, which contains no nitrogen, is used to estimate fuel NOx
ume is also included in the calculation of the BZHR. A correc- for the selected burner. Nf conversion is then calculated based
tion is made for re-radiation from any refractory on the floor, on the nitrogen content. The correlation of Nf conversion can
walls, or the burner throats. An example of the effects of boiler then be used to project fuel NOx for fuels containing differing
design on NOx is shown in Figure 19.8, where the same heat amounts of nitrogen. Thermal NOx is determined from NOx
input is placed in two drastically differently sized boilers. with natural gas. The correlation with BZHR is used to project
The correlation of thermal NOx with BZHR rate has been the measured thermal NOx for the selected burner from one
developed using an extensive database of gas and oil fired boiler to another. Total projected NOx emissions for the burner
utility boilers. Figure 19.9 shows the NOx of various boilers can then be determined by adding the fuel NOx and thermal
included in the database on oil and gas, respectively. NOx of NOx contributions for the fuel and boiler of interest. The
industrial boilers with comparable degrees of NOx control amount of air preheat can also have a dramatic effect on NOx,
were found to be consistent with the BZHR rate correlations. CO, and particulate emission rates, as well as on flame stability.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 555 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 555

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 19.9 The NOx vs. BZHR of various boilers included in the database on oil and gas, respectively.

NOx Generation with Firing Natural Gas and #6 Oil (0.55 Nf)
vs. Adiabatic Fame Temperature
1200
(NOx),
Natural Gas
NOx Calculated @ O2 = 3%, ppm

1000
(NOx),
#6 Oil

800
Thermal
(NOx), #6 Oil
600

400

200
(NOx) Nf

0
2600 2800 3000 3200 3400 3600
Temperature, F

FIGURE 19.10 NOx generation with firing natural gas and No. 6 oil (0.5% Nf) vs. adiabatic flame temperature.

19.2.2 Excess Air increase in NOx slows down, reaches a maximum, and then
19.2.2.1 Theoretical Effect of Excess Air on NOx NOx is reduced as excess air is further increased.
It is well-known that thermal NOx formation primarily According to some experimental studies, no maximum is
depends on the flame temperature, excess air in the flame, attained and the NOx = f(α) dependence approaches the sta-
and residence time. However, the flame temperature is also bilized section of the exponential curve. A difference between
dependent on excess air. As shown in Figure 19.11, demon- the two mentioned contradicting results brings different
strating the experimental data for natural gas combustion approaches to optimized low-NOx combustion evaluations.
under regular boiler conditions, with an excess air factor α That is why it is important to discuss the existence of the
value close to unity, increased excess air causes NOx to rise NOx maximum and to consider it in some detail. For sim-
considerably; then, as excess air is further increased, the plicity, one can consider a natural gas (containing no bound
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 556 Wednesday, March 23, 2005 7:40 AM

556 The John Zink Combustion Handbook

Nox vs. excess O2

(TGM-94 boiler, equipped with 21 burners, at ~94% load)
400
Natural Gas,
0% FGR
350
NOx @ O2 = 3%, ppm

300 #6 Oil,
0% FGR

250
Natural Gas,
FGR = 14%
200
#6 Oil,
FGR = 14%
150

100
0 0.5 1 1.5 2 2.5 3
O2, %

FIGURE 19.11 Excess O2 influence on NOx formation.

nitrogen) combustion process, as a result of which NOx is the fuel is not completely burned, indicating the need to not
formed only from nitrogen in the combustion air. It can be only improve the combustion process but also to check the

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
seen in Figures 19.10 and 19.11 that with a deficiency of validity of the methods used to measure ICP and NOx in the
oxidant, NOx depends more on α than on temperature; and given α range.
when α > 1 (mostly when α > 1.2), when the rate of com-
bustion reactions increases, the effect of temperature proves
19.2.2.2 Empirical Evidence of the Effect of
to be the dominant one. Therefore, despite the fact that the
Excess Air on NOx
theoretical combustion temperature in the region of α >1
systematically decreases with increased α, the NOx increases Detailed empirical data sets were obtained under identical
as long as the thermal NOx formation is not made more operational conditions at approximately full (~94%) load
difficult due to a considerable decrease in the temperature while firing natural gas and No. 6 oil (containing 0.22 to
level. In this case, there is also a decrease in the NOx in the 0.32% Nf) with preheated air on the two neighboring 165-
flue gas volume due to dilution by the excess air. When firing MWe utility boilers of the TGM-94 model (~1,100,000 lb/hr or
a fuel oil that contains Nf , it is unlikely that the nature of the 500,000 kg/hr superheated steam flow), installed at the same
NOx vs. O2 curve considered above will change; but in all power plant. The furnaces were balanced draft and had air
probability, the downward branch of the curve will be flatter. in-leakage of 8 to 10%. The two boilers were almost identical
When comparing the well-known empirical dependency except for one difference: the first boiler was equipped with
of the ICP component concentrations on α with the empirical 21 burners, rated at ~100 × 106 Btu/hr (30 MWt) heat input
NOx data as a function of α, confirmed with the theoretically each, having a single swirled air flow channel, and the second
calculated dependencies on α shown in Figure 19.10, it was boiler was equipped with nine burners, rated at ~230 × 106
found that the excess air factor α value at which furnace ICP Btu/hr (67 MWt) heat input each, having two air flow channels,
losses virtually disappear almost coincides with the values one of which is a swirled portion consisting of ~15% of the
of α corresponding to the NOx maximum. It follows that the total air flow. On both boilers, the burners were installed in
NOx dependence on excess air has the form of an extreme three rows on the front wall (3 × 7 and 3 × 3, respectively). The
function with a maximum NOx value corresponding to that boilers were equipped with an FGR system designed to supply
α value at which virtually complete fuel combustion is up to 14% FGR flow (isolated from the air flow), supplied
attained under the given conditions. Hence, if NOx and α directly to the furnace through slots located on the target wall,
are determined by reliable and sufficiently accurate methods, opposite the lower burner row. The test data while firing either
the absence of an experimentally established maximum on or both fuels, both with and without FGR implementation and
the test curve NOx vs. O2 (or α) leads to an assumption that shown in Figures 19.11 and 19.12 for the 21- and 9-burner
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 557 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 557

Nox vs. excess O2

(TGM-94 boiler, equipped with 9 burners, at ~94% load)
500
Natural Gas,
0% FGR
450

400 #6 Oil,
0% FGR
350
Natural Gas,
FGR ~13%
300
#6 Oil,
FGR ~ 13%
250

200

150

100
0 0.5 1 1.5 2 2.5 3
O2, %

FIGURE 19.12 NOx vs. excess O2 (the TGM-94 boiler equipped with 9 burners, at ~94% load).

FIGURE 19.13 NOx vs. relative steam flow at the TGM-94 boilers (natural gas, O2 = 1.2–1.6%).

boilers confirm the well-known dependencies of NOx concen- a significant NOx difference of 100 to 120 ppm (NOx is ~20 to
trations on O2 and load, respectively. 24% higher than for the boiler equipped with nine burners).
Differences can be seen between the maximum NOx num- The increased temperature resulted in a stronger dependency
bers, and, accordingly, between shapes of the curves, obtained of NOx on O2 on the boiler equipped with nine burners, both
on these two boilers. These discrepancies are associated with with and without FGR, as demonstrated in Figure 19.12.
distinctions in the combustion processes related to the differ- The above data, as well as other test data obtained from
ences between the mentioned burner designs, heat inputs, and 150 to 800 MWe utility boilers, under similar operational
numbers of burners. The measured flame temperatures while conditions, illustrate a significantly lower NOx level on No. 6
firing gas under full load with all other conditions being equal oil (even while containing Nf of up to ~0.7%) in comparison
show that on the boiler equipped with nine burners, the max- with NOx numbers on gas. An explanation of this fact
imum flame temperature is ~160°F (70°C) higher, resulting in contradicting the empirical data obtained on industrial and
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

558 The John Zink Combustion Handbook

NOx vs, relative steam flow at the TGM-94 boilers (#6 Oil, O2 = 1.2-16%)

450

400

350 9 Burners,
NOx @ O2 = 3%, ppm

FGR = 13%
300

250 21 Burners,
0% FGR
200
9 Burners,
0% FGR
150
21 Burners,
FGR = 14%
100

0
0 20 40 60 80 100
Relative Steam Flow, %

FIGURE 19.14 NOx vs. relative steam flow at the TGM-94 boilers (No. 6 oil, O2 = 1.2–1.6%).
NOx vs. steam flow with firing natural gas on the TGME-206 boiler equipped
with Typical Venturi-type LNBs at O2 = 0.81-1%
100

80
NOx @ O2 = 3%, ppm
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

60
FGR = 6-7%

40
FGR = 12-

0
50 60 70 80 90 100
relative steam flow, %

FIGURE 19.15 NOx vs. relative steam flow with firing natural gas on the TGME-206 boiler equipped with Todd Com-
bustion low-NOx Dynaswirl burners at O2 = 0.8–1%.

comparatively smaller utility boilers (where NOx numbers range, CO concentrations were in the 50- to 150-ppm range
are higher while firing oil) is presented in Section 19.3.2.3. while firing gas. When firing No. 6 oil on both boilers, over
the entire load range and under the same O2 , CO and opacity
19.2.3 Boiler Load Influence on NOx did not exceed 100 ppm and 10%, respectively.
The relationship between NOx and load was investigated on Figure 19.15 shows the data of combined (load and FGR)
many utility boilers over their full load ranges, while firing influence on NOx on a 200-MWe boiler of the TGME-206
either or both fuels. A detailed investigation was performed on model, having no furnace air in-leakage, equipped with
the two TGM-94 boilers described above (see Section 19.2.2). 12 low-NOx venturi-style burners installed in two rows on
The test results obtained at ~94, 75, 50, and ~30% loads, the rear wall. The boiler was tested at loads ranging from
with and without FGR, are shown in Figures 19.13 (gas) and 100 to 53% while firing natural gas at comparatively low O2
19.14 (No. 6 oil). With O2 = 1.2–1.6%, over the entire load (0.8 to 1%) and with 6 to 7% and 12 to 14% FGR flows. The
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 559 Wednesday, March 23, 2005 7:40 AM

Boiler Burners
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- 559

FIGURE 19.16 NOx vs. load with firing natural gas on utility burners.

Degree of the power function NOx = f (load) vs. bounded nitrogen in #6 oil

1.3

1.25

1.2
degree number

1.15

1.1

1.05

1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Fuel Bound Nitrogen, %

FIGURE 19.17 Degree of the power function NOx = f (load) vs. bounded nitrogen in No. 6 oil.

FGR was mixed with the combustion air upstream of the the NOx number at full load) on relative load. It is seen that
windbox. Under all above conditions, CO concentrations relative NOx is a power function of relative load to the ~1.25
were less than 20 ppm. degree. While firing natural gas, significant changes in the O2
range, FGR flow level, preheated air and FGR temperatures,
Confirmation of these empirical dependencies on many util-
and other operational conditions provide comparatively small
ity boilers, equipped with various numbers, designs, and
derivations from the mentioned average degree number.
arrangements of burners, while firing gas and oil with preheated
air (500 to 700°F, or 260 to 370°C), is shown in Figures 19.13 With firing No. 6 oil under similar operational conditions
to 19.15. All available data obtained on 150- to 800-MWe on utility boilers, in general, the power function degree num-
boilers, firing natural gas in the load range of 100 to 30%, have ber depends on Nf concentration in the oil: the higher the Nf,
been generalized in Figure 19.16 as a dependence of relative the lower the degree number. Corresponding test data,
NOx (defined as a ratio of NOx numbers at current loads to obtained at 17 boilers of 150 (three), 165 (eight), 210 (one),
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 560 Wednesday, March 23, 2005 7:40 AM

560 The John Zink Combustion Handbook

Relative NOx vs. relative load on industrial boilers

firing gas and #6 oil with ambient air
1

0.8
relative NOx concentration

0.6

0.4
Natural Gas

#6 Oil
0.2

0
0 20 40 60 80 100
Relative Steam Flow, %

FIGURE 19.18 Relative NOx vs. relative load on industrial boilers firing natural gas and No. 6 oil with ambient air.

FIGURE 19.19 Relative NOx vs. relative load on industrial boilers firing natural gas and No. 6 oil with preheated air.

and 300 (five) MWe utility boilers at full load with an O2 385,000 lb/hr (10,000 to 175,000 kg/hr) steam flow boilers,
range of 0.6 to 1.2%, are shown in Figure 19.17. With the with firing both gas and No. 6 oil (0.22 to 0.49% Nf) in the --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

exception of one 150-MWe boiler, all boilers were equipped 100 to 25% load range, are similar to a linear dependence
with FGR systems of various design. Increasing FGR flow from as indicated in Figure 19.18. On preheated air (300 to 500°F
0 to 14% (on the 165- and 210-MWe boilers) and to 18% (on the [150 to 260°C]), all curves are power functions with the
other boilers) influences deviations from the average test data ~1.25 and (1.1–1.2) degree numbers on gas and No. 6 oil,
curve but does not change the established dependence. respectively, as in Figure 19.19.
Similar dependencies have also been established on indus- A clear conclusion was made based on the above empirical
trial single-burner and multi-burner boilers, while firing test data: as load decreases (i.e., as the flame temperature is
either gas and/or oil, with ambient and preheated air, and reduced due to BZHR), the NOx level also decreases. A
without staged combustion. With ambient air, the shapes of reduction in the original NOx level (full load) limits the
the experimental curves, established on the 22,000 to potential opportunity to achieve a required NOx reduction
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 561 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 561

FIGURE 19.20 Effect of furnace cleanliness on NOx emissions.

at lower loads, resulting in a reduction in the effectiveness provide the required NOx reduction; but sustaining this NOx
of all NOx reduction methods. This conclusion was made reduction requires the improved maintenance and optimized
based on evaluations of data obtained on utility boilers, while operational protocols. With a balanced air flow distribution,
using all known NOx reduction methods. For example, on balanced fuel flow distribution, and a set of proven low-NOx
the TGME-206 boiler with 12 venturi burners, ~14% FGR burners, the combustion system’s foundation is established
implementation provides ~70 and ~56% NOx reduction at for the desired NOx reduction.
100 and 80% loads, respectively. For industrial boilers with However, the NOx reduction provided by the low-NOx
ambient air, this effectiveness reduction is a little more pro- burner is only the first part of the process. The boiler must be
nounced, due to the significantly lower flame temperature properly maintained to sustain the NOx reduction. Proper main-
level in industrial boiler furnaces. tenance must address boiler cleanliness, the percentage of air
in-leakage, and oil heater maintenance. The maintenance pro-
19.2.4 Boiler/System Condition Impacts on tocols affecting boiler cleanliness include proper soot blowing
Combustion and NOx Formation practices on a daily basis and boiler washes during outages.
A new philosophy for NOx reduction is emerging among
multi-burner power boiler users. This philosophy views the 19.2.4.1 Boiler Cleanliness
entire boiler as a combustion system where NOx reduction The impact of boiler cleanliness on NOx became very appar-
can be accomplished and maintained in a three-step process: ent when the results of NOx testing while firing residual fuel
1. installation of new, low-NOx burners to deliver the oil, on the same unit, showed a dramatic 23% increase in NOx
required NOx reduction as compared to results of NOx testing performed just months
2. improved maintenance protocols to sustain the achieved earlier, with the same atomizer and at the same unit condi-
NOx reductions tions. A visual inspection of the boiler indicated that it was
3. optimized operational protocols to achieve the best overall well-seasoned (or dirty). A waterwall wash was performed
combustion and boiler performance
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

during the next available outage, about a week later. After the
Boiler design parameters such as air preheater (APH) design, outage, another NOx test was performed, and the results of
superheater (SH) and reheater (RH) configurations, and the this testing showed a 25% decrease in NOx as compared to the
amount of refractory in the furnace also have major impacts pre-outage results. The effect of furnace cleanliness on NOx is
on the amount of NOx reduction achievable on a given boiler. shown in Figure 19.20.
The cornerstone of this NOx reduction philosophy is still The second indication of the impact of boiler cleanliness on
— as it has been for years — the low-NOx burner. Proven NOx came when a sister unit started up after the low-NOx
low-NOx burners, along with a physical windbox model, burner retrofit. The startup testing measured residual fuel oil
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 562 Wednesday, March 23, 2005 7:40 AM

562 The John Zink Combustion Handbook

FIGURE 19.21 Effect of HRA cleanliness on NOx emissions.

Effect of Air In-Leakage on Perceived Burner Performance

0.8 20

0.7 15
NOx (lb m/MBtu)

Opacity (%)
0.6 10

0.5 5

0.4 0
0 0.5 1 1.5 2 2.5 3
Excess O2 (%)
PRV #4 NOx Economizer Outlet PRV #4 NOx Top of Furnace PRV #3 NOx Economizer Outlet
PRV #3 NOx Top of Furnace PRV #4 Opacity Economizer Outlet PRV #4 Opacity Top of Furnace
PRV #3 Opacity Economizer Outlet PRV #3 Opacity Top of Furnace

FIGURE 19.22 Effect of air in-leakage on the burner performance.

(RFO) NOx higher than any that had been previously measured NOx on the sister unit fell to a level equal to those measured
with identical low-NOx burners and atomizers at the first unit. with identical low-NOx burners and atomizers at the first
As the data sets from the units were analyzed, it became unit. Figure 19.21 exemplifies the impact of heat recovery
apparent that the relative level of furnace cleanliness could area cleanliness on NOx.
be correlated to the flue gas temperature entering the air
preheater (APHGIT). It was also found that the second unit 19.2.4.2 Furnace Air In-leakage Influence
had an extremely high APHGIT. The unit’s soot-blowing The impact of air in-leakage on burner performance is a phe-
practices were analyzed to determine the level of heat nomenon of balanced draft boilers. There is a significant dif-
recovery area (HRA) cleanliness, and it was found that ference between NOx generation conditions existing in forced-
many of the soot blowers were out of service. Upon mini- draft (pressurized) and balanced-draft furnaces. In pressurized
mizing the number of soot blowers out of service, the RFO furnaces under typical single-stage (unbiased) conditions, all
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Boiler Burners 563

AIRFLOW RESULTS AT CONTRA COSTA #7

10
PERCENT DEVIATION IN A REFLOW

-5

-10

-15

-20

-25

-30

-35
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
1
2
3
4
5
6
7
8
9

MODEL BEFORE MODEL AFTER UNIT AFTER BURNER NUMBER

FIGURE 19.23 Improvement of mass flow distribution to burners (differences within ±2%).

the combustion air flow enters the furnace through the burn- than the control room O2 measurements taken at the econo-
ers, and the O2 required for complete combustion can be mizer outlet. In many cases, the mentioned difference
minimized. In balanced-draft furnaces, there is usually an air between O2 measurements is higher. This air in-leakage
in-leakage of at least 3 to 5% (it sometimes exceeds 10%), between the furnace exit and the economizer outlet can shift

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
sometimes leading to sub-stoichiometric combustion air flow the NOx vs. O2 and opacity vs. O2 curves upward along the
coming through the burners. In-leakage air participates in the O2 axis, as shown in Figure 19.22.
combustion process as well, providing complete combustion
at slightly higher O2 levels for many applications. 19.2.4.3 Fuel Oil Temperature
If the in-leakage location is within the furnace, the air Problems with the main fuel oil heaters can reduce the maxi-
deficit in the burners provides a self-stage combustion that mum fuel oil temperature, therefore raising the minimum vis-
complicates ICP burning out on one hand, but can reduce cosity attainable. The increased viscosity will increase the
NOx on the other hand (again depending on the in-leakage Sauter mean diameter (see Chapter 8) of the atomized droplets,
location). With other conditions remaining constant in pres- thus causing higher opacity. The increased droplet size will
surized and balanced furnaces, an NOx reduction of up to increase the excess O2 required for complete combustion. NOx
15 to 18% is available. This was established by testing three will increase due to the increased excess oxygen levels. The
sets of identical utility boilers: (1) 210-MWe boilers with viscosity required by most burner vendors is 80 to 100 SSU.
12 burners on the rear wall, (2) 300-MWe opposed fired boil-
ers with 16 burners, and (3) 150-MWe opposed boilers with
19.3 CURRENT STATE-OF-THE-ART
six burners. A comparison was made of NOx data measured
in these similar utility boiler sets, one operating with a bal- CONCEPTS FOR MULTI-BURNER
anced draft and one operating with a forced draft, while firing BOILERS
gas at full load and with ~0% FGR flow. 19.3.1 Combustion Optimization
However, if the in-leakage location is downstream of the 19.3.1.1 Windbox Air Flow Modeling
furnace exit, in-leakage can have detrimental effects on the Extensive experience in the application of oil and gas firing
perceived performance of the burners. O2 measurements taken equipment to a wide range of boiler designs has led to the
at the top of the furnace can be approximately 0.8% lower conclusion that, especially on multiple burner installations, it
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 564 Wednesday, March 23, 2005 7:40 AM

564 The John Zink Combustion Handbook

12:00
5000 Before Modeling
After Modeling
4000
10:30 1:30
3000

2000

1000

9:00 0 3:00

7:30 4:30

6:00
PRIMARY PERIPHERAL VELOCITY DISTRIBUTION

FIGURE 19.24 Improvement of peripheral air flow distribution to burners (deviations ±10%).

is imperative to achieve proper air distribution to each burner starved for air. This starved burner will generate a high CO
in order to control flame shape, flame length, excess air level, concentration and, consequently, the total O2 must be raised
and overall combustion efficiency. to minimize the formation of CO in that burner. By equalizing
Proper air flow distribution consists of an even combustion the air flow to each burner and ensuring that the fuel flow is
air mass flow distribution, even burner entrance peripheral equal, the O2 can be lowered until the CO starts to increase
flow distribution, the elimination of tangential velocities equally for all burners. Lower O2 has additional benefits of
within each burner, as well as an even combustion air O2 lower NOx formation and higher thermal efficiency. The goal
content by balancing the FGR distribution to each burner. is to bring the mass flow differences for each burner (in the
Considering that air in the combustion process accounts for model) to within ±2% of the mean, as shown in Figure 19.23.
approximately 94% of the mass flow, numerous observations Flame stability is probably the most important aspect of
on boiler combustion systems have shown that correct air the model that appeals to the boiler owner. Flame stability is
distribution and peripheral entry condition are key factors in enhanced in the model by controlling two parameters: perim-
the achievement of high performance (low NOx, low O2, and eter air inlet distribution and inlet swirl number (flame sta-
low CO). The concept of equal stoichiometry at each burner bility is primarily controlled in the burner design but must be
results in the minimal O2, NOx, and CO. The most direct way supported by proper inlet conditions). The equalization of the
to achieve this is to ensure equal distribution of air and fuel peripheral air velocity at the burner inlet will result in equal
to each burner. Equal air distribution is difficult because it mass flow of air around and through the periphery of the

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
requires a reliable and repeatable flow measuring system in swirler. The flame stabilizer will tend to equalize any remain-
each burner, as well as a means to correct the air flow without ing flow deviations because of the high velocity developed in
disrupting the peripheral inlet distribution or adding swirl to this region of the burner throat. The result of this equal air
the air flow. mass flow distribution through and around the swirler will be
The purpose of each objective relates to a specific burner a fully developed and balanced air vortex at the center of the
performance parameter as described below. outlet of the swirler. Flame stability and turndown of the
To achieve the lowest emissions of NOx, CO, opacity, and burner depend on the condition of this vortex and its attach-
particulates, at the minimum excess O2, equalization of the ment to the swirler. Unequal peripheral inlet velocity distri-
mass flow of air to each burner is required. Mass flow devi- butions result in an asymmetrical vortex, leading to a flame
ations should be minimized to enable lower post-combustion that has poor combustion performance and is more sensitive
O2, CO, and NOx concentrations. The lowest post-combustion to operating conditions, turndown may be limited, combus-
O2 concentration possible is constrained by the burner most tion induced vibrations may be experienced, FGR may cause
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 565 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 565

IMPROVED FGR DISTRIBUTION THROUGH MODELING

% FGR AT EACH BURNER (NORMALIZED TO

35.0%
Before Modeling
After Modeling
30.0%

25.0%
25%)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

20.0%

15.0%

10.0%
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1617 18 19 20 21 22 2324 25
BURNER #

FIGURE 19.25 Improvement of FGR flow distribution to burners.

flame instability at lower loads, light-off by the ignitor may reactants — and hence the products — in the combustion
be more difficult, and flame scanning may exhibit increased process. The increased mass, as well as the increased reactant
sensitivity. The goal is to reduce peripheral air flow deviations diffusion time requirement, reduces the overall flame temper-
to ±10%, as shown in Figure 19.24. ature. The burner with the least amount of FGR will theoret-
The swirl number is an indication of the rotational flow ically have the highest flame temperature, and will therefore
entering the burner. The creation of swirling air is a funda- have the highest NOx. Likewise, the burner with the highest
mental requirement of all burners. Louvered burners create amount of FGR will theoretically have the lowest NOx. How-
this swirl by rotating the entire air mass. Unfortunately, this ever, due to the exponential nature of the NOx/temperature
creates a problem at high turndown rates. At low loads relationship, given an equal deviation (e.g., ±5%), the higher
(e.g., 10%), excess O2 is typically 11 to 13%. By swirling the NOx values from the low FGR burners will outweigh the
entire air mass, the fuel is diluted to the point where flame lower NOx values from the high FGR burners. Minimizing
stability becomes marginal. Swirling air entering louvered the FGR deviations, as shown in Figure 19.25, will even out
burners (not created by the burner louvers) can cause differing the flame temperatures and therefore minimize the NOx for-
burner-to-burner register settings to match swirl intensity at mation from each burner.
each burner. The differing register positions consequently
Equal inlet velocities and elimination of swirl through each
affect the air mass flow at each burner.
burner are crucial to burner performance. Because low-NOx
An axial flow burner operates on the principle of providing
burners rely on injection of fuel at precise locations within
axial air flow through the burner and developing a controlled
burner air flow, it is imperative that the proper air flow be
limited vortex (swirl) of primary air at the face of the smaller
present at these locations. Likewise, for optimum perfor-
centrally located swirler. This concept maintains a stable flame
mance, the only swirl present must be that created by the
at the core of the burner by limiting dilution at high turndown
burner itself.
rates. The secondary air that passes outside the swirler, how-
ever, is most effective if it is not swirling (which is the concept No burner can be expected to simultaneously correct imbal-
behind “axial flow” burners). Swirling secondary air increases ances in the draft system and precisely control fuel/air mixing
the dilution of the fuel and limits turndown. The goal of both to minimize NOx formation. Air flow modeling prepares the
the louvered burner and the axial flow burner is to eliminate air flow for the burner, allowing the burner to precisely control
any tangential velocities entering the burner. fuel/air mixing for maximum NOx reduction. This approach
The thermal NOx from a burner increases exponentially has freed burner designers to focus solely on NOx control,
with an increase in flame temperature. The introduction of thereby increasing the effectiveness of known control tech-
FGR into the combustion air increases the overall mass of the niques to their maximum extent.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 566 Wednesday, March 23, 2005 7:40 AM

566 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 19.26 A scaled, physical, aerodynamic simulation model.

To achieve these goals, a scaled, physical, aerodynamic sim-

ulation model, similar to the one shown in Figure 19.26, can
be performed based on the physical dimensions and flow rates
within the field unit. The goals of the model are to enhance
flame stability; increase turndown; improve flame appearance;
reduce NOx, CO, opacity, and particulate at the minimum O2
level; and minimize mal-distributions of air flow that can occur
during both normal and unusual operating conditions. A scale
model allows for full observation and photographic recording
of the air flow within the scale model version of the new
windbox/burner configuration and existing ductwork.
The goals of the model are primarily accomplished by
installing secondary air duct and windbox baffles. The mod-
FIGURE 19.27 Flame-to-flame similarity of appearance. eler determines the location of baffles and turning vanes
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 567 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 567

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 19.28 Premixing the FGR flow with the combustion air upstream of the windbox.

within the combustion air/FGR supply system. As an added equal fuel pressures (assuming that equal fuel pressures
criteria, the windbox modifications must provide the minimal means equal fuel flow), the burner that is –5% in air flow
amount of impact on overall combustion air/FGR supply sys- will be running at an O2 level that is approximately 1% lower
tem pressure drop. This is to minimize the effects on existing than average.
fan performance. The major constraint in achieving these
For further balancing, the unit should first be brought to
objectives is the air distribution arrangement and windbox
maximum continuous rating (MCR) conditions, and O2
internal dimensions. The result of a windbox model can be
should be lowered to a point where the CO is in the 200 to
seen in the flame-to-flame similarity of appearance, as shown
400 ppm range on gaseous fuel, or opacity is in the 12 to
in Figure 19.27.
14% by EPA Method 9 for oil. The burner with the highest
CO/opacity should be found and the fuel to that burner
19.3.1.2 Fuel Flow Balancing Techniques
reduced slightly, resulting in a reduction in the overall
Fuel flow balance is just as important as air flow in reducing
CO/opacity level. This is an iterative process and should be
O2. A rough balance of fuel flow distribution is relatively
repeated until any further fuel adjustments either show no
easy to achieve by balancing pressure drops in the fuel
effect or increase the CO/opacity level.
header, or in other words, equalizing the fuel pressure at each
burner, thus ensuring that each burner is receiving the same The trick to this method of burner optimization is to find
amount of fuel. the burner with the highest CO/opacity. Ideally, on a multi-
Once the air flow has been balanced by a windbox model burner boiler, there should be a measurement grid at the
and the fuel pressures at each burner have been equalized, economizer outlet that has been mapped out for burner strat-
the unit is in the proper starting condition for the concept of ification and measurement probes placed along the burner
fuel balancing to be taken a step further to reduce O2 and centerlines. CO measurements taken from this grid, at the
maximize boiler efficiency. This goes back to the concept conditions specified above, typically give a clear indication
that there is probably one burner that is limiting the O2 of the burner with the lowest amount of stoichiometric air.
reduction. Even when the field data indicate that the air mass Lacking a mapped-out grid, when the boiler is brought to the
flow is within ±5% of average, and all the burners indicate conditions outlined above, a visual inspection of the flames
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 568 Wednesday, March 23, 2005 7:40 AM

568 The John Zink Combustion Handbook

will typically indicate the “problem burner” as the one that There are also applications where a comparatively cold exit
has the “dirtiest” or “sootiest” flame. flue gas (240 to 320°F [120 to 160°C]) is taken from the ID
fan outlet and supplied, for example, to the FD fan inlet or to
19.3.2 Methods to Reduce NOx Emissions the windbox. Usually, this flue gas contains much more excess
O2, especially on the boilers equipped with air preheaters hav-
Combustion modification techniques that have been devel-
ing comparatively large air leakage characteristics. Empirical
oped to reduce NOx emissions include low excess air opera-
data show that if this air leakage can be minimized, this “cool”
tion, fuel and air staged burner design, staged combustion
FGR can provide 25 to 30% greater relative NOx reduction as
(off-stoichiometric or biased firing), reduced air preheat, flue
compared to the “hot” FGR, but it has a more severe impact
gas recirculation, fuel-induced recirculation, reburn, and
on boiler thermal efficiency reduction. FGR flow can also be
water tempering. As discussed in Section 19.2.2, low excess
induced into the combustion air forced-draft fan inlet, allowing
air operation limits the oxygen availability in the combustion
the use of FGR without requiring a separate FGR fan.
zone and is highly effective in controlling fuel NOx forma-
tion and, to a lesser extent, thermal NOx. FGR can be taken from the furnace zone adjacent to the
The NOx problem forced the development of various com- burner exit, and can be induced back into the combustion air
pletely different furnace designs, usually implementing low- flow, also allowing the use of FGR without requiring a sep-
NOx burners and FGR systems, sometimes containing over- arate FGR fan. Unlike regular, comparatively cold FGR flows,
fired air ports for supply of air and/or air/FGR mixtures to this flue gas has temperatures of at least 2000 and 2200°F
the furnace space located above the burners. There are single- (1100 and 1200°C) with gas and oil firing, respectively; it
wall burner utility boilers where air ports are located on the consists primarily of unreacted air (O2 ~ 16–19%) and ICP.
target wall. Also, there are low-NOx industrial and utility This method is completely different from the ones described
boilers containing special furnace devices for steam/water above because its effectiveness cannot be associated with
injection in the combustion zone and for fuel injection in the significant changes in the flame temperature conditions or
post-combustion zone (a kind of reburning). lowering O2 concentration in the combustion air. The effec-
tiveness of this method relies on interactions between reagents
Actually, any of the mentioned methods is capable of
participating in NO formation and both radicals and ICP that
reducing NOx by up to 35 to 40%, but it cannot satisfy the
are present in high concentrations, which slow the NO for-
strictest of today’s NOx requirements. These applications
mation reactions rates, thereby reducing NOx output.
require implementation of at least two NOx reduction meth-
ods. It is important to note that, as with the example of boiler Flue gas can be recirculated in a number of ways: directly
load and FGR in Section 19.2.3, the efficiency of the imple- to the furnace through slots located under, above, around, or
mented methods depends on their sequence. The first NOx between the burners; through over-fired air ports or ports
reduction method is much more efficient than the second one. located on the target wall; premixing the FGR flow with the
This conclusion has been confirmed on many utility boiler combustion air upstream of the windbox (called premix FGR),
retrofits where up to four NOx reduction methods were used. typically accomplished via a sparger section as depicted in
That is why a combination using more than three methods on Figure 19.28; mixing the FGR at the burner exit (called ple-
the same boiler is typically inefficient. num FGR); and premixing the FGR with the gaseous fuel (on
oil, the flue gas would pass through the fuel gas piping). This
19.3.2.1 NOx Reduction by FGR Implementation technique of flue gas entrainment is called fuel dilution and
It is well-known that thermal NOx can be effectively con- uses no additional fan power because it is typically induced
trolled by reducing the flame temperature. The most effective into the fuel stream by the fuel pressure itself, or by pressure
technique is the use of FGR, which when mixed with the energy supplied via an intermediate media such as steam if
combustion air acts as a diluent to decrease the flame temper- the available fuel pressure is not high enough. Also, FGR can
ature, and sometimes (at high FGR rates) even increases the be supplied to the boiler hopper; this method is typically used
flame luminescence. FGR with lower temperatures will be only for steam temperature control.
more effective in flame temperature reduction. Various levels of NOx reductions can be achieved at the same
Usually, FGR is taken after an economizer where temper- FGR rates using these different techniques. Figure 19.29 (illustra-
atures range between 500 and 800°F (260 to 430°C), the O2 ting test data from 150, 165, 200, and 300 MWe boilers) and
range is typically between 1 and 3%, and actual ICP concen- Figure 19.30 (800 MWe boiler) show full load gas firing test
trations are negligible. Usually, an FGR fan is required to data received on utility boilers equipped with different burner
supply FGR flow to the combustion air flow or to the furnace.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
designs, numbers, and arrangements, with O2 = 0.8–1.4%.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 569 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 569

Relative NOx Concentration vs. FGR Flow Rates with Firing Gas at Full Load
1

0.8
Relative NOx Concentration

0.6

0.4
TGM-94, FGR into slots against the
burners
TGME-206, FGR into combustion
0.2 air
TGMP-314, FGR into peripheral air
TGME-114, FGR into slots below
the burners
0
0 5 10 15 20
FGR flow rate, %

FIGURE 19.29 Relative NOx concentration vs. FGR flow rate with firing natural gas in the utility boilers at full load.

Relative NOx Concentration vs. FGR Flow Rate

with Gas Firing in the 800 MW e boiler at full load
1
Relative NOx Concentration

0.8

0.6
two-stage
combustion

0.4

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
single-stage
combustion

0.2

0
0 5 10 15 20 25
FGR flow rate, %

FIGURE 19.30 Relative NOx concentration vs. FGR flow rate with firing natural gas in the 800 MWe boiler at full load.

It should be noted that all curves shown in these two figures Figure 19.29 shows that the most effective conventional
are exponential, meaning that it would make sense to deter- FGR supply system for low NOx is to provide a high-quality
mine an optimal operational FGR rate. This optimal limit will mixture of the FGR and combustion air, as with the venturi-
be different for each boiler application because it depends on style, low-NOx burners installed at the 200 MWe TGME-206
the boiler design (and accordingly on FGR impact on the boiler (single-wall burner installation), where 10, 12, and 18%
boiler operational parameters). The data presented show that FGR flows provided 54, 63, and 74% NOx reduction, respec-
the typical optimal FGR rate range for NOx reduction is 14 to tively. A comparatively high NOx reduction efficiency is also
18%. The NOx reductions experienced through FGR imple- reached on the 300 MWe TGME-314 boiler model (16 burners
mentation, as described above, explain why it is the most installed in two rows on opposed firing walls), where straight
frequently implemented NOx reduction technique on newly FGR flow is mixed with the burner air flow. Methods used to
designed and low-NOx retrofit boilers. supply FGR directly into the furnace, implemented on the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 570 Wednesday, March 23, 2005 7:40 AM

570 The John Zink Combustion Handbook

NOx reduction versus Flue Gas Flow

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
16
14
12
Percent Flue Gas

10
8
6
4
Fuel Dilution
2 FGR
0
0 10 20 30 40 50 60

Percent NOx Reduction

FIGURE 19.31 NOx reduction vs. flue gas flow.

TGM-114 (opposite wall burner installation, 150 MWe) and Recent developments have shown that, when firing gaseous
the TGM-94 (single wall burner installation, 165 MWe), are fuels, using fuel dilution (mentioned above) to induce flue gas
dramatically less effective. into the gaseous fuel is truly the most effective method of FGR
Figure 19.30 demonstrates the test data obtained on the supply (~25% more effective, as shown in Figure 19.31). As
800 MWe boiler of model TGME-204 (36 low-NOx burners per the test data, while natural gas firing at 2.5 to 4.5% O2,
installed in three rows on two opposed firing walls) with full NOx can be reduced by 66% with 12 to 13% fuel dilution flow.
load single- and two-stage gas combustion (with 33% air The reason why fuel dilution provides more efficient NOx
supplied through over-fired air ports located above the top reduction as compared to conventional FGR supply methods
burners) at O2 = 0.8–1.2%. As previously discussed, it is only is as follows.
natural that the NOx reduction provided by premixing FGR When natural gas (conditionally CH4) is fired with com-
with combustion air flow is less effective with two-stage bustion air, without the use of FGR or fuel dilution, every
combustion when the flame temperature level is significantly CH4 molecule encounters a certain quantity of oxygen mol-
lower. However, the 46 and 57% NOx reductions achieved ecules that is dependant on the excess O2 level, the mixing
with 18 and 24% FGR flow rates, respectively, during two- quality of the fuel with the air, and the diffusion rates of the
stage gas combustion can still be considered a comparatively CH4 and O2 at the flame front region, as shown in Figure
high efficiency because it is the second NOx reduction 19.32(a). The combustion reaction rates (including reactions
method implemented (see Section 19.3.2). At full load on this for incomplete combustion product burnout, prompt NOx,
boiler, the total NOx reduction exceeds 80% (from 600–620 and thermal NOx generation) are dependent on the actual
to 90–95ppm), and CO < 200 ppm at O2 = 0.8–1.2%. This excess O2, the combustion air temperature, and the fuel/air
NOx reduction was achieved utilizing a combination of low mixture quality. These variables also determine, along with
NOx burner retrofit, two-stage gas combustion, premixed the given burner/furnace design, the heat release and heat
FGR, and water injection in the primary and secondary com- distribution within the furnace, which in turn determines the
bustion air flows. The same scope of work was repeated on shape of the temperature profile curve, as well as the maxi-
three neighboring 800 MWe boilers of the same design mum temperature value and its location within the given
installed at this power plant. On all four units, total NOx furnace. Concurrently, these specific features of the temper-
reductions of 81 to 85% were achieved. ature conditions have a major impact on the determination of
Similar full load NOx reduction data has been obtained on the above combustion reaction rates.
packaged industrial boilers equipped with a single venturi- When CH4 is fired with combustion air containing typical
style, low-NOx burner, firing gas with ambient air. The data FGR, with all other combustion conditions being the same as
are presented in Figure 19.31. in the previous example, every CH4 molecule will encounter
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 571 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 571

FIGURE 19.32 Full load NOx reduction data has been obtained on packaged industrial boilers equipped with a single
venturi-style, low-NOx burner, firing gas with ambient air. (a) When natural gas (conditionally CH4) is fired with combustion
air, without the use of FGR or fuel dilution, (b) when CH4 is fired with combustion air containing typical FGR, with all other
combustion conditions being the same as in the previous example; (c) when CH4 is diluted with flue gas at the 1:1 ratio, with
all other combustion conditions being the same as in the previous examples.

a lower quantity of oxygen molecules than it would without temperature reduction caused by this phenomenon is depen-
FGR. The FGR impedes the diffusion of O2 toward the flame dent on the FGR temperature; relatively high temperature
front, as depicted by the increased combustion product con- FGR flows will provide less flame temperature reduction than
tent on the air side (slight purple coloring on the air side) in relatively low temperature FGR flows. The combination of
Figure 19.32(b). The difference between the quantity of oxy- decreased O2 concentration and lowered flame temperature
gen values in these two cases depends on both the FGR level results in reduced reaction rates of all the combustion
composition and the FGR flow rate. For example, for a boiler reactions mentioned above, including those reactions respon-
design operating under a positive furnace pressure (i.e., no sible for NOx generation in the given furnace.
air in-leakage in either the furnace or the convective path) When CH4 containing fuel dilution is fired with combustion
and providing an equal distribution of the combustion prod- air in the same furnace, combustion conditions change signif-
ucts in both the furnace exit cross-section as well as in the icantly from the examples previously discussed; because the
convective path cross-sections, the FGR composition will be CH4 is only ~6% of the mass flow within the combustion
similar to the combustion product composition leaving the process, the fuel dilution has a much greater dilution effect on
furnace. Because the combustion air makes up ~94% of the the CH4. If the fuel dilution rate is the same as the FGR rate
mass flow within the combustion process, by supplying an previously discussed (11%), the dilution is equivalent to a
FGR flow rate of ~11%, the oxygen concentration in the 2 fuel dilution:1 fuel molecular ratio, as evidenced by the
combustion air/FGR mixture is reduced by the same 11% heavy orange coloring on the fuel side in Figure 19.32(c). Due
(being only slightly diluted; this is why there is only a slight to this phenomenon, when a diluted fuel fires, the flame tem-
purple coloring on the air side in Figure 19.32(b)). Also, the perature becomes significantly less in comparison with the
flame temperature level in the furnace is reduced primarily second case, related to natural gas firing in the combustion air
due to the addition of the FGR mass. The magnitude of the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- containing the same FGR flow. The reduced flame temperature
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 572 Wednesday, March 23, 2005 7:40 AM

572 The John Zink Combustion Handbook

influences all of the above combustion reactions rates, includ- issues of combustion (especially in the case of firing natural
ing reactions that are responsible for NOx generation in the gas in ambient air) can be met — even at low FGR amounts.
given furnace. This is the explanation for why the NOx reduc- However, these problems are not typically experienced when
tion efficiency of fuel dilution is so much greater than that of FGR flow rates are below 14 to 18%.
FGR. On the other hand, the mentioned flame temperature Alternative techniques to reduce flame temperature involve
reduction (associated with the substitution of FGR with fuel reduced air preheat and water or steam injection. Water or
dilution of the same percentage and temperature) reduces heat steam injection will result in efficiency losses that are unre-
release in the furnace, increases the furnace exit flue gas tem- coverable, while reduced air preheat can be achieved with no
perature, and increases heat release in the boiler convective loss in efficiency if steam surface changes are made.
path, especially in the superheater and reheater sections. Also,
the flue gas temperature will be increased throughout the con- 19.3.2.2 Multi-stage Combustion on Utility Boilers
vective path, including the stack flue gas temperature, which Staged combustion involves delaying the mixing of fuel and air,
reduces the boiler thermal efficiency. and is effective for both thermal and fuel NOx control. Typi-
Fuel dilution can also be implemented in combination with cally, staged combustion creates an initial fuel-rich combustion
FGR premixed in the combustion air and/or steam injection. zone with air added downstream to complete combustion.
A combination of fuel dilution and steam injection can pro- Staged combustion can be achieved using low-NOx burners
vide a total NOx reduction of up to ~82%. (internal staging), over-fired air ports, operating existing burn-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Low flue gas recirculation rates (<20%) typically have a ers with biased fuel firing, or with burners-out-of-service
relatively minor impact on boiler performance (in many cases, (BOOS). Biased fuel firing for a large, wall-fired unit is imple-
on utility boilers it can be compensated for by increasing the mented by operating the lower burners with reduced combus-
turbine’s efficiency; on industrial boilers, this loss is unrecov- tion air and/or increased fuel to produce a fuel-rich zone in the
erable). However, high FGR rates (>20%) can significantly lower portion of the furnace. The upper burners are operated
impact boiler performance, especially the redistribution of with a corresponding increase in air or decrease in fuel to
heat flux between radiant surfaces and high-temperature con- complete fuel burnout. An extreme form of staging involves
vective heat exchange surfaces of utility boilers and, accord- operation with an upper row of burners out of service or its
ingly, on superheated/reheated steam temperature. Excessive equivalent over-fired air ports. In this case, fuel is shut off to
use of flue gas recirculation can result in significantly these burners and the burners are operated with air only.
increased superheated and reheated steam temperatures, or The principle of stagewise fuel combustion involves the
significantly increased superheat and reheat attemperation arbitrary division of the flame into two or more stages. In
spray flows, the latter affecting boiler efficiency. In addition, most cases, in the first high-temperature stage, combustion
power expenses recalculated for thermal efficiency losses are takes place with an excess air factor of less than unity; in
estimated at 0.03 to 0.05% per 1% FGR. These facts explain subsequent stages, which have a comparatively low temper-
why FGR was (before NOx requirements were established) ature level, secondary combustion of the previous ICPs occurs
typically only used to control and regulate reheated and super- with relatively high excess air. Thus, NOx formation is
heated steam temperatures at minimum loads. A typical older retarded in the first stage because there is insufficient oxygen
boiler would be designed to provide the required steam in the reaction zone, and in the subsequent stages due to the
temperature and flows without FGR at full load; but at partial relatively low temperature of the flame. This has established
loads (usually 70%), FGR was switched on to maintain the advantages of stagewise combustion as an NOx reduction
steam temperatures. FGR flow rates increased as loads con- method over other methods.
tinued to decrease. Modern boiler designs incorporate FGR There are various methods of performing staged combus-
for both NOx reduction and steam temperature control. At tion in furnaces. The best known and most studied methods
higher loads (100 to 80%), FGR is typically used for NOx are based on the following three options:
reduction; however, as above, at partial loads, FGR is typi-
1. gas/oil combustion with a significant air deficit in all
cally used for steam temperature control.
burners and supplying a certain part of the total air flow
Very large amounts of premixed FGR, above 45%, can directly to the furnace through ports or slots located above
affect flame stability with conventional burner designs, while the burners
plenum FGR has been tested on some burners up to 70% with 2. redistribution of air flows between the burners located at
no noticeable effect on flame stability. Depending on the FGR different rows
supply method, the FGR/air mixing quality will be different, 3. redistribution of fuel flows between the burners located at
and, accordingly, flame stability, vibration, and other negative different rows
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 573 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 573

Overfire air flow rate influence on NOx reduction and CO emissions

with firing gas in the TGM-94 boiler at ~94% load and O2 ~1.2%
1000 100
900 90
800 80
700 70
CO @ O2 = 3%, ppm

NOx reduction, %
600 60
500 NOx reduction 50
400 40
300 30
200 20
100 CO
10
0 0
0 5 10 15 20 25 30 35
relative over fire air flow rate, %

FIGURE 19.33 Overfire air flow rate influence on NOx reduction and CO emission with firing natural gas in the TGM-94
boiler at ~94% load and O2 ~1.2%.

The mentioned options are directed to get a “vertical” fuel/air provides reductions in opacity, carcinogenic substances, and a
unbalance in the furnace: excess fuel and air deficit in the small CO reduction from ~38 to ~24 ppm. However, NOx
lower part and, on the contrary, excess air and fuel deficit in increases very quickly; on average, 1% of the secondary air flow
the top part of the combustion zone. However, when a boiler brings a relative NOx increase of ~1.5%. The boiler data from
changes over to stagewise fuel combustion with a “vertical” this testing indicated that this method can be applied to appli-
unbalance in the fuel/air ratio, concentrations of all ICPs, cations requiring increased reheated/superheated steam temper-
including benz(a)pyrene and other carcinogenic substances, atures. Moreover, this method increases the boiler thermal
increase. Much test data exist confirming the above conclu- efficiency and reduces operational power expenses due to a
sion. An example is shown in Figure 19.33, where test data decrease in the air resistance of the air path. On the above boiler,
obtained on a model TGM-94 boiler are presented. It can be ~15% secondary air flow increased the total efficiency by 1.1%.
seen that a 50% NOx reduction while firing gas is reached at The data indicate that an optimally designed three-stage fuel
~18% over fire air flow, but CO has increased to ~200 ppm. A combustion system could be developed with simultaneous
further increase in the overfire air flow to 25% results in a “vertical” and ”horizontal” imbalances of the fuel/air ratio.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
~60% NOx reduction; however, the CO has increased to Using calculations explaining the above data as a basis for
~400 ppm. Increases in CO concentrations and other ICPs are preliminary design work, a new furnace was developed. The
typical of “vertical” staged combustion. The data for this furnace had three burner rows and slots above the upper row
boiler also indicated that operation while firing gas at ~18% located on the furnace front wall. FGR could either be pre-
overfire air flow reduced the boiler efficiency by ~0.85%. mixed with the combustion air going through the burners and
Another system of stagewise combustion, with a “horizontal” slots; supplied directly into the furnace through slots located
unbalance of the fuel/air ratio, gives rise to substantial reductions on the rear wall, opposite the lower burners; or in both loca-
of ICP emissions. However, it results in increased NOx. This tions. The primary air flow was supplied through the burners,
method of stage combustion utilizes the interaction of fresh and secondary and tertiary air flows were redirected to the
combustion air jets directed from ports or slots located opposite slots located on the rear and front walls, respectively.
the burners, with fuel radicals and already formed ICP in flames Tests were performed over a boiler load range of 100 to
generated by burners receiving sub-stoichiometric air. It inten- 40%, α = 1.05–1.11 firing gas, and at full load, α = 1.03–1.1,
sifies the combustion process and especially burn out while firing firing No. 6 oil. To obtain reliable data on the NOx reduction
oil, as illustrated in Figure 19.34 where test data were obtained effectiveness of stagewise combustion, tests were carried out
on a TGM-84 (~930,000 lb/hr or 420,000 kg/hr steam flow) co- in alternate modes, with measurements being made during
generating boiler firing No. 6 oil. The data show that this method single-stage combustion under the same load, excess air
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 574 Wednesday, March 23, 2005 7:40 AM

574 The John Zink Combustion Handbook

FIGURE 19.34 “Horizontal” imbalance intensifies the combustion process and especially burnout while firing oil.

FIGURE 19.35 Data comparing single- and three-stage gas combustion.

factor, FGR, and other combustion conditions. Comparative air flows of 15 and 25%, respectively. This NOx reduction is
tests when firing oil were carried out with oil supplied from accompanied by a corresponding CO increase from ~135 to
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

the same tank. All results were compared with the baseline 165 ppm during single-stage combustion to 260–290 and
conditions of single-stage combustion, in which the second- 350–400 ppm, accordingly. Under the above conditions, oper-
ary and the tertiary air flow dampers were fully throttled, it ation with FGR was not required at any point over the entire
being conditionally assumed that their relative flows were ~0. load range. The decrease in thermal efficiency was 0.3 to
The test data comparing single- and three-stage gas com- 0.4%, much less dramatic than either two-stage combustion
bustion are shown in Figure 19.35. It is seen that transferring (with a “vertical” imbalance) or FGR fan operation. The same
from single- to three-stage combustion at ~10% secondary retrofit was performed on six neighboring TGM-94 boilers
air flow yields NOx reductions of ~40 and ~63%, with tertiary installed at this power plant.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 575 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 575

19.3.2.3 Special Features of Low-NOx Gas/Oil When firing oil containing Nf, there are two sources of
Combustion on Utility Boilers NOx formation: air nitrogen and fuel nitrogen. In addition
It was shown in Sections 19.1.2 and 19.2.2.2 that there are some to the two mechanisms available for gas firing, a third mech-
utility boilers where NOx is higher while firing natural gas than anism is available as well: oxidation of the partially
that experienced while firing No. 6 oil. For example, data destructed nitrogen-containing compositions that occur in the
obtained on the TGM-94 boilers presented in Figures 19.11 to beginning of the flame, whose concentrations decrease
19.14 demonstrates that the NOx while gas firing is ~24 to slightly with increasing temperature.
27% higher than NOx data obtained while firing No. 6 oil Figure 19.10 shows calculated NOx curves for both nitro-
under the same load (~94%), O2 (1.4 ± 0.2%), and other gen sources, along with their summed values. It shows that
operational conditions. oxidation of 0.5% Nf can generate NOx of up to ~200 ppm.
This phenomenon (i.e., NOx while firing gas is higher than In considering the processes in the comparatively cold fur-
NOx data obtained while firing No. 6 oil) has been experienced naces designed for industrial and small utility boilers (50 to
during testing of many former U.S.S.R. utility boilers in the 100 MWe), the above 200 ppm will far outweigh the thermal
size range of 165 to 800 MWe. In some cases, NOx while firing NOx formation. Because these processes are occurring at
gas exceeded the NOx while firing No. 6 oil (even containing relatively low temperatures, they indicate relatively low NOx
0.67 to 0.73% Nf) by ~50%. On the contrary, on utility boilers levels from these boilers and an oil firing NOx level much
in the size range of 25 to 80 MWe, the NOx while firing No. higher than the gas firing NOx level. In the 120 to 150 MWe
6 oil exceeded the NOx while firing gas by ~25%. Moreover, boilers, temperatures can reach 2900 to 3100°F (1600 to
as Nf values are increased (from 0.22 to 0.73%), the greater is 1700°C). Figure 19.10 indicates that the curves calculated for
the margin between the NOx while firing No. 6 oil and the gas and oil firing intersect within this temperature interval,
NOx while firing gas. Likewise, on industrial boilers, the NOx indicating relatively equal NOx levels while firing either gas
was almost always higher while firing No. 6 oil, and differences or oil. For units greater than 165 MWe, the flame temperature
of 24 to 36% and 12 to 18% were typically experienced with can reach between 3200 and 3500°F (1800 to 1900°C). This
ambient and preheated air, respectively. However, intermediate temperature increase results in a gas firing NOx level much
results were obtained while testing utility boilers in the size higher than the oil firing NOx level.
range of 100 to 150 MWe, where differences between the NOx The above explanation can be confirmed with the data
numbers for the two fuels are less than ±10%. The controversial presented in Figure 19.36, where the ratio between NOx
data can be explained in terms of the well-known influence of numbers obtained with gas and oil firing are considered as a
the flame temperature on NOx formation. function of the furnace space heat release. Furnace designs
While firing natural gas containing no Nf, NOx can only for increasingly higher capacity utility boilers are usually
be generated by the oxidation of nitrogen contained in the (with very few exceptions) accompanied by a significantly
combustion air. This NOx generation can only occur by one increasing heat release as related to the furnace volume. Thus,
of two mechanisms: (1) thermal NOx formation due to flame the design of these boilers allows the analysis of the ratio
temperatures of at least 3000°F (1650°C), mostly in the burner between NOx numbers and this design parameter. Despite
zone heat release (BZHR) space — its concentration rises experimental data deviations from the average empirically
with temperature according to an exponential dependence; established curve, the strong character of this dependence
and (2) prompt NOx formation in the beginning of flame cannot be doubted.
where the temperature level is significantly lower; under usual

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
combustion conditions, its concentration is much less than
the thermal NOx concentration. Prompt NOx has a slight
dependency on temperature. Under certain furnace conditions
19.4 LOW-NOx BURNERS FOR
required for low-NOx combustion, the thermal NOx concen- PACKAGED INDUSTRIAL
tration can be reduced to the prompt NOx concentration level BOILERS
— or even less. However, these conditions have little influ- There are a few state-of-the-art low-NOx burner (SLNB)
ence on prompt NOx. designs developed for typical industrial packaged boiler
While firing natural gas (conditionally consisting of only applications, including those of several burner manufacturers
methane), the maximum combustion temperature level depends (e.g., Todd Combustion, Coen, Forney, Alzeta, Peabody,
only on excess air and the combustion air temperature, as shown Pillard, Babcock Hitachi, and others) and consulting compa-
in Figure 19.10. The corresponding NOx concentration curves, nies (e.g., IGT, J.Lang, RJM, EPT, and others) in both the
calculated for excess air factor α = 1, are shown in Figure 19.10. United States and other countries. A schematic for a typical
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 576 Wednesday, March 23, 2005 7:40 AM

576 The John Zink Combustion Handbook

Ratio between NOx values with firing #6 oil and gas

vs. furnace space heat release of utility boilers
1.6
1.5
1.4
(NOx) oil/ (NOx) gas ratio

1.3
1.2
1.1
1
0.9
0.8
0.7
0.6
0 2 4 6 8 10 12 14 16
furnace space heat release, KW / cub. ft

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 19.36 The ratio between NOx numbers obtained with gas and oil firing are considered as a function of the
furnace space heat release.

FURNACE FRONT WALL

IFGR

COOLER OXYGEN RICH

ZONE REDUCES
FUEL SPRAY THERMAL NOx
OIL INLET

GAS INLET
LOW EXCESS AIR
ZONE REDUCES
FUEL NOx
OIL ATOMIZER
AIR SWIRLER
STAGED INJECTOR GAS
AIR INLET
TERTIARY
AIR INLET
PRIMARY / SECONDARY

FIGURE 19.37 The Todd Combustion low-NOx gas/oil burner.

packaged SLNB is depicted in Figure 19.37. This burner was the inner gas poker assembly implemented in the prototype.
developed using a standard venturi-type low-NOx burner as a Also, the SLNB utilizes internal furnace flue gas recirculation
prototype. By comparing the schematic for this burner with the (IFGR) instead of the external flue gas recirculation that was
schematic for a typical standard venturi-type low-NOx burner used by the prototype and other standard low-NOx burners.
shown in Figure 19.5, it can be seen that whereas the air chan- The SLNB shown in Figure 19.37 has been installed and
nels are the same as in the standard low-NOx burner, the tested at four industrial retrofit boilers where all existing
SLNB has two completely independent gas supply units: the auxiliary equipment, including fans, dampers, and combus-
center fire gas (CFG) and a set of outer gas injectors, replacing tion control systems, was kept. Before installation, the burner
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 577 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 577

NOx vs. heat load

NOx corrected @ O2 = 3%, ppm

0
0 20 40 60 80 100
heat load, MMBTU/h

FIGURE 19.38 NOx vs. heat load with firing natural gas with ambient air in the packaged industrial boiler equipped with
the Todd Combustion low NOx burner.

and windbox were modeled in order to provide equal air The emission test data sets obtained on this boiler, while firing
distribution. The SLNBs have designed heat inputs ranging No. 6 oil (0.22% Nf) at full load, are shown in Figure 19.40. At
from 28 to 134 × 106 Btu/hr (8.2 to 39.3 MWt). At full load, O2 = 3.5%, the following data are recorded: NOx ~ 200 ppm,
the burners usually operate at moderate gas pressures CO ~ 100 ppm, and opacity < 10%. At partial loads, NOx is
(< 6 psig) and at reasonable register draft loss (RDL < 7 in. sufficiently less at the same O2, CO, and opacity.
or 180 mm w.c.). Test data obtained on two industrial boilers A significant NOx reduction (~15 to 20%) is realized
(one with ambient air and one with preheated air) are through the use of 4 to 5% IFGR while firing gas with ambient
described and compared below. air. The IFGR contains relatively high ICP concentrations and
The first application was a 70,000 lb/hr (32,000 kg/hr) B&W has a temperature of about 1800 to 2000°F (980 to 1100°C).
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FM 103-70 boiler generating 250 psig (17 barg) saturated steam, The same relative NOx reduction due to IFGR was experienced
firing natural gas with ambient air. The burner was designed while firing No. 6 oil, as shown in Figure 19.41. This figure
for 84.8 × 106 Btu/hr (25 MWt) at full load. Over a 100 to 25% indicates an NOx reduction of ~20%. Although the IFGR tem-
load range, the installed burner has demonstrated excellent perature is higher as compared with gas firing (~200°F against
flame stability, high reliability, and highly efficient operation 2000°F), relatively higher ICP concentrations contained in this
while firing both natural gas and No. 6 oil as fuels. Measure- flow provide the mentioned NOx reduction effect. For both
ment has shown that the noise level is less than 85 dBA. fuels, over the entire tested load range, there was no significant
Vibration of boiler/burner parts is at regularly accepted levels. influence of IFGR on CO. There was also no influence on
opacity while firing No. 6 oil.
NOx emission test data obtained on the above boiler is
The second installation of the SLNB shown in Figure 19.37
illustrated in Figure 19.38. These data sets indicate that under
was on a B&W FM boiler designed to generate 100,000 lb/hr
optimal combustion conditions, the SLNB typically generates
(45,000 kg/hr) superheated steam at 415 psig (28 barg) and
NOx in the 23 to 28 ppm range and CO < 200 ppm (typically
600°F (320°C), while firing natural gas with ~415°F (213°C)
< 100 ppm). Due to the advanced gas staging provided by
preheated air. The specific features of this application as com-
this burner, O2 concentration changes in the range of 2.6 to
pared to the prior application are the following:
5.1% had practically no impact on NOx and a relatively low
impact on CO (Figure 19.39). At 25% load, the emission data • 60% higher heat input (134 × 106 Btu/hr vs. 84.8 × 106
were close to the data obtained at full load, as shown in Btu/hr, or 39 MWt vs. 25 MWt )
Figures 19.38 and 19.39. These results were numerously • 45% hotter furnace (BZHR: 245 × 103 Btu/ft2-hr vs. 168
repeatable over a 7-month testing period. × 103 Btu/ft2-hr)
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 578 Wednesday, March 23, 2005 7:40 AM

578 The John Zink Combustion Handbook

FIGURE 19.39 CO vs. heat load with firing natural gas with ambient air in the packaged industrial boiler equipped with the
Todd Combustion low NOx burner.

NO x, O 2 and opacity vs relative load

with 5-7 % IFGR and CO = o
15

200 12
NO @ O2=3%, ppm

N0x
150 9
opacity %
opacity
O2, %
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

100 6
02

50 3

0 0
40 50 60 70 80 90 100
relative load, %

FIGURE 19.40 NOx, O2, and opacity vs. relative heat load with firing No. 6 oil with ambient air in the packaged industrial
boiler equipped with the Todd Combustion low NOx burner.

• pre-heated air (415 vs. 80°F, or 213 vs. 27°C) while CO was kept in the range of 30 to 90 ppm. This
• excess air 15% (vs. 23%) reduction effect is very similar to the ambient air data where
• significantly lower CO level required (100 vs. 200 ppm IFGR provided a 15 to 20% NOx reduction. These data con-
at O2 = 3%) firm that the effectiveness of IFGR, established with ambient
• required NOx = 0.22 lb/106 Btu (182 ppm at O2 = 3%) air, is similar to preheated air.
• available gas pressure < 10 psig (vs. ~15 psig, or 1 barg)
The full load test data shown in Figure 19.42 have con-
• available register draft loss = 7.25 in. (18 cm) w.c.
firmed a previous conclusion based on the data obtained on
For this application, the IFGR was used to reduce NOx ambient air, that the ratio between injector and total gas flow
from 160–180 ppm to 120–150 ppm (14 to 18% reduction), rates is the major parameter dictating NOx reduction under
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 579 Wednesday, March 23, 2005 7:40 AM

Boiler Burners
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
579

FIGURE 19.41 Internal FGR impact on NOx with firing No. 6 oil with ambient air in the 30–100% load range.

FIGURE 19.42 NOx and CO emissions vs. injector gas flow rate/total gas flow rate ratio, at the boiler equipped with the
Todd Combustion low-NOx burner firing natural gas with preheated air at full load.

the natural gas staged combustion conditions formed with NOx on O2 does not work for the gas staged combustion
SLNB implementation on packaged boilers — not only on conditions formed by an SLNB. This confirms the conclusion
ambient air but also on preheated air. made based on the test data for ambient air.
There was an unsuccessful attempt to find an O2 influence Unlike the typical well-documented test data obtained with
on NOx in the data, meaning that a regular dependence of conventional low-NOx burners, indicating an NOx level that
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 580 Wednesday, March 23, 2005 7:40 AM

580 The John Zink Combustion Handbook

In developing a new burner technology, the control of CO

and VOC emissions in addition to NOx was targeted, and an
ultra-low emissions burner, shown in Figure 19.43, was the
result. In addition to reliably controlling NOx emissions
below 9 ppm, CO emissions are kept below 25 ppm, and VOC
emissions are kept to less than 3 ppm.

19.5.2 Implementation of the NOx Formation

Theory for Ultra-low-NOx Emissions
The relationship between temperature, stoichiometry, and
NOx formation was used as the basis for design of an ultra-
low emissions burner for natural gas and other nitrogen-free
fuels. Thermal NOx emissions are the major source of NOx
FIGURE 19.43 An ultra-low emissions burner. from these fuels, with NOx being created through the follow-
ing high-temperature reactions of atmospheric nitrogen and
oxygen from the combustion air (see Chapter 6):
increases with load, this was not experienced during any of
the SLNB testing. For the preheated air application, the full N + O 2 = NO + O
load NOx was actually slightly lower as compared to the data N + OH = NO + H
at partial loads. In any case, the test data described above
have shown that in the 4:1 regular emission turndown, burner N 2 + O = NO + N
characteristics have matched today’s requirements for both
ambient and preheated air. Thermal NOx formation can be reduced by controlling the
peak flame temperature. While thermal NOx is dependent to

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
some extent on oxygen availability, if the temperature can be
19.5 ULTRA-LOW EMISSION GAS lowered sufficiently, thermal NOx from a natural gas flame
BURNERS can be reduced to less than 1 ppm. Figure 19.44 shows the
relationship between the adiabatic flame temperature and
19.5.1 Background thermal NOx formation.
NOx formation fundamentals were used to design an ultra-
In attaining ultra-low-NOx levels, prompt NOx also
low-NOx burner with a gas injection and mixing system
becomes a significant emissions source. Under fuel-rich con-
radically different from all other commercially available
ditions, particularly when stoichiometry is under 0.6, both
low-NOx burners. The goal was to produce a burner capa-
HCN and NH3 can be formed through the extremely rapid
ble of reliable operation with single-digit NOx generation
reaction of CH with N2 to form HCN and N. The following
while firing nitrogen-free gaseous fuels, such as natural gas.
reactions are considered:
A new gas mixing approach was incorporated into an estab-
lished burner geometry that had been optimized over the years CH + N 2 = HCN + N
at the International Flame Research Foundation (IJmuiden,
The Netherlands) to provide an extremely stable flame. N + H 2 = NH + H
Previous low-NOx burner designs focused primarily on the
NH + H 2 = NH 3
reduction of NOx through techniques to reduce the formation
of thermal NOx. Techniques such as fuel-air staging, flue gas HCN + O 2 = NO + HCO
recirculation, and steam injection serve to lower NOx emis-
sions into the 20 to 30 ppm range by lowering the peak flame Below a stoichiometry of 0.5, almost all NOx formed is
temperatures. These methods do little to address the other attributable to prompt NOx. The rate of formation of
mechanism for NOx formation — prompt NOx — and thus prompt NOx is very rapid, being complete in under 1 ms.
are not able to reliably reach single-digit NOx levels. Because Although prompt NOx is temperature sensitive, the tem-
some of these techniques also rely on delaying combustion perature sensitivity is not as great as with thermal NOx.
and lowering reaction temperatures, incomplete combustion Unlike thermal NOx, simply lowering the peak flame tem-
leads to increased CO and VOC emissions. peratures will not reduce the prompt NOx into the single-digit
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 581 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 581

FIGURE 19.44 Relationship between the adiabatic flame temperature and thermal NOx formation.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 19.45 HCN and NH3 formation at three flame temperatures.

range. As indicated by the curves in Figure 19.45, depicting complete combustion, oxygen availability is limited, peak
HCN and NH3 formation at three flame temperatures, under flame temperature is lowered, and thermal NOx formation is
fuel-rich conditions and a temperature of 2400°F (1300°C), reduced. It can be further reduced through other techniques,
20 ppm of prompt NOx still remain. To further control the such as the addition of FGR or steam injection. However, the
formation of prompt NOx, it is necessary to take steps in the 20 ppm of prompt NOx created in the initial fuel-rich zone
burner design to minimize the formation of sub-stoichiomet- remains. It is this prompt NOx formation that has prevented
ric regions within the flame. conventional low-NOx burners from achieving sub-10 ppm
NOx levels.
19.5.3 Ultra-low Emissions Burner Design By “starting over” with NOx formation fundamentals, it
Most conventional low-NOx burners utilize staged combus- was determined that the most direct method of achieving very
tion to delay the mixing of fuel and air. By creating an initial low NOx emissions from a natural gas flame is to: (1) avoid
fuel-rich combustion zone and adding air downstream to fuel-rich regions with their corresponding potential for
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 582 Wednesday, March 23, 2005 7:40 AM

582 The John Zink Combustion Handbook

FIGURE 19.46 A nearly uniform fuel/air mixture at the ignition point.

prompt NOx, and (2) lower the flame temperature to reduce 19.5.4 Ultra-low Emissions Burner
thermal NOx to the desired level. To accomplish this, a burner
Operation
design that avoids fuel-rich regions by rapidly mixing gaseous
The basic ultra-low emissions burner consists of a parallel-
fuel and air near the burner exit was developed. The rapid
flow air register with no moving parts. Combustion air pre-
mixing results in a nearly uniform fuel/air mixture at the
mixed with FGR enters the register, and the entire mixture
ignition point (Figure 19.46), which virtually eliminates
then passes through a set of axial swirl vanes. These vanes,
prompt NOx formation. This rapid and complete combustion
which are attached to a central gas reservoir, have hollow
is also what results in the virtual elimination of both CO and
bases that are perforated for gas injection. Thus, the swirl
VOC formation by the burner. Thermal NOx is then mini-
vanes also serve as the gas injectors and provide the burner’s
mized using FGR, which is mixed with combustion air
near-perfect fuel/air mixing (Figure 19.47).
upstream of the burner, to control flame temperature. In effect,
For burner heat inputs greater than about 40 × 106 Btu/hr
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

the burner performs like a pre-mix burner with one important

(12 MWt), a second parallel-flow air sleeve surrounds the
distinction: because the fuel is added inside the burner, just
basic burner register. The outer sleeve contains a second set
upstream of the refractory throat, the extremely small pre-
of gas injector vanes attached to an outer gas reservoir
mixed volume eliminates the possibility of flashback inherent
(Figure 19.48). These vanes, however, do not impart any swirl
in pre-mix burner designs.
to the air flow. Air flow through the inner and outer burners
Contrary to conventional LNB theory, increasing excess air is designated as “primary” and “secondary” air flow, respec-
reduces NOx formation in the ultra-low emissions burner. tively. Both the primary/swirled and secondary/axial zones
Because the burner employs near-perfect mixing, the fuel operate with the same, near-perfect mixing.
already has access to all of the oxygen required at the ignition In addition to sub-9 ppm NOx, another benefit of the rapid
point, and increasing excess air just serves to reduce the peak mix design is an extremely stable flame. Swirler geometry,
flame temperatures. Therefore, excess air has the same cool- burner internal geometry, and quarl expansion are matched
ing effect as FGR, which provides major advantages in high to promote internal recirculation of a large amount of hot
excess-air combustion applications where FGR is impractical combustion gases. This enables the burner to operate at lower
or unavailable. In boiler applications where FGR is available, flame temperatures and NOx levels than other burners, with
it is preferred due to its lower impact on the boiler efficiency a “blow-off” point of about 3 ppm NOx. The ultra-low emis-
than high excess air. The rapid mix design is also what allows sions burner flame remains stable at 60% FGR; therefore, the
the burner to operate with preheated combustion for increased 25 to 30% FGR rate typically necessary for sub-9 ppm NOx
efficiency and still retain its single-digit emissions perfor- does not begin to approach the burner’s performance limits.
mance, by simply increasing the FGR or excess air rates to The rapid combustion also results in a very short flame length.
compensate for the higher air temperature. It is approximately half that of a staged combustion burner,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 583 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 583

which eliminates the potential for flame impingement, one of

the most common problems experienced with conventional
LNB retrofits.
For oil firing, the ultra-low emissions burner uses a con-
ventional center-mounted atomizer gun assembly. The burner
operates as a conventional, staged combustion LNB when
firing oil. While there is no rapid mixing, fuel staging is
provided by advanced atomizer designs, and air staging is
provided by the burners’ primary and secondary air flows.
This allows emissions performance on oil firing equivalent to
any other low-NOx burner available.
NOx emission data from the ultra-low emissions burner fir-
ing into a 4 × 106 Btu/hr (1.2 MWt) firetube boiler for ambient,
300°F (150°C) preheat, and 500°F (260°C) preheat as a func-
tion of FGR rate are shown in Figure 19.49. NOx without FGR
was a function of air preheat level and ranged from approxi-
mately 80 ppm with ambient air to 200 ppm with 500°F
(260°C) preheat. The FGR rate required to produce a given
NOx level varied with air preheat level; but, independent of the
preheat level, NOx could be reduced to approximately 3 ppm.
Similar results were obtained in a 25,000 lb/hr (11,000 kg/hr)
watertube boiler (30 × 106 Btu/hr or 9 MWt). The NOx emissions FIGURE 19.47 Swirl vanes also serve as the gas injec-
without FGR were higher for the larger burner but when enough tors, and provide the burner’s near-perfect fuel/air mixing.
FGR was added to reduce NOx below 20 ppm, both size burners
exhibited very similar performance. Similar characteristics were
also observed with air preheat on a 100 × 106 Btu/hr (30 MWt)
test furnace. Again, the larger burner produced higher NOx
without FGR; but once FGR was added, the NOx from the
4 × 106 Btu/hr and 100 × 106 Btu/hr (1.2 and 30 MWt) burners
were almost identical for a given FGR rate.
Based on this test work, the ultra-low emissions burner has
been applied to boilers ranging from 25,000 to 230,000 lb/hr
(11,000 to 100,000 kg/hr) or 30 to 275 × 106 Btu/hr (9 to
80 MWt). Repeatable performance has been observed across
the various applications and size ranges, whether using FGR
or excess air for NOx reduction, and single-digit emissions
performance has been consistently achieved. Figure 19.50
shows the actual performance data taken from the applica-
tion of an ultra-low emissions burner on a new 230,000 lb/hr
(100,000 kg/hr) “A” type Nebraska Boiler. The ultra-low
emissions burner can also be used on two-burner applications
where NFPA 8501 guidelines are being followed. An exam-
ple of this is shown in Figure 19.51.
Repeatable performance of this burner design has been
demonstrated across a wide range of applications and sizes,
and single-digit emissions performance has been consistently
achieved. The operational performance of the burner has
shown that ultra-low emissions technology has been proven
as a reliable alternative to the control of combustion emissions FIGURE 19.48 Outer sleeve contains a second set of
through the use of flue gas cleaning technologies, such as
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
gas injector vanes attached to an outer gas reservoir.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 584 Wednesday, March 23, 2005 7:40 AM

584 The John Zink Combustion Handbook

FIGURE 19.49 NOx emission from the ultra-low emissions burner firing into the firetube boiler for ambient, 300°F preheat,
and 500°F preheat as a function of FGR rate.

FIGURE 19.50 Data taken from the application of an ultra-low emissions burner on a new 230,000 lb/hr “A” type Nebraska boiler.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
SCR or SNCR. In addition, the use of an ultra-low emissions Figure 19.53), the most common of which is the “Y-Jet,” to
burner can be accomplished at much lower capital installation “internal mix” atomizers. The atomizing medium is typically
costs, lower annual operating costs, and without the added either steam or pressurized air.
complication of ammonia handling and emissions. Internal mix atomizers typically require a relatively large
amount of atomizing medium consumption, on the order of
0.2 to 0.3 lb medium per 1 lb fuel, but provide excellent
19.6 ATOMIZERS FOR BOILER atomization quality, with a typical SMD of 75 µm at a fuel
BURNERS viscosity of 100 SSU. The supply pressure for the atomizing
Boiler burners are designed to utilize the many types of medium must be kept above the oil pressure to prevent oil
mechanical, dual fluid, and occasionally rotary cup oil atom- from entering the atomizing medium supply lines. Operating
izers presently used in most power plants. Most of these pressures for internal mix atomizers range between 85 and
atomizers are fitted onto oil “guns” similar to the one shown 300 psig (5.8 and 20 barg) for the oil, with the atomizing
in Figure 19.52. For dual fluid, these range from steam assist, medium pressure typically regulated to a differential pressure
air blast, or other “external mix” atomizers (depicted in 15 to 30 psig (1 to 2 barg) higher than the oil. The turndown
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch19-Frame Page 585 Wednesday, March 23, 2005 7:40 AM

Boiler Burners 585

FIGURE 19.52 Atomizers are fitted onto oil “guns.”

The two main types of mechanical atomizers are simplex

atomizers and spill-return atomizers. Oil enters a mechanical
atomizer through slots that expel the oil tangentially into a
circular “whirling” chamber. The vortex created by the tan-
gential entry into the whirling chamber is essential for good
atomization quality. For a spill-return flow atomizer, a portion
of the supply flow will return back into the oil gun through
a series of holes located at the back of the atomizer. The
remaining oil exits the atomizer through either a single hole
or a series of holes located at the front end (furnace side) of
the atomizer. The “return” oil will be recirculated through the
oil supply system. Simplex atomizers are simply a spill-return
FIGURE 19.51 The ultra-low emissions burner can atomizer without the return flow. Burner performance has
also be used on two-burner applications where NFPA 8501 been improved using specially designed return flow atomizers
guidelines are being followed. to maximize combustion efficiency and minimize NOx.
These atomizers produce good atomization quality at
ratio for these atomizers, depending on the oil supply pres- full load, with a typical SMD of 125 µm at a fuel viscosity
sure, can be as high as 10:1. of 100 SSU. However, atomization degrades rapidly with
these atomizer types and limiting turndown. Operating oil
External mix atomizers such as the Y-Jet require much less
pressures for simplex-type atomizers range between 300
steam consumption, on the order of 0.05 to 0.07 lb medium
and 600 psig (20 and 40 barg). The turndown ratio for these
per 1 lb fuel. However, the atomization quality is slightly
atomizers, depending on the oil supply pressure, is very low
reduced with a typical SMD of 125 µm at a fuel viscosity of
— typically 2:1.
100 SSU. Due to the external mixing of these atomizers, there
is very little chance for oil to enter the atomizing medium The delivered flow supplied by a spill-return atomizer is
supply lines. Therefore, these atomizers do not require that dependent on the return “back” pressure, which is essentially
the supply pressure for the atomizing medium be kept above the same pressure the atomizer would require if it were
the oil pressure. Operating pressures for Y-Jet type atomizers operating in simplex mode. The differential between supply
range between 85 and 300 psig (5.8 and 20 barg) for the oil, and return oil pressures, which can range between 100 and
with the atomizing medium pressure typically regulated to a 350 psig (7 and 24 barg) allows these atomizers a greater
constant pressure of 85 to 150 psig (5.8 to 10 barg). The turndown range — up to 4:1. Operating supply pressures for
turndown ratio for these atomizers, depending on the oil sup- spill-return type atomizers typically range between 400 and
ply pressure, can be as high as 10:1. 10,000 psig (30 and 700 barg).
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

586 The John Zink Combustion Handbook

FIGURE 19.53 Steam assist, air blast, or other “external mix” atomizers.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 20
Flares
Robert Schwartz, Jeff White, and Wes Bussman

TABLE OF CONTENTS

20.1 Flare Systems .......................................................................................................................................... 590

20.1.1 Purpose...................................................................................................................................... 590
20.1.2 Objective of Flaring .................................................................................................................. 590
20.1.3 Applications .............................................................................................................................. 591
20.1.4 Flare System Types ................................................................................................................... 591
20.1.5 Major System Components....................................................................................................... 593
20.2 Factors Influencing Flare Design ............................................................................................................ 594
20.2.1 Flow Rate .................................................................................................................................. 594
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

20.2.2 Gas Composition....................................................................................................................... 594

20.2.3 Gas Temperature ....................................................................................................................... 595
20.2.4 Gas Pressure Available.............................................................................................................. 596
20.2.5 Utility Costs and Availability.................................................................................................... 596
20.2.6 Safety Requirements ................................................................................................................. 597
20.2.7 Environmental Requirements.................................................................................................... 597
20.2.8 Social Requirements ................................................................................................................. 598
20.3 Flare Design Considerations ................................................................................................................... 598
20.3.1 Reliable Burning ....................................................................................................................... 599
20.3.2 Hydraulics ................................................................................................................................. 599
20.3.3 Liquid Removal......................................................................................................................... 600
20.3.4 Air Infiltration ........................................................................................................................... 601
20.3.5 Flame Radiation ........................................................................................................................ 601
20.3.6 Smoke Suppression................................................................................................................... 603
20.3.7 Noise/Visible Flame.................................................................................................................. 604
20.3.8 Air/Gas Mixtures ...................................................................................................................... 605

589
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 590 Tuesday, March 15, 2005 9:36 AM

590 The John Zink Combustion Handbook

20.4 Flare Equipment..........................................................................................................................................................605

20.4.1 Flare Burners.................................................................................................................................................605
20.4.2 Pilots, Ignitors, and Monitors........................................................................................................................614
20.4.3 Knockout Drums ...........................................................................................................................................618
20.4.4 Liquid Seals ..................................................................................................................................................619
20.4.5 Purge Reduction Seals ..................................................................................................................................620
20.4.6 Enclosed Flares .............................................................................................................................................622
20.4.7 Flare Support Structures ...............................................................................................................................623
20.4.8 Flare Controls................................................................................................................................................624
20.4.9 Arrestors........................................................................................................................................................629
20.5 Flare Combustion Products .........................................................................................................................................629
20.5.1 Reaction Efficiency .......................................................................................................................................630
20.5.2 Emissions ......................................................................................................................................................631
20.5.3 Dispersion .....................................................................................................................................................632
References ....................................................................................................................................................................................633

20.1 FLARE SYSTEMS 1. The composition of the gases handled by a flare often
vary over a much wider range.
During the operation of many hydrocarbon industry plants,
2. Flares are required to operate over a very large turndown
there is the need to control process conditions by venting
ratio (maximum emergency flow down to the purge flow
gases and/or liquids. In emergency circumstances, relief rate).
valves act automatically to limit equipment overpressure. For 3. A flare burner must operate over long periods of time
many decades of the last century, process vents and pressure without maintenance.
relief flows were directed, individually or collectively, to the 4. Flare burners operate at high levels of excess air as com-
atmosphere unburned. Gases separated from produced oil pared to other burners.
were also vented to the atmosphere unburned. The custom of 5. Many flare burners have an emergency relief flow rate
unburned venting began to change in the late 1940s when that produces a flame hundreds of feet long with a heat
increased environmental awareness and safety concerns cre- release of billions of Btu per hour (Figure 20.3).
ated the desire to convert vents to continuously burning flares.
Burning brought about the need for pilots and pilot ignitors 20.1.1 Purpose
and the need for awareness of the design factors and consid- The wide range of applications for flares throughout the
erations imposed on a system by a flame at the exit. In many hydrocarbon and petrochemical industries challenges plant
cases, the desirable flaring of the gases was accompanied by owners and designers as well as the flare equipment designer.
objectionable dense black smoke as shown in Figure 20.1. The purpose of this chapter is to provide an understanding of
In addition to their development of flare pilots and ignition the design considerations and factors influencing flare system
systems, industry pioneers John Steele Zink and Robert and equipment design. The most frequently used flaring tech-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Reed1 invented the first successful smokeless flare burner niques and associated equipment are also discussed.
(Figure 20.2) in the early 1950s. This invention was an
important point in the transition from unburned vents to 20.1.2 Objective of Flaring
flaring and from vent pipes to burners specifically designed Regardless of the application, flare systems have a common
for flare applications. prime objective: safe, effective disposal of gases and liquids…at
While the combustion fundamentals discussed earlier in an affordable cost. There should be constant awareness that
this book continue to apply, flare burners differ from process flare system design and operation must never compromise the
and boiler burners in several respects, including: prime objective.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 591 Tuesday, March 15, 2005 9:36 AM

Flares 591

FIGURE 20.1 Typical early 1950s flare performance.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 20.2 An early model smokeless flare. FIGURE 20.3 Major flaring event. (Note that the stack
height is over 300 ft (90 m).)

20.1.3 Applications
Within the hydrocarbon and petrochemical industries, from wide range of flare applications and site conditions, often
the drilling site to the downstream petrochemical plant and at requires that the flare system be custom designed.
many facilities in between, flares are utilized to achieve the
prime objective. Individual flare design capacity can range 20.1.4 Flare System Types
from less than 100 to more than 10 million lb/hr. Material Flares for service in the hydrocarbon and petrochemical
released into a flare system is often a mixture of several con- industries are generally of the following types or combina-
stituents that can vary from hydrogen to heavy hydrocarbons tions thereof:
and may at times include inert gases. Some of the heavy
hydrocarbons may be in the gaseous state when released into • single point
the system, but will be condensed as they cool. • multi-point
While this chapter focuses on flares and flaring in the • enclosed
hydrocarbon industry, many of the subjects discussed also
relate to flare applications in other industries. 20.1.4.1 Single-Point Flares
The design requirements for a given facility are seldom Single-point flares can be designed with or without smoke
identical to those of any other facility. This variation, plus the suppression equipment and are generally oriented to fire
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 592 Tuesday, March 15, 2005 9:36 AM

592 The John Zink Combustion Handbook

FIGURE 20.4 Typical elevated single-point flare. FIGURE 20.6 A grade-mounted, multi-point LRGO
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

flare system.

FIGURE 20.5 Typical pit flare installation. FIGURE 20.7 Elevated multi-point LRGO flare system.

upward, with the discharge point at an elevated position rel- 20.1.4.3 Enclosed Flares
ative to the surrounding grade and/or nearby equipment Enclosed flares are constructed so as to conceal the flame
(Figure 20.4). Occasionally, a single-point flare is posi- from direct view. Additional benefits can be the reduction in
tioned to fire more or less horizontally, usually over a pit or noise level in the surrounding community and minimization
excavation (Figure 20.5). Horizontal flares are generally lim- of radiation. Capacity can be the system maximum, but is
ited to drilling and production applications where there is a often limited to a flow rate that will allow the connected facil-
high probability of nonrecoverable liquids. ity to start up, shutdown, and operate on a day-to-day basis
without exposed flame flaring. Multiple enclosed flares are
sometimes used to achieve the desired hidden flame capacity.
20.1.4.2 Multi-point Flares
Each of the two units shown in Figure 20.8 is designed for
Multi-point flares are used to achieve improved burning by 100 metric tons/hr of waste gas from ethylene furnaces dur-
routing the gas stream to a number of burning points. For ing startups.
refinery or petrochemical plant applications, multi-point
flares are usually designed to achieve smokeless burning of 20.1.4.4 Combination Systems
all flows. Such flares often divide the multiple burning A common combination is an enclosed flare of limited capac-
points into stages to facilitate better burning. The multiple ity paired with an elevated flare (Figure 20.9) that is sized for
burning points can be arranged in arrays located at or near the maximum anticipated flow to the system. Such a pairing
grade (Figure 20.6) or at an elevated position (Figure 20.7). results in a flare system that only has an exposed flame
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 593 Tuesday, March 15, 2005 9:36 AM

Flares 593

during major upset or failure events. Other pairings, such as

an elevated flare with a multi-point flare, have also been used
(Figure 20.11).

20.1.5 Major System Components

Each flare system type has its own set of required compo-
nents. In addition, systems may include components that are
noted below as being optional. These optional components
address special needs, such as smoke suppression or liquid
removal. Optional equipment can also help reduce cost or aid
system operation. It should be noted that there is a difference
between the flare burner, or burners, required for each type of
flare system.

20.1.5.1 Single-Point Flares

For a single-point flare, the major required and optional com-
ponents are:

1. flare burner, with or without smoke suppression capability:

a. one or more pilots
b. pilot ignitor(s)
FIGURE 20.8 Multiple ZTOF installation in an ethylene
c. pilot flame detector(s)
plant.
2. support structure, piping, and ancillary equipment
3. purge reduction device (optional)
4. knockout drum (optional)
5. liquid seal (optional)
6. auxiliary equipment:
a. smoke suppression control (optional)
b. blower(s) (optional)
c. flow, composition, heating value, or video monitor
(optional)

20.1.5.2 Multi-point Flares

For a multi-point flare, the major required and optional com-
ponents are:

1. two or more multi-point flare burners

2. pilot(s), pilot ignitor(s), and pilot flame detector(s)
3. if elevated, support structure and ancillary equipment
4. a fence to limit access and reduce flame radiation and
visibility (optional)
5. knockout drum (optional)
6. liquid seal (optional)
7. piping
8. auxiliary equipment:
a. staging equipment and instrumentation (optional)
b. smoke suppression means where very large turndown
is required
c. flow, composition, heating value, or video monitoring FIGURE 20.9 Combination ZTOF and elevated flare
(optional) system. --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

594 The John Zink Combustion Handbook

20.1.5.3 Enclosed Flares costs, and can lead to shorter service life. Understatement can
For an enclosed flare, the major required and optional com- result in an ineffective or unsafe system.
ponents are: Flow rate obviously affects such things as the mechanical
1. flare burners, with or without smoke suppression capability
size of flare equipment. Its influence, however, is much
2. pilot(s), pilot ignitor(s), and pilot flame detector(s) broader. For example, increased flow generally results in an
3. enclosure/structure with protective refractory lining increase in thermal radiation from an elevated flare flame,
4. a fence to limit access which in turn will have a direct impact on the height and
5. knockout drum (optional) location of a flare stack.
6. liquid seal (optional) The maximum emergency flow rate can occur during a
7. piping and optional heat shielding major plant upset such as the total loss of electrical power or
8. auxiliary equipment: cooling water. However, some processes have their maximum
a. staging equipment and instrumentation (optional) flow rate under less obvious emergency conditions such as
b. flow, composition, heating value, or video monitoring partial loss of electrical power whereby, for example, pumps
(optional) continue to supply feedstock to a disabled section of the plant.
The duration of the maximum flow rate can affect flare
20.2 FACTORS INFLUENCING system design in a number of ways. For example, the length
of time a worker is exposed to heat from the flare flame can
FLARE DESIGN
affect the choice of allowable heat flux. Usually, a very short
As one approaches the specification of a flare system, there
duration relief into a flare system can result in a relatively high
must be an awareness of certain factors that influence size,
allowable radiation. In contrast, a very long duration, high flow
safety, environmental compliance, and cost. Major factors
relief may require a lower design allowable radiation level.
influencing flare system design2 include:
In the past, the maximum flow rate was sometimes deter-
• flow rate mined by summing the flow rates of each of the connected
• gas composition relief devices. This approach resulted in an unrealistically large
• gas temperature maximum flow rate because the assumption that all the con-
• gas pressure available nected devices relieve simultaneously is often false. Modern
• utility costs and availability plant design and analysis tools such as dynamic simulation
• safety requirements allow the process designer to define more appropriately the
• environmental requirements maximum flow rate to the flare. Careful attention to the design
• social requirements of control and electrical power systems can significantly
Information regarding each of these factors is normally reduce flare loads as well.3
available to the plant designer and/or the plant owner. These In addition to the maximum flow conditions, it is also
factors define the requirements of the flare system and should important to explicitly define any flow conditions under which
be made available to the flare designer as early in the design the flare is expected to burn without smoke. These flow con-
process as possible. ditions can come from process upsets, from incidents such as
In reviewing the list of factors, it can be seen that the first a compressor trip-out, or from various operations of the plant,
four factors are all determined by the source(s) of the gas including startup, shutdown, and blowdown of certain equip-
being vented into the flare header. The next factor is related ment. Attempts to shortcut the establishment of factually
to the design of the facility itself and its location. Safety, based smokeless burning scenarios by setting the smokeless
environmental, and social requirements all relate to regulatory flow rate as a percentage of the maximum emergency flow
mandates, the owner’s basic practices, and the relationship rate can lead to disappointment or needless expense.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

between the facility and its neighbors. A discussion of each

factor is provided below. 20.2.2 Gas Composition
Gas composition can influence flare design in a number of ways.
20.2.1 Flow Rate The designer should be given the gas composition for each of
The flare system designer relies heavily on the flow data pro- the flow conditions identified above and for any special gases
vided. Therefore, the data must realistically reflect the various that may be in use, such as pilot fuel or purge gas. By studying
flow scenarios. Overstatement of the flows will lead to over- the gas composition, its combustion characteristics and the
sized equipment, which increases both capital and operating identity of potential flue gas components can be determined.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 595 Tuesday, March 15, 2005 9:36 AM

Flares 595

For example, the composition reveals the relative presence

of hydrogen and carbon. The weight ratio of hydrogen to
carbon in a gas is one of the parameters that can indicate the
smoking tendency of the gas.4 The influence of the weight
ratio of hydrogen to carbon, often referred to as the H/C ratio,
on the smoking tendency is illustrated in Figure 20.10. Figure
20.10 shows the flame produced by burning three different
gases using the same flare equipment and operating condi-
tions. The flame produced by burning a 25 MW well head
natural gas (H/C = 0.27) is clean, as shown in Figure 20.10(a).
The flame in Figure 20.10(b) reflects the burning of propane
(H/C = 0.22). Note that the propane flame has some trailing
smoke, has a yellow color much closer to the flare burner,
(a)
and is more opaque when compared to the natural gas flame.
The dense black smoke and dark flame shown in Figure
20.10(c) was produced by burning propylene (H/C = 0.17).
Note that a portion of the flame is being cloaked or shrouded
by the smoke. The fact that the smoke hides part of the flame
must be accounted for when calculating the radiation from
the flare flame.
Composition analysis also will reveal the presence of non-
hydrocarbons such as hydrogen sulfide or inerts. Such gases
might require special metallurgies or design considerations
such as ground level concentration analysis. Composition com-
bined with the flow rate allows determination of the volume
flow or mass flow of gases to be handled by the flare system.
The practice of defining a stream by its bulk properties (b)
alone (molecular weight [MW], lower heating value [LHV],
upper/lower explosive limits [UEL/LEL], etc.) can disguise
safety hazards or prevent equipment and operating cost reduc-
tions that would otherwise be recognized by the flare designer.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

For example, a 28 to 30 MW gas could be ethane, ethylene,

nitrogen, carbon monoxide, air, or even a mixture of hydrogen
and xylene. A flare system to handle some fixed amount of
28 MW gas would have very different design and operating
characteristics, depending on the actual gas composition
involved. Ethylene may tend to smoke, but remains lit and
stable at high exit velocities. Carbon monoxide, on the other
hand, will not smoke, but can be difficult to keep lit even at
moderate to low exit velocities. Radiation, possible relief gas
(c)
enrichment, purge requirements, and the potential for con-
densation at ambient temperatures are other examples of the FIGURE 20.10 Comparison of the flame produced by
impact that gas composition can have on flare design. burning (a) 25 MW well head natural gas, (b) propane, and
(c) propylene.

20.2.3 Gas Temperature

In addition to the impact of relief gas temperature on thermal or two-phase flow necessitates liquid removal equipment to
expansion, gas volume, and metallurgical requirements, there is avoid a higher smoking tendency and/or the possibility of a
the more subtle effect of gas temperature on the potential for burning liquid rain. Condensation at low or no flow condi-
some components of the gas to condense. Possible condensation tions will result in formation of a vacuum condition in the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 596 Tuesday, March 15, 2005 9:36 AM

596 The John Zink Combustion Handbook

mine the resulting gas temperature at the flare. Attention to

such details can result in a reduced cost for the stack.

20.2.4 Gas Pressure Available

The gas pressure available for the flare is determined by
hydraulic analysis of the complete pressure relief system
from the vent or pressure-relieving devices to the flare burner.
Each major flowing condition should be analyzed to deter-
mine the pressure at each relief or vent in each branch of the
flare header. This pressure is usually limited by the lowest
allowable back-pressure on any relief valve in the system.
The limit applies to all flowing conditions, regardless of
whether or not the limiting relief valve contributes to the
flowing condition under study.
In most flare systems, much of the system pressure drop is
due to flare header piping losses with little pressure drop
remaining for the flare burner. Such system designs may not
maximize the value of the gas pressure in promoting smoke-
less burning. Smokeless burning can be enhanced by convert-
ing as much of the gas pressure available as possible into gas
momentum. In addition, redistributing the system pressure
drop to provide more pressure at the flare tip can reduce the
overall system cost.
Another benefit of taking a greater pressure drop across the
flare burner is the increase in gas density in the flare header,
FIGURE 20.11 Combination elevated LRGO and Util- which can lead to a smaller flare header size and reduced
ity flare system. piping cost. More pressure at the flare tip generally means a
smaller flare burner and, consequently, lower purge flows. The
enhancement of smokeless burning and the decrease in purge
flare header and the resulting potential to draw air in through
gas requirements both reduce the day-to-day operating cost.
the flare tip or through piping leaks. Liquid seals are some-
Both capital and operating costs can be reduced in this manner.
times used to address this hazard. However, gas temperature
Available gas pressure at the flare can be defined as total
can affect liquid seal design and operation. Hot gases will tend
pressure at the flare inlet, or as static pressure in a specific size
to boil off the seal fluid, sometimes very suddenly. On the
inlet pipe. Static pressure is the pressure applied by the gas to
other hand, extremely cold gases present a freezing scenario
the walls of the pipe. That is, this is the pressure sensed by a
that could completely block waste gas flow. pressure gage mounted on a simple nozzle in the side of the
While a flare stack may appear to be unrestrained and pipe. This is also the pressure that determines gas density. Total
therefore free to expand, there can be mechanical design pressure is the sum of the static pressure and the velocity
challenges as a result of large gas temperature variations. pressure at a given point in the piping (e.g., the stack inlet).
Header piping growth, relative movement of utility piping, When static pressure is used to define available gas pressure, --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

and stack guy wire tensioning are just three areas where the plant designer should also specify the anticipated inlet size.
problems can arise. Both high and low temperatures have the
potential to create issues that affect the design of the stack. 20.2.5 Utility Costs and Availability
In cases where the relief gas source pressure is extremely In many cases, the momentum of the gas stream alone is not
high, the plant designer should account for cooling by expan- sufficient to provide smokeless burning. In such cases, it is
sion across the relief or vent valve. Where the gas temperature necessary to add an assist medium to increase the overall
at the source is significantly different from ambient, it is momentum to the smokeless burning level. The most com-
advisable to estimate the heat loss or gain through the flare mon medium is steam, which is injected through one or
header walls from the source to the flare stack and to deter- more groups of nozzles. An alternative to steam is the use of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 597 Tuesday, March 15, 2005 9:36 AM

Flares 597

a large volume of low-pressure air furnished by a blower. Most specifications call for a maximum radiation level of
Local energy costs, availability, reliability, and weather con- 1500 Btu/hr/ft2 (4.73 kW/m2) for emergency flaring condi-
ditions must be taken into account in selecting the smoke- tions. Some specifications define an additional radiation level
suppression medium. limit of 500 Btu/hr/ft2 (1.58 kW/m2) for unprotected individ-
Purge and pilot gas must be supplied to the flare at all uals during long-duration flaring events. Special consider-
times. The amount of each gas required is related to the size ation should be given to radiation limits for flares located
of the flare system. The composition of the purge gas and/or close to potential public access areas along the plant boundary
the composition of the waste gas can also influence the purge where public exposure could occur. See Section 20.3.5 for
gas requirement. Pilot gas consumption can be affected by more detail on flame radiation.
the combustion characteristics of the waste gases. The gases Reliable ignition at the flare tip is one of the most funda-
used for purge gas and to fuel the flare pilots should come mental safety requirements, ensuring that gases released to
from the most reliable source available. the flare are burned in a defined location. Dependable burn-
Purge gas can, in principle, be any noncorrosive gas that ing also ensures destruction of potentially toxic releases. The
does not contain oxygen and does not go to dew point at any prime objective (Section 20.1.2) demands reliable burning
expected conditions. An attractive option for purge gas may of the flare. The subject is covered in more detail in Sections
be a mixture of nitrogen and a non-hydrogen-containing fuel 20.3.1 and 20.4.2.
gas such as natural gas or propane. For example, a 300 Btu/scf
Hydraulics of the flare system determine back-pressure on
mixture of nitrogen and propane can be effective as a purge
relief valves. Improper initial system sizing or subsequent
medium. Such a mixture presents a number of benefits when
additions to the flare relief loads can prevent a unit from
compared to fuel gas alone, including:
achieving its maximum relief rate when necessary and create
• reduced CO2 emissions an over-pressure risk in the plant. Section 20.3.2 provides
• potential cost savings if nitrogen is less expensive further discussion on the effect of relief valve selection on
• higher reliability because either supply alone can func- the design of the flare system.
tion as purge gas Prevention of air infiltration should be a consideration when
• reduced wear-and-tear on the flare burner developing operations and maintenance plans for the flare sys-
tem and connected equipment. Air sources include the flare tip
20.2.6 Safety Requirements exit, loop seals on vessels, low point drains, high point vents,
Almost every aspect of flare design involves some safety con- and flanges. These issues are discussed in Section 20.3.4.
cerns. Safety concerns include thermal radiation from the
flare flame, reliable ignition, hydraulic capacity, air infiltra-
20.2.7 Environmental Requirements
tion, and flue gas dispersion. Certain aspects of safety are
dictated by the basic practices of the owner. For example, the Flares can affect their environment by generating smoke,
owner’s safety practices usually set the allowable radiation noise, or combustion products. Regulatory agencies sometimes
from the flare flame to people or equipment. Therefore, it is define limits in some or all of these areas. In many cases, it is
not surprising that the allowable radiation level will vary necessary to inject an assist medium such as steam in order to
from owner to owner. achieve smokeless burning and to meet smoke emission regu-
A common point of variation involves the treatment of solar lations. The injection of the steam and the turbulence created
radiation relative to the allowable level. Experience has shown by the mixing of steam, air, and gas cause the emission of
sound. The sound level at various points inside and outside the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

that solar radiation need not be considered in the majority of

designs. In practice, consideration of solar radiation is a com- plant boundary is often subject to regulation.
plex issue that does not lend itself to a simple solution. The Other environmental concerns are the reaction efficiency and
solar radiation question involves a number of variables and is flue gas emissions. Pioneering tests conducted by the John Zink
site specific. By way of example, it would be appropriate to Company established that a properly designed and operated
include solar radiation in the design basis if there is substantial flare burner will have a combustion efficiency of more than
likelihood that a worker can become exposed to the maximum 98%.6 Emissions of NOx, CO, and unburned hydrocarbons
flare radiation and the sun’s radiation in an additive manner. (UHCs) were also determined during these tests. NOx, CO,
There are several sources for guidance on the allowable and UHC emission factors for flares are available in AP-42.7
radiation level. The most widely referenced is American These emission factors are widely accepted by regulatory
Petroleum Institute (API) Recommended Practice (RP) 521.5 authorities as a basis for flare permit emissions estimates. For
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 598 Tuesday, March 15, 2005 9:36 AM

598 The John Zink Combustion Handbook

requirements, but might be objectionable to the neighbors due

to light or noise.
The John Zink Company recognized the need to reduce
elevated flaring more than 30 years ago and invented the
world’s first successful enclosed ground flare for the elimina-
tion of day-to-day visible flaring. Since that time, many plants
have included a Zink Thermal Oxidizer Flare (ZTOF) in their
flare system. The general arrangement of a flare system that
incorporates a ZTOF for day-to-day flaring rates and an ele-
vated flare for emergency rates is shown in Figure 20.12. The
liquid seal in the system acts to divert flow to the ZTOF until
it reaches its maximum capacity. Any additional flow will pass
through the liquid seal and be burned at the elevated flare tip.
FIGURE 20.12 General arrangement of a staged flare An installation of such a system is shown in Figure 20.9.
system, including a ZTOF and an elevated flare.

20.3 FLARE DESIGN

CONSIDERATIONS
Having received information on the system defining factors
set forth above, the flare designer must now apply his/her
expertise to the following design considerations. A given
project may require inclusion of all or only a few of these
points, depending on the nature of the system information
disclosed and the scope of the project. The flare designer
must consider how decisions relating to each factor will
affect the entire flare system as well as all the other factors.
The prime objective of safe, effective disposal of gases and
liquids should be used as a guiding principle as each appro-
priate consideration is incorporated into the overall flare
design. The design considerations are:
1. reliable burning
2. hydraulics
FIGURE 20.13 John Zink Co. test facility in Tulsa, 3. liquid removal
Oklahoma. 4. air infiltration
5. flame radiation
6. smoke suppression
emissions estimates of SOx, it is often assumed that 100% of 7. noise/visible flame
the available sulfur is converted to SO2. 8. air/gas mixtures
Successful selection and operation of flare equipment
20.2.8 Social Requirements require a clear understanding of these design considerations.
The success and cost-effectiveness of a flare design are depen-
Most emergency flare systems include a flare stack that is the dent on the skill and experience of the flare expert and his/her
tallest, or one of the tallest, structures in the plant. As a result, access to the latest state-of-the-art design tools and equip-
the flare flame is visible for great distances. Although the plant ment. A key development tool is the ability to conduct flare
owner has complied with all environmental regulations, the tests at the high flow rates experienced in real plant opera-
flare system may not meet the expectations of the plant’s neigh- tions. Facilities capable of conducting tests of this magnitude
bors. Public perception of the purpose and performance of the (Figure 20.13) represent a substantial capital investment and
flare can place more stringent requirements on the flare design. take on the characteristics of a complex process plant flare
For example, a smokeless flame may meet the regulatory
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
system. Insight into each consideration is set forth below.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 599 Tuesday, March 15, 2005 9:36 AM

Flares 599

20.3.1 Reliable Burning velocity can help improve smokeless performance. It is

Venting of waste gases can happen anytime during plant important to note that discharge velocity can be constrained
operation. Therefore, an integrated ignition system is by the gas pressure available or concerns about flame stability.
required that can immediately initiate and maintain stable In some circumstances, such as VOC control, the discharge
burning throughout the period of waste gas flow. Stable burn- velocity may be limited by regulation. Early research on
ing must be ensured at all flow conditions. An integrated igni- flare system design suggested limiting discharge velocity to
tion system includes one or more pilot(s), a pilot ignitor, pilot 0.2 Mach due to stability concerns. It was later suggested that
monitor(s), and a means to stabilize the flame. a discharge velocity of 0.5 Mach or higher could be used if
In principle, all flares should have a continuous pilot flame proper flame stabilization techniques were employed. Flame
to ensure reliable burning. This is especially true of refinery, stabilization techniques have been successfully employed for
petrochemical, and production field flares because they cannot exit velocities of Mach 1 or greater.
be shut down unless and until the entire plant shuts down. In Waste gas composition can significantly affect the allow-
addition, such flares may be online for weeks or months able exit velocity. For example, a properly designed flare
before an unpredictable event that creates an immediate need burner can maintain stable burning of propane at Mach 1 or
for reliable ignition. Notable exceptions are landfill flares or greater. On the other hand, if the propane is mixed with a
biogas flares that operate continuously at substantial rates and large quantity of inert gas, the maximum exit velocity must
include flame monitoring systems that automatically shut off be limited to a much lower Mach number in order to ensure
waste gas flow in case of flame failure. Discontinuous pilots stable burning.
should only be considered in cases where all of the following
conditions apply: 20.3.2 Hydraulics
• The main flame remains lit and stable without a pilot at Most flare systems consist of multiple relief valves discharg-
all design conditions. ing into a common flare manifold or header system. A key
• The main flame is monitored. item influencing the flare system design is the allowable
• The flare is shut down automatically on main flame failure. relief valve back-pressure. The system pressure drop from
• The flare shutdown does not create a safety hazard in each relief valve discharge through the flare tip must not
the plant. exceed the allowable relief valve back-pressure for all system
The number of pilots required can vary, depending on the flow conditions. The allowable back-pressure is typically
size and type of the flare burner and its intended use. Flare limited to about 10% of the minimum relief valve upstream
pilots are usually premixed burners designed such that pilot set pressure for conventional relief valves. The allowable
gas and air are mixed together at a point remote from the flare relief valve back-pressure can be increased by the use of bal-
burner exit and delivered through a pipe to the pilot tip for anced pressure relief valves. Balanced valves can accept a
combustion. This ensures that the pilot flame is not affected back-pressure of about 30% of upstream set pressure in most
by conditions at the flare burner exit (e.g., the presence of cases. Where there is a wide variation in the allowable relief
flue gas, inert gas, or steam). Pilot gas consumption varies valve back-pressures, it may be economical to use separate
according to the specific flaring requirements. However, there high- and low-pressure flare headers.
is a practical lower limit to the pilot gas consumption. Increasing the allowable relief valve back-pressure can have
A pilot monitor is often required to verify the pilot flame. several effects on the flare system components, including:
As a safety consideration, pilot ignition is usually initiated
• smaller manifold and header piping
from a position remote from the flare stack. Either a flame
• smaller knockout and liquid seal drums
front generator or direct spark pilot ignition can be used,
• smaller flare size, giving lower purge rates and enhanced
depending on the system requirements. Further discussion of
operating life
this important safety aspect is provided in Section 20.4.2.
• significant reduction or elimination of utilities required
There is a complex relationship between flare tip exit veloc-
for smokeless burning through the utilization of
ity, gas composition, tip design, and the maintenance of stable increased pressure energy at the flare tip
burning. There are a number of advantages in using the high-
est exit velocity possible, including minimum equipment size As mentioned in Section 20.2.4, each major flowing con-
and optimal flame shape. In addition, because high discharge dition should be analyzed to verify that no relief source is
velocity tends to improve air mixing with a resultant reduction over-pressured. In some applications, a large number of dif-
in soot formation, one can see that maximizing discharge ferent flowing conditions can occur. To simplify the process
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

600 The John Zink Combustion Handbook

MW T
Veq = Q × (20.1)
29 520
where Veq = Volumetric equivalent, SCFH
Q = Waste gas flow, SCFH
MW = Waste gas molecular weight
T = Waste gas temperature, R

Veq is the volumetric flow of air that would produce the same
velocity pressure as the waste gas flow in the same size line.
While this method gives general guidance, it should not
replace a more thorough hydraulic analysis.
Properly utilized, a higher allowable pressure drop for the
(a) flare system provides an opportunity for capital cost savings,
operating cost savings, and reduced downtime due to longer
equipment life. While the capital cost savings are most appar-
ent on entirely new flare systems, all of these savings can be
realized on existing systems as well.

20.3.3 Liquid Removal

Inherent in many flare systems is the potential for either liquid
introduction into, or the formation of hydrocarbon or water
vapor condensate in, the flare header. Allowing this liquid-
phase material to reach the combustion zone may make opera-
tion more difficult. For example, hydrocarbon droplets small
enough to be entrained by waste gas and carried into the flame
usually burn incompletely, forming soot, and, as a result,
(b) reduce the smokeless capacity of the flare. If the droplets
become larger, they may be able to fall out of the main flame
envelope. In addition, events have been reported where a
mostly liquid stream has been discharged from the flare.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Figure 20.14 shows an offshore flare that received liquids
in the manner described above. The last two photos, (b) and
(c), were taken only a few minutes apart and illustrate how
rapidly this situation can deteriorate.
Incorporation of a properly designed and operated knockout
drum into the flare system can minimize these problems. There
are three basic types of knockout drums that can be incorporated
into a flare system: a horizontal settling drum, a vertical settling
drum, and a cyclone separator. For more information on each
of these types of knockout drums, refer to Section 20.4.3.
(c)
Regardless of the knockout drum concept, the holding
FIGURE 20.14 Liquid carryover from an elevated capacity of the drum should be carefully considered. An
flare. (a) Start of flaring event. (b) Liquid fallout and flaming overfilled knockout drum can obstruct gas flow to the flare,
rain from flare flame. (c) Flaming liquid engulfs flare stack. resulting in over-pressure to upstream systems. In the
extreme case, an overfilled knockout drum can result in blow-
ing large volumes of liquids up the flare stack. Liquid draw
of identifying those cases that are likely to govern the hydrau- off capacity must be adequate to prevent overfilling of the
lics, a comparative measure of flow rates is useful. The drum. In addition, a backup pump and drive means should
volumetric equivalent, or Veq, is one measure used to identify be considered. Liquid recovered from the knockout drum
the hydraulically controlling case: must be carefully disposed of or stored. Flare header piping
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 601 Tuesday, March 15, 2005 9:36 AM

Flares 601

must be sloped properly to prevent low point pockets where negative pressure. Operation of the flare system under negative
liquids can accumulate. pressure greatly increases the potential of air infiltration into
the header system through leaks, open valves, or flanges, or
20.3.4 Air Infiltration through the tip exit by decanting in the stack. Such leakage is
Infiltration of air into a flare system can lead to flame burnback, known to have occurred during the servicing of relief valves.
which in turn could initiate a destructive detonation in the sys- Installation of a liquid seal in the system can produce
tem. Often, burnback can only be observed at night. Air can positive flare header pressure although the pressure down-
enter the flare system by one or more of the following scenarios: stream of the seal is negative. This greatly reduces the poten-
• through stack exit by buoyant exchange, wind action, or tial of air leakage into the system. Because a liquid seal can
contraction also be a barrier to air entering the header from the flare stack,
• through leaks in piping connections locating the liquid seal in the base of the stack offers maxi-
• as a component of the waste gas mum protection of the header system. In this position, the
liquid seal can also be designed to isolate the flare ignition
Prevention measures are available to address each of the air
source from the flare header and the process units.
infiltration mechanisms.
Oxygen-containing gases should be segregated from the
Purge gas is often injected into the flare system to prevent
main flare system. Waste gases that contain oxygen present
air ingression through the stack exit. The quantity of purge
a special design challenge. The risk of flashback in systems
gas required is dependent on the size and design of the flare,
handling such gases can be minimized through the use of
the composition of the purge gas, and the composition of any
flame/detonation arrestors, special liquid seals, and/or the use
waste gas that could be present in the system following a vent
of specialized flare burners. The presence of more than a trace
or relief event. In general, the lower the density of the gas in
amount of oxygen (more than 1% by volume) in a waste gas
the flare stack, the greater the quantity of purge gas necessary
stream creates a separate design consideration discussed in
for the safety of the system. The purge gas requirement can
Section 20.3.8.
be reduced using a conservation device such as a John Zink
Airrestor or Molecular Seal. The cost and availability of the 20.3.5 Flame Radiation
purge gas will guide the choice of such a device. As the waste gases are burned, a certain portion of the heat
Contraction of gas in the flare system occurs due to the produced is transferred to the surroundings by thermal radia-
cool-down following the flaring of hot gases. The rate of tion. Safe design of a flare requires careful consideration of
contraction will accelerate dramatically if the cooling leads the thermal radiation. The radiation limits discussed in Sec-
to condensation of components of the contained gas. Contrac- tion 20.2.6 can become the basis for determining the height
tion risk can be minimized by use of the Tempurge system.8 of the flare stack and its location. For a given set of flare flow
Tempurge senses conditions in the flare header and initiates conditions, the radiation limits can usually be met by adjust-
the introduction of extra purge gas to offset contraction. ment of the flare stack overall height and/or the use of a lim-
An elevated flare stack filled with lighter than air gas will ited access area around the flare. The flare height and/or size
have a negative pressure at the base created by the difference of the limited access area can affect the economics of the
in density between the stack gas and the ambient air. The gas plant. For plants with limited plot area (or for ships), an
density in the stack is related to the molecular weight of the enclosed flare can be employed to meet radiation restrictions.
gas and its temperature. Equation (20.2) defines the pressure Water spray curtains have also been used to control radiation
at the base of a flare stack at very low flow conditions such on offshore platforms.
as purge.
In Chapter 3, radiative heat transfer was described in the-

Pbase =
(
27.7 H ρgas − ρamb ) (20.2)
oretical terms. Radiation from a flame to another object is
determined by:
144
• flame temperature
where Pbase = Static pressure at base of stack, in. w.c. • concentrations of radiant emitters in the flame (e.g.,
H = Height of stack above inlet, ft CO2, H2O, and soot)
ρgas = Gas density in stack, lb-m/ft3 • size, shape, and position of the flame
ρamb = Density of atmospheric air, lb-m/ft3 • location and orientation of the target object relative to
the flame
If a negative static pressure exists at the base of the stack, • characteristics of the intervening space between the
then at low flows the entire flare header system will be under --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- flame and the object
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 602 Tuesday, March 15, 2005 9:36 AM

602 The John Zink Combustion Handbook

vary a great deal from point to point within a flare flame,

which creates problems in predicting beam lengths and emis-
sivity. The detailed shape of a flare flame is much more
convoluted and chaotic than any simple geometric approxi-
mation can represent. Wind fluctuations cause the flame to
move constantly, so concentrations, temperature, and relative
positions are always changing. Atmospheric absorption and
scattering depend on transient and unpredictable weather con-
ditions such as ambient temperature, humidity, fog, rain, etc.
To overcome these difficulties, engineers have historically
estimated radiation by treating flare flames as point sources,
using heat release as a basis for emissive power and empirical
radiant fractions in lieu of true radiant emissivity. Figure
20.16 shows the general geometry assumptions that affect the
point source approach. The classic API radiation equation
represents this approach in its simplest form:

FIGURE 20.15 Thermogram of a flare flame. τFQ

K= (20.3)
4 πD 2

Calculations based on theory may be feasible within the where K = Radiation, Btu/hr-ft2
well-defined confines of a furnace operating at a steady con- τ = Atmospheric transmissivity
dition. Unfortunately, most of these factors cannot be accu- F = Radiant fraction
rately defined for a flare flame in the open air. Q = Heat release, Btu/hr
Because temperature appears in the radiation equation to D = Distance from the heat epicenter to the object, ft
the fourth power, it is clearly a dominant factor. Despite its
importance, the temperature of a flare flame is extremely Many of the complexities of the full theoretical treatment
difficult to measure or estimate. An error of only 10% in are lumped into the empirically determined radiant fraction.
absolute temperature affects the calculated radiant heat trans- This factor includes flame temperature effects, gas and soot
fer by over 40%. Observers have noted variations in local emissivity, mean beam length, and other flame shape issues.
flame temperature as high as 1000°C (1800°F) between the The distance factor disguises a number of subtleties that arise
core and the cooler outer surface of an open burning flame. as a result of flame shape prediction, including flame length,
Figure 20.15 shows a thermogram of a flare flame. In the flame trajectory, and position of the heat epicenter. Neverthe-
thermogram, white represents the highest temperature and less, this type of simplified approach has been used in one
dark blue the lowest. Thus, only the small, bright yellow zone form or another to estimate radiation from flare flames for
is at a high temperature. The temperature falls rapidly as one many years.
approaches the outer edge of the flame envelope. Several published methods are available for preliminary
The temperature of a flame is influenced by its interaction estimation of flare radiation and stack heights. An article by
with its surroundings. The availability of ambient air causes Schwartz and White9 presents a detailed discussion of flare
the outer portions of the flame envelope to cool. In addition, radiation prediction and a critical review of published meth-
the flame will radiate both to cold outer space and to relatively ods. Based on Example 2 in the referenced paper, Figure 20.17
warmer objects on Earth. Therefore, it is not surprising that provides a visual comparison of the stack heights determined
observations indicate peak flare flame temperatures far less by each of several radiation methods and the relative equip-
than the calculated adiabatic flame temperature. To approach ment cost associated with each stack height. Plant designers
flare radiation from a theoretical basis, local flame tempera- and users alike must be cognizant that traditional methods of
ture, which varies substantially throughout the flare flame, calculating radiant heat intensities are neither consistently
would need to be predicted with greater accuracy than present conservative nor consistently optimistic. Long ago, the John
tools allow. Zink Company recognized the limitations and risks associated
The other factors listed are also very difficult to determine. with the traditional methods and undertook the development
The concentrations of substances that are radiant emitters of proprietary methods for radiation prediction. The latest
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Flares 603

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 20.16 API radiation geometry.

prediction methods capture the effect of flare burner design, Briefly, energy transformation entails conversion of the inter-
gas quantity and composition, various momenta, smokeless nal energy (pressure) of the waste gas to kinetic energy (veloc-
burning rate and smoke formation on the flame shape, and ity). Designs for high-pressure flares (5 to 10 psig or more) exist
radiant characteristics. that require no supplemental assist medium. Systems employ-
ing this technique have been very successful and enjoy low
operating cost and an excellent service life.
20.3.6 Smoke Suppression
Steam injection is the most common technique for adding
Smokeless burning is a complex issue that involves many of momentum to low-pressure gases. In addition to adding
the system defining factors discussed in Section 20.2. In choos- momentum, steam also provides the smoke suppression ben-
ing the best smoke suppression method, the flare designer is efits of gas dilution and participation in the chemistry of the
guided by his/her experience in interpreting the job-specific combustion process. The effectiveness of steam is demon-
information received relative to each of these factors. strated in the series of photographs shown in Figure 20.18.
Smokeless burning, in general, occurs when the momentum In frame (a), there is no steam injection; in frame (b), steam
produced by all of the employed energy sources educts and injection has just begun; and in frame (c), steam injection has
mixes sufficient air with the waste gas. For smokeless burn- achieved smokeless burning.
ing, a key issue is the momentum of the waste gas as it exits Some plants have steam available at several different pres-
the flare burner. In some cases, the waste gas stream is avail- sure levels. There is often an operating cost advantage to using
able at a pressure that, if properly transformed, can provide low-pressure steam (30 to 50 psig). The plant designer must
the required momentum. If the waste gas pressure (momen- balance this operating cost advantage against the increased
tum) is not adequate for smokeless burning, the flare designer piping costs associated with low-pressure steam. Also, while
must enlist assistance from another energy source (e.g., steam the flare may achieve the design smokeless rate at the maxi-
or low-pressure air). In some cases, a combination of energy mum steam pressure, steam consumption at turndown firing
sources can be effective. rates below the maximum may be higher than expected.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 604 Tuesday, March 15, 2005 9:36 AM

604 The John Zink Combustion Handbook

FIGURE 20.17 Comparison of stack height and relative cost for various radiation calculation methods.

Because most flaring events involve relatively low flow rates, considered to constitute noise. Flaring sound is generated
performance under these turndown conditions must be care- by at least three mechanisms:
fully considered.
1. by the gas jet as it exits the flare burner and mixes with
Low-pressure (0.25 to 1.0 psig) air is utilized in cases where
surrounding air
gas pressure is low and steam is not available. The supplied
2. by any smoke suppressant injection and associated mixing
air adds momentum and is a portion of the combustion air
3. by combustion
required. Figure 20.19 shows another series of photographs
that illustrates the effectiveness of air assistance. Frame (a) Upstream piping and valves associated with the source of
shows the flare with no assist air. The blower is turned on in the relief gas may also create substantial noise levels that are
frame (b), but because the blower requires some time to reach carried along the flare header and exit through the flare tip.
full speed, the complete effect of air injection is not seen until At the maximum smokeless flaring rate of a steam- or air-
frame (d). The gas being flared in this case is propylene. assisted flare burner, gas jet noise is usually a minor contrib-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Generally, the blower supplies only a fraction of the com- utor. The noise generated by the second mechanism can be
bustion air required by the smokeless flow condition. For most mitigated by use of low noise injectors, mufflers, and careful
designs, 15 to 50% of the stoichiometric air requirement is distribution of suppressant. The Steamizer™ flare burner
blown into the flame. The remainder of the required air is shown in Figure 20.20 is of low noise design with additional
entrained along the length of the flare flame. noise reduction coming from a muffler concept first developed
for use on enclosed ground flares. Careful design can reduce
20.3.7 Noise/Visible Flame flaring noise levels by a factor of 75% or more (6 dB or more).
The energy released in flare combustion produces heat, Where the light from a flare flame is objectionable, an
ionized gas, light, and sound. Most plants are equipped enclosed flare is a good selection. A properly sized enclosed
with elevated flares that by their nature broadcast flaring flare can eliminate visible flame for all cases except emer-
sound to the plant and to the surrounding neighborhood. In gencies. An equal benefit of an enclosed flare is the reduction
some cases, the sound level becomes objectionable and is of flaring noise.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 605 Tuesday, March 15, 2005 9:36 AM

Flares 605

20.3.8 Air/Gas Mixtures

Waste gas streams containing air/gas mixtures can generally
be divided into two types. The first type is comprised of sys-
tems that are expected to contain air/gas mixtures. Examples
include landfills, gasoline loading terminals, and medical
equipment sterilization facilities. The second type is poten-
tially more dangerous in that air is not expected in the com-
position of vents and reliefs. An example is venting air from a
vessel or tank at the beginning of a pre-startup purging cycle.
Flare systems that handle air/gas mixtures usually involve
a number of special safety considerations. The special con-
siderations, which relate primarily to the increased risk of
flashback, include: (a)
1. safety interlocks to prove purge gas and pilots on startup
2. automatic shutdown on loss of purge gas or pilots
3. higher than normal purge rates to maintain burner exit
velocity and prevent burnback
4. limited turndown range to maintain burner exit velocity
and prevent burnback
5. use of detonation and/or flame arrestors
6. special operational practices

20.4 FLARE EQUIPMENT

While evaluating the general design considerations set forth
above, the flare designer must also begin equipment selection.
Overall design considerations and specific equipment selec- (b)
tions are interrelated aspects of the system design process.
The various major system components listed in Section 20.1.5
are discussed here in more detail.

20.4.1 Flare Burners

Although they are installed at the end of the flare system,
flare burners are among the first items considered in the sys-
tem design. At this point it should be clear that substantial
benefits are attached to the use of most of the available pres-
sure at the flare burner. The exit of the flare burner is the point
where the flow determinate pressure drop usually occurs.
Designs range from simple utility flares to enclosed multi-
point staged systems, and from non-assisted to multiple
(c)
steam injectors to multiple blower air-assisted designs.
Regardless of the tip design, adequate ignition means must FIGURE 20.18 Effectiveness of steam in smoke sup-
be available to ensure that the prime objective is achieved. pression: (a) no steam, (b) starting steam, and (c) smokeless.

20.4.1.1 Non-assisted or Utility

The simplest types are the non-assisted, or utility, flare burn- relief gases. A typical non-assisted flare is shown in Figure
ers. These burners consist essentially of a cylindrical barrel 20.21. Utility flares are usually flanged for ease of replace-
with attachments for enhanced flame stability (flame reten- ment. The horizontally mounted versions are usually referred
tion means) and pilots to initiate and maintain ignition of the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- to as burn pit flare tips (Figure 20.5). In both cases, a turbu-
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 606 Tuesday, March 15, 2005 9:36 AM

606 The John Zink Combustion Handbook

(a) (b)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

(c) (d)
FIGURE 20.19 Effectiveness of air in smoke suppression: (a) no blower air, (b) start blower, (c) air flow increasing,
and (d) smokeless.

lent diffusion flame is produced. The flame may be an 20.4.1.2 Simple Steam Assisted
attached or detached stable flame. The exit velocity of a flare The first smokeless flares were adaptations of simple utility
burner is dependent on the waste gas composition, the spe- flares. This basic design has been improved over the years
cific design of the flare burner, and the allowable pressure with multi-port nozzles to reduce steam injection noise, opti-
drop. In some cases, the exit velocity can safely reach mized injection patterns to improve steam efficiency, and
Mach 1. It should be noted that some flares or flare relief optional center steam injection to reduce damaging internal
cases are subject to regulations that limit the exit velocity. burnback. Figure 20.23 shows a modern example of this
Optional features that can extend equipment service life design. A steam manifold, often referred to as the upper steam
include windshields and refractory lining. The Zink double manifold or ring, is mounted near the exit of the flare tip. The
refractory (ZDR) severe service flare tips (Figure 20.22) use steam ring can be designed for steam supply pressures nor-
refractory linings internally and externally to protect the tip mally ranging from 30 to 150 psig. Several steam injectors
against both internal burning and flame pulldown outside the extend from the manifold and direct jets of steam into the
tip. Alternatively, center steam is sometimes used to help avoid waste gas as it exits the flare tip. The steam jets inspirate air
internal burning instead of an internal refractory lining to extend from the surrounding atmosphere and inject it into the gas
tip life. This approach is most effective in climates where freez- with high levels of turbulence. These jets also act to gather,
ing is not an issue. Center steam is a relatively inefficient means contain, and guide the gases exiting the flare tip. This prevents
to control smoke because it does not entrain air, which is nor- wind from causing flame pulldown around the flare tip.
mally an essential part of any smoke suppression strategy. Injected steam, educted air, and relief gas combine to form a
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 607 Tuesday, March 15, 2005 9:36 AM

Flares 607

FIGURE 20.21 Typical non-assisted flare.

FIGURE 20.20 Steamizer™ steam-assisted smokeless flare.

mixture that burns relief gas without smoke. The maximum

smokeless rate achievable with a given flare burner depends
on a number of factors, including the gas composition, the
amount of steam available, and the gas and steam pressures.
However, there are inherent design limitations in this type
of flare tip. The steam injectors are located close to the exit of
the flare tip, so it becomes very difficult to muffle the steam
noise produced by the high-pressure jets. Any muffler for the FIGURE 20.22 Zink double refractory (ZDR) severe
upper steam noise would need to be able to withstand direct service flare tip.
flame impingement in adverse winds and it could tend to inter-
fere with air being drawn in by the steam jets. Furthermore, as
the tip size increases, it becomes more difficult for the steam/air core of the waste gas exit (Figure 20.25). The presence of
mixtures to reach the center of the flame. Finally, the perimeter these tubes, properly distributed across the tip exit, increases
of the flare tip only increases linearly with tip size, while the the effective perimeter available for air access to the waste
flow area of the flare tip increases with the square of tip size, gas (Figure 20.26).
as shown in Figure 20.24. Therefore, as flare tip size increases,
the need for air (a function of the flow area) quickly outstrips The external-internal tubes start outside the wall or barrel
the ability to educt air (a function of the perimeter.) This fun- of the flare tip, pass through openings in the wall, and turn
damental characteristic of a simple steam-assisted flare limits upward, terminating at the tip exit. Welding seals the point
the maximum effective size of such a flare burner. where the tube penetrates the wall. Steam jets inject steam
into the inlet of these tubes, inspirating large amounts of air.
20.4.1.3 Advanced Steam Assisted The steam/air mixture exits the tubes at high velocity, deliv-
To overcome the limitations and other shortcomings of the ering momentum, dilution steam, combustion air, and turbu-
simple steam-injected design, an advanced steam-assisted lence into the base of the flame. Recent innovations have
flare was invented that uses multiple steam injection points.10 increased the effectiveness of the tubes allowing greater
In addition to the upper steam manifold, a set of external- smokeless capacities. New, enhanced steam injectors have
internal tubes is utilized to deliver a steam/air mixture to the increased the air eduction efficiency. Upper steam injection
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

608 The John Zink Combustion Handbook

As the capacity, size, and smokeless capability of flare burn-

ers have increased, more emphasis has been placed on sophis-
ticated design tools that can predict noise and smokeless
performance. Design tools of this type can be developed and
validated using large-scale tests in a facility such as shown in
Figure 20.13.

20.4.1.4 Low-Pressure Air-assisted

Not all plants have large amounts of steam available for use
by the flare. Some plants prefer not to use steam to avoid
freezing problems; others cannot commit water to make
steam for smoke control; and still others choose not to install
a boiler. To meet this need, a series of air-assisted flare
FIGURE 20.23 Simple steam-assisted flare. designs were invented.11
Generally, the air-assisted flare burner consists of a gas
burner mounted in an air plenum at the top of the flare stack
(Figure 20.28). Relief gas is delivered to the burner by a gas
riser pipe running coaxially up the center of the flare stack.
Low-pressure air is delivered to the burner from one or more
blowers located near the base of the flare stack. The air flows

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
upward through the annular space between the flare stack and
the gas riser.
The first air-assisted flare applications were associated with
operations some distance from the main plant or totally remote
from plant utilities support. Early air flares were often designed
to flare small to moderate flow rates. The success of these flares
led to the use of air assist on flares of greater capacity. More
recently, air-assisted flares have come into use as the flare for
large process facilities. The Flame Similarity Method and the
FIGURE 20.24 Perimeter:area ratio as a function of related near field mixing region models discussed in Chapter 8
tip size for a simple steam-assisted flare. are examples of the design tools necessary for cost-effective
application of air-assisted flares. Today, air flare designs are
available with demonstrated tip life spans of 5 to 10 years.
further enhances smokeless combustion through increased
Smokeless rates above 150 × 106 standard cubic feet per day
turbulence and mixing and by mitigating adverse wind effects.
(SCFD) are available for saturated hydrocarbons such as pro-
The advanced steam-assisted flare design incorporates sev- duction facility reliefs.
eral smoke suppression strategies: increased perimeter, higher Figure 20.29 shows an example of the latest air flare design.
momentum, more combustion air, greater turbulence for mix- Waste gas exits the burner in one or more narrow annular jets,
ing, dilution and chemical interaction by steam, and molding each surrounded by assist air. This design makes good use of
of the flame to resist wind effects. Each of these strategies the perimeter:area ratio concept discussed above in the con-
helps to reduce smoke; in combination, they produce some text of steam-assisted flares.
of the highest smokeless rates available in single-point flares.
New flare systems can achieve smokeless rates of more 20.4.1.5 Energy Conversion
than 500,000 lb/hr (230,000 kg/hr) of gases that are generally In the smokeless flaring discussions above, the focus cen-
considered difficult to burn cleanly. An example of a state- tered on adding energy from an outside source to boost the
of-the-art flare burner design is shown in Figure 20.27. overall energy level high enough to achieve smokeless burn-
Improved muffler designs and redistribution of steam can give ing. An advantage is gained if an outside source is not
noise levels much lower than earlier models. In some cases, required. This is the case with energy conversion flare burn-
the steam jet noise can be totally neutralized through injector ers. Such burners are also referred to as high-pressure flare
design and the use of new muffling techniques. burners or multi-point flare burners. Where they can be
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 609 Tuesday, March 15, 2005 9:36 AM

Flares 609

FIGURE 20.25 Schematic of an advanced steam-assisted flare.

employed, the use of energy conversion flare burners can pro-

vide a significant reduction in flare operating cost.
There are two distinct groups of energy conversion burn-
ers. The first group is distinguished by having a single inlet
and relatively close grouping of the waste gas discharge
points. The other group employs a means beyond energy
conversion to achieve smokeless flaring. In both groups, an
underlying principle is the conversion of the static pressure
of the waste gas, at the burner, to jet velocity and ultimately
into momentum.
Another concept employed by both groups is the division
of the incoming gas stream into multiple burning points. This
concept is sometimes referred to as the “firewood principle”
because of an illustrative analogy. The “firewood principle” FIGURE 20.26 A comparison of the perimeter:area
considers the situation where a tree has been cut down and ratio for simple and advanced steam-assisted flares.
is to be used as fireplace fuel. If a large section of the tree
trunk is placed in the fireplace, it will be difficult to ignite
and burn because the ratio of fuel to surface (air exposure) is Applications of flare burners that employ energy conver-
very large. This situation can be improved by splitting the sion to achieve smokeless burning require special attention to
trunk section into smaller pieces, thus gaining a better flow turndown cases. Some flares, usually those associated
fuel:surface ratio. Obviously, a balance must be reached with oil production activities, operate at or near their maxi-
between the effort expended to split the wood (think of it as mum design flow most of the time. Energy conversion flares
cost) and the improved ability to burn (think of this as smoke- for production applications, such as the commercially avail-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

less burning ability). able Hydra flare shown in Figure 20.30, offer a controlled
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 610 Tuesday, March 15, 2005 9:36 AM

610 The John Zink Combustion Handbook

FIGURE 20.27 State-of-the-art Steamizer™ flare burner

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

and muffler.

flame shape and reduced flame radiation. These features are

particularly attractive for platform-mounted flares.
The opposite is normally true of flare burners serving refin-
eries or petrochemical plants. These flares usually have a very
small load (purge gas plus any leakage) with an occasional
short-duration intermediate load. In energy conversion flares,
the gas pressure vs. flow characteristic follows the same rules FIGURE 20.28 Air-assisted smokeless flare with two
as an orifice. Above critical pressure drop, the flow/pressure blowers in a refinery.
characteristic varies with the ratio of the absolute pressures.
At or below the critical pressure drop, the flow/pressure char-
acteristic follows a square relationship. For example, a flow or group of burners is controlled by a valve that operates in
reduction of 50% will reduce the gas pressure drop to 25% an on/off manner as directed by a control system. The valve
of the full flow value. to the first stage is usually open all the time. The control
The turndown ratios experienced by refinery and petro- system principle is to proportion the number of burners in
chemical flares would reduce the gas pressure to such a service to the gas flow. In effect, this allows the burners to
degree that energy for smokeless burning would not be operate within a certain pressure range (see Figure 20.31) so
present. This problem has been overcome using a staging that at least the minimum energy level for smokeless burning
concept.12 The staging concept divides the burners into stages is always present.
or groups of burners, with the first stages having a smaller The operation of a multi-point flare can be spectacular, as
number of burners than the last stages. The flow to each stage shown in Figure 20.31. This series of photographs illustrates
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 611 Tuesday, March 15, 2005 9:36 AM

Flares 611

the addition of burner stages as the gas flow to the flare

increases. In the final frame, the flare achieves a smokeless
burning rate of more than 550,000 lb/hr (250,000 kg/hr). The
commercially available LRGO flare system shown is sur-
rounded by a fence to exclude personnel and animals from the
flare area. In cases where land space is limited, an enclosing
fence, such as shown in Figure 20.32, can be employed to
reduce radiation to the surroundings and to reduce visibility and
noise. The largest flare system in the world employs an LRGO
system to handle more than 10 million pounds of waste gas per
hour, smokelessly.
Design of an energy conversion flare system, either single
point or multi-point, involves issues that are not a consider-
ation for other flare burner types. Questions such as burner or
gas jet spacing must be resolved. For example, the ability of
a given burner to be lit and to light its neighbors is of para-
mount importance. Figure 20.31(f) captures this cross-lighting
feature in progress. However, spacing burners solely on the
basis of cross-lighting may restrict air flow and hinder smoke-
less burning. Gas properties are even more critical when the
design depends on energy conversion alone to achieve smoke- FIGURE 20.29 Annular air flare.
less burning.

20.4.1.6 Endothermic
Some flare applications involve gases with a high inert gas
content. When the inert content is high enough, the combus-
tion reaction becomes endothermic, meaning that some
external source of heat is required to sustain the reaction.
Crude oil recovery by CO2 injection, incinerator bypasses,
coke ovens, and steel mills are examples of activities that
generate gases that require additional fuel to maintain the
main burner flame. Such gases often contain significant
amounts of toxic materials such as H2S, NH3, CO, or various
gases normally sent to an incinerator. Flaring has been rec-
ognized for many years as an adequate method of disposing
of such gases. Substantially complete destruction of such
gases protects the community and the environment.
The earliest endothermic flares consisted of simple non-
assisted flare tips with fuel gas enrichment of the waste gas
upstream of the flare to ensure that the mixture arriving at the
tip was burnable. This system was simple but imposed a high
fuel cost on the facility. An alternative design supplied a
premixed supplemental fuel/air mixture to an annulus around
the flare tip. Combustion of this mixture supplied heat and
ignition to the waste gas as it exited the flare tip. This design
had a limited supplemental fuel gas turndown before burn-
back occurred in the annulus, thus requiring a full on or off
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

operation of the supplemental gas.

Today’s high energy costs provide an incentive to reduce
such fuel usage. The RIMFIRE® flare burner (Figure 20.33) FIGURE 20.30 Hydra flare burner in an offshore location.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 612 Tuesday, March 15, 2005 9:36 AM

612 The John Zink Combustion Handbook

(a) (e)

(b)
(f)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

is a modern endothermic flare with an air-assisted supple-

mental gas burner surrounding the waste gas exit.13 Using
supplemental fuel to build a strong “forced-draft” ignition
flame, the amount of enrichment gas required to sustain
ignition is reduced. The amount of supplemental fuel
required depends on the flare burner size and service.
The RIMFIRE® flare was originally developed for the CO2
(d) injection fields in the western United States. Vents and reliefs
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 613 Tuesday, March 15, 2005 9:36 AM

Flares 613

FIGURE 20.32 Multi-point LRGO system with a FIGURE 20.34 OWB liquid flare test firing 150 gpm.
radiation fence. (Courtesy of Maria Celia, Setal Engineering,
Sao Paulo, Brazil).

FIGURE 20.35 Forced-draft Dragon liquid flare.

requirement for other designs. Today, the RIMFIRE® burner

is used in a broad range of applications where enrichment of
the waste gas is required.

20.4.1.7 Special Types

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Sections 20.4.1.1 through 20.4.1.6 have described the wide

variety of flare burners available to the flare designer today.
In general, these burners are used to burn gaseous waste
FIGURE 20.33 A RIMFIRE® endothermic flare. material and utilize certain smoke suppression techniques.
There are, however, other flares that are designed to burn
liquid hydrocarbons or to use a liquid as the smoke suppres-
from the oil recovery system sent highly inert materials to the sant. Some examples are given below.
flare. In this service, a non-assisted tip would require enrich- Figures 20.34 and 20.35 show two examples of liquid flares
ment of the relief gases to an LHV of about 300 Btu/scf. having quite different designs. The OWB flare (Figure 20.34)
Using the RIMFIRE® flare burner, the LHV requirement was has multiple burners combined into a single unit. At the time
reduced to approximately 180 Btu/scf. of the photograph, the flare was burning oil with a heat release
The total fuel requirement, enrichment plus supplemental of about 1 billion Btu/hr. The OWB tips allow a modular
fuel, for the RIMFIRE® is substantially lower than the fuel approach to design for a specific capacity. In addition, a very
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 614 Tuesday, March 15, 2005 9:36 AM

614 The John Zink Combustion Handbook

with a burning rag attached, or shooting a signal flare over

the top. These methods, however, were not reliable or safe. In
1949, the John Zink Company developed the first pilot to
light and continually burn the vented gases from a flare.
Since then, there have been a number of improvements in
pilot design, ignition, and monitoring. This section discusses
pilots and methods used for pilot ignition and monitoring.

20.4.2.1 Pilots
A flare pilot is a premixed burner system designed to operate
over a narrow heat release range. As a burner, the pilot must
(1) meter the fuel and air, (2) mix the fuel with the air, (3) mold
the desired flame shape, and (4) maintain flame stability.
Typically, pilots consist of four fundamental parts: a mixer or
venturi, a gas orifice, a downstream section that connects the
FIGURE 20.36 Poseidon flare: water-assisted Hydra. mixer and the tip, and a tip as illustrated in Figure 20.37. All
components of a pilot are carefully designed to work together
as a system to achieve proper performance. A change in any
component will affect the balance of the system and hence the
operation of the pilot.
In operation, the pressure energy of the pilot fuel is used
to aspirate ambient air into the mixer inlet, mix the gas and
FIGURE 20.37 Fundamental pilot parts. air, and propel the mixture through the downstream section
and out the pilot tip. The key goals for a properly designed
pilot itself are to:
wide turndown can be obtained by simply turning off flow to
some of the tips. Typical applications are oil well testing and 1. be capable of reliable ignition
spill cleanup. 2. provide pilot flame stability
The Dragon flare (Figure 20.35) uses one or more burners 3. prevent the pilot flame from being extinguished
and is equipped with a blower to improve mixing for smoke- 4. provide a long service life
less burning. This flare is employed in destruction of surplus
or off-spec product or waste oil. An application goal of the pilot is to ignite the waste gas exit-
The success of steam as a smoke suppressant has prompted ing the flare burner.
the use of water to prevent smoke. Over the years, a number If the volume of air aspirated into the pilot falls outside the
of designs using water for smoke suppression have been flammability limits of the pilot fuel gas, the pilot will not
invented. In certain situations, water can be used with great operate properly. For example, methane requires 5.7 to 19
benefit. A case in point is the offshore installation of a Poseidon volumes of air per volume of fuel in order to burn. If a pilot
flare shown in Figure 20.36. This flare utilizes water to is operating below this volumetric air:fuel ratio limit, air
enhance smokeless burning, reduce thermal radiation, and external to the pilot must be available in order to burn the
decrease noise. The installation shown in the photograph fuel. If the pilot tip is engulfed with inert gases from the flare,
achieved a 13-dBA reduction in noise and a 50% drop in due to purge gas or flue gas from the flare flame, then the
radiation compared to the previous conventional flare.14 pilot cannot be lit, nor will it burn. Conversely, if a pilot is
operating with methane and the volumetric air:fuel ratio is
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

20.4.2 Pilots, Ignitors, and Monitors above or near 19, the pilot will be difficult to light and may
Prior to 1947, venting of unburned hydrocarbons to the atmo- be unstable in windy conditions.
sphere was an industry practice. After 1947, regulations The operational environment of a flare pilot requires that
required hydrocarbons to be burned or “flared.” Early meth- the pilot be able to withstand rain, wind, heat from the flare
ods to light a flare included hoisting a burning, oily rag to the flame, and direct flame contact. Common pilot problems are
top of the flare, shooting over the top of the flare an arrow failure to light and burn with a stable flame and flashback.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 615 Tuesday, March 15, 2005 9:36 AM

Flares 615

FIGURE 20.38 Conventional flame-front generator.

20.4.2.2 Pilot Ignition the air/fuel mixture should flow for about 20 to 30 seconds
The most common method of lighting elevated flare pilots is before attempting to light the pilot. Each attempt to light the
by the use of a flame-front generator (FFG). An FFG is a pilot in this example should allow 20 to 30 seconds for the
device designed to produce a fireball, which travels inside a line to refill with the air/fuel mixture.
pipe from the point of ignition to the pilot, thereby lighting
the pilot. There are three fundamental types of FFGs: con- 20.4.2.2.2 Slipstream FFG
ventional, slipstream, and self-inspirating. The slipstream FFG directs a portion of the air/fuel mixture
generated by the pilot venturi to a tube located adjacent to the
20.4.2.2.1 Conventional FFG pilot, as shown in Figure 20.39. The slipstream travels
A conventional FFG is illustrated in Figure 20.38. A combi- through the tube and exits near the pilot tip. A high-energy
nation venturi mixer/ignition chamber is connected by a 1-in. discharge ignitor probe is used to ignite the mixture, generat-
pipe to the pilot. The pipe length can be 5000 ft (1500 m) or ing a fireball within the slipstream line that in turn ignites the
more, but the pipe must be 1 in. (2.5 cm) and no heavier than pilot. The main advantages that this system has over the con-
sch. 80. The ignition sequence starts by flowing air and gas to ventional FFG system are quick pilot relight, no flame-front
the venturi mixer. In this system, the flow rates of air and gas lines, and no compressed air required. The main disadvan-
are each controlled and monitored using a needle valve and tages are that critical components are located at the flare tip

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
pressure gage. The ignition chamber is located immediately and, therefore, inaccessible without a flare shutdown, and
downstream of the mixing zone. A spark plug in the ignition that the electrical wire leading from the transformer to the
chamber can ignite the air/fuel mixture. The resulting fireball spark probe is limited to approximately 750 ft (230 m). Many
travels through the pipe until it reaches the pilot. The fireball of the pilots equipped with a slipstream FFG are also
then ignites the air/fuel mixture generated by the pilot as it equipped with a conventional FFG and the associated piping.
exits the pilot tip. In this case, the conventional FFG is used as an installed
It is important to let the FFG line fill completely with the backup ignition system.
flammable air/fuel mixture before the spark is generated. If
the FFG line is not completely filled with the air/fuel mixture, 20.4.2.2.3 Self-inspirating FFG
the fireball will extinguish before it reaches the pilot and will The self-inspirating FFG is a system in which an air/fuel
not light the pilot. Long FFG lines may take minutes to mixture is generated at grade using an eductor system, as
completely fill with a flammable air/fuel mixture, depending shown in Figure 20.40. This eductor is separate from the
on the air/fuel mixture flow rate and FFG line size. For exam- main pilot venturi mixer. A spark, generated just down-
ple, if the velocity of the air/fuel mixture in the FFG line is stream of the ignitor eductor, creates a fireball inside the
50 ft/s (15 m/s) and the FFG line is 1000 ft (300 m) long, ignition line that leads to the pilot tip. The main advantage
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 616 Tuesday, March 15, 2005 9:36 AM

616 The John Zink Combustion Handbook

FIGURE 20.39 Slipstream flame-front generator.

20.4.2.3 Monitors
Verification that a flare pilot is burning is an important and in
some cases mandatory requirement. The pilot’s remote loca-
tion and inaccessibility during flare operation make flame
verification difficult. A brief review of flare pilot monitoring
methods illustrates the difficulties.
Most pilot fuels produce a low luminosity flame because
the gas mixture at the pilot tip contains close to 100% of the
air required by the fuel. It can be very difficult to see a pilot
flame during the day. Viewing at night is generally more
successful. If the pilot is ignited using a conventional FFG,
opening the fuel valve of the FFG can enhance visual sighting,
day or night. The added fuel will produce a larger and more
luminous flame at the pilot. After the pilot flame has been
sighted, the extra fuel should be shut off.
By nature, one immediately associates flame with heat. In
FIGURE 20.40 Self-inspirating flame-front generator. fact, a flame produces heat, ionized gas, light, and sound.
The technique for verifying a pilot flame by sensing the
flame-generated heat with a thermocouple has been used for
that this system has over the conventional FFG system is many years. In the thermocouple technique, the thermo-
that compressed air is not required. The advantage over the couple junction is placed in a position to sense the heat
slipstream FFG is that the critical parts are accessible during generated by the pilot flame. A balance must be struck
flare operation. The disadvantage of this system, however, is between a high exposure to heat with possible rapid thermo-
that the maximum distance of the ignition line is limited to couple burnout and a lower exposure with a slower response
approximately 200 ft (60 m). The exact distance, however, time. The thermocouple is connected to a temperature switch
can vary, depending on the fuel pressure available, composi- or a computer that indicates pilot failure if the temperature
tion of the fuel, diameter and wall roughness of the ignition drops below a set point. In most cases, a shutdown is required
line, and ambient air density. to replace a burned out thermocouple.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Flares 617

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 20.41 SoundProof acoustic pilot monitor.

Other techniques have sought to verify a flame using flame the movement of the top of the flare stack may move the
ionization or optical scanning. The flame ionization method flame out of the sensor’s field of view.
requires two elements located in the pilot flame. The presence The newest flame monitoring technique detects the pilot
or absence of a flame is detected by a change in resistance flame remotely using the overlooked flame characteristic of
between the elements. Like a thermocouple, these elements sound generation.15 This system consists of an acoustic sensor
cannot be maintained during flare operation. The use of flame and a signal processor. The sensor listens to the pilot sounds
ionization to monitor flare pilots is limited. through the flame-front generator line much as a doctor uses
Optical sensing benefits from an accessible grade-level loca- a stethoscope to listen to a heart. Acoustic data is conveyed
tion. Most optical sensors employed on flare pilots use one or from the sensor to the signal processor via a cable. The signal
more infrared wave bands to sense the presence of a flame. processor analyzes the acoustic data and signals the pilot
Ultraviolet sensors, which are frequently used on process and flame status. An acoustic monitor system is shown in
boiler burners, are generally limited to use on enclosed ground Figure 20.41. An acoustic pilot monitor can distinguish its
flares. Optical methods may be unable to distinguish pilots connected pilot from nearby sound sources such as other
from the main flame or one pilot from another. In addition, the pilots, steam injection, and combustion of the flare. Weather
optical path can be obscured by heavy rain, fog, or snow, or conditions do not adversely affect the monitor.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 618 Tuesday, March 15, 2005 9:36 AM

618 The John Zink Combustion Handbook

20.4.3 Knockout Drums

There are three basic types of knockout drums that can be
incorporated into a flare system: a horizontal settling drum, a
vertical settling drum, and a cyclone separator.
Horizontal settling drums are large drums in which droplets
are allowed sufficient residence time to separate from the gas
by gravity. API RP-5215 provides detailed design guidelines
for this type of drum. Figure 20.42 shows a typical example

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
of this design. The pressure drop across these drums is rela-
tively low. Drums of this type are particularly useful for
removing liquids within or near the process units that may
send liquids to the flare header. It is common for the maxi-
mum liquid level to be at the drum centerline, thus allowing
50% of the total vessel volume to be used for temporary liquid
storage during a relief.
Vertical settling drums work in a similar fashion. In design-
ing vertical settling drums, careful attention must be focused
on droplet terminal velocity because this velocity determines
the drum diameter. Also, the volume available for storage of
liquid during a relief is limited by the elevation of the flare
header piping.
Any small droplets that pass through the knockout drum
can agglomerate to form larger droplets in the flare system
downstream of the knockout drum. Locating the knockout
drum very near the base of the flare stack, or incorporating
it into the stack base, can minimize this problem. Although
the pressure drop required for the settling drums is generally
FIGURE 20.42 Horizontal settling drum at the base of
an air assisted flare. low, the required drum diameter can become impractical to
shop-fabricate if the flow rate is high.
Elimination of very small liquid droplets cannot be
accomplished through a simple reduction in gas stream
velocity. Cyclone separation is best for small droplet
removal. Mist eliminators, utilizing centrifugal force, can be
very effective when incorporated into the base of the flare
stack. They are smaller in diameter than horizontal or ver-
tical settling drums and usually provide high liquid removal
efficiency at the expense of a greater pressure drop. The frost
on the outside of the drum in Figure 20.43 vividly illustrates
the liquid flow pattern.
Agglomeration of droplets downstream of a cyclone sep-
arator is generally less of a problem than it is for the settling
drum designs. The typical settling drum is designed to
remove droplets larger than 300 to 600 microns at the smoke-
less flow rate. Droplet sizes at the maximum flow rate can
be over 1000 microns in some cases. By comparison, the
FIGURE 20.43 Cyclone separator. (Note the frost droplets exiting the cyclone are much smaller, typically 20 to
indicating the flow path of the low-temperature liquid as it is 40 microns, and the droplet size remains low throughout the
removed.) operating range.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 619 Tuesday, March 15, 2005 9:36 AM

Flares 619

However, available volume for liquid storage in a cyclone

separator is generally small compared to the vertical settling
drum because a substantial length is required for the vapor space
in this design. When substantial liquid loads are anticipated,
horizontal settling drums are usually provided at the upstream
end of the flare header to catch most of the liquid volume. The
mist eliminator in the base of the flare stack removes the remain-
ing liquid and minimizes problems with agglomeration.

20.4.4 Liquid Seals

A liquid seal is a device that uses a liquid, such as water or
glycol/water mix, as a means of providing separation of a gas
(or vapor) conduit into an upstream section and a downstream
section. The physical arrangement of a typical liquid seal is
shown in Figure 20.44. Gas flows through the seal when the
gas pressure on the upstream side of the seal is equal to or
greater than the pressure represented by the seal leg submer-
gence plus any downstream back-pressure. The submergence
depth used depends on the purpose of the seal. Present practice
involves submergence depths ranging from about 2 in. (5 cm) FIGURE 20.44 Schematic of a vertical liquid seal.
to over 120 in. (300 cm).
Liquid seals are found in many types of combustion sys-
tems, including flares, due to the fact that liquid seals can be Figure 20.12. The liquid seal in the line to the elevated flare
used to accomplish any of a variety of goals: diverts all flows to the enclosed flare until the pressure drop
• prevent downstream fluids from contaminating the caused by gas flow through the enclosed flare system exceeds
upstream section the submergence depth.
• pressurize the upstream section
• divert gas flow 20.4.4.4 Control Valve Bypass
• provide a safe relief bypass around a control valve A control valve in flare header service can represent a safety
• arrest a flame front or detonation hazard should it fail to open when required. A liquid seal
bypass around the control valve protects the plant against
20.4.4.1 Prevent Upstream Contamination possible failure of the control valve to open during a relief.
One of the most frequent uses of liquid seals is to prevent air
When the upstream pressure reaches the submergence depth
infiltration into the downstream section from propagating to
plus any back-pressure from the elevated flare, waste gas will
the upstream section. Properly operated liquid seals provide a
begin to flow through the liquid seal to the elevated flare —
safeguard against the formation of an explosive mixture in
whether the control valve is open or not.
the flare header by acting as a barrier to backflow.

20.4.4.2 Pressurize Upstream Section 20.4.4.5 Liquid Seals as Arrestors

As discussed in Section 20.3.4, a negative pressure can exist Liquid seals that are used as flame arrestors generally fall
at the base of the flare stack at low flow conditions. By inject- into three categories as follows:
ing purge gas upstream of the liquid seal, the upstream sec- 1. seals designed to handle incoming combustible gases that
tion is pressurized to a level related to the submergence do not contain air (or oxygen); example: a refinery flare
depth. As a result, any leaks in the upstream section will flow 2. seals designed to handle incoming vapor streams that
gas out of the flare header rather than air into the header. contain a mixture of combustible gases and air (or oxy-
gen); example: the fuel vapor/air stream produced during
20.4.4.3 Diverting Gas Flow tank truck or barge loading operations
Liquid seals are often used to divert gas flow in a preferred 3. seals designed to handle either ethylene oxide (ETO) or
direction. An example is a staged flare system involving an acetylene; example: the vent seal for the ETO gases from
enclosed flare and an elevated flare such as the one shown in
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
a sterilizer used for treating medical supplies
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 620 Tuesday, March 15, 2005 9:36 AM

620 The John Zink Combustion Handbook

of alternative designs have been developed16 that improve on

the basic idea. Several of the designs that have been used over
the years are shown in Figure 20.46. Today, designs are avail-
able that can convert an existing horizontal knockout drum
into a combination liquid seal/knockout drum with enough
submergence depth to enable flare gas recovery.
The two fluids normally used for a liquid seal are water
and glycol/water mix. Whenever possible, water is preferred.
The use of hydrocarbon liquids (other than glycol) is strongly
discouraged.

20.4.5 Purge Reduction Seals

Air infiltration into a flare system through the flare burner was
FIGURE 20.45 “Smoke signals” from a surging liquid discussed in Section 20.3.4. Most systems are designed to
seal. combat air infiltration into the tip and riser by purging. The
amount of purge gas required to prevent air from entering into
the system can be quite large, especially in the case where
20.4.4.6 Design Factors light gases are present. A high purge rate may pose several
Once the purpose of the liquid seal has been established, the disadvantages. First is the cost of the purge gas; second, the
designer can produce a suitable liquid seal design. Several heat from the combustion of the purge gas can be damaging to
factors influence general liquid seal design, including: the flare burner; and third, burning more gas than is absolutely
necessary increases the emissions level of the plant. Adding a
• seal vessel orientation (horizontal or vertical) purge reduction device to the flare system can mitigate these
• seal vessel diameter disadvantages. Normally, such devices are installed just above
• seal leg submergence depth or immediately below the flare burner-mounting flange so as
• configuration of the end portion of the seal leg (seal tip to maximize the air exclusion zone.
or seal head)
Purge reduction devices are intended to improve the effec-
• space above the liquid level
tiveness of the purge gas so that the amount required to
• type and size of outlet
protect the system can be reduced. Purge reduction devices,
• seal fluid selection
often referred to as seals, are based on the use of either of
A complete discussion of all these factors is beyond the two basic strategies: (1) density difference (sometimes called
scope of this text. a density seal), or (2) trap and accelerate (sometimes called a
Liquid seals have been known to cause pulsating flows to velocity seal). The discussion below addresses the principle
flares. Pulsating flow, in turn, makes smoke control difficult of each seal type and its advantages and disadvantages.
and creates a fluctuating noise and light source that can A velocity seal is shown in Figure 20.47. The principle of
become a nuisance to neighbors. Figure 20.45 shows the the velocity seal is to trap air as it enters the flare tip, reverse
“smoke signal” effect that can occur in extreme cases of such its direction, and carry it out of the tip with accelerated purge
flow patterns. The fundamental cause of this pulsating flow gas. Tests on flare stacks, large and small, have demonstrated
is the bi-directional flow that occurs when the displaced liquid that air enters a flare tip along the inner wall of the tip. In the
in the vessel moves to replace gas bubbles released at the gas velocity seal, a shaped trap is placed on the inner wall of the
exit of the seal head or inlet pipe. As the submergence depth tip. The trap intercepts the incoming air and turns it back
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

increases, the buoyant forces acting on the liquid increase and toward the tip exit. At the same time, the shape of the trap
the potential for violent movement by the seal fluid grows. acts to accelerate the purge gas. The accelerated purge gas
Larger vessel diameters also increase the potential for liquid and outflowing air meet at the exit of the seal device and flow
sloshing, which is another driver of pulsation. out the tip. Without the accelerated purge gas, the trap will
Properly designed internals can reduce such pulsations by only delay air entry, not reduce it.
controlling the bi-directional gas flow and movement of the Compared to a density seal, a velocity seal is relatively
liquid. Robert Reed produced some of the earliest internals small and has low capital cost. The velocity seal will reduce
for liquid seals in the late 1950s. Since that time, a number the purge gas requirement but the reduction is tempered by
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 621 Tuesday, March 15, 2005 9:36 AM

Flares 621

FIGURE 20.46 Various liquid seal head types: “A” Beveled end, “B” Sawtooth, “C” Slot and triangle (after API RP-521),
“D” Arms with ports on upper surfaces, “E” Downward facing perforated cone, “F” Upward facing perforated cone.

the amount of oxygen allowed below the seal. A velocity seal interface is all that is required. This purge rate is much lower
requires more purge gas than a density seal. An additional than the rate required for a velocity seal (which will have some
disadvantage of a velocity seal occurs when purge gas flow level of oxygen below it). A density difference as small as
is interrupted. In this event, the oxygen level in the riser nitrogen to air is sufficient for the seal to function. The lighter
begins to increase almost immediately. (or heavier) the purge gas, the more effective the seal becomes.
The arrangement of a density-type purge reduction device Tests have shown that the oxygen level below a properly
is illustrated in Figure 20.48. As the gas flows upward through purged density-type seal will be zero. If the purge gas flow to
the riser, it is directed through two annular 180° turns, thus a density seal is interrupted, air will begin to penetrate the gas
forming spaces where lighter- or heavier-than-air gases are by diffusion. However, the diffusion process is slow and a
trapped. The density difference between the trapped purge gas significant time will pass before air enters the riser.
and air forms a barrier to air movement. Only diffusion will The density seal requires the smaller purge gas rate and
allow the air to work its way through the barrier. Thus, a purge enjoys the lower operating cost. A lower purge gas requirement
rate sufficient to constantly refresh the gas at the gas/air
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
also means less heat around the flare tip and lower emissions.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 622 Tuesday, March 15, 2005 9:36 AM

622 The John Zink Combustion Handbook

dense that it forced plant shutdown. In another case, smoking

caused the shutdown of a major highway and noise broke
windows at a great distance.
A different design concept yielded the first successful
enclosed ground flares in 1968 when two units designated as
ZTOFs (Zink thermal oxidizer flares) were placed in service
by Caltex.17 These enclosed ground flares and dozens of addi-
tional units constructed over the ensuing years led to the
development of the modern enclosed ground flare.
Enclosed ground flares use a refractory-lined combustion
chamber to contain the entire flame, rendering it invisible to
neighbors. A schematic diagram of an enclosed ground flare
is shown in Figure 20.49. The combustion chamber is gener-
ally cylindrical, but can be rectangular, hexagonal, or other
shapes formed from flat panels. Cylindrical sections are gen-
FIGURE 20.47 Airrestor velocity-type purge reduction erally favored. Flat panels are used in some cases to reduce
seal. shipping costs or to optimize field assembly.
A ZTOF is essentially a giant, direct-fired air heater. Air
required for combustion and for temperature control enters the
combustion chamber by natural draft after passing through a
burner opening. Elevated temperatures in the combustion cham-
ber reduce the density of the flue gases inside and produce draft
according to Eq. (20.2). This draft is the motive force that drives
combustion products out the top of the stack and draws air in
through the burner openings. Optimizing the use of the available
energy is an essential part of the proper design of a ZTOF.
Most ZTOFs are designed to handle substantially more
than the stoichiometric air requirement. The excess air is
used to quench the flame temperature. This reduces the
required temperature rating for the refractory lining, which
is a significant part of the overall cost of the system. Although
the quenched flue gas temperature may be 1600°F (900°C)
or lower, the refractory lining — at least in the lower section
of the stack — should be selected for a higher service tem-
perature because local temperatures may be higher than the
final flue gas temperature.
When a ZTOF is used as the first stage of a flare system,
FIGURE 20.48 Molecular Seal density-type purge reduc- there is the potential to deliver more waste gas to the ZTOF
tion seal. than it is designed to handle. Overfiring the ZTOF can result --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

in flame and/or smoke out the top of the stack. There are
usually two safeguards to prevent this from happening. First,
However, the physical size of the density seal makes its capital when the pressure drop created by the gas flow through the
cost larger. ZTOF system exceeds the setpoint of the diversion device
(whether liquid seal or valve), excess gas automatically flows
20.4.6 Enclosed Flares to the elevated flare. Second, most ZTOFs are equipped with
The desire to hide flaring activities dates back to the 1950s. thermocouples to monitor the stack temperature. When the
Flare vendors and users tried for several years to design stack exit temperature exceeds the design level, a temperature
enclosed ground flares and failed, sometimes spectacularly. switch initiates an automatic shutdown, either partial or total,
In one case, the smoke generated by a ground flare was so of the ZTOF burner system. Gas flow is sent to the other parts
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 623 Tuesday, March 15, 2005 9:36 AM

Flares 623

FIGURE 20.49 Schematic of a ZTOF.

of the plant flare system until the cause of the overfiring 20.4.7 Flare Support Structures
condition can be identified and corrected. The combination of the heat released at maximum design

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
ZTOFs have been designed with capacities ranging from flow and the owner’s instructions on allowable incident radia-
less than 100 lb/hr (45 kg/hr) to more than 100 metric tons/hr. tion poses a design challenge that is often solved by elevating
Combustion chambers vary from 3 ft (1 m) to more than 50 ft the flare burner. (See Sections 20.2.6 and 20.3.5 for discus-
(15 m) across and may be over 100 ft (30 m) tall. sion of the factors and design considerations that influence
To maximize the benefit of the available combustion volume, the determination of the required height.) Once the height has
ZTOFs are usually equipped with multi-point burner systems. been calculated, the design focus turns to the selection of the
As discussed in Section 20.4.1.5, breaking up the gas flow into type of flare structure to use.
many small flames improves burner performance. ZTOF sys- In principle, there are three basic support structure concepts
tems frequently operate at pressure levels consistent with liquid plus a variant that can be very useful in certain circumstances.
seal depths. As a result, the available energy from the waste The concepts and the variant are:
gas is reduced. Thus, staging the ZTOF burner systems is even
more important to maintain good performance at turndown • self-supported
• guy wire supported
conditions. When steam-assisted burners are used in ZTOFs,
• derrick supported
steam efficiencies are substantially higher than open air flares,
• derrick with provision for lowering the riser and flare
resulting in lower day-to-day steam consumption.
burner
On small units, adjusting the air openings feeding air into
the combustion chamber can control temperature in the com- A self-supported structure (Figure 20.50) requires the least
bustion chamber. Temperature control is common in landfill land space and can easily accommodate a liquid seal or
flares, biogas flares, and vapor combustors in gasoline loading knockout drum, or both, in the base section. Varying the
terminals. Proper temperature control minimizes emissions diameter and thickness of the structure at various elevations
from these units, which in some cases run continuously. absorbs wind loads. Potential undamped vibration is avoided
The windfence used to manage air flow into the combustion by varying the length and diameter of sections of the structure.
chamber can be designed to muffle combustion and steam Generally, self-supported structures are not cost-effective at
noise generated by the burners. The refractory-lined combus- heights above about 250 ft (76 m).
tion chamber may also absorb high-frequency noise and Perhaps the most common means of supporting an ele-
serves to block the direct line-of-sight path for noise trans- vated flare burner is a riser that is held in line by guy wires
mission from the flames in the chamber. By providing clean, (Figure 20.51). Usually, there are three sets of guy wires
quiet, invisible disposal of day-to-day reliefs, use of a ZTOF spaced 120° apart. The number of guy wires arranged verti-
allows plant operation in harmony with its neighbors. cally at a given location is dependent on the height of the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 624 Tuesday, March 15, 2005 9:36 AM

624 The John Zink Combustion Handbook

With the exception of demountable derricks, any flare

burner support structure with a height of more than about
50 ft (15 m) above grade typically includes a 360° platform
for use during maintenance. If a 360° platform is used, it is
often located just below the flare burner-mounting flange. The
structure will also provide support for ladders, required step-
off platforms, and utilities piping. Some flare structures, due
to their height and location, may require aircraft warning
markings such as paint or lights.
A number of factors enter into the selection of the structure:
physical loads, process conditions, land space available, cost

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
of land, availability of cranes, and the number of risers to be
considered. The selection process can be simplified by using
the guide shown in Figure 20.54. The guide asks a series of
questions that can be answered “yes” or “no,” with the answer
influencing the next question. While the yes/no answers
appear to lead to an absolute answer, there are subtleties that
can promote an alternative. For example, the desire to locate
a liquid seal or knockout drum, or both, in the base of the
stack may make a self-supported design attractive.
The guide refers to situations in which there will be one
waste gas riser (R1) or two waste gas risers (R1 and R2). If
there are two risers, the guide questions the size of the second
riser as compared to R1/3. Typically, a second riser with a
size of R1/3 or less will be a small-capacity flare serving a
vent system or incinerator bypass. Such a small flare, often
referred to as a “piggyback flare,” will be supported by the
main flare or its support structure. If the second riser is greater
FIGURE 20.50 Self-supported flare. than R1/3, it will be treated, for structural design purposes,
as a second major flare. Cases involving more than two risers
are good candidates for a demountable derrick structure.
structure, wind loads, and the diameter of the riser. Guy wire
supported-structures require the greatest land space commit-
ment. Overall heights can reach 600 ft or more. 20.4.8 Flare Controls
Where land area is of high value or limited availability, Flare systems are often associated with flare headers that col-
a derrick structure can be employed. The derrick itself lect gases discharged from relief valves and other sources.
(Figure 20.52) acts as a guide to keep the riser in line. In A flare is called upon to operate properly during upset and
general, derricks are designed with three or four sides and malfunction conditions that impact control systems through-
have been utilized at heights greater than 650 ft (200 m). out the plant, including power failure and instrument air
Flare burners on very tall support structures and flares failure. Therefore, controls on flare systems must be used
located in remote areas are difficult to maintain or replace with discretion to ensure that the flare will continue to
due to the limited number of cranes that can service the operate safely even if its controls fail. Flare controls can
required elevation. In such cases, a derrick variation often help provide effective smokeless performance, low noise
referred to as a demountable derrick is employed. The design operation, and other desirable characteristics during nor-
of the demountable derrick (Figure 20.53) allows the riser mal day-to-day operation.
and attached flare burner to be lowered to the ground, either Many of the controls used in flare systems are associated
as a single piece or in multiple sections. An additional advan- with pilots, ignition, and pilot monitoring and have already
tage of a demountable derrick is its ability to support more been discussed in Section 20.4.2. This section discusses steam
than one full-size riser. control, burner staging, level controls, and purge control.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 625 Tuesday, March 15, 2005 9:36 AM

Flares--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
625

FIGURE 20.51 Guy wire-supported flare.

626 The John Zink Combustion Handbook

FIGURE 20.52 Derrick-supported flare. FIGURE 20.53 Demountable derrick.

20.4.8.1 Typical Steam Control Valve a manual valve is installed around the control valve and its
Reliable steam control is an important part of the smoke sup- block valves. A pressure gage should be installed down-
pression strategy for steam-assisted flares. The simplest stream of the control valve station to provide the operator
steam control system consists of a manual valve that an with a tool for diagnosing control issues and a guide for
operator uses to adjust steam flow to the flare tip. Most plants manual control, when needed.
prefer not to dedicate an operator to manage the steam use of Most steam-assisted flares require a minimum steam flow
their flares. Instead, steam control valves are equipped with for two reasons. First, a minimum steam flow keeps the steam
remote positioning equipment that allows an operator in the line from the control valve to the flare burner warm and ready
control room to adjust steam flow while performing other, for use. It also minimizes problems with condensate in that
more profitable duties. Figure 20.55 depicts a typical steam line. Second, a minimum steam flow keeps the steam mani-
control valve station. fold on the flare burner cool (“cooling steam”) in case a low
The steam control valve on a flare can operate almost flow flame attaches to the steam equipment. To maintain the
completely closed for extended periods of time. As a result, minimum steam flow, a second bypass line is installed with
wear on the valve seat becomes a maintenance issue. To a metering orifice sized for the minimum flow and a pair of
allow for removal and maintenance while the flare is in block valves for maintenance of the orifice.
operation, block valves are recommended both upstream and Steam traps are mandatory wherever condensate can accu-
downstream of the control valve. To operate the flare smoke- mulate in the steam piping. Many steam injector designs use
lessly during control valve maintenance, a bypass line with relatively small orifices, at least in part to reduce audible
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Flares 627

FIGURE 20.54 Flare support structure selection guide.

hand by-pass

pressure
gage
control valve

strainer to
flare

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
steam
supply
steam orifice steam
trap trap

cooling steam by-pass

FIGURE 20.55 Steam control valve station.

noise. Therefore, a steam line strainer is recommended. If the periodic observations by an operator in the control room
orifices are very small, all stainless steel steam piping may looking at a video image from a camera aimed at the flare.
be appropriate. Any smoking condition will be quickly corrected by an
increase in steam flow to the flare. However, when the gas
20.4.8.2 Automatic Steam Control flow begins to subside, the flare flame continues to look
As the flow or composition of waste gas sent to the flare “clean” to the operator. Therefore, some time may pass
varies, the amount of steam required for smoke suppression before the operator reduces the steam flow. As a result, this
changes. Many plants adjust the steam requirement based on method of smoke control tends to result in oversteaming
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 628 Tuesday, March 15, 2005 9:36 AM

628 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 20.56 Staging control valve assembly.

of the flare, which in turn produces excessive noise and available to handle emergency relief loads. Such staging valve
unnecessary steam consumption. assemblies usually do not include a bypass device.
Optical sensing systems are available to monitor the con- Staging valves, especially on the last stages, of a multipoint
dition of the flare flame and adjust the steam flow continu- staged plant flare system are usually designed to fail open when
ously. Automatic optical sensing equipment can effectively the bypass device is a rupture disk. If the bypass device is easily
control steam flow to maintain a consistent flame appearance reclosed or resets automatically, the staging valve can be
with minimum steam usage and minimum noise. designed to fail closed. The bypass device, shown in the figure
as a rupture disk, can also be a relief valve or a liquid seal.
20.4.8.3 Typical Staging Control Valve
20.4.8.4 Level Controls
Energy conversion flare types such as the LRGO discussed Flare systems often include vessels such as knockout drums or
in Section 20.4.1.5 are designed to operate smokelessly liquid seals that can contain liquid levels that must be moni-
when the gas pressure is above a certain level. Two key oper- tored and/or controlled for safe operation. Liquid level is con-
ational goals for such systems using energy conversion burn- trolled in knockout drums to prevent overfilling, as discussed in
ers are to: (1) maintain the gas pressure above the minimum Section 20.3.3. In some cases, it is also important to prevent
level required for smokeless performance, and (2) prevent too low a level. When all the liquid in a knockout drum is
back-pressure from exceeding the allowable design level. To removed, it becomes possible for waste gases in the flare
achieve these goals, a staging control system is used that header to migrate into the drain system, creating a possibly
starts and stops flow to various groups of burners based on explosive mixture and a serious safety hazard. Instrumentation
the incoming gas flow. Depending on the application, the generally consists of one or more level switches or transmitters
staging valve used to accomplish this can be installed in any often mounted together with gage glasses to simplify setpoint
of a number of possible configurations, ranging from a sin- adjustments and to allow visual monitoring or manual control.
gle valve to a complex system of bypasses and block valves. Liquid seal level control presents a number of challenges
Figure 20.56 shows a typical staging valve assembly. not found in other level control applications. In normal oper-
The main staging valve can be either fail closed or fail ation, when gas is flowing through the liquid seal, the surface
open, depending on the safety considerations governing the of the liquid is violently agitated. Small-scale wave action,
system as a whole. Generally, the staging valves for an spraying, and foam generation also create special require-
enclosed ground flare are fail closed when another flare is ments for liquid seal level control systems.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 629 Tuesday, March 15, 2005 9:36 AM

Flares 629

Liquid seals in flare service can accumulate a certain

amount of hydrocarbon condensate. Such condensates are
generally lighter than water and affect level control and safety.
The presence of such condensates in the liquid seal creates
the potential to generate hydrocarbon droplets in the waste

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
gas flowing to the flare tip. As discussed in Section 20.3.3,
this can become a safety hazard. To protect against this haz-
ard, flare liquid seals are often equipped with hydrocarbon
skimming systems that remove accumulated condensate from
the liquid surface, in some cases automatically.
Loop seals are used to prevent gases from escaping a vessel
while allowing liquids to be removed automatically. Hydro-
carbon skimming systems on flare liquid seals often utilize
loop seals, such as shown in Figure 20.57, to provide constant
removal of liquids. Loop seal design guidance is provided by
API RP-521.5 Some additional concerns include:

• Loss of loop seal fluid by evaporation or overpressure

can result in waste gas entering the sewer system and/or
escaping to atmosphere through the anti-siphon break at
the top of the outboard loop.
• Freezing of loop seal fluid can result in overfilling the FIGURE 20.57 Loop seal.
vessel and/or accumulation of hydrocarbon condensate
in the vessel.
• Elevation of the top of the outboard loop must be at or
below the controlled liquid level. • scale or debris carried by high waste gas velocities in the
flare header
• If liquids of differing densities are anticipated, the eleva- • gas compositions that include compounds prone to poly-
tion of the outboard loop must be low enough to allow merize
the lightest liquid to push out the heaviest liquid.
• two-phase flow carrying liquids or condensate into the
arrestor
20.4.8.5 Purge Controls
There are some limited circumstances where the use of a dry
Purge gas injection is one of the most important safety fea-
arrestor may be acceptable. For example, an arrestor may be
tures in a flare system. A common method for controlling the
used on systems that handle relatively clean, dry material or
flow of purge gas is the use of a metering orifice and a supply
systems that can be easily shut down for maintenance
of purge gas with regulated pressure. A typical arrangement
should an over-pressure be detected. As with all flare system
is shown in Figure 20.58. Safety features should include an
issues, careful attention to the safety aspect of the prime
effective strainer to prevent pluggage of the metering orifice,
objective is required.
a flow transmitter with an alarm for low flow condition, and a
supply pressure substantially higher than any anticipated
flare header pressure.
20.5 FLARE COMBUSTION
PRODUCTS
An industrial flare is the most suitable and widely used tech-
20.4.9 Arrestors nology for disposing of large quantities of organic vapor
Dry arrestors such as flame arrestors or detonation arrestors releases. In the 1980s, the U.S. Environmental Protection
have limited application in plant or production flare systems. Agency (U.S. EPA) conducted several tests to determine the
This is due to the concern that the small passages of an arrestor destruction and combustion efficiency of an industrial flare
could become plugged, leading to increased back-pressure on operating under normal conditions. Based in part on these
the relieving or venting source. In the worst case, the source tests, the U.S. EPA made several rulings concerning the
could become over-pressured. A few of the sources of con- design of a flare. The purpose of this section is to discuss the
cern regarding plugging are: results of these tests and U.S. EPA rulings on flare design.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 630 Tuesday, March 15, 2005 9:36 AM

630 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 20.58 Purge control station.

20.5.1 Reaction Efficiency CO 2

%CE = × 100 (20.5)
The terms combustion efficiency and destruction efficiency CO 2 + CO + UHC
have frequently and mistakenly been considered synony-
mous. In fact, these two concepts are quite different. A flare
operating with a combustion efficiency of 98% can achieve a where %CE is the percent combustion efficiency. Notice that
destruction efficiency in excess of 99.5%. even if no unburned hydrocarbons escape the flame, the com-
bustion efficiency can be less than 100% because CO repre-
sents incomplete combustion. It is evident from Eqs. (20.4)
20.5.1.1 Definition of Destruction and
and (20.5) that the combustion efficiency will always be less
Combustion Efficiency
than or equal to the destruction efficiency.
Destruction efficiency is a measure of how much of the origi-
nal hydrocarbon is destroyed, that is, broken down into non-
hydrocarbon forms, specifically carbon monoxide (CO), car- 20.5.1.2 Technical Review of Industrial Flare
bon dioxide (CO2), and water vapor (H2O). The destruction
Combustion Efficiency
efficiency can be calculated by a carbon balance as follows:
In 1983, the John Zink Company and the Chemical Manufac-
turers Association (CMA) jointly funded a research project
CO 2 + CO
%DE = × 100 (20.4) aimed at determining the emissions of flares operating under
CO 2 + CO + UHC
normal, real-world conditions.18 Various mixtures of crude
propylene and nitrogen were used as the primary fuel, with
The term %DE is the percent destruction efficiency, and CO2, waste gas lower heating values (LHVs) varying from approxi-
CO, and UHC are the volume concentrations of carbon dioxide, mately 80 to 2200 Btu/scf and flow rates up to 3000 lb/hr. Tests
carbon monoxide, and unburned hydrocarbons (as methane)
were conducted using steam-assisted, air-assisted, and non-
in the plume at the end of the flare flame, respectively. Notice
assisted flares. These tests concluded that flares operating with
that if no unburned hydrocarbons escaped the flame, the
a normal, stable flame achieve combustion efficiencies greater
destruction efficiency would be 100%.
than or equal to other available control technologies.
Combustion efficiency is a measure of how much of the
original hydrocarbon burns completely to carbon dioxide and The U.S. EPA ruled that a flare can achieve a combustion
water vapor. Using the carbon balance approach, combustion efficiency of 98% or greater if the exit velocity of the organic
efficiency can be calculated as: waste stream, at the flare tip, is within the following limits:19
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 631 Tuesday, March 15, 2005 9:36 AM

Flares 631

• Non-assisted and steam-assisted flares: Find: The maximum exit velocity of the gas at the flare
– If 200 Btu/scf < LHV < 300 Btu/scf, tip to achieve a combustion efficiency of 98% or
greater according to the U.S. EPA ruling.
 T (° F ) + 460 
Vmax ( ft s) = 60 × 
Solution: Because the LHV of the fuel is 450 Btu/scf, Eq.
 (20.6)
 520  (20.7) is used to determine the maximum velocity:

– If 300 Btu/scf < LHV < 1000 Btu/scf,  450 + 1209.6  

Vmax ( ft s) = anti log10 
  849.1  
LHV + 1209.6 
Vmax ( ft s) = anti log10  160 + 460 
 849.1  × = 107.4 ft/s (20.11)
 520 
 T (° F ) + 460 
×  (20.7)
 520 
20.5.2 Emissions
– If LHV > 1000 Btu/scf, Industrial flares have been endorsed by the Clean Air Act
Amendments to be one of the acceptable control technolo-
 T (° F ) + 460  gies that can effectively destroy organic vapors. The U.S.
Vmax ( ft s) = 400 ×   (20.8) EPA AP-42 guidance document 7 suggests that a properly
 520 
operated flare, with a combustion efficiency of 98% or
greater, will emit UHC, CO, and NOx at the following rates:
• Air-assisted flare:
– If LHV > 300 Btu/scf, • UHC = 0.14 lb/MMBtu fired
• CO = 0.37 lb/MMBtu fired
(329.5 + LHV )  T (° F ) + 460 
Vmax ( ft s) = ×  (20.9) • NOx = 0.068 lb/MMBtu fired
11.53  520 
In 40 CFR 60.8c, the regulations indicate that during periods
20.5.1.2.1 Hydrogen Enrichment
of startup, shutdown, or malfunction, emissions above the
Because hydrogen has a lower volumetric LHV than organic
regulated limit may not be considered a violation. This
gases commonly combusted in flares, the U.S. EPA amended
includes UHC, CO, and NOx. This could also be interpreted
40 CFR 60 to include an allowance for hydrogen content.
to mean that exit velocity limits do not apply under these
The EPA believes that hydrogen-fueled flares, meeting the
conditions. However, according to 40 CFR 60.10(c), states can
maximum velocity limitation as shown below, will achieve a
make their own rules regarding flare operations as long as they
combustion efficiency of 98% or greater.
are more stringent than the U.S. EPA ruling. As regulations
are in a constant state of flux, the reader should determine the
 T (°C) + 273 
( )
Vmax ( m s) = X H2 − K1 × K2 × 
 293
 (20.10)

current regulations for the plant site in question.

where Example 20.2

Vmax = Maximum permitted velocity, m/s Given: Following the conditions of Example 20.1,
K1 = Constant, 6.0 vol% hydrogen assume that the flare is properly designed and
K2 = Constant, 3.9 (m/s)/vol% hydrogen operated according to the U.S. EPA ruling.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

XH2 = Vol% hydrogen, on a wet basis Find: The pounds of UHC, CO, and NOx emitted in
1 year if the flare is burning the waste gas at a
Equation (20.10) should only be used for flares that have rate of 10,000 lb/hr (4500 kg/hr), 50 times per
a diameter of 3 in. or greater, are non-assisted, and have a year, for 1 hour during each event (neglect the
hydrogen content of 8.0% (by volume) or greater. emissions contribution from the pilots and purge
gas). The density of the gas is 0.05 lb/scf.
Example 20.1
Given: A steam-assisted flare is burning a gas with an Solution: First determine how many Btus are released in
LHV of 450 Btu/scf and 160°F (71°C). 1 year (yr):
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 632 Tuesday, March 15, 2005 9:36 AM

632 The John Zink Combustion Handbook

HR BTU 1 scf 10000 lb such factors as the wind speed, time of day, cloudiness, and
= 450 × × type of terrain. The Gaussian dispersion model was one of the
year scf 0.05 lb hr
first models developed to estimate GLC. This model assumes
1 hr 50 Events that the concentration of the pollutants, in both the crosswind
× ×
Event year and vertical directions, takes the form of a Gaussian distribution
about the centerline of the plume and is written as follows:
Btu
= 4500 × 10 6 (20.12)
yr
 Q   −H2   − y2 
C=  exp  2σ 2  exp  2σ 2  (20.16)
The pounds of UHC, CO, and NOx emitted in 1 year are then  Uσ z σ y π   z   y
calculated as follows:
where
Btu 0.14 lb lb C = Predicted GLC concentration, g/m3
UHC = 4500 × 10 × = 630
6
(20.13)
yr 1 × 10 6 Btu yr Q = Source emission rate, g/s
U = Horizontal wind speed at the plume
Btu 0.37 lb lb centerline height, m/s
CO = 4500 × 10 6 × = 1670 (20.14)
yr 1 × 10 6 Btu yr H = Plume centerline height above ground, m
σy and σz = Standard deviations of the concentration
distributions in the crosswind and
Btu 0.068 lb lb
NOx = 4500 × 10 6 × = 310 (20.15) vertical directions, respectively, m
yr 1 × 10 6 Btu yr
y = crosswind distance, m (see Figure 20.59)

20.5.3 Dispersion This Gaussian dispersion model was derived assuming a

If a flare fails to properly dispose of toxic, corrosive, or flam- continuous buoyant plume, single-point source, and flat terrain.
mable vapors, it could pose a serious health hazard to person- Beychok22 discusses the shortcomings of Gaussian dispersion
nel in the vicinity and the community downwind of the models. Beychok suggests that it is realistic to expect Gaussian
release. Sax20 provides extensive information on many com- dispersion models to consistently predict real-world dispersion
pounds that are sometimes sent to a flare. It is important for plume concentrations within a factor that may be as high as 10.
owners to have a dispersion analysis of their flare performed. Gaussian dispersion models, however, are useful in that they
A dispersion analysis is a statistical method used to estimate can give a rough and fairly quick estimation and comparison
a downwind concentration of a gas vented to the atmosphere of pollutant levels from elevated point sources.
or emitted from a flare flame. Dispersion models are widely The accuracy of a Gaussian dispersion model depends on
used in the industry and have been used in the past to size how well one can determine the plume rise, H, at any given
flare stack heights, estimate worst-case scenarios from emer- downwind distance and dispersion coefficients, σy and σz .
gency releases, and determine potential odor problems. A standard atmospheric stability classification method,
Mathematical modeling of stack gas dispersion began in known as the Pasquill-Gifford-Turner classification, is widely
the 1930s. At that time, these models were somewhat simpli- used in GLC models. This method categorizes the stability

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
fied. Today, however, through the advent of computers, these of the atmosphere into six classes that vary from very unstable
models have become more sophisticated and able to capture (class A) to very stable (class F). An atmosphere that is stable
much more detail of the dispersion problem. The purpose of has low levels of turbulence and will disperse a pollutant more
this section is to discuss the general concepts used for esti- slowly than an unstable atmosphere. The dispersion coeffi-
mating the ground-level concentration (GLC) of a pollutant cients, σy and σz , are dependent on the amount of turbulence
emitted from a flare. in the atmosphere and are, therefore, related to the atmo-
When a pollutant is emitted from a flare, it is dispersed as spheric stability class. For more information on the equations
it moves downwind by atmospheric turbulence and, to a lesser describing the dispersion coefficients, see Turner.23
extent, by molecular diffusion, as illustrated in Figure 20.59. The plume height is defined as the vertical distance from
The GLC of a pollutant downwind of the flare depends on how the plume centerline to grade, as illustrated in Figure 20.59.
fast the pollutant is spreading perpendicular to the direction of There are several variables that can affect the plume height.
the wind and on the height of the plume above the ground.21 These variables are divided into two categories: emission
The rate at which a pollutant is dispersing, in turn, depends on factors and meteorological factors. The emission factors
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 633 Tuesday, March 15, 2005 9:36 AM

Flares 633

Gaussian distribution
of pollutants

Dispersion of Pollutants

z
Wind
Direction

Plume
Height
x
Stack
Height
y
Grade

FIGURE 20.59 Geometry for dispersion calculations.

include the (1) stack gas exit velocity, (2) stack exit diameter, 4. R.D. Reed, Furnace Operations, 3rd ed., Gulf Publish-
(3) stack height, and (4) temperature of the emitted gas. The ing, Houston, TX, 1981.
meteorological factors include the (1) wind speed, (2) air
5. API RP-521, 4th ed., American Petroleum Institute,
temperature with height, (3) shear of the wind with height,
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Washington, D.C., March 1997.

(4) atmospheric stability, and (5) terrain. None of the equa-
tions reported in the literature for estimating plume heights, 6. M.R. Keller and R.K. Noble, RACT for VOC — A
however, take into account all the emission and meteorolog- Burning Issue, Pollution Engineering, July 1983.
ical factors. For a review of these equations, see Moses,
Strom, and Carson.24 7. U.S. EPA, Compilation of Air Pollutant Emissions Fac-
GLC analysis is very complex because the results can tors, Vol. 1, Stationary Point and Area Sources, AP-42,
depend on so many variables, as briefly discussed above. In 4th ed., 1985, Supplement F: Section 11.5, 9/91, Indus-
the past, engineers and scientists have described GLC mod- trial Flares.
eling as an art rather than a science. However, this paradigm 8. J.S. Zink, R.D. Reed, and R.E. Schwartz, Temperature-
is shifting due to more sophisticated computer models.25 Due Pressure Activated Purge Gas Flow System for Flares,
to the complexity of these models, one should consult an U.S. Pat. 3,901,643, August 26, 1975.
expert when requiring GLC analysis.
9. R.E. Schwartz and J.W. White, Flare Radiation Predic-
tion: A Critical Review, John Zink Company, Tulsa,
REFERENCES OK, 1996.

10. R.D. Reed, Flare Stack Burner, U.S. Pat. 3,429,645,

1. J.S. Zink and R.D. Reed, Flare Stack Gas Burner, U.S. February 25, 1969.
Pat. 2,779,399, January 29, 1957.
2. R.E. Schwartz and S.G. Kang, Effective design of 11. R.D. Reed, R.K. Noble, and R.E. Schwartz, Air Powered
emergency flaring systems, Hydrocarbon Engineering, Smokeless Flare, U.S. Pat. 3,954,385, May 4, 1976.
February 1998. 12. R.D. Reed, J.S. Zink, and H.E. Goodnight, Smokeless
3. S.H. Kwon et al., Improve flare management, Hydro- Flare Pit Burner and Method, U.S. Pat. 3,749,546,
carbon Processing, July 1997. July 31, 1973.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch20-Frame Page 634 Tuesday, March 15, 2005 9:36 AM

634 The John Zink Combustion Handbook

13. R.E. Schwartz and R.K. Noble, Method and Apparatus 19. U.S. EPA, Code of Federal Regulations, Title 40,
for Flaring Inert Vitiated Waste Gases, U.S. Pat. Part 60, Standards of Performance for New Stationary
4,664,617, May 12, 1987. Sources.
14. W.R. Bussman and D. Knott, Unique concept for noise 20. N.I. Sax and R.J. Lewis Sr., Dangerous Properties of
and radiation reduction in high-pressure flaring, Off- Industrial Materials, 7th ed., Van Nostrand Reinhold,
shore Technology Conference, Houston, TX, May New York, 1989.
2000. 21. M. Miller and R. Liles, Air modeling, Environmental
15. R.E. Schwartz, L.D. Berg, and W. Bussman, Flame Protection, September 1995.
Detection Apparatus and Methods, U.S. Pat. 5,813,849, 22. M. Beychok, Error propagation in stack gas dispersion
September 29, 1998. models, The National Environmental Journal, January/
16. J.S. Zink, R.D. Reed, and R.E. Schwartz, Apparatus for February 1996.
Controlling the Flow of Gases, U.S. Pat. 3,802,455, 23. B. Turner, Workbook of Atmospheric Dispersion Esti-
April 9, 1974. mates, U.S. Environmental Protection Agency, 1970.
17. H. Glomm, Anordnung und Betrieb von Notabblase- 24. H. Moses, G.H. Strom, and J.E. Carson, Effects of
systemen (blow down systems), Rohrleitungstechnik in meteorological and engineering factors on stack plume
der Chemishen Industrie, 199, 18-28, 1967. rise, Nuclear Safety, 6(1), 1-19, 1964.
18. Chemical Manufacturers Association, A Report on a 25. C. Seigneur, Understanding the basics of air quality
Flare Efficiency Study, March 1983. modeling, Chemical Engineering Progress, 68, 1992.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Castable Refractory
Lining
Stack

Brick Lining W/Castable

Refractory Back-Up Lining

Thermal Oxidizer

Waste Liquid Gun

Waste Gas Stream

Castable Refactory
Floor
Fuel

Controls & Instrumentation Package

Air
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Chapter 21
Thermal Oxidizers
Paul Melton and Karl Graham

TABLE OF CONTENTS

21.1 Introduction............................................................................................................................................. 638

21.2 Combustion Basics.................................................................................................................................. 639
21.2.1 Material and Energy Balance.................................................................................................... 639
21.2.2 Oxidizing/Reducing Combustion Processes ............................................................................. 639
21.2.3 NOx Formation ......................................................................................................................... 639
21.2.4 Carbon Monoxide ..................................................................................................................... 639
21.2.5 Acid Gases ................................................................................................................................ 639
21.2.6 Particulate ................................................................................................................................. 640
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

21.3 Basic System Building Blocks................................................................................................................ 640

21.3.1 Burners...................................................................................................................................... 640
21.3.2 Furnace/Thermal Oxidizer/Incinerator/Combustion Chamber ................................................. 644
21.3.3 Refractory ................................................................................................................................. 645
21.3.4 Catalytic Systems...................................................................................................................... 646
21.3.5 Flue Gas Processing Methods................................................................................................... 647
21.4 Blowers ................................................................................................................................................... 670
21.5 Control Systems and Instrumentation ..................................................................................................... 675
21.6 System Configurations ............................................................................................................................ 677
21.6.1 Non-acid Gas Endothermic Waste Gas/Waste Liquid System.................................................. 677
21.6.2 Non-acid Gas Exothermic Waste Gas/Waste Liquid System.................................................... 681
21.6.3 Acid Gas Systems ..................................................................................................................... 682
21.6.4 Salts/Solids systems .................................................................................................................. 686
21.6.5 NOx Minimization or Reduction Systems................................................................................ 688
21.7 Conclusion .............................................................................................................................................. 689
21.8 Nomenclature .......................................................................................................................................... 689
References ................................................................................................................................................................ 689

637
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 638 Wednesday, September 3, 2003 8:12 AM

638 The John Zink Combustion Handbook

21.1 INTRODUCTION during this process, some elements in wastes (e.g., sulfur and
chlorine) form compounds (e.g., sulfur dioxide and hydrogen
Although improvements are continually made to the effi-
chloride) that, if present in sufficient quantities, must be recov-
ciency of the many chemical and mechanical manufacturing
ered or removed from the combustion products by post-oxida-
methods used to produce an ever-increasing number of com-
tion methods to meet various federal, state, and local air quality
pounds and products, unwanted by-products result from vir-
guidelines. Other elements or compounds in wastes (e.g.,
tually all methods. Such by-products can exist in the vapor,
sodium, sodium chloride, catalysts, or other inert solids) will
liquid, or solid phase. Many are hydrocarbons, although
produce particulates that, if present in sufficient quantity, will
some non-hydrocarbon materials also exist. Regardless, the
also have to be removed by post-oxidation treatment. If nitrogen
many different by-products must be safely contained or
is organically bound to any waste compound, special staged
destroyed to prevent potential environmental damage.
oxidation methods may have to be used to prevent formation
The by-products come from many different industrial sec- of excessive amounts of NOx. The same methods can be used
tors. Petroleum and natural gas production and refining, petro- to break down existing NOx that is part of a waste.
chemical manufacturing, pulp and paper production,
A number of factors determine the design of thermal oxida-
agricultural chemicals, pharmaceuticals, distilling, automo-
tion systems. Process variables such as the waste composition
biles, plastics molding, and carbon fiber and fiber optics
and flow rate affect the size, materials of construction, and
production are just a few of the diverse areas that produce
stability of the system. Economic considerations often impact
wastes. The by-products can be the remains of the normally
decisions of capital expenditure vs. operating costs, as is the
less than 100% efficient chemical processes used to create
case with determining the feasibility of heat recovery systems.
hydrocarbon-based products. The wastes can be impurities
Regulations set the required destruction efficiencies, emission
and catalysts in feedstocks that are not consumed during the
rates, and acceptable ground-level concentrations. Very few
process. Manufacturing processes in widely varying fields
applications have identical specifications. Thus, many systems
often require significant ventilation resulting in contaminated
are custom designed to satisfy the process, economic, and reg-
air that must be treated. By-products can vary from a few
ulatory requirements of a particular application.
parts per million (ppm) in air or water to nearly 100% con-
Regardless of the specific design, most thermal oxidation
centration of a hydrocarbon.
systems consist of some or all of the following components:
Many methods of elimination are available. Smaller
1. a device and method to supply oxygen to the process to
amounts of nonsoluble solids and liquids can be put into sealed
initiate and sustain oxidation (burner or catalyst)
drums and isolated in secure landfills. Larger amounts of
2. a vessel (combustion chamber/thermal oxidizer) to con-
liquids, primarily contaminated water, have been injected into
tain the waste hydrocarbons during oxidation
deep wells. For all practical purposes, vapors cannot be stored 3. a heat recovery (heat exchanger or boiler) and/or flue gas
and must be treated as generated. Activated carbon, for exam- conditioning system
ple, can be used to adsorb organic materials from vapor. Strip- 4. emission control equipment (filters, scrubbers, etc.) to
ping and absorption can also remove contaminants from liquid treat the flue gas prior to discharge to the atmosphere
and vapor streams. Filtration methods are used to remove solid 5. an elevated exhaust point (stack) through which the flue
materials from vapors and liquids. None of these methods, gas can be dispersed into the atmosphere
however, actually destroys the waste material. Chemical and 6. control hardware and logic to automatically maintain and
biological treatments are used to destroy organic waste but monitor the various process parameters to ensure safe
may not be the most cost-effective alternative for rapidly and operation
efficiently treating large amounts of material. The most effec- The purpose of this chapter is to provide a better under-
tive method of rapidly eliminating a high percentage of hydro- standing of the use of thermal oxidation to destroy fume and
carbon contaminants is to oxidize the organic materials at an liquid wastes. To accomplish this, the following are needed.
elevated temperature (at or above 1500°F/800°C). Such high-
temperature oxidation is known as combustion or thermal • An explanation of the basic practical thermal oxidation
oxidation. For some contaminated air streams, effective oxi- and post-oxidation processes and components that can
be combined into complete systems to destroy hydro-
dation can also be achieved at lower temperatures using a
carbon wastes and treat the combustion products to
catalyst to increase the oxidation reaction rate.
achieve required emission limits.
High-temperature thermal oxidation quickly and efficiently • Examples of practical complete system configurations
destroys hydrocarbon-based waste materials, converting the that can be applied to treat different waste compositions
carbon and hydrogen to carbon dioxide and water. However, and combinations.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Thermal Oxidizers 639

21.2 COMBUSTION BASICS must be completely oxidized. The final oxidation is completed
The prerequisites for, the actions during, and the results of at temperatures significantly lower than flame temperature,
combustion must be known to design a thermal oxidation thereby minimizing NOx formation, although SOx is formed.
system equipment that will achieve the destruction and
removal efficiencies needed to protect the environment. 21.2.3 NOx Formation
In Chapter 6, NOx formation is discussed at length. Basically,
21.2.1 Material and Energy Balance three mechanisms for NOx formation are present. Thermal
To correctly choose and design the components of a thermal NOx is formed by the high-temperature reaction of nitrogen
oxidation system, a certain amount of process information with oxygen. Formation increases exponentially with increas-
must be developed. For the burner, the designer must know ing operating temperature. At greater than 2000°F (1100°C),
the amount and properties of fuel/waste and the amount of air it is usually the primary source of NOx if the waste does not
required to provide stable, effective combustion. For the com- contain nitrogen compounds. Prompt NOx is formed by hun-
bustion chamber, the operating temperature, volume, and dreds of rapid reactions between nitrogen, oxygen, and hydro-
properties of the combustion products must be determined. carbon radicals, intermediate species formed during the
For post-combustion treatment of the combustion products combustion process. Prompt NOx can be a large contributor in
such as heat recovery or flue gas conditioning, the mass flow lower temperature combustion processes. Fuel NOx is
and composition of the combustion products must be known. formed by the excess-oxygen combustion of organic com-
For emission control applications such as acid gas removal or pounds containing nitrogen. However, as noted in the preced-
particulate removal, the amount of the pollutant must be ing section, NOx formation can be sufficiently reduced by
known to predict the level of removal required. By completing initially utilizing the substoichiometric combustion process to
material and energy balances at different equilibrium points destroy the waste.
throughout the overall process, all of this information can be
generated. This information is then used to configure systems 21.2.4 Carbon Monoxide
and to design and dimension the individual components. Carbon monoxide (CO) is produced by all combustion reac-
tions in relatively small amounts, especially if proper burner
21.2.2 Oxidizing/Reducing Combustion design is followed. As the products of combustion travel
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Processes through the high temperature combustion chamber, part of

Oxidation reactions occur during both excess oxygen (excess that CO will be oxidized by the excess O2 in the combustion
air) and sub-stoichiometric (starved air) processes. These pro- products. By the time the combustion products exit the cham-
cesses are used separately and in combination to destroy ber, the resulting CO is at levels well below 100 ppmv. If
hydrocarbon wastes. The high-temperature, excess O2 process large amounts of CO are produced because of poor combus-
is by far the most commonly used. The excess O2 process con- tion at the burner, only part of it will be oxidized as the gases
verts virtually all of the carbon and hydrogen in hydrocarbon pass through the combustion chamber, and the outlet concen-
wastes to harmless CO2 and H2O. To be certain the conversion tration will be much greater than 100 ppmv. A high level of
is maximized, combustion air supplied to this process is CO in the combustion products exiting a combustion cham-
greater than the theoretical amount needed to oxidize the com- ber is a direct indication of poor combustion.
pounds. If sulfur is present, the excess air also provides oxy- Some wastes contain CO as a component. Because the
gen to convert it to SO2/SO3 (SOx). Additionally, some recognized auto-ignition temperature of CO is greater than
nitrogen is converted to NO/NO2 (NOx). These compounds 1200°F (650°C), very little is oxidized in lower temperature
have a detrimental effect on the environment and are regulated systems (less than 1400°F or 800°C) using lower residence
with respect to stack emission levels. times (smaller vessels) often used to minimize equipment costs.
The sub-stoichiometric (reducing) process operates with
less than 100% of the theoretical amount of O2 needed to 21.2.5 Acid Gases
oxidize the elements of the hydrocarbon compounds. Some Many waste gases and liquids contain sulfur or chlorine
of the oxygen is combined with the carbon and hydrogen to compounds. When these compounds are oxidized, the sulfur
form CO2 and H2O. However, partial reactions also occur, is converted to SO2/SO3 and, with sufficient moisture, the
forming CO and H2 as well as other products of incomplete chlorine converts mostly to HCl and a small amount of free
combustion (PICs). SOx and NOx production are significantly Cl2. To meet emission limits, these compounds must be
reduced, but eventually the incomplete combustion products removed from the flue gas. HCl can be recovered or
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 640 Wednesday, September 3, 2003 8:12 AM

640 The John Zink Combustion Handbook

removed, while SO2 is normally just removed. Although pilot to provide the initial source of ignition; assemblies to
sulfur and chlorine are the most common acid-producing introduce the fuel, waste, and air; and the means to ensure
components, phosphorus, fluorine, and bromine are occa- flame stability once lit. Burners are used over a wide range of
sionally encountered. Phosphorus can cause corrosion prob- heat releases and can burn gas and/or liquid fuels and com-
lems in heat recovery equipment because of high dew points. bustible waste streams. The mixing of fuel/waste with com-
HF is very reactive and will attack virtually every part of a bustion air is accomplished by the combination of air velocity
system until neutralized. Bromine in organic compounds is through the burner (usually referred to as “pressure drop”)
very difficult to convert to HBr for easy removal. Much of it and the velocity and distribution of the combustible material
goes to Br2 in normal oxidizing conditions. A special pro- as it is introduced into the burner.
cess must be used to achieve high conversion to HBr to
allow high removal efficiency. 21.3.1.1 Pilots
A pilot is essentially a very small burner that provides the igni-
21.2.6 Particulate tion source for the main burner fuel. Pilots normally utilize nat-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Particulate can exist in waste streams as inert solids that ural gas or propane gas that is mixed with air in the pilot
remain after the waste material is oxidized. Common exam- assembly. Two methods of mixing the air and fuel are used. An
ples would be NaCl in wastewater and catalyst material car- inspirated air pilot utilizes pilot gas pressure drop through an
ried over in the off-gas stream from a catalytic cracking unit. orifice at the entrance of a venturi assembly to educt air into
Organic materials can contain elements that remain after the the venturi throat where it mixes with the pilot gas. A pressur-
organic waste is burned. Examples include elements such as ized air pilot must have air available at a pressure greater than
sodium or silicon in compounds catalyst materials such as the operating pressure of the burner. Pressurized air is metered
cobalt, manganese, and nickel. Depending on the point of by use of an orifice or valve into a small mixing chamber
introduction into the system, the particulates formed will where it mixes with the pilot gas, also metered into the cham-
vary in size from several microns to sub-micron in diameter. ber through an orifice. The premixed air and fuel then travel
through a tube to a special high-temperature pilot tip where it
is ignited.
There are two common methods for pilot ignition. The most
21.3 BASIC SYSTEM BUILDING
simple ignition method utilizes a high-voltage (> 6000 V)
BLOCKS electric spark to ignite the air/fuel mixture at the pilot tip.
A simple thermal oxidation system may consist of only com- This arrangement is inexpensive, but heat exposure over time
bustion components, that is, a burner mounted on an integral can damage the spark delivery hardware and lead to ignition
vertical combustion chamber and stack. A complicated difficulties. A common alternative is the flame-front ignition
system may include the combustion components and all system, in which a spark, located outside the burner, ignites
possible heat recovery and flue gas treatment components an air/fuel mixture flowing to the pilot tip through a steel
such as boilers, hot oil heaters, waste preheaters, flue gas pipe of about 1 in. (2.5 cm) diameter. The spark initiates a
conditioning equipment, acid gas and particulate removal flame front that travels through the flowing mixture in the
equipment, and a stack. A catalytic system with a preheater, pipe finally emerging near the area of the pilot tip, igniting
burner, catalytic oxidation chamber, and stack lies some- the pilot flame. Only the open pipe from which the flame
where between the simple and complicated systems. Each of front emerges to ignite the pilot air/fuel mixture is exposed
the components is a stand-alone process block. When neces- to heat. The end of that flame-front pipe is high-temperature
sary, the blocks are combined to build a complete thermal stainless steel and is far less likely to be damaged and result
oxidation system. To properly utilize the building blocks, an in ignition difficulties.
explanation of the components as well as the advantages and
disadvantages is necessary. 21.3.1.2 Fuel Introduction
To be considered a fuel, the material must have sufficient heat-
21.3.1 Burners ing value to sustain stable combustion once ignited. The mate-
The burner is the component required to mix and ignite the rial can be a gas or a liquid. To quickly and efficiently burn, a
fuel (and waste, if capable of sustained combustion) and air, gas need only be mixed with the appropriate amount of air, but
and to provide a stable flame with appropriate shape and a liquid must first be atomized into fine droplets and then
combustion characteristics throughout the design operating mixed with air. Unlike the pilot, main burners used in thermal
range of the system. The basic parts of a burner include a oxidizers almost never use premixed air and fuel. Therefore, a
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 641 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 641

method must be employed to quickly mix the main fuel and assist medium such as steam or air is utilized. Droplet sizes
air. The most effective method is to separate the main fuel flow, similar to that achieved in a mechanical tip at 200 to 300 psig
whether gas or atomized liquid, into smaller “jets” of flow (14 to 20 barg) can be achieved in an assist medium gun at
using a tip or multiple tips with orifices (ports) drilled at the only 60 to 100 psig (4 to 7 barg) pressure drop for both fluids.
proper size and orientation. This serves two purposes: (1) more Lower pressure operation also has the important advantage that
individual jets provide more fuel surface area exposed to the it is accomplished with larger liquid passages through the tip,
combustion air, regardless of the jet velocity; and (2) as the jets making it less susceptible to plugging.
of fuel exit the tip at significant velocity, combustion air is
drawn into the rapidly dispersing fuel jet. Thus, air is quickly
mixed with the fuel. Once the fuel/air mixture is ignited, a 21.3.1.3 Waste Introduction
mechanism is required to provide flame stability, that is, con- Waste gases and waste liquids that are not capable of burning
tinuous ignition of the fuel/air mixture near the point of fuel as a stable fuel are usually introduced downstream of the
introduction. This is usually accomplished by establishing an burner. However, although a waste does not have sufficient
airflow disturbance adjacent to some of the fuel exit ports. The heating value to sustain stable combustion, it is still classified
flow disturbance creates localized recirculation of a portion of as endothermic or exothermic for a specific operating tem-
the reacting flame constituents, thus continuously igniting the
perature. Waste is considered to be endothermic if the hydro-
main fuel as the fuel mixes with air.
carbon content (heating value) is small and much more than a
minimum amount of auxiliary fuel must be burned to main-
21.3.1.2.1 Gas Tips
tain the required operating temperature in the combustion
For fuel gas or higher-pressure combustible waste gas introduc-
chamber. If the heating value is high enough that a cooling
tion, specially designed tips made of heat-resistant alloys are
utilized. They are mounted on the end of pipes, which are often medium must be added to control the maximum operating
removable through the front of the burner, so the tips can be temperature, the waste is considered exothermic. A waste can
externally accessed for maintenance and replacement. Based on be exothermic at a lower operating temperature but endother-
the type and amount of gas to be introduced and the amount of mic for a higher operating temperature. Waste liquids are nor-
gas pressure available at the tips, a specific number of firing mally available at higher pressure and can be atomized into the
ports are drilled into each tip. Smaller ports, known as ignition system using hardware (tips, etc.) similar to that used for
ports, are also drilled into each tip. A very important purpose of liquid fuels. Waste gases, on the other hand, are normally
the firing and ignition ports is to direct and shape the gas dis- available only at lower pressures, so injection hardware must
charge from the tips, thereby directing and shaping the flame. be designed for the lower pressure drop. Waste gas injection
For lower heat release burners, a single gas tip, located at the “tips” are often simply open pipes. As with fuel gas, more
center of the burner, is often used. For higher heat release burn- pipes will better distribute the waste gas and mix it more rap-
ers, multiple gas tips, arranged symmetrically around the cir- idly with available oxygen. Endothermic wastes may not sup-
cumference of the burner, are used, just as multiple ports on a port stable combustion but a significant amount of organic
single tip are used, to increase the rate of mixing of the fuel and material (heating value) may still be present. If so, liquid and
air thereby increasing the oxidation reaction rate. Rapid oxida- gas waste injection hardware may have annular spaces
tion of fuel is important because it must be burned before non-
around them for local introduction of air to react with the
flammable wastes, such as contaminated water, can be
organic material. The air entering through the annulus also
introduced into the system. The gas pressure drop through the
serves to cool the hardware as well as providing some or all
tips is usually in the range of 10 to 25 psig (0.7 to 1.7 barg).
of the air required for oxidation. Positioning of the waste
21.3.1.2.2 Liquid Tips injection hardware is critical. Low organic content wastes,
Liquid fuel and liquid waste tips serve the same purposes as whether liquid or gas, should be introduced downstream of
gas tips but are more complex. Mechanical atomization, the burner flame zone so as not to impair oxidation of the
requiring a 200 to 300 psig (14 to 20 barg) liquid pressure drop fuels and result in formation of carbon monoxide, unburned
at the tip, can be used. However, the turndown for mechanical hydrocarbons, or worse, soot. If the waste is air that is only
atomization is only about 3:1. At that point, the atomized slightly contaminated with organic material, it may be used
liquid droplets become larger than preferred for optimum burn- as the combustion air source, and introduced directly
ing. To maintain the small droplet size over a larger operating through the burner. This is the method of waste introduction
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

range and reduce the amount of liquid pressure needed, an for a catalytic system.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 642 Wednesday, September 3, 2003 8:12 AM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
642 The John Zink Combustion Handbook

FIGURE 21.1 Typical natural-draft burner.

21.3.1.4 Low Pressure Drop Burners systems require at least one blower to move the gas through
Low pressure drop burners are designed to operate with a the system. Figure 21.1 illustrates a typical natural-draft
very low pressure drop/air velocity through the combustion burner configuration.
zone. Generally, the motive force that “pulls” the air through
the burner is created by the buoyancy of hot flue gas in the
21.3.1.5 Medium to High Pressure Drop
stack relative to the cooler atmospheric air. This “pull,” or
(Forced-Draft) Burners
“draft,” created depends on stack height, flue gas tempera-
ture, flue gas flow rate, and stack diameter. Because the draft Medium to high pressure drop burners operate at much
is a natural result of system configuration, it is designated higher pressure drops than natural-draft burners because the
“natural” draft, and burners that utilize this motive force are combustion air supplied to the burner is pushed, or “forced,”
natural-draft burners. Natural-draft systems typically gener- into the system by a blower or compressor. Thus, such burn-
ate 0.15 to 1.5 in. (3.8 to 38 mm) w.c. (water column) nega- ers are known as forced-draft burners. Generally, the
tive pressure at the stack base. The majority (75 to 80%) of medium pressure drop burner would be designed for 1.0 to
the drop occurs as the air passes through the throat of the 8.0 in. (2.5 to 20 cm) W.C. pressure drop. A high pressure
burner into the combustion zone. At such low pressure drop, drop burner may be designed for 30 in. (76 cm) W.C. or
the velocity of the air at the burner throat can be no more than more pressure drop. With much more energy available to
15 to 50 ft/s (4.6 to 15 m/s). As a result, the burner flow-area- provide fuel and air mixing, the heat-release-to-flame-length
vs.-air-flow ratio is relatively large, and air flow control and ratio can be much greater for both gas and liquid fuels. Low
distribution can be more difficult. With little energy available pressure combustible waste gases can be more easily intro-
in the form of combustion air velocity, and because a short duced because the greater energy of the air provides a large
flame is usually desired, multiple gas tips are normally uti- portion of the fuel-to-air mixing. Combustible waste liquids
lized with a fuel gas fired natural-draft burner. For liquid that are more difficult to burn are more easily burned in a
fueled natural-draft burners, an increased number of ports higher pressure drop burner. With more pressure available,
and wider-angle port orientations for the liquid tip are the flow-area-vs.-air-flow ratio is smaller, so the burner is
methods used to shorten the flame. smaller. Also, proper combustion air flow distribution is
Natural-draft systems are impractical if any equipment more easily achieved. Some of the available pressure can be
(e.g., boiler or scrubber) with substantial pressure drop is used to impart an angular velocity component to the com-
required downstream of the furnace. Hence, most incinerator bustion air to further aid in mixing. Figures 21.2 and 21.3
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 643 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 643

Combustion Air Low Btu

Waste Gas

Pilot

Atomizing
Steam

Waste Liquid
(>6,000 Btu/lb)
Fuel Gas

FIGURE 21.2 Typical medium pressure drop burner.

FIGURE 21.3 Typical high pressure drop burner.

are typical configurations for medium and high pressure center of the burner, through a stabilizer cone, with gas tips
drop burners, respectively. located at the circumference of the circular throat of the
burner. This configuration is shown in Figure 21.2. By sepa-
21.3.1.6 Combination Gas and Liquid Fuel Burners
rating the fuels, better air mixing and faster burning occurs. If
Combination fuel burners are virtually the same as dedicated
gas or liquid burners. The difference is the special care the gas tips are located near the liquid tip, the gas will con-
required when locating the liquid and gas tips. A typical consume the air more readily, delaying mixing of air with the
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

figuration would be for the liquid injection point to be at the atomized liquid. This would delay the oxidation reaction of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 644 Wednesday, September 3, 2003 8:12 AM

644 The John Zink Combustion Handbook

bons. Additionally, modifying design factors to reduce size

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
TABLE 21.1 Typical Thermal Oxidizer Operating Conditions
Operating Retention can increase undesirable emissions, such as NOx.
Temp. Time Typical Some complex hydrocarbons in waste require longer resi-
Waste Type (°F) (seconds) DRE
dence Times and/or higher Temperatures than a simple sulfur
Lean gases containing 1200–1400 0.3–0.6 >99 compound such as H2S. Longer Time means a larger vessel,
hydrocarbons or sulfur compounds
which increases capital cost. Higher Temperature means more
Lean gases containing common 1600–1800 1 >99.9
chlorinated solvents fuel usage, unless the waste is exothermic, which increases
Liquid streams 1600–2000 1.0–2.0a >99.99 operating cost.
Halogenated hydrocarbon liquids 1800–2000 1.5–2.0a >99.99 For the same operating temperature, a waste stream con-
Dioxins and polychlorinated 2200 2 >99.9999
biphenyls (PCBs) taining primarily water with a hydrocarbon contaminant will
a
require more time for the mixture to vaporize and raise the
Extra time for liquids to allow for droplet evaporation.
hydrocarbon to its oxidation point than would a vapor waste
stream containing the same hydrocarbon contaminant. Again,
the liquid resulting in a longer flame zone. In general, the a longer time means a larger, more expensive T.O.
faster the gas and liquid fuels are burned, the sooner non- If the flow in the T.O. is not relatively Turbulent or contains
combustible wastes can be introduced to the system. areas of stagnant flow, a longer residence Time will be required
for the available oxygen to come into contact with the organic
material so it can be oxidized and destroyed. This poor mixing
21.3.2 Furnace/Thermal Oxidizer/ of waste with air delays the onset of oxidation. A larger, more
Incinerator/Combustion Chamber costly vessel would be needed to achieve the desired DRE.
Although combustible fuels and wastes are introduced through T.O. designers are careful to avoid such areas of stagnant flow
the burner in high-temperature oxidation systems, noncombus- to minimize wasted volume whenever possible.
tible wastes must be introduced downstream of the flame zone. A high DRE requirement may result in the need for longer
Once all the fuel, waste, and air are combined, several inti- residence time and/or higher operating temperature, assuming
mately connected and simultaneous conditions must exist to the mixing is adequate. The greater the T.O. temperature, the
achieve the required destruction removal efficiency (DRE). greater the oxidation reaction rate. Again, a higher tempera-
The mixture must be (1) exposed to a sufficiently high Tem- ture may require additional fuel, adding to the operating cost.
perature (2) for an adequate period of Time (3) in a relatively A higher temperature can also result in the need for more
Turbulent environment to enable the oxidation reactions to expensive refractory material. On the other hand, the
reach the degree of completion needed to achieve the waste increased reaction rate could reduce the residence time needed
destruction efficiency required. These conditions are known and result in a smaller volume T.O., thus decreasing capital
as the “Three Ts” in the combustion industry. The vessel that cost. For wastes with high heating values, intentionally
provides the environment for all these conditions is known by increasing the operating temperature of a T.O. in order to
various designations as a furnace, thermal oxidizer, combus- reduce T.O. size can be economically attractive.
tion chamber, or incinerator. For consistency, this vessel is Increasing the temperature has the above-mentioned pros
referred to as a thermal oxidizer (T.O.). To provide the environ- and cons, but an overall drawback is that higher temperatures
ment needed for the three conditions to be optimized, the T.O. may lead to higher nitrogen oxide (NOx) emissions.
must have the correct volume and geometry. It is obvious that for T.O. design, it ultimately becomes a
trade-off between the capital cost of increased residence time
21.3.2.1 Size and the costs and problems associated with an increase in
The optimal residence Time, or working volume, needed to operating temperature.
meet the required DRE in the T.O. is a function of many factors, Table 21.1 shows typical T.O. operating conditions for a
the most important of which are waste composition, waste variety of cases. Although ranges of operating conditions have
characteristics, degree of Turbulence, DRE required, and T.O. been developed by testing, experience with actual operating
operating Temperature. The working volume is generally systems allows specific conditions to be set on a case-by-case
considered to be from the point at which the final amount of basis, depending on past experience with similar waste streams.
waste is introduced, to the point nearest the incinerator outlet
where the operating temperature is measured. Size can also 21.3.2.2 Flow Configuration
be influenced by the capital and/or operating costs, which are The orientation and flow configuration of the T.O. must
not directly related to the destruction efficiency of hydrocar- accommodate the user’s space restrictions, the characteristics
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 645 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 645

of the waste being burned, the downstream flue gas treat- Refractories used in T.O. vessels are primarily ceramic
ment requirements, and again, provide the most cost-effec- materials made from combinations of high-melting oxides
tive solution. The T.O. can be arranged for vertical up-flow, such as aluminum oxide or alumina (Al2O3), silicon dioxide

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
horizontal flow, or vertical down-flow of the combustion or silica (SiO2), or magnesium oxide or magnesia (MgO).
products. Vertical up-flow is preferred for the simplest situa- Refractories containing primarily alumina and silica are
tions when the combustion products are vented directly to “acid” refractories and are by far the most common type used
atmosphere with no downstream treatment needed. This is for T.O. linings. Refractories containing large amounts of
also the arrangement of choice if the equipment located MgO are “basic” refractories and are used for their good
immediately downstream of the T.O. requires an elevated resistance to specific reactive ash components, particularly
entry. Horizontal flow is the most utilized configuration alkali metal compounds, which result from burning some
when a heat recovery boiler or other equipment with side inorganic salt-laden wastes. Because MgO refractories are
entry near grade is located immediately downstream of the significantly more expensive than the alumina and silica mate-
T.O. Vertical down-flow is required in many systems when a rials, the T.O. configuration is often optimized to allow use
waste is burned that contains ash-forming materials or salts of and to maximize the life of the less-alkali-resistant alumina
to prevent the accumulation of these materials in the T.O. and silica refractories.
On-line solids collection and removal equipment must be Refractories can be further divided into “hard” and “soft”
installed at the base of the T.O. for this case. categories, which applies to their state when ready for service.
Most T.O. vessels are cylindrical in design with length-to- Hard refractories can be further categorized as bricks, plastics,
inside diameter of refractory (L:D) ratios ranging from 2:1 to or castables. Brick refractory is available in a wide variety of
4:1. For a cylinder of a specific volume, the surface area is compositions ranging from high-alumina-content aluminosil-
the least when the L:D is to 1:1. As L:D increases for the icates to magnesites. The binding material in brick refractory
same volume, the surface area of the cylinder increases. From can be calcium cement based or phosphoric acid based. A
an L:D of 1:1 to 2:1, the increase is almost 26%; from 1:1 to brick lining is held in place by gravity and/or the compressive
3:1, the increase is more than 44%; and from 1:1 to 4:1, the forces resulting from proper placement (as in the construction
increase is almost 59%. The result is that the cost of the of an arch). The linings must be installed in a vessel by skilled
cylinder increases as L:D increases. Also, the greater the L:D, craftsmen and require more time to install, especially if special
the longer the T.O., thus requiring a larger plot area for a shapes have to be assembled by cutting bricks. Because of its
horizontal configuration. A positive result of greater L:D is a high density and low porosity (good penetration resistance to
smaller cross-sectional flow area, increasing flue gas velocity. molten or refractory-attacking materials), brick typically offers
As discussed in the previous section, the greater velocity helps the best abrasion and corrosion resistance of any refractory.
minimize dead zones in the T.O., improving mixing and the However, the high density results in the brick usually being
DRE, which enables use of a smaller working volume. A heavy (120 lb/ft3 or more) and the insulating value being lower,
smaller volume reduces equipment costs. Another important resulting in greater lining thickness to achieve the same ther-
consideration that affects L:D is that flame or liquid impinge- mal resistance. An additional consequence of a thicker lining
ment on the T.O. will cause incomplete combustion and is that a larger, more expensive T.O. shell is required to main-
refractory deterioration. A larger diameter or length may be tain the needed inside diameter.
required to maximize DRE or equipment life, regardless of Plastic refractories have similar alumina and silica content
the oxidation reaction rate. as brick and are so-called because the binder in plastic
refractories — usually a water-wetted clay but also available
with a phosphoric acid base — is not set and the material
21.3.3 Refractory is very malleable or “plastic” in the ready-to-install condi-
Virtually all T.O. vessels are internally lined with heat-resis- tion. Once in place, however, the binder is set by exposure
tant refractory material. Installation of a refractory lining pro- to air or to heat. Plastic refractories are shipped ready-to-
vides three important consequences. install in sealed containers to provide shelf life. Once
opened, the material must be used immediately. Because the
1. The steel shell is protected from the high-temperature refractory is in a “plastic” state, it can be forced or
environment inside the T.O. “rammed” into place and can be formed into almost any
2. An extremely hot external surface is avoided. shape needed. It is held in position by an anchor system that
3. The oxidation process is insulated against heat loss so that consists of a metal piece welded to the shell and a prefired
the vessel is a reasonably adiabatic chamber. anchor (a special ceramic refractory shape) held by the metal
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 646 Wednesday, September 3, 2003 8:12 AM

646 The John Zink Combustion Handbook

portion attached to the shell. The anchor extends through Common problems related to all refractories include the
the lining to the surface of the refractory. Also, because following:
plastic refractory also is dense (140 lb/ft3 or more), has low
1. Improper operation or design can lead to thermal shock
porosity, and is relatively easy to install, it is often used in or erosion damage.
difficult-to-brick places when a high-temperature, corrosive 2. Normal acid gases (SO2 or HCl) in the combustion prod-
environment is expected. The total cost of plastic refractory, ucts can raise the gas dew-point temperature to as high
based on material and installation, insulating capabilities, as 400°F (200°C). Excessive refractory thickness, rain, or
and erosion and corrosion resistance is usually less than that extremely cold weather can result in a furnace shell temp-
of comparable composition brick, especially if the final erature below the dew point, which can result in acid
shape needed is unusual. condensation and its associated corrosion. To minimize
weather-related effects, a ventilated, sheet-metal rain
Castable refractories consist of fireclay or high-alumina- shield is often used to prevent rain contact and limit exter-
content aggregates that are held together in a matrix of nal convective heat transfer.
hydraulic calcium aluminate cement. Castable refractories 3. Flame impingement on refractory surfaces can result in
are the least expensive to install among the hard refractories. higher-than-expected temperatures, frequent temperature
Castable refractory is shipped in bags like dry cement, fluctuations, and locally reducing conditions, all of which
mixed with water prior to installation (a variable that can can shorten refractory life.
affect its final properties), and either poured or gunned 4. Liquid impingement on hot refractory will cause spalling
(slightly dampened and blown through a nozzle) into place. and erosion, which decrease refractory life.
The castable refractory is held in place by alloy steel anchors 5. At higher temperatures, salts (those containing Na, Ca,
that are welded to the furnace shell. The castables used for K, etc.) and alkaline-earth oxides (e.g., K2O, Na2O, CaO,
and MgO) will react with most acid refractories. The result
incinerator applications usually weigh between 50 lb/ft3 and
of these reactions can be a loss of mechanical strength,
120 lb/ft3. Compared to the other hard refractories, castables
crumbling, or even a “fluxing” (i.e., liquefaction) of the
generally have the best insulating properties and the poorest exposed surface. In any case, refractory life is shortened.
corrosion and erosion resistance. To minimize lining thick-
ness, a layer of insulating castable refractory is often
21.3.4 Catalytic Systems
installed as a backup to the brick layer, which is the internal
In a typical thermal incinerator, waste destruction occurs in
surface of a T.O. The brick provides the resistance to high
the flame or T.O. because of high-temperature, gas-phase
temperature, corrosion, and abrasion, while the castable pro-
oxidation reactions. In a catalytic unit, waste destruction
vides the insulation qualities needed to reduce overall refrac-
occurs within a catalyst bed at much lower temperatures via
tory thickness.
surface oxidation reactions. The lower operating tempera-
Soft refractories are composed of ceramic fibers formed ture required in a catalytic unit is its advantage because the
into a blanket, a soft block module, or stiff board. They lower temperature reduces the need for auxiliary fuel to
remain soft when in service. The blanket and board are maintain furnace temperature, thus lowering the operating
usually held in place with stainless or other high temperature cost. Another advantage to lower operating temperature cat-
alloy anchors (or pins) welded to the inside T.O. wall. They alytic oxidation is that NOx formation during oxidation is
are easily installed by pressing onto the steel shell with the reduced. The lower temperature also eliminates the need for
pins projecting through. Self-locking washers are then internal refractory lining, reducing the shell diameter. How-
placed on the pins to keep the material from coming loose. ever, use of a stainless steel shell and external insulation is
The block modules have an internal frame that is attached necessary and cancels any reduced operating cost. A major
to an anchor welded to the shell. Soft refractories are much limitation of catalytic oxidation is that the catalyst is suscep-
lighter (usually less than 12 lb/ft3), are much better insula- tible to damage from certain compounds in the waste or
tors, and can be heated rapidly without fear of damage from overheating. Most waste streams for which catalytic
because of thermal shock. Thermal shock is the rapid ther- oxidation is considered are contaminated air streams that
mal expansion of the surface of hard refractory. That layer have more than enough O2 to complete combustion. It is pos-
then separates and falls off, reducing the refractory thick- sible to treat an inert gas stream, such as nitrogen contami-
ness. Soft refractories are limited to 2300°F (1260°C), are nated by a small amount of hydrocarbon, but enough air
susceptible to erosion, and do have poor resistance to alkali must be blended with the gas stream prior to entering the
liquids and vapors. Ceramic fiber refractory is very cost catalyst bed to give approximately 2% O2 in the flue gas
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

effective in certain applications. after oxidation has occurred. This greatly increases the
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 647 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 647

volume and overall capital and operating costs. Another lim- maintained above the ignition temperature. As the waste
itation is that a DRE greater than 99% requires a significant moves through the catalyst bed, oxidation occurs and the gas
amount of catalyst, which increases the capital cost. Recu- and catalyst temperatures rise. The temperature rise in the
perative heat exchangers are often used downstream of the catalyst bed depends on the heating value of the waste stream.
catalytic unit to preheat the incoming waste gas to further If subjected to temperatures between 1200°F (650°C) and
reduce the fuel requirement. 1350°F (730°C) for very long, many catalysts will begin to
Catalysts used in catalytic oxidation systems actually con- suffer significant damage as a result of sintering. Sintering is
sist of a ceramic or stainless steel base material (carrier or the melting and coalescence of the active catalyst material,
support structure) covered with a thin coating of catalyst which results in a loss of available catalyst surface area and,
material. The more surface available for contact with the consequently, a loss of catalytic activity. The rate of sintering
waste gas, the greater the amount of oxidation reaction. The increases rapidly with increasing temperature. A catalyst that
catalyst material is generally one of two types: noble metal shows the first signs of damage at 1200°F (650°C) will likely
or transition metal oxide. The noble metals are generally be severely damaged in a matter of hours at 1500°F (820°C).
preferred. Catalyst type is also based on the ability of the Therefore, for long-term operation and best DRE, the catalyst
catalyst to resist chemical deactivation (poisoning) from com- bed needs to be maintained above the temperature at which
pounds present in waste streams. Typical compounds respon- high-rate reactions occur but below the temperature at which
sible for poisoning are HCl, HBr, HF, and SO2 (reversible significant sintering occurs. Typical catalyst outlet tempera-
poisons) and elements such as Pb, Bi, Hg, As, Sb, and P tures are in the range of 600°F (320°C) to 1000°F (540°C).
(irreversible poisons). While catalysts have been formulated Destruction efficiency in a catalytic incinerator depends on
that will retain their activity in the presence of many of these the waste gas composition, catalyst type and configuration,
poisons, there is no single catalyst that is best for all appli- waste gas temperature at the entrance to the catalyst bed, and
cations. An additional reversible situation is fouling by fine the amount of time the contaminant is exposed to the catalyst
particulate, which could be fine rust particles, refractory dust, (catalyst surface area). Changes in destruction efficiency are
or particulate in the waste stream. For this reason, refractory achieved by changing the amount of catalyst, or by changing
is not used upstream of the catalyst bed, and the vessel mate- the waste flow rate for a given amount of catalyst, either of
rial upstream of the catalyst is often made of stainless steel. which changes the effective exposure (or residence) time. Res-
Fine particulate quite simply covers the surface of the catalyst, idence time in a catalyst bed is often expressed as its inverse
reducing the amount of surface area available for reaction. and is called space velocity (volumetric flow rate of waste,
When the DRE has degraded too much, the particulate can SCFH/catalyst bed volume, ft3). Typically, catalytic units are
often be removed by removing and washing the catalyst designed with space velocities of less than 30,000, inlet tem-
blocks or by washing in place. peratures less than 700°F (370°C), and outlet temperatures less
There are generally two types of catalyst carrier media: than 1200°F (650°C). Practical catalytic systems typically
ceramic beads and honeycomb monoliths. Virtually all new achieve destruction efficiencies of 90 to 99%.
applications utilize catalysts that are applied onto honeycomb Catalytic oxidizers must be configured to provide well-
monoliths because they require less pressure drop and allow mixed, uniform waste gas flow at the catalyst bed entrance
more flexibility in furnace design and orientation. If a bead and to avoid flame impingement on the catalyst bed by the
catalyst is used, the flow must usually be in a vertical heat-up burner. As noted previously, units constructed from
(up or down) direction. The honeycomb monoliths can be catalyst-coated monoliths can be oriented in any flow direc-
installed in any orientation and are usually found in horizon- tion but are usually mounted in horizontal flowing units to
tal flow catalytic oxidizers, which are easier to maintain due facilitate catalyst loading and maintenance. Systems using
to access. either a fixed or a fluidized bed of beads are mounted verti-
Below some minimum threshold temperature, all oxidation cally. Figure 21.4 shows a typical horizontal system with a
catalysts become ineffective. At higher temperatures, the oxi- preheat exchanger.
dation rate increases rapidly until the rate becomes limited
only by the catalyst surface available for interaction with the
waste gas. The temperature at which this rapid increase occurs 21.3.5 Flue Gas Processing Methods
varies, depending on the hydrocarbon, but is typically In some incineration systems, the flue gas does not require
between 400°F (200°C) and 700°F (400°C) and is usually treatment to reduce emissions of acid gases or particulate, and
referred to as the “ignition” temperature. Catalytic units are heat recovery is not economical. In such cases, the hot flue
designed such that the inlet temperature to the catalyst bed is gas leaving the furnace is vented directly to the atmosphere.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

648 The John Zink Combustion Handbook

FIGURE 21.4 Typical horizontal system with a preheat exchanger.

For many systems, however, some form of flue gas cooling is products flow over the outside of the tubes while the steam is
utilized, either by heat recovery and/or conditioning of the generated on the inside of the tubes. Cooling flue gas from
flue gas before it enters downstream equipment. In general, 1800 to 500°F (1000 to 260°C) with a boiler can result in
flue gas cooling is accomplished indirectly or directly. Indi- substantial steam production. Adding an economizer down-
rect cooling is achieved by heat transfer from a higher temper- stream of the boiler will recover even more heat by reducing
ature mass to a lower temperature mass through the use of the flue gas temperature to about 350°F (180°C). (An econo-
heat recovery devices such as boilers, recuperative preheat mizer is a lower temperature heat exchanger used to heat the
exchangers, heat-transfer fluid exchangers, or regenerative boiler feedwater from its normal supply temperature of about
preheat exchangers. Heat recovery devices remove heat from 220°F (100°C) before it is injected into the boiler.)
the flue gas to lower the temperature but do not change the
mass flow rate. Direct cooling is accomplished by adding a 21.3.5.1.1.1 Watertube boiler There are several important
cooler material directly to the flue gas to complete the neces- differences between the watertube and firetube boilers. A water-
sary heat transfer. Adding cooling material to the flue gas tube boiler (Figure 21.5) is generally less expensive to build for
increases the total mass flow rate as well as reduces the over- applications that require high steam pressure (i.e., > 700 psig
all total gas temperature. The added material can be water, air, or 48 barg) and/or large steam flows (i.e., > 50,000 lb/hr or
or recycle flue gas, depending on the downstream equipment. 23,000 kg/hr). Extended surfaces (finned tubes) and superheaters
Emission control procedures include wet and dry particulate are more easily incorporated into watertube boilers, often
removal, wet and dry acid gas removal, and NOx removal. resulting in smaller space requirements. Most importantly, the
heat transfer surfaces of a watertube boiler are accessible to
21.3.5.1 Cooling by Heat Recovery soot blowers (high-pressure steam or air lances) used for peri-
Heat recovery can be in the form of steam with either odic cleaning of the flue gas side of the tubes to prevent loss
firetube or watertube boilers, recuperative heat exchangers, of efficiency due to fouling by nonmolten particulate resulting
process oil heaters, or regenerative preheat systems. Such from waste combustion. Thus, for an application that requires
heat transfer devices act indirectly on the flue gas so that as high-pressure steam production from a large flow of combus-
its temperature is reduced, only the heat content is changed tion products containing significant amounts of particulate, the
— not the composition. boiler design of choice is the watertube. The typical flue gas
pressure drop through a watertube boiler is 2 to 6 in. (5 to
21.3.5.1.1 Boilers 15 cm) w.c. Thus, the watertube boiler is also used when
There are two basic types of boilers: firetube and watertube. pressure drop must be minimized.
In a firetube boiler, the combustion products pass through the
inside of the boiler tubes while water is evaporated on the 21.3.5.1.1.2 Firetube boiler Firetube boilers have the
outside. Conversely, in a watertube boiler, the hot combustion important advantage that virtually all the surfaces in the boiler
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Thermal Oxidizers 649

FIGURE 21.5 Watertube boiler.

are maintained at the steam temperature. Thus, there are no its latent heat of vaporization, as well as the sensible heat at
cold spots on which acid can condense. For this reason, fire- the exit temperature.
tube boilers are well-suited to those applications where low
21.3.5.1.2 Recuperative Preheat Exchanger
to medium pressure steam is to be produced from an acidic
If no steam is needed, waste heat boilers are not a viable heat
flue gas. The typical flue gas pressure drop through a firetube
recovery option. However, if a low heating value waste gas is
boiler is 8 to 12 in. (20 to 30 cm) w.c. Figure 21.6 shows a
being treated and a large amount of auxiliary fuel is needed
typical firetube boiler. to maintain the operating temperature for the required DRE, a
There are times when the flue gas contains molten particu- preheat exchanger can be used to minimize the auxiliary fuel
late but it is cost-effective to cool the flue gas to “freeze” requirement by transferring heat from the flue gas to the
molten particulate so the remaining heat can be recovered in incoming waste gas or combustion air. Figure 21.7 shows a
a boiler. Even if the flue gas is cooled to 1200°F (650°C), for typical all-welded shell-and-tube heat exchanger. The fur-
example, significant heat recovery is still available. It is impor- nace exhaust flows through the tube in the exchanger while
the waste gas or combustion air flows around the tubes inside
tant to understand that the choice of a quenching medium used
the shell. Up to 70% of the energy released in the furnace can
upstream of a boiler will affect both the size of the equipment
be recovered economically by this method. Normal recovery
and the heat recovery. If clean, recycled flue gas from the outlet
efficiencies are in the 55 to 60% range. Structural limitations
of the system is used for quenching, a relatively large mass is
(thermal expansion) typically constrain the hot flue gas tem-
required because of its high initial temperature (350 to
perature to no more than 1600°F (870°C). However, more
500°F/180 to 260°C). The flow through the rest of the system expensive U-tube type heat exchangers exist that can tolerate
could be twice the mass of the flue gas from the T.O. If water higher temperatures.
is used to quench, a much smaller mass is required, resulting Recuperative exchangers can also be of plate-and-frame
in a much smaller mass flow through the rest of the system. type construction. This type can withstand higher temperature
However, using recycled flue gas for the quenching medium expansion differences because of its non-welded construction.
results in much greater heat recovery efficiency than if water However, some leakage of waste gas into the clean combus-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

quench is used because the added water leaves the system with tion products will occur, increasing the unburned waste
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 650 Wednesday, September 3, 2003 8:12 AM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
650 The John Zink Combustion Handbook

FIGURE 21.6 Firetube boiler.

FIGURE 21.7 Typical all-welded shell-and-tube heat exchanger.

Thermal Oxidizers 651

FIGURE 21.8 Regenerative preheat exchanger.

hydrocarbon emission to the point that the apparent DRE is ings make these systems attractive for large waste flows that
not as required. Also, although the initial leakage may be low have little heating value.
enough when the exchanger is first put into service, it will
In operation, the waste gas flows into the system through
likely increase with time, particularly in systems that operate
a hot bed of packing (Chamber 1) before it enters the T.O.
intermittently (shut down daily or weekly). To avoid leakage
The incoming waste gas temperature is monitored at the hot
problems, the more expensive shell-and-tube type heat
end (nearest the T.O.) of Chamber 1. The flue gas exiting the
exchangers with all-welded construction should be used when
T.O. flows through an identical but cooler bed of packing
high DRE is required.
(Chamber 2) before it is vented to the atmosphere. When the
21.3.5.1.3 Regenerative Preheat Exchanger packing in Chamber 2 has absorbed heat to the point that the
A T.O. system with regenerative preheat exchange consists of exit gas temperature rises above a preset maximum, typically
a refractory-lined T.O. connected to three or more vessels 300 to 350°F (150 to 180°C), the hot gas is redirected to the
containing a ceramic packing (often ceramic scrubber bed third bed of packing (Chamber 3), which was out of service
packing) that alternately functions to preheat the waste and to and is cool. At the same time, the incoming waste gas is
cool the flue gas exiting the T.O. section. Figure 21.8 shows switched from Chamber 1 to Chamber 2, which is now the
the configuration of such a system. This system is primarily hot bed, to pick up stored heat before flowing into the T.O.
for contaminated air stream. The temperature of the gases is Chamber 1, which is now temporarily out of service, was in
measured at the inlet and outlet of each ceramic-packed ves- preheat service when the flows were switched. Consequently,
sel. Numerous valves must be used to control the direction of Chamber 1 is filled with untreated waste gas. The waste gas
flow at all times during operation. The system is somewhat in Chamber 1 is purged into the T.O. with “cool” recycle flue
larger and more expensive than a recuperative system, but it gas while it is out of heat exchange service. If the most
is much more efficient. Up to 95% heat recovery is possible if recently used incoming bed was not purged, or if only two
the incoming hydrocarbon content is very low. However, the beds were used, a bed full of waste gas would be vented at
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

normal rates are in the 85–90% range. The potential fuel sav- each flow reversal. The result would be similar to the plate
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 652 Wednesday, September 3, 2003 8:12 AM

652 The John Zink Combustion Handbook

FIGURE 21.9 Organic fluid transfer system configuration.

exchanger problem above in that the apparent DRE would out of the stack into a duct connected to the hot oil heat
not be high enough to meet the emission regulations. Thus, exchanger. An induced-draft blower is located downstream of
at least three beds are required to allow the cool inlet bed to the hot oil heater to “pull” the flue gas from the stack and
be purged before it becomes an outlet bed. In practice, some through the exchanger. The cool flue gas exiting the exchanger
large systems are constructed of five or more beds to over- is then pushed through a cool duct and injected into the hot
come shipping restrictions and to allow prefabrication. stack at least two stack diameters about the base-connection
centerlines. A hot oil temperature controller monitors the fluid
21.3.5.1.4 Organic Heat-Transfer Fluid Heat Exchangers temperature and modulates the flue gas flow to maintain the
Many plants use organic heat-transfer fluids (hot oil) to pro- desired temperature.
vide a controlled heat source for plant processes. If the T.O. To cool the incoming hot flue gas temperature, a flue gas
system can be operated with or without waste so that a contin- temperature controller monitors the hot oil heater inlet
uous heat source is available, the flue gas can be used to heat temperature and operates a valve that regulates the amount
this fluid in a shell-and-tube design exchanger. However, there of ambient air drawn into the stack-to-oil heater duct.
may be periods when the system is generating more flue gas This particular configuration can be used for any similar
than is needed because the heat demand of the hot oil is heat transfer device when more flue gas is available than the
reduced. Also, during normal operation, the temperature of exchanger can process.
the flue gas entering the exchanger may be limited to some
maximum because of the properties of the fluid. For such 21.3.5.2 Cooling Without Heat Recovery
cases, the system may have to be designed to bypass some or Flue gas cooling by means other than heat removal is often
all of the hot flue gas directly to the atmosphere and to cool necessary. When the flue gas has to be processed to remove
the flue gas by mixing in some ambient air to avoid coking the emissions, it must be cooled to a temperature that will not
organic fluid. Figure 21.9 shows the necessary configuration. harm the downstream equipment nor reduce the efficiency of
Briefly, the waste is fired into a horizontal T.O. that is the downstream equipment. If particulate is to be removed
connected directly into the base of a hot vent stack. Another by a dry process such as an electrostatic precipitator, the flue
connection at the base of the stack allows flue gas to be drawn gas must usually be cooled (conditioned) to below 650°F
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Thermal Oxidizers 653

(340°C). If it is to pass through a baghouse, it will have to be flue gas is used to provide the cooling instead of water. As
conditioned to 400°F (200°C) or less. For wet particulate noted in an earlier heat recovery section, although the flue
removal or wet acid gas removal, the flue gas will likely gas flow when using air or cooled recycled flue gas will be
have to be cooled (quenched) to its saturation temperature greater than with water, and the downstream equipment size
for treatment. will be larger, increasing equipment cost, the value of the
additional heat recovered will likely exceed the additional

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
21.3.5.2.1 Conditioning Section cost of larger equipment in only a few years. Also, because
The flue gas exiting a T.O. is at such a high temperature that of the higher flue gas temperature, the entire conditioning
often it must be cooled before entering downstream equip- section would be refractory lined.
ment to prevent damage to that equipment. In some cases,
flue gas may contain molten droplets of material that must be If the flue gas is being quenched to low temperature (400 to
cooled below their melting point (frozen) so they will not 600°F, or 200 to 320°C), the vessel is usually made of carbon
adhere to downstream boiler tubes or other cooler surfaces. steel with internal refractory lining for part of its length at
In other words, the “condition” of the flue gas must some- the hot inlet end, and external insulation but no refractory for
times be altered before it can be further treated. Removing the rest of its length. External insulation is used at the cooler
heat with heat recovery devices (indirect conditioning) has end to prevent condensation in the cool unlined portion.
already been discussed. As noted, removing heat reduces the The hot flue gas is usually passed through a reduced
temperature but does not change the mass flow rate or com- diameter, refractory-lined section to increase the velocity
position. This section reviews direct conditioning heat trans- just before the atomized water is injected. This is done to
fer methods that reduce the temperature and also change the improve mixing and heat transfer, which increases the evap-
mass flow rate as a result of adding a cooling material to the oration rate of the atomized water droplets traveling through
flue gas to which heat is transferred. the conditioning section. The conditioning section outlet
The cooling medium that adds the least mass to the flue temperature is continuously monitored and the water flow
gas is water. Each pound of water sprayed into the flue gas adjusted to maintain the desired temperature. Although the
absorbs almost 1000 Btu as it vaporizes (heat of vaporization), temperature of the conditioned flue gas may be well above
as well as sensible heat. For example, assuming no other heat the saturation temperature, if the flue gas has to be cooled
losses occur, cooling 10,000 lb/hr (4500 kg/hr) of flue gas to less than 400°F (200°C), so much cooling water may
from 1800 to 600°F (980 to 320°C) requires about 2655 lb/hr cause the particulate to become damp and stick to the outlet
(1200 kg/hr) of water at 70°F (21°C). Cooling the same duct between the conditioning section and the dry particulate
amount of flue gas with 70°F (21°C) air requires about removal device. Also, the chances of condensation on the
26,190 lb/hr (11,900 kg/hr) of air — almost 10 times the walls of the outlet duct and dry particulate removal device
cooling mass. If the flue gas was only cooled to 1200°F increase as the water content of the flue gas increases. To
(650°C), the water required is about 1110 lb/hr (500 kg/hr) reduce the possibility of such problems, the flue gas is
and the air needed is about 6150 lb/hr (2800 kg/hr) — only sometimes cooled the last 100 to 150°F (38 to 66°C) by the
about 5.5 times more mass. Minimizing the flue gas flow to addition of ambient air.
downstream equipment is normally desired so the smallest
Figure 21.10 is a general representation of a vertical, down-
size, lowest cost equipment can be used.
flow conditioning section that can use water, air, etc. as the
Because 100% of the water injected should vaporize, the
cooling medium.
conditioning section can be oriented in any direction. How-
ever, the vertical up or down flow configuration is usually
utilized. This configuration also allows features to be used 21.3.5.2.2 Saturation Quench Section
(e.g., a hopper at the base of the unit, regardless of flow If hot flue gas is to be treated by a wet particulate or acid gas
direction) to collect and remove water online if any of the removal device, it is usually best to complete the heat transfer
removable spray tips fail to properly atomize the water. The portion by quenching to full saturation before the subsequent
hopper will also provide storage volume for particulate that mass transfer process is initiated. Although this is not always
may not exit the conditioning section. done, most mass transfer equipment is designed for fully sat-
If the purpose of the conditioning section is to cool flue urated flue gas. The saturation temperature could typically be
gas to only 1200°F (650°C) to freeze molten particulate as high as 210°F (100°C) and usually is no less than about
before the flue gas goes into a waste heat recovery device, 135°F (57°C), depending on the composition and tempera-
more heat will be available for recovery if air or cool recycle ture of the flue gas when it enters the quench section.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 654 Wednesday, September 3, 2003 8:12 AM

654 The John Zink Combustion Handbook

Three basic saturation quench configuration options exist:

1. direct spray contact quench section
2. submerged quench section
3. combination adjustable plug-type quench and venturi
scrubber section with an integral droplet separator

A description of the three options follows, along with advan-

tages and disadvantages of each. In each configuration a
steep temperature gradient will exist between the furnace exit
gas and the quenched gas. As waste rates to the incinerator
change, the location of this gradient will shift slightly, pre-
senting a challenge to the hardware designer because radical
temperature variations may eventually result in refractory
damage. For this reason, a hot-to-cool water-cooled metal
interface is often used in this area.

21.3.5.2.2.1 Direct spray contact quench As shown in Fig-

ure 21.11, the hot flue gas flows downward through an annular
overflow assembly (weir), then through a brick-lined water
spray contact duct (contactor tube) fitted with recirculating
water spray guns, then through a downcomer tube (down-
comer), and into the water collection and droplet separator
vessel (quench tank).
FIGURE 21.10 Vertical, down-flow conditioning section. The weir is the hot-to-cool junction in the system. It is used
to feed fresh (make-up) water into the quench system and to
form a wetted-wall in the contactor tube and downcomer. The
wetted-wall provides a cooling film in the contactor tube and
downcomer. Material of construction for the weir is stainless
steel unless more severe service (chlorinated waste) requires
use of a more corrosion-resistant metal such as hastelloy,
zirconium, or inconel.
The contactor tube, located directly under the weir, is a
short, brick-lined spray tower into which a large quantity of
recycled water (approximately 4 times the calculated amount
needed for saturation) is injected. The spray guns must provide
efficient atomization and penetration across the full cross-
section of the contactor tube if quenching is to be effective
and virtually complete saturation of flue gases achieved. The
high-temperature flue gas is cooled, primarily through vapor-
ization of atomized recycle water, to near the adiabatic satu-
ration temperature. If the flue gas contains particulate, some
will drop out in the quench tank. The recycle water may absorb
some acid gas as long as the equilibrium concentration of acid
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

in the water is less than in the flue gas. The contactor tube
shell can be constructed of carbon steel with a corrosion bar-
rier, special corrosion-resistant alloys, or fiberglass-reinforced
plastic (FRP).
After the nearly saturated flue gas and extra water exit the
contactor tube, they pass through the downcomer, which
FIGURE 21.11 Direct spray contact quench. extends to the water level in the quench tank. The downcomer
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 655 Wednesday, September 3, 2003 8:12 AM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Thermal Oxidizers 655

FIGURE 21.12 Submerged quench.

is smaller in diameter than the contactor tube, thus increasing removes the smaller pieces that might plug the atomizing spray
the velocity and mixing energy of the flue gas and atomized tips. A continuous blowdown from the hydroclone carries out
recycle water to complete the cooling and saturation of the the collected solids, but the strainer must be manually cleaned
flue gas. The flue gas makes a 180° direction change at the periodically. Blowdown from the base of the conical section
downcomer outlet as it turns to travel upward in the annular can be automatic or manual.
space between the downcomer and the quench tank wall to The advantages of the direct spray contact quench are:
the quench tank outlet. The higher velocity through the down-
comer and subsequent 180° direction change, along with the 1. faster, more efficient cooling of the flue gas as a result of
the large heat transfer surface area of the atomized water
low velocity in the annulus, allow water droplets to be sepa-
2. very low pressure drop across the entire section, usually
rated from the saturated flue gas. The downcomer should be
less than 2 in. (5 cm) w.c.
constructed of a corrosion-resistant metal such as stainless
3. efficient water droplet separation from the flue gas
steel or other more exotic alloys as required for the service.
4. large, open flow area unlikely to be affected by any
The quench tank can be made of carbon steel with a corrosion
obstruction
barrier, special corrosion-resistant alloys, or FRP.
Although the primary purpose of the quench tank is as the Disadvantages include loss of cooling flow as a result of
collection sump for recycling water, it also captures some of pump failure and potential plugging of the spray tips. How-
the particulate coming from the T.O. Thus, the configuration ever, a hydroclone separator is used, plugging is minimized.
of the quench tank must allow for continuous, if necessary, Also, utilizing two recycle pumps with auto-start on the spare
removal of solids and continuous withdrawal of recycle water. when flow drops below a minimum greatly reduces the poten-
To accomplish this, the base of the quench tank should be tial loss of flow.
conical, with a solids blowdown nozzle at the base and the
suction point for the recycle pump(s) in the conical section. 21.3.5.2.2.2 Submerged quench This configuration, Figure
Furthermore, the recycle line should be equipped with a dual- 21.12, is mechanically similar to the direct-spray design, but
basket strainer upstream of the pump(s) and a cyclone sepa- instead of atomizing water to provide a large contact surface
rator (hydroclone) downstream of the strainer. The strainer area between the water and hot flue gas, this method divides
openings are sized to remove only large pieces from the recy- the hot gas flow and “bubbles” it through the water. To accom-
cle water that might damage the pumps, while the hydroclone plish this, the hot flue gas enters the quench section, traveling
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 656 Wednesday, September 3, 2003 8:12 AM

656 The John Zink Combustion Handbook

The quench tank serves as the source of the water for

quenching. Although a significant amount of water is enter-
ing the vessel through the downcomer, the internal design of
the quench tank ensures proper internal recirculation of the
water and prevention of excessive gasification of the water
in any location. The bubbling flue gas creates significant
turbulence and splashing at the surface, making droplet sep-
aration in the quench tank more difficult, and often a second
vessel is used as a separator to remove droplets before the
cooled flue gas travels to the next flue gas treatment section.
The separated water usually flows back into the quench tank
by gravity from the separator. As with the direct spray
quench, if particulate enters the quench section, some will
remain in the quench tank, and some of the acid gas will be
absorbed by the recycle water. The configuration of the base
FIGURE 21.13 Adjustable-plug venturi quench. of the quench tank and recycle pumping system components
(strainer and hydroclone) should be similar to that of the
direct spray quench to allow for continuous removal of solids
downward through an annular overflow section (weir),
from the bottom of the vessel as well as continuous with-
directly into a downcomer tube (downcomer) that extends
drawal of particulate from the recycle water to prevent filling
several feet below the liquid surface in the separator vessel
the overflow weir with solids. The quench tank (and separa-
(quench tank). The lower, cylindrical portion of the down-
tor, if used) can be made of carbon steel with a corrosion
comer contains a number of smaller holes through which the
barrier, special corrosion-resistant alloys, or FRP, depending
hot gas exits the downcomer. As the flue gas bubbles upward
on the service.
through the water, cooling occurs. Often, an additional droplet
separator vessel is used. The total pressure drop across this The primary advantages of the submerged quench are:
section is usually between 24 and 30 in. (61 and 76 cm) w.c.
1. No spray tips are required.
The weir is again the hot-to-cool interface in the system. 2. There is a large flow area that is unlikely to be affected by
It is used to feed fresh (make-up) water into the quench system obstructions.
and, unlike the spray quench, is also where the large volume
3. Even with pump failure and loss of recirculation water,
of recirculated water from the quench tank is introduced. All
little or no downstream equipment damage is likely to
the water flows down the wall of the downcomer, flooding occur because the high liquid level in the quench tank
and, therefore, cooling the downcomer wall. The material of will prevent hot gas from traveling downstream.
construction for the weir is usually stainless steel for non-
acid gas service and hastelloy, zirconium, or inconel for acid Disadvantages include high pressure drop across the sec-
gas, such as chlorinated hydrocarbon service. tion and, often, another vessel must be used for effective
The downcomer is usually an extension of the weir for this droplet separation.
process. It is open-ended and extends several feet below the
water level in the quench tank. Although the downcomer is 21.3.5.2.2.3 Adjustable-plug venturi quench This down-
open at the bottom, the flue gas actually passes through holes ward flow-oriented device, Figure 21.13, combines both heat
cut in the circumference of the lower portion of the down- and mass transfer. The upper portion of this one-piece section
comer, well below the water surface but above the open end. utilizes the familiar hot-to-cool interface, an annular overflow
As the flue gas passes through the openings, it is distributed weir followed by a short converging inlet tube, then the adjust-
evenly and separated into smaller volumes that further divide able plug/throat section and an outlet tube. After passing
into smaller bubbles, providing the contact surface area needed through the adjustable-plug venturi, the quenched flue gas
to quickly evaporate water and quench the flue gas to satura- and water then travel through a separator/quench tank and on
tion temperature. The openings must be properly sized and to the next section.
located at the proper depth below the surface for the down- Once again, a weir is the hot-to-cool interface in the system.
comer to function properly. The downcomer is fabricated from It is used to feed fresh (make-up) water into the quench system
the same material as the weir. and recirculate some of the needed large volume of recycle water.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Thermal Oxidizers 657

A converging duct, wetted by the weir overflow, provides an needed, adding to the system pressure drop. Another potential
entrance for the flue gas to the throat section. A tapered, ver- disadvantage is that pieces of refractory material falling from
tically adjustable plug varies the throat area, changing the gas the refractory-lined duct or T.O. above the venturi could block
velocity and pressure drop across the section. A large volume the annular space in the venturi section. This is not as likely if
of recycle water is added to the center of the throat through a the adjustable plug is operated in the automatic mode to main-
pipe that directs the water onto the center of the adjustable tain a preset pressure drop. In the case of material falling from
plug. The water and flue gas pass through the annular space, above, blocked flow area would cause an increase in pressure
between the plug and the throat wall. The high velocity (up to drop above the setpoint, resulting in the plug being moved
500 ft/s, or 150 m/s) at the throat provides the energy needed upward to open the annular space, allowing most pieces to flush
to atomize and mix the water with the flue gas, cooling the flue through the throat. Another potential disadvantage is that very
gas to saturation. Downstream of the throat, a 90° elbow directs large pieces will not pass through the plug /throat annulus as
the quenched flue gas and water to a separator/quench tank. they would with the direct spray or submerged quench designs.
A shaft extends from the plug, through the bottom of the elbow,
and is attached to an actuator, which is used to automatically 21.3.5.3 Particulate/Acid Gas Removal
adjust the position of the plug to maintain a set pressure drop Equipment used to remove pollutants from T.O. system flue
regardless of flow variation. The entire venturi assembly can gas streams are known as air pollution control (APC) devices.
be fabricated using stainless steel, acid-resistant metals, or, in The most common pollutants that result from burning liquid
some cases, a combination of FRP and metals. and gaseous wastes, which require removal, are particulate
The separator/quench tank usually has a tangential inlet to matter and acid gases such as SOx and HCl. The most com-
help separate the droplets and water more effectively from monly used equipment for the removal of these pollutants are
the saturated flue gas. Recycle water and solids removed from dry removal devices such as baghouses and electrostatic pre-
the flue gas collect in the separator/quench tank, so vessel cipitators, and wet removal devices such as venturi scrubbers
configuration and recycle pumping configuration should also and packed columns. This section provides a brief summary
be very similar to that of the two previously described quench of the design considerations associated with these devices.
processes. The separator/quench tank can be built of the same More detailed information can be found in References 1–3.
materials noted above.
Because particulate in the flue gas is more effectively 21.3.5.3.1 Particulate Removal: Dry
wetted (resulting in more being captured) in this quench sec- The most common methods of dry particulate removal are
tion than the two previously described, this method of quench filtration and electrostatic attraction collection. Reasonable,
is sometimes used when particle removal is required or cost-effective efficiencies are achieved even when sub-
desired, especially if the particulate is relatively large and micron particles must be removed. High pressure drop
easy to remove. Greater acid gas absorption is also more likely cyclonic separation is also available but is rarely used with
with this method of quench because of the extra mixing combustion systems.
energy at the throat. Dry particulate recovery is often preferred if the material
The total pressure drop across this section could vary recovered is to be re-utilized. If a wet process is used to
between 20 and 70 in. (51 and 180 cm) w.c., depending on recover catalyst particles, the particles have to be separated
whether the device is used primarily for quench or if partic- from the water and dried before any purification or refining
ulate removal is needed. Advantages of the adjustable-plug process can begin. Dry removal minimizes the volume and
venturi quench are: weight of material that must be handled after recovery. Much
less water is used in the dry removal process making dry
1. Quenching and scrubbing in one section reduces plot
removal more attractive in locations where water is scarce or
space.
expensive. For the same removal efficiency, dry removal
2. No spray tips are required.
requires much less pressure drop than wet removal of the
Disadvantages include (1) higher pressure drop than other same removal efficiency.
quenching methods, (2) loss of cooling water flow as a result
of pump failure, (3) more complexity of design, including the 21.3.5.3.1.1 Filtering device (baghouse) A common method
moving plug and shaft, with a shaft seal that must be main- of dry particle recovery is by collection on the surface of fabric
tained, and (4) the solids removal efficiency of the combination bags (baghouses). The principal design parameters for a bag-
quench/venturi may not be sufficient to meet the particulate house, assuming a particle size distribution for the particulate
emission limit and a second particulate removal device may be is known, are fabric type, air-to-cloth ratio, and cleaning
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

658 The John Zink Combustion Handbook

FIGURE 21.14 Baghouse.

method used. Gas to be cleaned enters the baghouse, flows of bags is set by the desired gas:cloth ratio, which usually
through the bags from the outside (depositing the particles falls between 2 and 5 ACFM/ft2.
on the outside surfaces of the bags), flows inside the bags up
The choice of a cleaning method (pulse-jet, shaker, or flow
to the clean gas plenum, and out to a stack or to another
reversal) can be determined by the fabric strength and dura-
treatment device. Many bag fabrics are available. The fabric
bility, but is often a compromise between capital and operat-
weave is tight enough that some of the particles are initially
ing expenses. The pulse-jet technology has relatively high
captured in or on the surface. Once a base coat of particulate
energy requirements but can usually operate at higher filtra-
(filter cake) has been collected, an even finer filter medium
tion rates, thus requiring a smaller filter area. Conversely, the
than the original fabric now exists, allowing high efficiency
shaker and flow reversal technologies typically have lower
capture of even small particles (up to 99.5% for 0.1-micron
energy requirements and lower filtration rates, leading to
particles and up to 99.99% for 1- to 10-micron particles) with
larger filter areas. Pulse-jet is the cleaning method normally
a relatively low pressure drop of about 6 in. (15 cm) w.c. As
used for T.O. systems. To minimize air usage, bags are not
the particulate accumulates, the pressure drop increases and
all cleaned simultaneously. Also, care must be taken to avoid
eventually the bags must be cleaned. The commonly used
excessive cleaning, which removes the filter cake, reducing
cleaning methods are pulse-jet, shaker, or flow reversal. The
collection effectiveness until it builds up again.
dislodged dust falls to the bottom of the baghouse and is
removed during operation through special valves. Baghouses are sometimes compartmentalized so that
None of the baghouse design parameters is independent, valves can isolate the compartment or section of the baghouse
and all are based on testing and previous experience. Fabrics being cleaned. While this action requires increasing the num-
may be woven or felted and can be made from a variety of ber of bags by 50% and adding inlet and outlet manifolding
materials. Fabrics differ in their particle capture efficiencies, and valves, it does reduce the amount of material being drawn
corrosion resistance, erosion resistance, temperature range, back to the filter immediately after cleaning and increases the
pressure drop, strength, durability, and the ease with which period of time needed between cleaning each section.
they can be cleaned. The fabric is chosen based on exhaust In general, baghouses are used only in applications where
gas conditions, type and size distribution of the particles to the gas is dry and the temperature is below 450°F (230°C).
be filtered, particle loading, and cleaning method. A typical T.O. system application would utilize a pulse-jet
A typical bag used in a T.O. system would be a 6-in. (15-cm) baghouse operated at 400°F (200°C), with a Nomex fabric
diameter by 10-ft (3-m) long cylinder closed at the bottom, designed for an air-to-cloth ratio of between 2 and
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

supported by an internal wire cage and suspended from a tube 5 ACFM/ft2 and a pressure drop of 5 to 10 in. (13 to 25 cm)
sheet forming the top of the dirty gas chamber. The number w.c. Figure 21.14 is a general representation of a baghouse.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 659 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 659

Baghouses are usually used in non-acid gas service and are The force that moves the particles to the collection plates
therefore usually constructed of carbon steel. For non-halogen results from the charge on the particles and the strength of
acid gas service, stainless steel has also been used. Any mate- the electrical field between the emitters and collectors. For
rial can be used as long as it can withstand occasional short- smaller particles, the electrical field strength must be greater
term temperature excursions to more than twice the design to remove the same percentage of smaller particles as bigger
flue gas temperature. Bag filter material can be polyester, particles. The force can be up to several thousand times the
polyaramid, cellulose, fiberglass, Nomex, Gortex, or any acceleration of gravity so that the particles move rapidly to
other proven fiber that meets the operational criteria. the collection surfaces. Because most of the particles retain
The advantages of using a baghouse are: a portion of their negative charge even after contacting the
collecting plate, some remain on the plates until a physical
1. The particle removal efficiency is high.
2. The pressure drop to collect the material is low. action dislodges them.
3. The material is collected “dry” and does not have to be The weight of the particulate on the plates causes some of
separated from water. the particulate to fall into the collection hoppers, but the
4. Although it contains some moisture, the recovered mate- remainder must be dislodged by vibration (or rapping) of the
rial is basically at its final volume and weight. plates and emitters. Once collected in the hoppers at the
bottom of the chamber, the particulate is removed by the same
The disadvantages are:
means as particulate is removed from baghouses.
1. The baghouse is relatively expensive and occupies a lot Design considerations include:
of plot space.
2. Field construction may be required for large flue gas 1. electrical characteristics of the particulate (i.e., how well
flows. it will accept and hold a charge [the particles must have
3. The large baghouse surface area requires extensive insu- a resistivity in the range of 104 to 1010 ohm-cm for efficient
lation to minimize acid gas or moisture condensation. removal by electrostatic means])
4. The maximum treatable flue gas temperature is about 2. gas and particle velocity (very important with sub-micron
450°F (230°C). particulate), including gas velocity in the unit, drift veloc-
ity of the particulate induced by the electric field, and
21.3.5.3.1.2 Electrostatic precipitator An electrostatic particle settling velocity
precipitator (ESP) is a device that removes particles from a 3. gas distribution
gas stream by means of electrostatic attraction. A high voltage 4. electrical sectionalization (i.e., the increase in power input
potential, usually applied to weighted vertical hanging wires in sequential zones or cells through the length of the ESP
to achieve the desired removal efficiency)
(emitters), causes the particles to be charged. Once charged,
5. particle re-entrainment
the particles are exposed to grounded collecting electrodes
(plates) to which the particles are attracted, separating The flue gas must also contain readily ionizable species
(precipitating) them from the flue gas. The particles must then such as O2, CO2, and SO2. Particle resistivity can be a strong
be separated from the collecting plate, while minimizing re- function of the flue gas temperature, composition, and mois-
entrainment, and removed from the ESP collection hopper. ture content. Thus, ESP performance can be quite sensitive
An ESP is a large, often rectangular-shaped, chamber con- to changes in upstream process conditions.
taining numerous flat parallel collecting plates with emitter As with the baghouse, most of the design parameters have
wires located midway between the plates (see Figure 21.15). been developed empirically and then fit to equations to help
The flue gas entering the ESP must be uniformly distributed the engineer develop the physical equipment design. The result
across the chamber for effective treatment. As the flue gas is a chamber containing the correct number and size (length and
passes between the plates, the high voltage potential (40 to width) of collector plates spaced appropriately to allow locating
50 kV) carried by the emitter wires creates a corona discharge, emitters between the plates. The capital cost of the ESP is
making a large number of both positive and negative gas ions. directly related to that physical information. Both sides of each
The positive ions are attracted to the negatively charged emitter collector plate functions as collecting surface area (CSA),
wires, leaving the space between the plates rich in negative which is also referred to as specific collector area (SCA) or
ions. Particles passing through the negative ion-rich space specific collector surface (SCS). This area is often expressed
quickly acquire a negative charge. Smaller particles are, how- in terms of surface area per 1000 ACFM of flue gas through
ever, more difficult to charge. (Note: Negative discharge elec- the ESP. Depending on the particle-size distribution and other
trodes are normally used for industrial ESPs because of the particle-related parameters, removal efficiencies of more than
higher potentials available and more predictable performance.)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
99% can be achieved at less than 300 ft2/1000 ACFM of
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 660 Wednesday, September 3, 2003 8:12 AM

660 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 21.15 Dry electrostatic precipitator.

collector area, but can be upward of 900 ft2/1000 ACFM for 3. field construction may be required for units with large
very high particulate loading. flows
Because the temperature of the flue gas entering the ESP 4. the large surface area encourages heat loss, leading to
is normally greater than 650°F (340°C), corrosion is not usu- potential acid gas dew-point problems with certain wastes
ally a problem if the exterior is well-insulated and sealed from 5. multiple stages may have to be used to achieve the high
rainwater intrusion. Therefore, materials of construction removal efficiency.
would be carbon steel or stainless steel for the emitters, and
carbon steel for the plates and casing. 21.3.5.3.2 Particulate Removal: Wet
Advantages of using an ESP include: The most common methods of wet particulate removal uti-
lized in T.O. systems are venturi-type devices and wet elec-
1. the desire for dry particle collection trostatic precipitators (WESPs) for very small particles.
2. relatively low energy usage Reasonable removal can be achieved with the simple ven-
3. very low pressure drop turi-type device if particles are larger than 1 micron. For
4. small particle removal smaller particles, multi-stage venturi-type devices with sub-
cooling can be effective at high pressure drop. To treat a flue
Disadvantages include:
gas containing a high percentage of very small particles, a
1. relatively high capital cost WESP is the better choice. The pressure drop is also much
2. relatively large space requirement less than with a venturi-type device.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 661 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 661

21.3.5.3.2.1 Venturi-type scrubber The common feature of and the droplet diameter. A greater relative velocity difference
all venturi-type devices is a constricted passage or “throat” improves impaction effectiveness. Similarly, a smaller water
that increases gas velocity to achieve a desired pressure drop. droplet also improves impaction effectiveness.
Flue gas pressure drops range from 20 to 100 in. (51 to The relative velocity difference between the particles and
250 cm) w.c. depending on various factors. If the flow rate the liquid droplets increases with higher flue gas pressure
varies significantly, the cross-sectional area of the throat must drop (i.e., increased energy consumption), which increases
be adjusted to maintain the necessary velocity (pressure drop). the velocity of the flue gas and the particles suspended in it.
Some configurations use a fixed throat but increase or decrease Smaller scrubbing water droplets can be produced by mechan-
recycle flue gas flow to maintain the constant flow rate neces- ically atomizing the water or by using an atomizing medium.
sary to sustain the design pressure drop. Liquid is injected The shearing effect of the high-velocity flue gas also atomizes
either in the throat or just upstream of the throat. Typical liquid the water more finely.
injection rates are in the range of 10 gal/1000 ft3 of gas. If the flue gas is not fully saturated with water vapor,
Because of the high scrubbing water flow rate, venturi scrub- scrubbing fluid will be vaporized until saturation is achieved
bers require a liquid recirculation system. The recirculation and removal performance will be poorer. Vaporization of the
system will include equipment to control the blowdown rate scrubbing fluid causes problems such as reducing the amount
to maintain the total (suspended and dissolved) solids content of scrubbing fluid available for particle removal, forming new
of the recycle water at about 5% by weight. Additional equip- particles from previously captured particulate matter as the
ment can also be utilized to cool the recirculated water and droplet evaporates, and reducing the diffusive capture because
control pH. of a net flow of gas away from the evaporating particle.
Typical particle removal efficiencies (when operated at sat- The vertical, downflow adjustable plug venturi described
uration with a relatively high pressure drop) are > 90% for previously utilizes mostly pressure drop to shear/atomize the
particles with aerodynamic diameters of ≥ 1 micron and ±50% water and to provide the relative velocity difference. Other
for particles with aerodynamic diameters of 0.5 to 1.0 micron. types use a combination of pressure drop and water atomiza-
Overall particle collection efficiency obviously depends on the tion to achieve the velocity difference and smaller droplet
particle size distribution and will range from 80% to above size. Figure 21.16 shows a horizontal venturi scrubber with
99%. For a properly designed, sub-cooled system with efficient atomized water injection upstream of the throat.
droplet separation, the overall particle collection efficiency Although particles larger than 1 micron are more easily
should be ≥ 99%. removed, smaller, sub-micron size, particles are much more
The basic principle of operation for the venturi-type par- difficult to capture. The capture effectiveness can be enhanced
ticulate-scrubbing device is to provide small water droplets by (1) sub-cooling saturated flue gas to below the saturation
that will capture (wet and surround with water by inertial temperature, and (2) using colder water for scrubbing. The
impaction) the particulate matter suspended in saturated flue purpose is to take advantage of thermophoresis and diffusio-
gas. Given time, the small water-encapsulated particle drop- phoresis effects to produce a directional preference in the
lets will then agglomerate (i.e., droplets contact other droplets Brownian motion toward the target droplet by these sub-
and combine to form larger droplets). The larger droplets can micron particles.
be separated from the flue gas downstream of the venturi by Basically, thermophoresis is the migration of a particle
a cyclonic separator, a mist eliminator, a settling chamber, or away from a higher temperature zone and toward a lower
by a combination of two or all three separation methods. temperature zone.
Assuming the flue gas is fully quenched to saturation, the Diffusiophoresis is a more complicated phenomenon that
overall removal efficiency depends how effectively the parti- occurs when a mixture of particles of varying weight exist in
cles are wetted, how much droplet agglomeration time is a gas stream and a concentration gradient within the gas
provided, and how effectively the larger droplets are separated stream occurs for the heavier particles. Diffusion of the
from the flue gas. As with other types of particulate removal heavier particles from the higher concentration zone to the
devices, many of the design parameters have been developed lower concentration zone occurs in accordance with Ficke’s
from empirical data. law. The net motion of the lighter particles is also altered
Inertial impaction of the particle into the droplet is the toward the low concentration zone due to the momentum
dominant mechanism for removal of larger particles with an imparted during collisions with the heavier particles traveling
aerodynamic diameter greater than 1 micron. Primarily, two in that direction. By injecting colder water and sub-cooling
things determine the effectiveness of capture: the relative the flue gas below saturation temperature, a temperature gra-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
velocity difference between the particle and the water droplet, dient and water vapor concentration gradient are created,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 662 Wednesday, September 3, 2003 8:12 AM

662 The John Zink Combustion Handbook

FIGURE 21.16 Horizontal venturi scrubber.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 21.17 Wet electrostatic precipitator.

resulting in a net increase in particle motion toward the target previously described dry ESP. It removes particles from a
water droplets.1,2 flue gas stream by means of electrostatic attraction. As
Advantages of the venturi scrubber include: before, particles must first be charged, and, as in the dry ESP,
vertically oriented emitter wires are used to generate the
1. reasonably good particulate removal performance
corona. Once charged, the particles are drawn to the oppo-
2. relatively simple design
3. usually lower capital cost sitely charged, grounded collecting electrodes, which for the
4. relatively low plot space requirement WESP are tubes through which the emitter wires hang. The
particulate collects on the inside surface of the tubes. The
Disadvantages are (1) high energy consumption (i.e., high particles are then removed from the collection surface while
flue gas pressure drop), (2) high removal efficiency for small minimizing re-entrainment. The primary process difference
particles requires additional cost of sub-cooling, (3) continu- is that the flue gas must be saturated when it enters the
ous blowdown to maintain a low total solids content would emitter/collector section.
be difficult if water was scarce, and (4) the removed solids
are in a large volume of water which has to be treated. Flue gas coming into the WESP first enters a chamber
under the vertically oriented collector tube section (see Figure
21.3.5.3.2.2 Wet electrostatic precipitator A wet electro- 21.17). The saturated flue gas is uniformly distributed before
static precipitator (WESP) functions very similar to the reaching the tubes. The collector tube section is similar in
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 663 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 663

construction to a shell-and-tube heat exchanger in that it has

an inlet and outlet tubesheet and a sealed casing (shell) around
the tubes. The emitter wires are positioned at the centerline
of each tube. Ambient-temperature water is circulated through
the shell side of the collector tube section, ensuring that the
collector tube temperature is less than the temperature of the
saturated flue gas. This causes water vapor in the flue gas to
condense on the tubes and flow downward into the chamber
under the collector tube section. As the wetted particles are
charged and move to the tube wall, they are washed down the
tube by the condensing water. Occasionally, the power is shut
off to part of the collector tube section and water is sprayed
downward through that section of tubes to further clean them.
The distribution section in the inlet chamber can also be
designed to absorb acid gases similar to a packed column. FIGURE 21.18 Simple packed column.
The design criteria for emitting and collecting electrode
areas in a wet ESP are similar to those for the dry ESP.
However, the potential for sparking/arcing is greater in the used primarily on large power boiler applications that have a
WESP, so proper spacing between emitters and tubes, or high flue gas volume. Some of the drawbacks are:
anything else, is very important.
Because the WESP is wet, materials of construction would 1. The reactant/adsorbent must be disposed of after use, so
by removing the acid gas, a solid waste is created.
be stainless steel or Hastelloy (or equivalent) for emitters and
tubes, and stainless steel or FRP for the housing and water 2. The removal efficiency is not high.
collection section. 3. The capital cost of equipment needed for dry inject is
Advantages of the WESP are: quite high.

1. high removal efficiency for sub-micron particles For these reasons, no detailed description of dry systems will
2. low gas-side pressure drop, usually less that 6 in. (15 cm) be covered.
w.c. (normally about 4 in. or 10 cm w.c.) if the distribution
section is designed to absorb acid gases as well as dis- 21.3.5.3.4 Acid Gas Removal: Wet
tribute flow
The most common methods of wet acid gas removal used in
3. can absorb acid gases T.O. systems are packed columns and the previously
4. “cool” service described venturi-type devices. Packed columns contain
5. less problems with re-entrainment because the wet parti- packing material that distributes the water over a large sur-
cles stick to the tube wall
face area for contact with the flue gas. The venturi-type
Disadvantages include (1) greater capital cost than most devices utilize many small water droplets to provide the large
other wet removal devices, (2) larger plot space required, amount of liquid surface area for contact by flue gas. Each
(3) more complicated operation than with other equipment, has its advantages and disadvantages.
(4) multiple stages may have to be used to achieve extremely
low particulate emission levels if the particulate loading is 21.3.5.3.4.1 Packed column A packed column is the
high, and (5) more maintenance may be required due to device of choice most often used to recover or remove acid
corrosion and complexity. gases from a flue gas stream. The device consists of a vertical,
usually cylindrical, vessel containing a section filled with
21.3.5.3.3 Acid Gas Removal: Dry packing material supported by internal hardware. Recirculated
Acid gases can be removed from flue gas by reaction with or water is pumped to the top of the packed section and flows
adsorption by dry alkaline materials such as limestone/lime. downward through the packing and collects in the base of the
For most waste incineration systems utilized by waste by- vessel. Flue gas, quenched to its saturation temperature, enters
product generators, dry removal is not cost-effective and the vessel under the packed section and flows upward through
does not achieve the removal efficiency needed for the the packing and out to the atmosphere or to another treatment
amount of combustion products generated. Dry removal is section. Figure 21.18 shows a simple packed column.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

664 The John Zink Combustion Handbook

The purpose of the packed column is to transfer compounds Packed columns using water to recover acid are commonly
from the flue gas to an absorbing liquid. For combustion sys- referred to as “absorbers,” and those operated with a caustic
tems, the compounds are acid gases such as HCl or SO2. The solution to remove acid are commonly referred to as “scrubbers.”
absorbing liquid is typically water or a weak caustic solution. Absorbers yield an acid solution and are typically used in
As the liquid flows downward through the randomly placed applications in which the acid solution can be used in a
packing, it is distributed over the large amount of packing process or when the waste water treatment plant can make
surface. The counter-current flow of flue gas traveling in the use of the acid solution to neutralize a caustic waste. Packed
open spaces between packing comes into contact with a large bed absorbers can be designed to produce acid purge streams
amount of liquid for a relatively long period of time. As the of up to 22% HCl by weight (the azeotropic maximum con-
flue gas flows through the absorbing liquid, soluble gases are centration), assuming the HCl vapor concentration in the flue
dissolved into it. Just as in wet particulate removal cases, the gas leaving the absorber is sufficiently high. The concentra-
flue gas must be saturated prior to coming into contact with tion of HCl in the flue gas determines the maximum strength
the absorbing liquid to eliminate evaporation during contact. of the acid blowdown stream. A typical 10-ft (3-m) deep
absorber section could remove about 99% of the HCl present
The rate of acid gas absorption at any point within the in the flue gas while producing a 2% HCl blowdown stream.
packed section is limited by the mass transfer rate across the Scrubbers not only remove acid gas from the flue gas, but
gas-phase boundary layer. Thus, a large mass transfer coeffi- also neutralize the dissolved acid. By adding NaOH to maintain
cient, a high concentration of the pollutant in the flue gas, a the pH between 6.8 and 8.0 in the absorbing liquid, the vapor
low concentration of the dissolved pollutant in the absorbing pressure of the acid gas in the outlet flue gas is greatly reduced,
liquid (i.e., low pollutant vapor pressure over the absorbing thus increasing the absorption rate. The advantages of the
liquid), and a large amount of interfacing contact area all scrubber over the absorber are a less corrosive blowdown
increase the rate of absorption. Counter-current flow of flue stream, and either a greater acid gas removal for the same
gas to the absorbing liquid, which puts the cleanest absorbing blowdown rate or a much lower blowdown rate for the same
liquid in contact with the cleanest flue gas, is the primary percentage of acid gas removal. A typical 10-ft (3-m)-deep
reason absorption can be so effective in a packed column. scrubber could remove 99.9% of the HCl present in the flue
The packing type, the flue gas velocity, the type of substance gas and produce a blowdown stream containing up to 5% by
(i.e., acid gas) to be dissolved into the absorbing liquid, and weight total solids, most of it being NaCl.
the type of absorbing liquid determine the mass transfer coef- The neutralized effluent from a scrubber also contains some
ficient. The amount of interfacial area is determined by the sodium hypochlorite, which results from NaOH reacting with
type of packing in the column and by how well the packing the free chlorine generated during the combustion process.
is “wetted.” Operational testing has been used to develop The sodium hypochlorite is a strong oxidant and may require
virtually all of the packing performance parameters. treatment to meet effluent requirements. A reducing agent
such as SO2 or a solution of sodium bisulfate can be added
The upstream incineration process determines the concen-
to the recirculating stream in the scrubber to reduce the
tration of acid gas in the flue gas entering the column, the sodium hypochlorite:sodium chloride ratio. It is important to
acceptable outlet gas concentrations are set by environmental remember that CO2 in the flue gas also reacts with NaOH.
regulations, and the liquid purge stream concentration is The reaction rate, however, is relatively low until the pH of
either specified by the customer or determined by design to the recirculating water becomes greater than 8.0. Above that
achieve adequate acid gas removal. The diameter of a packed concentration, the CO2/NaOH reaction rate increases signif-
column is typically designed to give a superficial gas velocity icantly, greatly increasing the consumption of NaOH. Proper
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

of 7 to 10 ft/s (2 to 3 m/s). This range of velocities is high pH control of the recycle water is an economic necessity.
enough to create enough pressure drop (about 0.3 to 0.5 in., In some applications, both acid production and a very high
or 8 to 13 mm w.c. per foot of bed depth) to prevent poor level of acid gas removal are required. This can be accom-
flue gas flow distribution (channeling), yet low enough to plished with a two-stage system. Figure 21.19 illustrates this
avoid flooding. A recirculation rate of 7 to 10 gpm of liquid application. The first stage is an absorber that removes 80 to
per square foot of bed cross-sectional area is usually required 95% of the acid gas and produces an acid blowdown stream.
to ensure adequate “wetting” of the packing. The remaining The second stage is a scrubber, which achieves the desired
variables, which are determined by the designer, are column level of acid gas removal and, with a significant portion of
height, packing type, and the type and temperature of absorb- the acid gas removed in the absorber section, produces a
ing liquid. relatively low volume blowdown stream. Such a system is
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 665 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 665

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 21.19 Two-stage acid gas removal system.

advantageous when (1) there is a need to achieve very high Common materials of construction for a packed column
levels of acid gas removal, and/or (2) there is a use for the assembly are FRP for the vessel, and ceramic, FRP, plastics
acidic blowdown stream or, because of the amount of acid (such as polypropylene, PVC, CPVC, Teflon, Kynar, etc.), or
gas to neutralize, it is more economical to neutralize it exter- some combination of these for the packing, internal support
nally with a less expensive reagent such as lime (CaO) or hardware, and water circulation equipment. If the recirculat-
slaked lime (Ca(OH)2). ing liquid is sub-cooled to enhance recovery/removal, the heat
exchanger would have to be built of similar corrosion-resis-
In some applications, the flue gas quench can be incorpo-
tant materials.
rated in the packed column. For this condition, the lower
The advantage of using a packed column include:
section of the vessel and the packing material must be able
to withstand both high temperatures and acidic conditions. A 1. high removal efficiency capability
high-temperature and acid gas-resistant lining must be used 2. low pressure drop/energy cost (< 10 in. or 25 cm w.c. drop
for a 10-ft- or 3-m-deep packed section)
in the lower portion of the vessel, and the packing material
3. no moving parts in the column itself
and internal support hardware must be made of ceramic or
graphite. Figure 21.20 represents this application. Disadvantages are (1) poor small particulate removal capa-
bility, (2) dependent on recirculation pumps, caustic feed
While the packed column is primarily designed for acid
pumps, etc. to operate, and (3) momentary loss of quench can
gas removal, it will also remove some particulate matter.
cause “meltdown.”
However, the basic mechanism of entrained particle removal
is inertial separation from the gas stream, followed by entrap- 21.3.5.3.4.2 Venturi scrubber The venturi scrubber described
ment in the absorbing liquid. Because the gas velocities earlier as a particulate removal device can also function as an
through the packed section are far too low for effective inertial acid gas removal device. One of the important factors for effec-
separation and entrainment of particles with an aerodynamic tive mass transfer is intimate contact between the flue gas and
diameter of less than 10 microns, a packed column is not an the liquid used to absorb the acid gases; the venturi scrubber
effective particulate removal device. can provide relatively good gas-to-liquid contact. However,
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 666 Wednesday, September 3, 2003 8:12 AM

666 The John Zink Combustion Handbook

FIGURE 21.20 Combination quench/two-stage acid removal system.

because of other factors such as acid gas solubility and solu- outlet the cleanest gas is exposed to recycle water that has
bility rates (which are affected by gas film and liquid film the highest concentration of acid gas, which further reduces
resistance), and the difference in concentration of the pollutant the effectiveness of the scrubber. Despite these drawbacks, a
in the flue gas compared to the concentration of the dissolved properly designed venturi system, including the separation
pollutant in the absorbing liquid (i.e., low pollutant vapor pres- equipment, can still effectively remove much more than 90%
sure over the absorbing liquid), the venturi scrubber, although of the acid gas from a flue gas stream.
good, is not the best overall choice for acid gas removal. Advantages of the venturi as an acid gas removal device
HCl is a good example of a highly soluble acid gas. It include:
easily absorbs into the water, creating hydrochloric acid. If 1. relatively effective for highly soluble acid gases
necessary, the HCl can then be reacted with sodium hydroxide 2. low capital cost
(NaOH) added to the recycle water to form NaCl. 3. low plot space requirement
SO2 is an example of an acid gas that is only moderately Disadvantages of the venturi as an acid gas removal device
soluble in water. Therefore, it must be in contact with the consist of (1) only moderately effective for lower solubility
recycle water for a longer period of time for high removal acid gases, (2) high operating cost (pressure drop), (3) short
efficiency. Adding NaOH to the recycle water will improve gas-to-liquid contact period, and (4) co-current flow, which
the solubility by decreasing liquid film resistance, but a longer minimizes the difference in concentration of the pollutant in
period of contact is still needed to achieve high removal the flue gas compared to the concentration of the dissolved
efficiency. The concentration of NaOH must be as low as pollutant in the absorbing liquid.
possible to avoid excessive reaction of NaOH with CO2.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Because of design, the venturi utilizes a relatively short 21.3.5.4 NOx Control Methods
gas/liquid contact time. The short period of time reduces the NOx can be controlled during or after the combustion pro-
amount of absorption, especially for moderately soluble acid cess. The most effective method will be determined by the
gases. Also, because the flue gas and the absorbing fluid have NOx emission allowed and by the capital and operational
to travel in the same direction (co-current flow), at the venturi costs of various methods.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 667 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 667

FIGURE 21.21 Three-stage NOx reduction process.

Two primary mechanisms for the formation of NOx in com- If the waste streams do not contain large quantities of
bustion systems exist. They are (1) high-temperature dissocia- noncombustible materials such as air, water, or other inerts,
tion of molecular nitrogen (N2) and molecular oxygen (O2) and a modified combustion process, utilizing sub-stoichiometric
subsequent reaction of the independent radicals to NOx (ther- oxidizing conditions, can be used very effectively. Although
mal NOx), and (2) any oxidation reaction of bound nitrogen many different post-combustion treatment methods are avail-
in organic compounds (i.e., acrylonitrile, ammonia, nitroben- able — including some wet techniques that are very efficient
zene, hydrogen cyanide, amines, etc.) to NOx. Thermal NOx — the most commonly used methods chosen for typical ther-
occurs just downstream of burners in the flame zone, where mal oxidation systems are selective noncatalytic reduction
peak temperatures occur. NOx from organic-bound nitrogen is (SNCR) and selective catalytic reduction (SCR).
formed at any location during the oxidation reaction.
21.3.5.4.1 Combustion Process Modification
Limiting, or preventing, the high-temperature formation
from occurring during combustion reduces NOx production For wastes and/or fuels with a high concentration of bound
by the first mechanism. This is basically accomplished by nitrogen, or for wastes that contain NOx, the single most effec-
reducing peak temperatures with different burner design vari- tive practical process modification is a form of staged-air com-
ations. Thermal NOx control methods using burners are dis- bustion. The most common implementation of staged-air
cussed in previous chapters and will not be covered here. It combustion is accomplished in a three-stage combustion process
is important to understand that many of the burner design (Reed, 1981),3 as shown in Figure 21.21. As noted above, this
methods can also be used to minimize thermal NOx formed method may not be the most cost-effective for NOx reduction if
in the burners of T.O. systems. the waste stream(s) containing the NOx-producing compounds
are mostly air, water, or inerts because of the large amount of
The second mechanism is much more difficult to prevent auxiliary fuel it would take to operate such a unit. If steam is
because combustion of organics with bound nitrogen in the needed in the plant, it is possible that the high fuel use could be
presence of excess oxygen at any temperature will result in justified based on the amount of steam generated. NOx reduc-
NOx production. In addition, the NOx produced will usually tion by this method is achieved in the first stage by combusting
be much greater than the burner-generated thermal NOx. the nitrogen-bound compound or NOx itself in a high-tempera-
Also, some waste streams contain NOx that is not destroyed ture atmosphere that is deficient in oxygen (i.e., sub-stoichio-
in an excess oxygen environment. For these conditions, reduc- metric or reducing). This results in dissociation of the organic
ing burner-generated NOx is virtually inconsequential. Either compound and reaction of the released nitrogen atoms to N2.
the combustion process has to be changed to reduce the NOx The flue gas is cooled in the second stage, resulting in a lower
exiting the combustion section, or the flue gas must be treated peak temperature in the third stage. The lowered peak tempera-
after the NOx is formed (post-combustion). ture minimizes reformation/formation of thermal NOx.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

668 The John Zink Combustion Handbook

Fuel, waste(s), and less-than-stoichiometric combustion cially if the waste contains significant heating value. Nitrogen
air are introduced into the first stage (reduction furnace) to in the chemical waste can be in the organic or inorganic form.
produce a high-temperature (2000 to 2800°F, 1100 to Examples of the organic form are HCN amines, nitriles, and
1500°C) reducing atmosphere. The excess combustible nitroaromatics; examples of the inorganic form are NH3 and
material in the high-temperature reducing zone provides the NOx. A small process stream containing some quantity of
driving force for the reduction of the oxides of nitrogen and NOx can be treated to produce a cleaner stream that contains
conversion of bound nitrogen to N2 instead of NOx. Some comparably less NOx than the original waste. This method
inert (low or no oxygen content) such as recycle flue gas, is not suitable for NOx reduction in flue gas from large
steam, or water can also be introduced into the reducing combustion processes, such as utilities. It also is not appro-
zone to allow more consistent control of the operating tem- priate for waste streams containing large quantities of air,
perature so as not to exceed the limits of the refractory. The
water, or inerts. Either case would require an excessive
optimum level of oxygen deficiency in this stage depends
amount of auxiliary fuel to operate the system.
on the waste composition. Although the oxygen is usually
supplied by combustion air, the oxygen in NOx and in waste Three-stage systems using the process described above
gas streams will also be utilized. The residence time in this have been operating in a variety of industrial applications
stage is usually in the range of 0.5 to 1.0 s. The primary for more than 30 years. Destruction efficiencies of incoming
components of the gas leaving this stage are CO, H2, CO2, components of more than 99.99% are achieved. Carbon mon-
N2, and H2O. Most of the fuel-bound nitrogen is converted oxide is usually in the 50 to 100 ppmv (dry, corrected to 3%
to N2, with the remaining present primarily as very low O2) range. NOx in the flue gases can vary from 50 to 200
levels of HCN, NH3, and NO. ppmv, dry, corrected to 3% O2, depending on the composition
The hot flue gas then enters the second process stage (the of the waste stream being treated.
quench chamber) by passing through a venturi mixing sec-
A two-stage process modification can also be used if the
tion. An inert cooling medium, as described above, is
injected through multiple openings in the venturi throat to NOx level does not have to be as low. The same first-stage
quickly mix with the flue gas and decrease the temperature reduction furnace is still used. The difference starts at the
to 1300 to 1750°F (700 to 950°C). The temperature must reduction furnace outlet. Instead of adding an inert cooling
be high enough so that rapid ignition of the combustibles medium and then adding just enough combustion air to
occurs by simply adding air, but low enough to limit the oxidize combustibles and maintain 1.5% excess O2, a large
temperature achieved in the final oxidation stage to less than amount of air is introduced to cause the oxidation reactions
2000°F (1100°C). Although cooling is rapid, sufficient time to occur and to limit the oxidation section outlet tempera-
must be allowed in this section to ensure that the bulk gas ture to 1800°F (1000°C) or less. This process is shown in
temperature is uniform. Figure 21.22.
As the cooled flue gas exits the quench section, air in
The high excess O2 (up to 10%) causes more equilibrium
excess of stoichiometric is introduced at the entrance into
NOx to form in the oxidation section than would form in
the third stage (the oxidation zone), again using a venturi
the lower O2 three-stage system. Also, the method of control
mixing section. In this final stage, the carbon monoxide,
is slightly different in that the two-stage system requires the
hydrogen, and any remaining hydrocarbons produced in the
use of a sometimes maintenance-intensive combustibles
first stage are oxidized. The flue gas cooling step performed
analyzer to measure and control the combustibles level in
prior to introduction of the oxidation air controls the peak
oxidizing operating temperature, thereby limiting formation the reduction furnace. For the three-stage system, the com-
of thermal NOx. These process steps often result in levels bustibles level is controlled by measuring the differential
of NOx less than 150 ppmv (parts per million by volume) temperature between the quench section and the outlet of
at excess oxygen conditions of less than 1 to 2% (dry). The the oxidizing section, thus avoiding the instrument mainte-
residence time in this stage is usually in the range of 0.5 to nance. Another minor drawback is that recycle flue gas
1.0 s. cannot be used to control temperature in the reduction fur-
The flue gas can then be treated by any of the previously nace, which reduces heat recovery efficiency in the event
discussed methods or exhausted directly to the atmosphere that heat recovery is used. Despite of these less positive
for dispersion. items, the two-stage system is still far better than a single-
A wide variety of bound-nitrogen-containing gas and liquid stage, oxidizing-only combustion process for wastes con-
wastes can be incinerated using this treatment method, espe-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
taining bound nitrogen.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 669 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 669

FIGURE 21.22 Two-stage NOx reduction process.

FIGURE 21.23 Selective noncatalytic reduction system.

21.3.5.4.2 Selective Noncatalytic Reduction formly into the hot flue gas in the presence of excess O2. The
The selective noncatalytic reduction (SNCR) process for overall net reaction, which occurs by way of a complex free
NOx reduction is one in which a compound, added at the end radical chain reaction, is:
of the combustion process zone (post combustion), selec-
tively reacts with NOx without the aid of a catalyst. The most
commonly used process utilizes ammonia (NH3) as the addi- 1 3
NO + NH 3 + O2 Æ N 2 + H2O
tive. Figure 21.23 shows such a system. It must be mixed uni-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
4 2
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 670 Wednesday, September 3, 2003 8:12 AM

670 The John Zink Combustion Handbook

Concurrently with the NO reduction reaction, NH3 is oxi- lower temperatures (400 to 850°F or 200 to 450°C) than the
dized to form NO following the overall reaction: noncatalyzed SNCR reactions. The specific reactions cited
for SCR include:
5 3
NH 3 + O Æ NO + H 2 O
4 2 2 4 NO + 4NH 3 + O 2 Æ 4 N 2 + 6H 2 O

Although the reduction and oxidation reactions are in com- 6NO + 4 NH 3 Æ 5N 2 + 6H 2 O

petition and are both very temperature sensitive, there is a 2NO 2 + 4 NH 3 + O 2 Æ 3N 2 + 6H 2 O
relatively narrow but maintainable range of NH3-to-NO and
temperature in which the balance between reduction and 6NO 2 + 8NH 3 Æ 7N 2 + 12 H 2 O
oxidation is favorable and both reactions proceed toward
NO + NO 2 + 2 NH 3 Æ 2 N 2 + 3H 2 O
completion. The temperature range for best reaction is 1600°F
(900°C) to 1900°F (1000°C), which is the basic operating
temperature range of a 99.99% DRE system. The amount of The NH3 must still be uniformly mixed with the flue gas
ammonia used is about 0.5 lb (0.2 kg) per pound of NOx, as upstream of the catalyst section using a multi-point grid for
NO2, so the overall NH3 volume is quite low relative to the best results. Catalyst materials are primarily vanadium and
total flue gas volume. To achieve the mixing required, the wolfram derivatives, but others are being used and developed
NH3 must be combined with a carrier gas such as steam or such as alumina/titania, rhodium, etc. See Figure 21.24 for a
air for injection. The total volume of the carrier gas should common configuration.
be about 1.5% of the total flue gas. For optimum results, it SCR can achieve as high as 95% reduction of NOx because
should be distributed uniformly by use of an injection grid to it enables reaction with NO2 as well as NO. The high reduction

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
be rapidly mixed. Because most (90 to 95%) of the NOx is achieved at lower temperature operation, such as might be
formed is NO, effective reaction with NH3 can reduce overall found downstream of a heat recovery section. A heat recovery
NOx significantly. or cooling section protects the catalyst from over-temperature
For waste combustion systems for which process modifi- damage. The NH3 slip is normally less than in the SNCR method.
cation, by burner modification or by staged combustion, is The catalyst surface can be fouled by particulate or pos-
not a viable alternative to reduce NOx, NH3 injection can be sibly poisoned by some materials, but the catalyst material
reasonably effective. The process is relatively predictable, is relatively resistant to those as well as erosion damage.
although some operational trial runs are usually needed to Also, in the unlikely event that a high temperature upset
establish the optimum operating conditions. High operational condition occurs, the catalyst could be damaged. Catalysts
temperature is necessary, but that fits in with the high DRE are usually guaranteed for between two and four years, at
required by most waste combustion systems. SNCR can be which time they must be replaced to maintain maximum NOx
used in systems containing particulates with little or no pro- reduction. However, if multiple layers/sections of catalyst
cess degradation. The NH3 injection process can also be are used, instead of replacing all the catalyst, replacement
utilized at the end of a modified combustion process, such as of sections can often be alternated and still maintain adequate
a staged combustion system, to further reduce NOx. removal efficiency.
NOx reduction of 50 to 70% is possible, but if that is still
greater than the allowable emission, NH3 injection is not the
complete solution to the situation. Even with optimized 21.4 BLOWERS
operation, some small amount of NH3 slip (nonreacted NH3 Blowers (also referred to as fans) are used to overcome the
in the flue gas) will occur. Also, 1 to 2% of the reduced NO pressure drop required to move air/flue gas through a T.O.
will produce N2O. Urea is also used as a reactant. It works system consisting of multiple flue gas treating processes.
in a similar manner with similar reduction capabilities. Although blowers are also used to move waste gases, this
section will discuss only combustion air and flue gas blowers.
21.3.5.4.3 Selective Catalytic Reduction Location of the blower is a major factor in blower selection. If
The selective catalytic reduction (SCR) process is very simi- a blower is located at the front end of a system, “pushing” flow
lar to SNCR except the use of a catalyst increases the reduc- through, the process is described as forced draft and the blower
tion of NOx. The increased reduction is primarily the result will be handling clean air. The forced-draft blower will draw in
of increased reaction between NH3 and NOx (both NO and air at ambient pressure and raise it to the pressure needed to
NO2). In addition, the increased reactions occur at much push combustion products through the system.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 671 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 671

FIGURE 21.24 Common catalyst configuration.

If a blower is located at the back end of a system, “pulling” Centrifugal blowers draw gas (air) into the center of the
flow through, the process is described as induced draft. At that blower wheel at the axis of rotation where it enters the spaces
point, the gas handled will be combustion products, cooled between the paddle-like blades that impart radial-outward, as
sufficiently for the fan material of construction (by heat recov- well as angular, velocity to the gas, generating pressure. Centri-
ery or quenching). For this case, the blower draws in condi- fugal blowers can generate well over 100 in. (250 cm) w.c.
tioned flue gas at a pressure below ambient (vacuum) and raises pressure, more than sufficient for almost any multi-component
it to just greater than ambient to exhaust it to the atmosphere. system. As the gas is centrifuged to the periphery of the hous-
Occasionally, a blower will be located at the front end, ing, more gas is drawn into the blade space.
“pushing” through part of a system, while a blower at the There are three distinct types of centrifugal blowers. The
outlet “pulls” through the rest of the components. In that case, difference is basically the blade configuration used. The three
the process is called balanced draft. Obviously, the same basic blade types are straight (or radial), forward curved, and
criteria for blower selection applies as noted above. backward curved. Figure 21.25a illustrates the types of fan
When the blower location is determined, the final selection wheel designs available. Each has its advantages and disad-
is based on inlet/outlet composition and volumetric flow rate, vantages relative to the other two. Table 21.2 summarizes
inlet/outlet pressure requirements, inlet/outlet temperature, those relative characteristics.
and flow, pressure, and temperature ranges during operation. Proper selection of a centrifugal blower is not accom-
Two general classes of blowers exist for moving gas vol- plished by any single straightforward formula. Experience,
umes: axial and centrifugal. Axial blowers use propellers to usage, and careful evaluation of each application are neces-
move the gas parallel to the axis of rotation of the blower. sary to ensure proper fan selection. It is always best to work
Although axial blowers are inexpensive to buy and install and with blower manufacturers to get the most cost-effective
occupy little space, they are very limited in pressure capability recommendation.
(less than 20 in. or 51 cm w.c.) and very noisy at higher Another aspect of proper blower selection is utilizing the
pressures. For these reasons, axial-type blowers are seldom best-suited method of flow control for the blower to match
used for multi-component T.O. systems and are not examined the combustion system. Three primary methods are used:
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

further in this chapter. discharge damper control, inlet vane control, and fan speed
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 672 Wednesday, September 3, 2003 8:12 AM

672 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

FIGURE 21.25a Fan wheel designs. (Robinson Industries, Zelienople, PA. With permission.)

Copyright CRC Press

Thermal Oxidizers 673

TABLE 21.2 Relative Characteristics of Centrifugal Blowers control. The pressure-flow-horsepower curves are shown in
Forward Backward Figure 21.25b.
Radial Curved Curved
A discharge damper consists of one or more sliding or
Efficiency Medium Medium High
Tip speed High Medium Medium
pivoting blades (such as a butterfly damper) that reduce flow
Sizea Small Medium Large area in a duct. Closing the damper increases resistance to flow
Initial costb Small Medium Large and reduces flow. However, when flow is reduced, the oper-
HP curve Medium rise Medium rise Power limiting
Accept corrosion coating Excellent Fair to poor Good (thin coat) ating condition of the blower (and the point on the fan curve)
Abrasion resistance Good Medium Medium is shifted to lower flow but also to a corresponding higher
Sticky material handling Good Poor Medium
High temperature capability Excellent Good Good
pressure. This pressure is greater than needed, so the pressure
Running clearance Liberal Medium Minimum req’d drop taken across the damper wastes it. The blower perfor-
Operation without diffuser Not as efficient Must use Good efficiency mance curve is not changed by using a discharge damper. The
Noise level High Medium Low
Stability/non-surge rangec Medium Poor Medium horsepower usage ratio is reduced, but by less than the flow
20%–100% 40%–100% 20%–100% ratio change (see Figure 21.25c).
(a) Size is based on fans at the same speed, volume, and pressure.
(b) Cost is based on fans at the same speed, volume, and pressure. Inlet vane control is accomplished with a special damper
(c) More a function of operating point along a curve than fan type. that consists of multiple adjustable vanes oriented radially
from the centerline of the damper, which is located at the

FIGURE 21.25b Radial blade operating curve for 1780 RPM, 70°F, and 0.075 lb/ft3 density. (Robinson Industries,
Zelienople, PA. With permission.)

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

674 The John Zink Combustion Handbook

FIGURE 21.25c Forward tip blade operating curve for 1780 RPM, 70°F, and 0.075 lb/ft3 density. (Robinson Industries,
Zelienople, PA. With permission.)

blower inlet. As the vanes are rotated (adjusted) closed, reduc- able speed control is that the volume of air flowing is
ing the amount of air allowed into the blower inlet, the enter- proportional to the blower speed, the pressure developed is
ing air is given an angular velocity vector (spin) in the proportional to the square of the speed, and the horsepower
direction of rotation of the blower wheel. This spin modifies required is proportional to the cube of the speed. Thus,
the basic characteristics of pressure output and power input,
unlike inlet vane control, for which the blower curves start
resulting in new and reduced pressure and horsepower char-
at the same low flow point for each vane setting and change
acteristics. As the vanes are further closed, the flow of air is
the end point for the high flow, variable speed control
further reduced but the spin is increased. This further reduces
the pressure and horsepower characteristics. Effectively, the results in a completely separate performance curve for each
inlet vanes change the blower performance curve so that the blower speed. The net effect is that the horsepower reduc-
horsepower reduction ratio is actually greater than the flow tion ratio is even greater than with inlet vane control (see
ratio change. A specially designed inlet box with a parallel Figure 21.25e).
blade damper directing the flow to one side of the box, effec-
As with blowers, the method of flow control is not neces-
tively providing rotation at the blower inlet, provides very
sarily based simply on cost or efficiency. Often, a combination
similar results (see Figure 21.25d).
Fan speed control for most combustion system blowers of two methods is needed, such as discharge dampers con-
is accomplished utilizing variable-speed drivers (motors or trolling flow to different parts of a combustion system while
turbines). Ideally, if a blower is controlled by varying its the pressure upstream of the dampers is maintained at a con-
speed, there is little wasted energy. The theory behind vari- stant point by use of an inlet vane damper.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Thermal Oxidizers 675

FIGURE 21.25d Backward curved blade operating curve for 1780 RPM, 70°F, and 0.075 lb/ft3 density. (Robinson
Industries, Zelienople, PA. With permission.)

21.5 CONTROL SYSTEMS AND 1. flame sensor (flame scanner, flame rod, or other method):
to ensure that either a stable flame is maintained (normal
INSTRUMENTATION operation) or that no flame is present (during purge)
All thermal oxidizer systems require some sort of control, if 2. fuel supply pressure switches: to ensure that the fuel sup-
only a flame failure switch to ensure waste and fuel shutoff ply is within the design range
when needed. Controls can be classified into flame safeguard 3. air supply flow or pressure switches: to be certain an
adequate supply of combustion air is available
and process control functions.
4. automatic shut-off block valves: for the fuel flow and
Flame safeguard requirements evolved from insurance and waste flow
general safety regulations for fuel-fired burners in general. Process controls are provided to keep the system operating
This part of a control system is usually designed to satisfy within boundaries to meet legal emission requirements and to
detailed rules published by the National Fire Protection Asso- protect the equipment from operational damage. Some form
ciation (NFPA), Industrial Risk Insurers (IRI), and Factory of automatic temperature control is normally used to adjust
Mutual (FM). They are meant to ensure that the fuel and fuel and quench (air, water, or steam) flow to the unit so that
waste flows to an incinerator are stopped if the flame is lost the waste is burned properly without exceeding the refractory
and that the furnace is fully purged of combustibles prior to lining temperature limits. In many cases, the outlet O2 con-
ignition, so that the potential for explosions is eliminated. centration is monitored and used to control the combustion
Generally, the major components of these systems include: air flow. Downstream flue gas treating equipment will be
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

676 The John Zink Combustion Handbook

FIGURE 21.25e Outlet damper flow control. (Robinson Industries, Zelienople, PA. With permission.)

FIGURE 21.25f Radial inlet damper/inlet box damper flow control. (Robinson Industries, Zelienople, PA. With permission.)
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Thermal Oxidizers 677

FIGURE 21.25g Blower speed control. (Robinson Industries, Zelienople, PA. With permission.)

built for much lower temperature than the T.O., so flue gas treatment sections. The choice of these components for any
cooling process parameters (boiler water level, water spray total system depends primarily on the nature of the waste
flow rate, etc.) and the resulting flue gas temperature are stream to be destroyed and the emission requirements for the
always monitored to avoid expensive thermal damage. Auto- flue gas ultimately exhausted to the atmosphere. This section
matic steps (system shutdown, hot gas diversion, etc.) are in is a discussion of suggested system configurations designed
place to deal with any failure. For scrubbers, the flow rate and for seven of the most common types of waste streams. The
pH of the circulating liquids are controlled automatically to types are:
ensure proper removal of acid gases. Less obvious are controls
1. non-acid gas endothermic waste gas/waste liquid
applied to deal with specific variations in incinerator waste 2. non-acid gas exothermic waste gas/waste liquid
feed streams. In some applications, the waste flow and com- 3. sulfur-bearing acid gas (includes pulp and paper)
position are expected to change abruptly. When waste flow 4. chlorine-bearing acid gas
and composition are expected to change abruptly, the control 5. down-fired “salts” (i.e., solids that melt)
method required to maintain effective system performance 6. vertical/horizontal combustible solids
(or even flame stability) could become very complex. The 7. NOx minimization or reduction
speed of analyzer or thermocouple response often plays a
major part in control system design. In these cases, control 21.6.1 Non-acid Gas Endothermic Waste
system (and burner) design experience is absolutely critical. Gas/Waste Liquid System
Endothermic waste gas and waste liquid are often handled
similarly. Examples are waste gas containing primarily inert
21.6 SYSTEM CONFIGURATIONS materials such as N2 or CO2 with some hydrocarbon contam-
The previous sections of this chapter covered the basics of ination, and waste liquid that is water with some small
the combustion process, the individual components that make amount of organic material. For these cases, a system would
up the combustion section, and the post-combustion flue gas be required to introduce significant amounts of fuel and air to
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

678 The John Zink Combustion Handbook

bring the waste components to the temperature needed to oxi- tions as both the combustion chamber, for residence time, and
dize them. The simplest system option is a vertical thermal the stack, to disperse the flue gas to the atmosphere. The
oxidizer with either a natural- or forced-draft burner. The flue operating temperature is maintained at the minimum possible
gas exhausts directly to the atmosphere. to achieve DRE so the least amount of fuel will be used. For
Often, it is economically desirable to recover heat from the discussion purposes, a waste gas stream of 100,000 lb/hr
flue gas of such systems by making steam from a waste liquid (45,000 kg/hr) of N2 at 80°F (27°C) with no hydrocarbon
or waste gas T.O. system or by preheating the waste gas content is considered. The operating temperature chosen is
stream to reduce fuel usage. Preheating a waste liquid is not 1500°F (820°C). To heat the waste to the chosen operating
usually cost-effective. temperature, the calculated fuel requirement, ignoring heat
Also, if the waste gas is air contaminated with a small losses, is slightly greater than 65 ¥ 106 Btu/hr (19 MW).
amount of hydrocarbon and is well below the lower flamma-
bility (or explosive) limit (LEL), standard thermal oxidation Recovering some of the fuel heating value put into a T.O.
is still used, but other methods such as catalytic oxidation system can be more cost-effective when compared to the sim-
and regenerative oxidation can be more cost effective. ple system, especially if a large amount of fuel is needed. If
In all cases, fuel firing capability is required to maintain a T.O. is to be located in a plant that needs steam for process
the correct outlet temperature except in the case of the cata- use, a system configuration represented by Figure 21.27 is
lytic system, where the fuel firing is often used only to heat better to use than the simple vertical T.O. system. The refrac-
up the unit prior to introducing the contaminated air. tory-lined T.O. itself is horizontal so flue gas can exit directly
Figure 21.26 shows a simple thermal oxidizer. Fuel and into a boiler (flue gas cooler), which can be either a firetube
combustion air enters the burner while the inert waste gas or or a watertube configuration. As noted previously in this chap-
inert waste liquid is effectively introduced past the fuel/air ter, firetube boilers are normally recommended for smaller
combustion zone. Because no heat recovery or flue gas treat- units with lower pressure steam needs, while watertube boilers
ment is needed, the most cost-effective design is a refractory- are the preferred type for larger systems with higher steaming
lined, vertical up-flow unit. The refractory-lined vessel func- rates or higher steam pressures. For the same N2 waste case

Castable Refractory
Lining
Stack

Brick Lining W/Castable

Refractory Back-Up Lining

Thermal Oxidizer

Waste Liquid Gun

Waste Gas Stream

Castable Refactory
Floor
Fuel

Controls & Instrumentation Package

Air

FIGURE 21.26 Simple thermal oxidizer.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Thermal Oxidizers 679

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 21.27 Thermal oxidizer system generating steam.

noted above, a boiler could produce as much as 44,500 lb/hr of fuel fired, the savings in fuel cost would be about
(20,000 kg/hr) of 150 psig (10 barg) steam by cooling the flue $1,000,000 per year compared to the simple vertical T.O. The
gas from 1500 to 500°F (820 to 260°C) (a recovery efficiency amount of heat transferred from the flue gas would drop the
of about 70%). By utilizing an economizer to further cool the temperature from 1500°F to about 900°F (820°C to about
flue gas to 350°F (180°C), about 50,000 lb/hr (23,000 kg/hr) 480°C), which corresponds to a heat recovery of about 42%.
of steam could be produced (a recovery efficiency of more Because less fuel is required, less combustion air
than 80%). Because the flue gas temperature is 500°F (260°C) (operating cost) is required, and the amount of flue gas is
or less, an economical, unlined carbon steel vent stack is used. reduced. This reduces the size of the T.O. (capital cost). The
If no steam is needed, the best system is one that reduces vent stack would have to have refractory for this example
the fuel required by transferring heat from the outlet hot flue unless it was fabricated using a heat-resistant alloy that could
gas from the T.O. to the waste gas. Figure 21.28 illustrates withstand the greater than 900°F (480°C) flue gas temperature.
such a system. The process is usually referred to as a recu- Overall, the value of the steam generated and the reduction
perative process because energy is being removed from the in fuel used during the first year will typically pay for the
flue gas and put back into the system by heating the waste additional equipment needed to make steam or preheat the
gas before it enters the T.O. Usually, the refractory-lined T.O. waste gas.
is horizontal for the same reasons as the boiler system. Some When the waste gas is contaminated air, the configuration
smaller systems are built utilizing a vertical up-flow T.O. with of the simple system and the boiler and preheat systems is
the preheat exchanger mounted on the top end of the T.O. nearly the same as when the waste gas is a non-oxygen-
and the stack on top of the exchanger. For the same waste bearing inert. The only real difference in the simple system
case described above (100,000 lb/hr or 45,000 kg/hr of N2), is that it is forced draft. Both the boiler and recuperative
preheating from 80 to 900°F (27 to 480°C) by transferring systems are already forced (or induced) draft to overcome the
heat from the flue gas would reduce the fuel needed to main- pressure drop of the heat transfer component. Because the
tain 1500°F (820°C) in the T.O. from more than 65 ¥ 106 waste contains the O2 needed to burn the fuel, no additional
Btu/hr (19 MW) to less than 29 ¥ 106 Btu/hr (8.5 MW). That combustion air is needed for any of the systems, so the fuel
is a fuel reduction of about 56%. At only $3.50 per 106 Btu/hr requirement for these cases is reduced.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 680 Wednesday, September 3, 2003 8:12 AM

680 The John Zink Combustion Handbook

FIGURE 21.28 Heat recovery thermal oxidation system.

For example, 100,000 lb/hr (45,000 kg/hr) of 80°F (27°C) attractive when comparing the sum of the capital and oper-
contaminated air with no hydrocarbon content can be heated ating costs. Those are catalytic oxidation with recuperative
to 1500°F (820°C) with only 39 ¥ 106 Btu/hr (11 MW) of heat recovery and thermal oxidation with regenerative heat
fuel. Cooling the flue gas to 350°F (180°C) can generate about recovery. The catalytic process (Figure 21.4), is usually used
30,000 lb/hr (14,000 kg/hr) of 150 psig (10 barg) steam. For for lower volumes of particulate-free contaminated air than
the recuperative case, by preheating the contaminated air from the regenerative process. For the catalytic process, the max-
80 to 900°F (27 to 480°C), the fuel required would be a little imum heating value of the contaminated air must be limited,
more than 17 ¥ 106 Btu/hr (5 MW). The amount of heat not just to remain below LEL, but also to prevent overheating
transferred from the flue gas would drop its temperature from the catalyst, which would be damaged rapidly at greater than
1500°F (820°C) to about 735°F (391°C), which corresponds 1300°F (700°C). The unlined but externally insulated cham-
to a heat recovery of about 54%. ber upstream of the catalyst receives and evenly distributes
However, if the hydrocarbon concentration in the air is too the approximately 600°F (320°C) preheated contaminated air
great, care must be taken to preheat the waste air to a safe to the catalyst section. Downstream of the catalyst, the cham-
margin below the lowest accepted auto-ignition point. If some ber is sometimes lined because the temperature after reaction
of the hydrocarbon oxidizes in the preheat exchanger, severe may be 1000 to 1300°F (540 to 700°C) before passing into
over-temperature damage can occur if the exchanger metal- the heat exchanger.
lurgy is not capable of the higher temperature. For a simple For the regenerative process, a short-term over-temperature
example, at a concentration of 1% methane (CH4) in air, the problem does not seem to be as great because the refractory
calculated temperature rise is greater than 450°F (230°C). If in the T.O. portion and the ceramic media in the heat recovery
preheat to 900°F (480°C) is intended, an additional tempera- chambers can withstand operating temperatures greater than
ture rise of 450°F (230°C) yields 1350°F (730°C). If the T.O. 2000°F (1100°C) to produce higher DRE. However, rapid
outlet temperature is 1500°F (820°C), the average tube tem- cycling of the recovery chambers will result in excessive wear
perature at the exchanger inlet is about 1450°F (790°C), 250°F and tear on the valves and ceramic media. Furthermore, the
(120°C) greater than the design. In general, the maximum temperature downstream of the chambers in the carbon steel
heating value of contaminated air for a T.O. system using ductwork can exceed the maximum allowable.
recuperative heat recovery is about 20 Btu/ft3. In general, the maximum heating value of contaminated
When the hydrocarbon content of the contaminated air is air for a catalytic system with recuperative heat recovery is
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

low enough, two other systems become more economically about 13 Btu/ft3. For a regenerative system, which has even
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 681 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 681

FIGURE 21.29 Bypass recuperative system.

greater heat recovery efficiency, the maximum heat content normal operation. If the flue gas from these wastes does not
is about 7 Btu/ft3. For greater organic content, more air can require treatment before discharge, such systems generally
be added to dilute the overall heat content but that increases consist of a simple natural- or forced-draft burner mounted on
the volumetric flow rate, which increases the capital cost of a vertical refractory-lined T.O. similar to the non-acid endo-
either type system. Also, some of the flue gas can be thermic system (Figure 21.26). If heat recovery in the form of
bypassed in a recuperative system so that less heat is avail- steam or hot oil is desirable, a boiler or oil heater can be used
able for heat transfer. Such a bypass system is shown by downstream of a horizontal refractory-lined T.O. configured
Figure 21.29. similar to the system shown in Figure 21.27. In any case, the
Also, depending on the DRE required, strong consideration maximum temperature in the T.O. section will have to be held
should be given to using a fired boiler that consists of a radiant below a set maximum by using a direct cooling medium.
and a convective section. It is essentially a “cold-wall” T.O. High-heating-value hydrocarbon wastes, whether gaseous
with a boiler. Heat recovery is greater and the NOx emission or liquid, have characteristics very similar to those of fuels.
will be reduced. They are generally as easy to burn and are typically injected
At times, the contaminated air may contain enough hydro- directly into the burner. The burner can be of low, medium,
carbon to be greater than 25% of LEL. In such cases, use of or high intensity, depending on the waste being burned, the
explosion safeguards such as flame arrestors, detonation destruction efficiency desired, and what, if any, post-combus-
arrestors, or liquid seals is recommended. tion treatment is utilized. Liquid wastes are atomized with
medium-pressure steam or compressed air and are nearly
always fired through the throat of the burner. Waste gases
21.6.2 Non-acid Gas Exothermic Waste with higher heating values can be fired through a single gun
Gas/Waste Liquid System or through multiple tips. Fuel is used only to heat up the
Incineration systems for the disposal of high-heating-value system in most cases.
hydrocarbon wastes (greater than 200 Btu/ft3 for gases and The high heating value of these wastes produces high flame
greater than 5000 Btu/lb for liquids) that do not have a sub- temperatures (≥ 2800°F or 1500°C). To achieve the desired
stantial halogen, ash, nitrogen, or sulfur content are relatively destruction efficiency, the combustion chambers for these
simple. The combustion of these wastes provides more than wastes are generally maintained at relatively high temperatures
enough heat to maintain the furnace above the desired oper- (≥ 1800°F or 1000°C) and have residence times of greater than
ating temperature so that no auxiliary fuel is required during or equal to 1 second. Destruction efficiency of organic com-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

682 The John Zink Combustion Handbook

pounds is 99.99% or greater when using 20 to 25% excess must be removed before the flue gas is exhausted into the
combustion air. To avoid damage to the unit, excessive tem- atmosphere. The systems that complete the destruction and
perature is controlled by cooling the products of combustion removal are acid gas systems.
with additional air, water, recycled flue gas, or steam injected
into the T.O. 21.6.3.1 Sulfur-Bearing Hydrocarbon Systems
Additionally, the high operating temperatures result in During the early stages of the oil and gas refining process,
excessive NOx formation. Therefore, being able to limit the sulfur compounds, primarily in the form of hydrogen sulfide
maximum local temperature at any point in the T.O. must also (H2S), are removed. The H2S is then converted to elemental
be a consideration. In some cases, the cooling medium can sulfur by the Claus process. The final “clean” by-product of
be injected into multiple locations, both in the burner portion that process is known as tail gas. Tail gas is mostly N2, CO2,
and in the T.O. to limit thermal NOx formation. Low NOx and water vapor. However, although the efficiency of the
burner techniques can also apply with these wastes. Claus process has improved over the years, some of sulfur
The refractory used in the T.O. partially depends on the compounds still remain. Because sulfur compounds have a
waste type and the operating temperature. For gases, a very strong odor, often likened to “rotten eggs,” even a small
castable refractory of sufficient thickness to protect the T.O. amount is detectable by the human nose. By destroying a
shell is adequate. A liquid-burning incinerator, however, can high percentage of the sulfur compounds, the concentration
be lined with a firebrick backed with an insulating castable in the flue gas is reduced to less than the detectable limit. A
to withstand potential impingement of flame or liquid. The majority of the thermal oxidation systems supplied for sulfur-
difficult-to-brick areas are usually still castable-lined but with bearing waste streams have been simple units for the treat-
a higher density material. Also, because excess heat is avail- ment of tail gas. The sulfur compounds in tail gas include
able, less concern is shown for minimizing heat loss through H2S, sulfur dioxide (SO2), carbonyl sulfide (COS), carbon
the T.O. Maintaining a low stack temperature for personnel disulfide (CS2), and elemental sulfur vapors. A small amount

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
protection may require thicker refractory or a personnel proof CO and hydrocarbon is also usually present. Thermal oxi-
tection shield. dizers are very effective for odor control of wastes containing
mercaptans and other odoriferous sulfur compounds.
Because the exothermic T.O. systems are usually forced-
Sulfur plant tail gas incinerators are generally designed to
draft, stack height is not required for them to provide draft
operate with natural draft. A stack tall enough to create the
in the simple system, and the exit gas velocity is often greater
necessary amount of draft is used to provide air flow to the
than 50 ft/sec. In addition, the high temperature of the flue
burner. The burners are designed for pressure drops from
gas exiting the simple unit has sufficient buoyancy to carry
0.25 in. (6.4 mm) w.c. to more than 1.0 in. (2.5 cm) w.c. Waste
the flue gas to high altitude. Consequently, a tall stack is not
heat recovery boilers are also utilized occasionally. For those
necessary for flue gas dispersion, and stack height becomes
cases, medium pressure drop, forced-draft burners are used.
a matter of customer choice based on surrounding structures
and/or dispersion modeling. The simple incineration process is nearly the same as the
simple endothermic configuration described earlier. Castable
As with the endothermic wastes, heat recovery can be
type refractory is usually adequate for the temperature and
performed with either a boiler or a hot oil heater. With heat
environment expected. As before, the lower portion of the
recovery, the flue gas is typically vented at less than 600°F
stack is the residence time section. One difference from the
(320°C) through an unlined steel stack. Although the buoy-
simple endothermic system is that an internal coating may
ancy is not as great because of lower flue gas temperature,
be applied to the vessel shell before the refractory is installed
exhaust dispersion is rarely an issue. Depending on the flue
to protect the steel shell from weak sulfuric acid attack.
gas temperature, external insulation of the stack may be desir-
Another alternative is to add an external rainshield around
able to prevent water condensation and possible corrosion on
the vessel to keep the shell temperature above the dew point
the inside of the stack.
to prevent condensation.
Because of the low auto-ignition temperatures (generally
21.6.3 Acid Gas Systems less than 700°F or 370°C), destroying sulfur-bearing com-
Acid gas systems are so called because the wastes treated pounds is very easy. Traditionally, a T.O. temperature of
contain components that, as a result of the oxidation reaction 1000 to 1200°F (540 to 650°C) and a residence time of 0.6 to
to destroy the waste, produce acid compounds such as 1.0 seconds was used for all sulfur plant tail gases. In the
SO2/SO3 and HCl. Emission of these acid compounds is lim- 1980s, the need to reduce the H2S content to less than 5 ppmv
ited by national and local air quality permitting agencies and required an increase in operating temperature to 1400°F
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 683 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 683

(760°C) in some cases. The higher temperature ensures a Because of the low temperatures and lower oxygen content,
higher degree of destruction of the sulfur compounds. It also NOx formation is fairly limited and is not normally a con-
ensures that the fuel use is greater. sideration. However, when operating at a T.O. temperature of
During the oxidation of sulfur compounds, a small only 1200°F (650°C), any carbon monoxide coming in with
amount (typically 1 to 5%) of the sulfur dioxide is further the tail gas will be only partially (40% or less) destroyed. At
oxidized to SO3. The extent of conversion depends on a 1400°F (760°C), substantially more is oxidized (more than
number of conditions such as the temperature profile the 80%), but CO destruction is still not high. At lower temper-
flue gas experiences, the amount of SO2 in the flue gas, the atures, increased residence time (large T.O.) can provide
potential catalytic action of alumina refractory material, etc. greater destruction, but at greater cost.
Once formed, SO3 reacts at a temperature between 450 and Pulp and paper plants also generate waste gas containing
650°F (230 and 340°C), with water in the gas stream, to sulfur compounds. Those wastes are handled in the same
form sulfuric acid. Sulfuric acid can cause several problems. manner. One difference in the equipment is in the material of
First, it raises the dew point substantially, sometimes to construction. While the refineries find carbon steel an accept-
temperatures above 400°F (200°C). Sulfuric acid condensa- able material, the pulp and paper industry often prefers stain-
tion can lead to rapid corrosion of carbon steel surfaces. less steel for many of their installations.
Second, when quenched rapidly, sulfuric acid can form an
extremely fine aerosol that is difficult to remove in a packed 21.6.3.2 Halogenated Hydrocarbon Systems
bed scrubber. Third, the fine aerosol can create a visible While sulfur is probably the most common acid gas waste
plume that has a bright blue-white hue. constituent found in petroleum refining, chlorine (Cl2) is
probably the most common halogen encountered in petro-
The flue gas treatment depends on the sulfur content of the chemical plant wastes. Because it is commonly found, this
waste and the regulatory requirements. Typically, tail gas chapter section discusses chlorinated hydrocarbon treatment
incinerators have such a small SO2/SO3 emission that no flue only. Of the other halogens, fluorine is also relatively com-
gas treatment is used. The maximum allowable ground level mon. It converts even more readily to hydrogen fluoride (HF),
concentration (GLC) of SO2 can be achieved by utilizing a while bromine and iodine have much lower reaction rates
tall stack. Stack heights of 100 to 300 ft (30 to 90 m) are (conversion to HBr and HI) and must be handled differently.
common. However, if the requirement is to meet a maximum Chlorine is added to many hydrocarbon feedstocks to
stack emission instead of a GLC, something else must be done. formulate numerous useful compounds. Wastes containing
For applications in which emission limits are very stringent chlorinated hydrocarbons can be in the form of gas or
or that produce higher SO2 concentration, a scrubber is used liquid. Waste gases can be air based, inert based, or organic
to remove a good portion of the SO2 before the flue gas is based; waste liquids can be organic or water based. In each
dispersed into the atmosphere. To accomplish this, the flue of these cases, the waste can contain chlorinated hydro-
gas from the T.O. must first be cooled by a boiler, then carbons such as vinyl chloride, methyl chloride, chloro-
quenched to its adiabatic saturation temperature by one of the benzene, polychlorinated biphenyls (PCB), etc. A typical
quenching methods described earlier, and then passed through incineration system for a relatively high-heating-value
a packed bed scrubber. If steam were not needed, the flue gas chlorinated waste consists of a horizontal T.O. to destroy the
would be quenched directly to saturation. SO2 is less soluble wastes, a firetube boiler to cool the flue gas for further treat-
than HCl in water, and therefore more difficult to remove. ment (and for heat recovery), a direct spray contact quench
Adding sodium hydroxide (NaOH), often referred to as “caus- to cool the flue gas to saturation, a packed column to remove
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

tic,” is normally used to enhance the removal efficiency and HCl, and a stack to vent the cleaned flue gas to the atmos-
to convert the SO2 sulfates and sulfites for further treatment. phere. Figure 21.30 shows this process. The boiler may not
The pH of the recirculated scrubbing solution must be no be used if no steam is needed. In that case, the T.O. could
greater than 8.0 or a significant amount of caustic will react be vertical up-flow with a 180° turn into the quench section.
with the CO2 in the flue gas, wasting caustic. Also, because Figure 21.31 would apply.
SO2 is more difficult to remove, taller packed beds are Depending on the composition/heating value, wastes can
required when compared to similar applications (same inlet be fired through the burner as a fuel or added peripherally
concentration and removal efficiency) scrubbing HCl. The into the T.O. The type of chlorinated hydrocarbon waste and
presence of a significant quantity of SO3 in the flue gases may the destruction efficiency required dictates the incinerator
also necessitate the use of a mist eliminator downstream of operating temperature, typically between 1600 and 2200°F
the scrubber. (870 and 1200°C). Residence time varies from 1.0 to
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 684 Wednesday, September 3, 2003 8:12 AM

684 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
FIGURE 21.30 Horizontal thermal oxidizer with firetube boiler and HCl removal system.

2.0 seconds. Generally, air-based waste streams have lower bustion products are scrubbed with sodium hydroxide. The
organic content and require lower destruction efficiency, presence of hypochlorite in the blowdown stream from the
which can be accomplished beginning at 1600°F (870°C). On caustic scrubber may require special treatment, as noted in
the other hand, wastes containing PCB require a destruction the section discussing packed column scrubbers. Cl2 forma-
efficiency of 99.9999%. This level of destruction efficiency tion can be minimized by shifting the reaction equilibrium
is usually accomplished at a temperature of 2200°F (1200°C) away from Cl2 formation and toward HCl formation. As can
with a residence time of up to 2.0 seconds. The majority of be seen from the reaction equilibrium vs. operating temper-
waste streams containing chlorinated hydrocarbons require a ature curve (Figure 21.32), the equilibrium can be shifted by
destruction efficiency of 99.99%, which is usually obtained increasing the T.O. temperature, by increasing the water vapor
at temperatures of 1800 to 2000°F (1000 to 1100°C) and at concentration, or by decreasing the excess oxygen concentra-
residence times of 1.0 to 1.5 seconds. tion. In practice, some excess O2 is required to maintain
The T.O. refractory can be as simple as a ceramic fiber highly efficient oxidation reactions so only changes to the
blanket for air-based fume streams with low operating temp- temperature and water concentration are used.
erature requirements, or as elaborate as a high alumina fire- A flue gas cooler/waste heat boiler is often used in chlori-
brick with an insulating firebrick backup for waste liquid nated hydrocarbon systems. As noted previously (Section
streams with high operating temperature requirements. Plastic 21.3.5.1.1), a firetube boiler is preferred to a watertube boiler
refractory is used in places where brick is not easily installed. because all of the heat transfer surfaces in a firetube boiler can
Castable refractory with a calcium oxide binder is generally be maintained at temperatures above the dew point (< 250°F or
avoided because the HCl in the flue gases can react with the < 120°C). Of course, to do this, the steam pressure in the
binder and cause the refractory to degrade. boiler must be high enough to have the saturation temperature
The oxidation process produces HCl and some free chlo- sufficiently above the dew point of flue gas. The material of
rine gas, along with the normal combustion products. It is construction for the boiler tubes is carbon steel, as it is for
important to limit the quantity of Cl2 produced because it is the rest of the boiler.
very corrosive at higher temperatures and also because it The flue gas must be quenched to saturation before the
forms sodium hypochlorite, a strong oxidant, when the com- absorption/scrubbing step to remove HCl. When quenching
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 685 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 685

FIGURE 21.31 Vertical thermal oxidizer with 180° turn quench section.

follows a boiler, it can be carried out in the bottom section tities of HCl, it is not cost-effective to use large quantities of
of the packed column. A two-stage column should be used water or caustic to remove the HCl. In such systems, a two-
for this case if HCl is being removed using caustic. With no stage removal system is used. The first stage is an absorber
waste heat boiler, the flue gas is quenched in a direct contact from which the majority of the acid is discharged as a con-
quench. Quench systems are often fabricated using reinforced centrated solution of HCl (up to about 20% by weight). The
plastic (FRP), with protection against hot flue gases provided
HCl and Cl2 remaining in the flue gas after it passes through
by dry-laid brick lining. Some users prefer metal fabrication
the absorber are removed in the downstream scrubber, where
such as Hastelloy.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

The quantity of HCl in the flue gas dictates whether a caustic is used as the scrubbing reagent. HCl removal effi-
single-stage absorber or a two-stage absorber/scrubber is ciency of as high as 99.9% can be achieved, although most
used. For smaller systems where a relatively small quantity applications require only about a 99% removal.
of HCl is present in the flue gas, a single-stage water absorber While oxidation of chlorinated hydrocarbons produces HCl
or a caustic scrubber can be used, with sufficient blowdown with a little free Cl2, fluorinated hydrocarbons produce even
to maintain the concentration of HCl in the recycle water low
less and are easier to remove than HCl. However, burning
enough to allow it to absorb the incoming HCl. If caustic
brominated compounds results in as little as half of the bro-
scrubbing is used, the blowdown should maintain the dis-
solved and suspended solids (mostly NaCl) at no more than mine being converted to HBr. Iodine compounds are worse.
5%, but preferably at 3%. Again, the pH should be kept below Free bromine and iodine are much more difficult to remove
8.0. The single-stage absorber produces a dilute acid stream than Cl2. Brominated and iodine compounds must be treated
and the single-stage scrubber produces a stream containing completely differently. Sub-stoichiometric oxidation is nec-
NaCl. For larger systems and those that contain larger quan- essary to drive the reaction toward HBr and HI.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 686 Wednesday, September 3, 2003 8:12 AM

686 The John Zink Combustion Handbook

are present, the system must be designed to handle molten

material. Some examples of organic solids containing non-
organic material are sawdust and rice hulls. Both can be
conveyed into a combustion system with air as the carrier,
and both have good heating value. However, both contain a
large amount of silica. When both are burned, the silica forms
very small molten particles of silicon dioxide in the flame
zone. When cooled, these fine particles of “fume silica” have

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
a tendency to collect inside equipment and are difficult to
remove from the flue gas.
If the solid material is an alkali metal inorganic salt such
as NaCl, Na2SO4, CaCl, or KCl, in water, for example, the
total system design must be based on molten material in the
flue gas. Because virtually all cases involve molten material,
the inorganic salt case is reviewed here.
Most of the wastes in this category are salt-contaminated
liquids. The waste streams are both organic and water based.
Because the waste is a hydrocarbon liquid or because the
water-based waste liquid contains a hydrocarbon, thermal
FIGURE 21.32 Chlorine reaction equilibrium vs. oper- oxidation is often the best method for disposal.
ating temperature. Unfortunately, because most salt-containing wastes are
liquid, they usually have to be oxidized at higher temper-
atures to rapidly achieve high destruction efficiency. This
21.6.4 Salts/Solids Systems presents a problem with the refractory in the T.O. Alkali
material attacks the binder in refractory materials. Higher
Solids system configurations are determined by the condition temperatures increase the rate of attack. The design must
of the material coming into and out of the combustion sec- balance the need for higher temperature (i.e., higher
tion. Either a liquid or a gas can carry the solid. If the material destruction efficiency) with the need to prolong refractory
has a high melting point (> 2400°F or 1300°C) and comes in life. Temperatures ranging from 1600 to 2000°F (870 to
with water, it will likely pass through the combustion section 1100°C) and residence times of 1.0 to 1.5 seconds are
virtually unaffected. If the solid is part of an organic liquid, commonly used.
it may melt in the flame zone, but will return to solid phase The system best-suited for molten salts is illustrated in
very quickly after leaving the flame zone. If flow patterns Figure 21.33. The incinerator is usually vertical, with the
and vessel orientation are wrong, the material will often col- burner mounted on the top and firing downward. The vertical
lect in the equipment. Regardless, if the particulate emission design is highly desirable because it does not allow the molten
excluded the allowable amount, a particulate removal device salts to accumulate on the refractory, as it would in a horizontal
must be employed. Examples are catalyst fines in organic or oxidizer. This is extremely important, as any accumulation of
aqueous liquids, metal-machining dust in air, titanium diox- salts in the oxidizer can drastically reduce refractory life.
ide (an opaque, white additive to many products) in water or Basic (MgO) brick has very good resistance to salt attack.
organic liquid, and some organic salts such as sodium However, it is very expensive, highly susceptible to thermal
acetate (which oxidizes to form organic combustion prod- shock and hydration, and will quickly erode if hit by a stream
ucts and sodium oxides, which have high melting points). of water. High alumina brick has been used with varying
If the material comes into the vessel as an organic solid, it results. In some cases, a 90% alumina brick, with its low
may very well be completely oxidized before it exits. How- porosity, has been found to be more effective against salt
ever, organic materials often contain non-organic compounds attack than a 60% alumina brick. Unfortunately, while the
that may be high-melting-point inerts or lower-melting-point higher alumina brick is two to three times the cost of the
materials such as silica (~2000°F or 1100°C), or alkali metals lower alumina brick, it does not last two or three times longer.
inorganic salts such as NaCl (~1500°F or 820°C). If the solid In general, using the lowest porosity, 60 to 70% alumina brick
is organic, certain configurations allow quicker, more com- and the proper burner/T.O. configuration provides the most
plete oxidation than others. If lower-melting-point materials cost-effective service.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 687 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 687

FIGURE 21.33 Molten salt system.

For organic wastes, moderate to low-pressure-drop burners be allowed to burn to near completion before injecting the
are preferred, because a high-pressure-drop burner, with its aqueous waste stream into the burner combustion products.
attendant turbulence, tends to centrifuge the salts in the waste Incomplete combustion of the auxiliary fuel can cause soot
toward the refractory, causing it to deteriorate more rapidly. formation and incomplete combustion of the waste.
A moderate-pressure-drop burner with no cyclonic action Although rare, waste heat recovery boilers have been used
tends to keep the salts in suspension and away from the in some systems. To utilize heat recovery, the flue gas must
incinerator walls. The less salt that contacts the refractory, be conditioned to a temperature below the melting point of
the slower the refractory degrades. The hydrocarbon liquid is the salt, freezing it before it contacts the cooler heat transfer
fired through the burner. surface. This makes the salt friable so that the accumulation
Water-based wastes are injected into the T.O. downstream is easier to remove from the heat transfer surface. A water-
of the burner. Because the water in these wastes must be tube-type boiler is used to allow online cleaning by soot
evaporated and the bulk mass raised to near the oxidation blowers. Figure 21.34 shows this configuration.
point before combustion of organic materials can begin, it is Before the flue gas can be vented to the atmosphere,
necessary to burn some amount of auxiliary fuel. The amount particulate matter must be reduced to the concentration
of auxiliary fuel burned depends on the operating temperature allowable by emission standards. This is accomplished with
and the waste composition. A similar moderate-pressure-drop a baghouse, an ESP a WESP, or a venturi scrubber. Each of
burner is also used for this condition. The auxiliary fuel must these particle removal devices has been discussed in previ-
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

688 The John Zink Combustion Handbook

FIGURE 21.34 Online cleaning with soot blowers.

ous chapter sections. As noted, proper flue gas conditioning vious sections. This section reviews only the three-stage
or quenching, for the baghouse or ESP, or for the WESP or system configuration.
venturi scrubber, must be accomplished upstream of any air Consider a waste stream consisting of chlorinated hydro-
pollution control equipment. carbon, amine, and other hydrocarbons. Oxidizing this waste
Excessive salt in the flue gas from a salts unit will form a mixture in a standard excess-air T.O. would produce high
nondissipating visible plume. Consequently, the flue gas from NOx and HCl emissions. A three-stage NOx system would
the saturated system often receives intense public scrutiny. provide the NOx reduction needed, but the HCl generated
For this reason, some operators choose to install equipment would be an issue. By simply adding a packed column scrub-
to eliminate the visible water vapor plume. This is accom- ber, which is capable of quenching the flue gas in the base
plished either by cooling and removing virtually all the water of the scrubber, to the end of the NOx system, the problem
vapor in the saturated flue gas, or by removing a substantial is solved. Figure 21.35 illustrates this configuration.
portion of it and then heating the flue gas stream so that the The combustion air blower, burner, reduction furnace,
plume becomes less visible. De-pluming is costly, both in quench section, oxidation air blower, and oxidation section
terms of capital and operating costs, and the results are not are very similar to those supplied for a “normal” three-stage
always satisfactory. NOx system. Carbon steel construction is still acceptable. It
For both water- and organic-based liquid streams, 99.99% is necessary to add a full-length rainshield or internal anti-
DRE is normally required. A particulate matter concentra- corrosion lining to the three vessels to protect against HCl
tion of 0.015 grains per dry standard cubic foot (DSCF) of corrosion. Also, the carbon steel duct between the boiler and
flue gas, corrected to a specific excess O2 level, is also the packed column and the entire recycle duct to the reduction
normally required. furnace and quench section must be externally insulated and
sealed from weather so the unlined ductwork steel tempera-
ture will stay well above the dew point. The recycle blower
21.6.5 NOx Minimization or should be carbon steel with sealed external insulation. The
Reduction Systems packed column should be FRP.
The theory behind the combustion-process-modification
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
The refractory is very similar to that in the “normal” NOx
method of minimizing or reducing NOx was covered in pre- system with the exception that the presence of HCl must be
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-ch21-Frame Page 689 Wednesday, September 3, 2003 8:12 AM

Thermal Oxidizers 689

FIGURE 21.35 Three-stage NOx system with packed column scrubber.

considered. The castable refractory should be reviewed to PCB Polychlorinated biphenyl

ensure that the material with the lowest CaO content is SCR Selective catalytic reduction
utilized without compromising the refractory strength or SCS Specific collector surface
operating capability. SNCR Selective noncatalytic reduction
Special consideration should be given to the instrumenta- T.O. Thermal oxidizer
tion used in the NOx reduction described here as it must be WESP Wet electrostatic precipitator
capable of operation in the anticipated HCl/Cl2 environment.

REFERENCES
21.7 CLOSING
The purpose of presenting this chapter was to attempt to pro- 1. C.D. Cooper and F.L. Alley, Air Pollution Control, A
vide a better understanding of the use of thermal oxidation pro- Design Approach, 2nd ed., Waveland Press, Prospect
cesses to destroy vapor and liquid wastes. Although specific Heights, IL, 1994.
design details were not included, written explanations and dia-
2. R.C. Flagan and J.H. Seinfeld, Fundamentals of Air
grams provided an overview of the multi-faceted subject.
Pollution Engineering, Prentice-Hall, Englewood
Cliffs, NJ, 1988.
3. W. Licht, Air Pollution Control Engineering, Basic
21.9 NOMENCLATURE
Calculations for Particulate Collection, Marcel Dekker,
APC Air pollution control
Inc., New York, 1980.
CSA Collecting surface area
DRE Destruction and removal efficiency 4. R.D. Reed, Furnace Operations, 3rd ed., Gulf Publishing,
ESP Electrostatic precipitator Houston, TX, 1981.
FRP Fiberglass-reinforced plastic 5. H.L. Gutzwiller, Fan Performance and Design, Robinson
LEL Lower explosive (flammability) limit Industries, Inc., January 2000.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendices

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendices
Appendix A — Physical Properties of Materials
Table A.1 Areas and Circumferences of Circles and Drill Sizes...................................................... 695
Table A.2 Physical Properties of Pipe .............................................................................................. 704
Table A.3 Physical Properties of Tubing .......................................................................................... 709
Table A.4 SAE Grades for Steel Bolts.............................................................................................. 711
Table A.5 ASTM Grades for Steel Bolts .......................................................................................... 712
Table A.6 Properties for Metric Steel Bolts, Screws, and Studs....................................................... 713
Appendix B — Properties of Gases and Liquids
Table B.1 Combustion Data for Hydrocarbons ................................................................................ 715
Table B.2 Thermodynamic Data for Common Substances .............................................................. 716
Table B.3 Properties of Dry Air at Atmospheric Pressure................................................................ 717
Table B.4 Properties of Gases and Vapors in English and Metric Units .......................................... 719
Appendix C — Common Conversions .................................................................................................................. 725

693 --`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendix A
Physical Properties of Materials
TABLE A.1 Areas and Circumferences of Circles and Drill Sizes
Drill Size Diameter (in.) Circumference (in.) Area (in.) Area (ft)

80 0.0135 0.042 41 0.000 143 0.000 000 9

79 0.0145 0.045 55 0.000 165 0.000 001 1
1/64" 0.0156 0.049 09 0.000 191 0.000 001 3
78 0.0160 0.050 27 0.000 201 0.000 001 4

77 0.0180 0.056 55 0.000 254 0.000 001 8

76 0.0200 0.062 83 0.000 314 0.000 002 2
75 0.0210 0.065 97 0.000 346 0.000 002 4
74 0.0225 0.070 69 0.000 398 0.000 002 8

73 0.0240 0.075 40 0.000 452 0.000 003 1

72 0.0250 0.078 54 0.000 491 0.000 003 4
71 0.0260 0.081 68 0.000 531 0.000 003 7
70 0.0280 0.087 96 0.000 616 0.000 004 3

69 0.0292 0.091 73 0.000 670 0.000 004 7

68 0.0310 0.097 39 0.000 755 0.000 005 2
1/12" 0.0313 0.098 18 0.000 765 0.000 005 3
67 0.0320 0.100 53 0.000 804 0.000 005 6

66 0.0330 0.103 67 0.000 855 0.000 005 9

65 0.0350 0.109 96 0.000 962 0.000 006 7
64 0.0360 0.113 10 0.001 018 0.000 007 1
63 0.0370 0.116 24 0.001 075 0.000 007 5

62 0.0380 0.119 38 0.001 134 0.000 007 9

61 0.0390 0.122 52 0.001 195 0.000 008 3
60 0.0400 0.125 66 0.001 257 0.000 008 7
59 0.0410 0.128 81 0.001 320 0.000 009 2

58 0.0420 0.131 95 0.001 385 0.000 009 6

57 0.0430 0.135 09 0.001 452 0.000 010 1
56 0.0465 0.146 08 0.001 698 0.000 011 8
3/64" 0.0469 0.147 26 0.001 73 0.000 012 0
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

55 0.0520 0.163 36 0.002 12 0.000 014 7

54 0.0550 0.172 79 0.002 38 0.000 016 5
53 0.0595 0.186 93 0.002 78 0.000 019 3
1/16" 0.0625 0.196 35 0.003 07 0.000 021 3

52 0.0635 0.199 49 0.003 17 0.000 022 0

51 0.0670 0.210 49 0.003 53 0.000 024 5
50 0.0700 0.219 91 0.003 85 0.000 026 7
49 0.0730 0.229 34 0.004 19 0.000 029 1

48 0.0760 0.238 76 0.004 54 0.000 031 5

5/64" 0.0781 0.245 44 0.004 79 0.000 033 3
47 0.0785 0.246 62 0.004 84 0.000 033 6

695
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Appendices-Frame Page 696 Tuesday, March 15, 2005 10:00 AM

696 The John Zink Combustion Handbook

TABLE A.1 (continued) Areas and Circumferences of Circles and Drill Sizes
Drill Size Diameter (in.) Circumference (in.) Area (in.) Area (ft)

46 0.0810 0.254 47 0.005 15 0.000 035 8

45 0.0820 0.257 61 0.005 28 0.000 036 7
44 0.0860 0.270 18 0.005 81 0.000 040 3
43 0.0890 0.279 60 0.006 22 0.000 043 2
42 0.0935 0.293 74 0.006 87 0.000 047 7

3/32" 0.0937 0.294 52 0.006 90 0.000 047 9

41 0.0960 0.301 59 0.007 24 0.000 050 3
40 0.0980 0.307 88 0.007 54 0.000 052 4
39 0.0995 0.312 59 0.007 78 0.000 054 0

38 0.1015 0.318 87 0.008 09 0.000 056 2

37 0.1040 0.326 73 0.008 49 0.000 059 0
36 0.1065 0.334 58 0.008 91 0.000 061 9
7/64" 0.1094 0.343 61 0.009 40 0.000 065 2

35 0.1100 0.345 58 0.009 50 0.000 066 0

34 0.1110 0.348 72 0.009 68 0.000 067 2
33 0.1130 0.355 00 0.010 03 0.000 069 6
32 0.1160 0.364 43 0.010 57 0.000 073 4

31 0.1200 0.376 99 0.011 31 0.000 078 5

1/8" 0.1250 0.392 70 0.012 27 0.000 085 2
30 0.1285 0.403 70 0.012 96 0.000 090 1
29 0.1360 0.427 26 0.014 53 0.000 100 9

28 0.1405 0.441 39 0.015 49 0.000 107 7

9/64" 0.1406 0.441 79 0.015 53 0.000 107 9
27 0.1440 0.442 39 0.016 29 0.000 113 1
26 0.1470 0.461 82 0.016 97 0.000 117 9

25 0.1495 0.469 67 0.017 55 0.000 121 9

24 0.1520 0.477 52 0.018 15 0.000 126 0
23 0.1540 0.483 81 0.018 63 0.000 129 4
5/32" 0.1562 0.490 87 0.019 17 0.000 133 1

22 0.1570 0.493 23 0.019 36 0.000 134 4

21 0.1590 0.499 51 0.019 86 0.000 137 9
20 0.1610 0.505 80 0.020 36 0.000 141 4
19 0.1660 0.521 51 0.021 64 0.000 150 3

18 0.1695 0.532 50 0.022 56 0.000 156 7

11/64" 0.1719 0.539 96 0.023 20 0.000 161 1
17 0.1730 0.543 50 0.023 51 0.000 163 2
16 0.1770 0.556 06 0.024 61 0.000 170 9

15 0.1800 0.565 49 0.025 45 0.000 176 7

14 0.1820 0.571 77 0.026 02 0.000 180 7
13 0.1850 0.581 20 0.026 88 0.000 186 7
3/16" 0.1875 0.589 05 0.027 61 0.000 191 7

12 0.1890 0.593 76 0.028 06 0.000 194 8

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

11 0.1910 0.600 05 0.028 65 0.000 199 0

10 0.1930 0.606 33 0.029 40 0.000 203 2
9 0.1960 0.615 75 0.030 17 0.000 209 5

Copyright CRC Press

Appendix A Physical Properties of Materials 697

TABLE A.1 (continued) Areas and Circumferences of Circles and Drill Sizes
Drill Size Diameter (in.) Circumference (in.) Area (in.) Area (ft)

8 0.1990 0.625 18 0.031 10 0.000 216 0

7 0.2010 0.631 46 0.031 73 0.000 220 4
13/64" 0.2031 0.638 14 0.032 41 0.000 224 8
6 0.2040 0.640 89 0.032 69 0.000 227 0

5 0.2055 0.645 60 0.033 17 0.000 230 3

4 0.2090 0.656 59 0.034 31 0.000 238 2
3 0.2130 0.669 16 0.035 63 0.000 247 5
7/32" 0.2187 0.687 22 0.037 58 0.000 261 0

2 0.2210 0.694 29 0.038 36 0.000 266 4

1 0.2280 0.716 28 0.040 83 0.000 283 5
A 0.2340 0.735 13 0.043 01 0.000 298 7
15/64" 0.2344 0.736 31 0.043 14 0.000 299 6

B 0.2380 0.747 70 0.044 49 0.000 308 9

C 0.2420 0.760 27 0.046 00 0.000 319 4
D 0.2460 0.772 83 0.047 53 0.000 330 1
E = 1/4" 0.2500 0.785 40 0.049 09 0.000 340 9

F 0.2570 0.807 39 0.051 87 0.000 360 2

G 0.2610 0.819 96 0.053 50 0.000 371 5
17/64" 0.2656 0.834 41 0.055 42 0.000 384 9
H 0.2660 0.835 67 0.055 57 0.000 385 9

I 0.2720 0.854 52 0.058 11 0.000 403 5

J 0.2770 0.870 22 0.060 26 0.000 418 5
K 0.2810 0.882 79 0.062 02 0.000 430 7
9/32" 0.2812 0.883 57 0.062 13 0.000 431 5

L 0.2900 0.911 06 0.066 05 0.000 458 7

M 0.2950 0.926 77 0.068 35 0.000 474 7
19/64" 0.2969 0.932 66 0.069 22 0.000 480 7
N 0.3030 0.951 90 0.071 63 0.000 500 7

5/16" 0.3125 0.981 75 0.076 70 0.000 532 6

O 0.3160 0.992 75 0.078 43 0.000 544 6
P 0.3230 1.014 74 0.081 94 0.000 569 0
21/64" 0.3281 1.030 8 0.084 56 0.000 587 2

Q 0.3320 1.043 0 0.086 57 0.000 601 2

R 0.3390 1.065 0 0.090 26 0.000 626 8
11/32" 0.3437 1.079 8 0.092 81 0.000 644 5
S 0.3480 1.093 3 0.095 11 0.000 660 5

T 0.3580 1.124 7 0.100 6 0.000 699 0

23/64" 0.3594 1.129 0 0.101 4 0.000 704 4
U 0.3680 1.156 1 0.106 4 0.000 738 6
3/8" 0.3750 1.178 1 0.110 5 0.000 767 0

V 0.3770 1.184 4 0.111 6 0.000 775 2

W 0.3860 1.212 7 0.117 0 0.000 812 7
25/64" 0.3906 1.227 2 0.119 8 0.000 832 2
X 0.3970
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- 1.247 2 0.123 8 0.000 859 6

Copyright CRC Press

698 The John Zink Combustion Handbook

TABLE A.1 (continued) Areas and Circumferences of Circles and Drill Sizes
Drill Size Diameter (in.) Circumference (in.) Area (in.) Area (ft)

Y 0.4040 1.269 2 0.128 2 0.000 890 2

13/32" 0.4062 1.276 3 0.129 6 0.000 900 1
Z 0.4130 1.297 5 0.134 0 0.000 930 3
27/64" 0.4219 1.325 4 0.139 8 0.000 970 8

7/16" 0.4375 1.3745 0.1503 0.001 044

29/64" 0.4531 1.4235 0.1613 0.001 120
15/32" 0.4687 1.4726 0.1726 0.001 198
31/64" 0.4844 1.5217 0.1843 0.001 280

1/2" 0.5000 1.5708 0.1964 0.001 364

33/64" 0.5156 1.6199 0.2088 0.001 450
17/32" 0.5313 1.6690 0.2217 0.001 539
35/64" 0.5469 1.7181 0.2349 0.001 631

9/16" 0.5625 1.7672 0.2485 0.001 726

37/64" 0.5781 1.8162 0.2625 0.001 823
19/32" 0.5938 1.8653 0.2769 0.001 923
39/64" 0.6094 1.9144 0.2917 0.002 025

5/8" 0.6250 1.9635 0.3068 0.002 131

41/64" 0.6406 2.0126 0.3223 0.002 238
21/32" 0.6562 2.0617 0.3382 0.002 350
43/64" 0.6719 2.1108 0.3545 0.002 462

11/16" 0.6875 2.1598 0.3712 0.002 578

23/32" 0.7188 2.2580 0.4057 0.002 818
3/4" 0.7500 2.3562 0.4418 0.003 068
25/32" 0.7812 2.4544 0.4794 0.003 329

13/16" 0.8125 2.5525 0.5185 0.003 601

27/32" 0.8438 2.6507 0.5591 0.003 883
7/8" 0.8750 2.7489 0.6013 0.004 176

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
29/32" 0.9062 2.8471 0.6450 0.004 479

15/16" 0.9375 2.9452 0.6903 0.004 794

31/32" 0.9688 3.0434 0.7371 0.005 119
1" 1.0000 3.1416 0.7854 0.005 454
1 1/16" 1.0625 3.3379 0.8866 0.006 157

1 1/8" 1.1250 3.5343 0.9940 0.006 903

1 3/16" 1.1875 3.7306 1.1075 0.007 691
1 1/4" 1.2500 3.9270 1.2272 0.008 522
1 5/16" 1.3125 4.1233 1.3530 0.009 396

1 3/8" 1.3750 4.3170 1.4849 0.010 31

1 7/16" 1.4375 4.5160 1.6230 0.011 27
1 1/2" 1.5000 4.7124 1.7671 0.012 27
1 9/16" 1.5625 4.9087 1.9175 0.013 32

1 5/8" 1.6250 5.1051 2.0739 0.014 40

1 11/16" 1.6875 5.3014 2.2365 0.015 53
1 3/4" 1.7500 5.4978 2.4053 0.016 70
1 13/16" 1.8125 5.6941 2.5802 0.017 92

Copyright CRC Press

Appendix A Physical Properties of Materials 699

TABLE A.1 (continued) Areas and Circumferences of Circles and Drill Sizes
Drill Size Diameter (in.) Circumference (in.) Area (in.) Area (ft)

1 7/8" 1.8750 5.8905 2.7612 0.019 18

1 15/16" 1.9375 6.0868 2.9483 0.020 47
2" 2.0000 6.2832 3.1416 0.021 82
2 1/16" 2.0625 6.4795 3.3410 0.023 20

2 1/8" 2.1250 6.6759 3.5466 0.024 63

2 3/16" 2.1875 6.8722 3.7583 0.026 10
2 1/4" 2.2500 7.0686 3.9761 0.027 61
2 5/16" 2.3125 7.2649 4.2000 0.029 17

2 3/8" 2.3750 7.4613 4.4301 0.030 76

2 7/16" 2.4375 7.6576 4.6664 0.032 41
2 1/2" 2.5000 7.8540 4.9087 0.034 09
2 9/16" 2.5625 8.0503 5.1572 0.035 81

2 5/8" 2.6250 8.2467 5.4119 0.037 58

2 11/16" 2.6875 8.4430 5.6727 0.039 39
2 3/4" 2.7500 8.6394 5.9396 0.041 25
2 13/16" 2.8125 8.8357 6.2126 0.043 14

2 7/8" 2.8750 9.0323 6.4918 0.045 08

2 15/16" 2.9375 9.2284 6.7771 0.047 06
3" 3.0000 9.4248 7.0686 0.049 09
3 1/16" 3.0625 9.6211 7.3662 0.051 15

3 1/8" 3.1250 9.8175 7.6699 0.053 26

3 3/16" 3.1875 10.014 7.9798 0.055 42
3 1/4" 3.2500 10.210 8.2958 0.057 36
3 5/16" 3.3125 10.407 8.6179 0.059 85

3 3/8" 3.3750 10.603 8.9462 0.062 13

3 7/16" 3.4375 10.799 9.2806 0.064 45
3 1/2" 3.5000 10.996 9.6211 0.066 81
3 9/16" 3.5625 11.192 9.9678 0.069 22

3 5/8" 3.6250 11.388 10.321 0.071 67

3 11/16" 3.6875 11.585 10.680 0.074 17
3 3/4" 3.7500 11.781 11.045 0.076 70
3 13/16" 3.8125 11.977 11.416 0.079 28

3 7/8" 3.8750 12.174 11.793 0.081 90

3 15/16" 3.9375 12.370 12.177 0.084 56
4" 4.0000 12.566 12.566 0.087 26
4 1/16" 4.0625 12.763 12.962 0.090 02

4 1/8" 4.1250 12.959 13.364 0.092 81

4 3/16" 4.1875 13.155 13.772 0.095 64
4 1/4" 4.2500 13.352 14.186 0.098 52
4 5/16" 4.3125 13.548 14.607 0.101 4

4 3/8" 4.3750 13.745 15.033 0.104 3

4 7/16" 4.4375 13.941 15.466 0.107 4
4 1/2" 4.5000 14.137 15.904 0.110 4
4 9/16" 4.5625 14.334 16.349 0.113 5
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

700 The John Zink Combustion Handbook

TABLE A.1 (continued) Areas and Circumferences of Circles and Drill Sizes
Drill Size Diameter (in.) Circumference (in.) Area (in.) Area (ft)

4 5/8" 4.6250 14.530 16.800 0.1167

4 11/16" 4.6875 14.726 17.257 0.1198
4 3/4" 4.7500 14.923 17.721 0.1231
4 13/16" 4.8125 15.119 18.190 0.1263

4 7/8" 4.8750 15.315 18.665 0.1296

4 15/16" 4.9375 15.512 19.147 0.1330
5" 5.0000 15.708 19.635 0.1364
5 1/16" 5.0625 15.904 20.129 0.1398

5 1/8" 5.1250 16.101 20.629 0.1433

5 3/16" 5.1875 16.297 21.135 0.1468
5 1/4" 5.2500 16.493 21.648 0.1503
5 5/16" 5.3125 16.690 22.166 0.1539

5 3/8" 5.3750 16.886 22.691 0.1576

5 7/16" 5.4375 17.082 23.221 0.1613
5 1/2" 5.5000 17.279 23.758 0.1650
5 9/16" 5.5625 17.475 24.301 0.1688

5 5/8" 5.6250 17.671 24.851 0.1726

5 11/16" 5.6875 17.868 25.406 0.1764
5 3/4" 5.7500 18.064 25.967 0.1803
5 13/16" 5.8125 18.261 26.535 0.1843

5 7/8" 5.8750 18.457 27.109 0.1883

5 15/16" 5.9375 18.653 27.688 0.1923
6" 6.0000 18.850 28.274 0.1963
6 1/8" 6.1250 19.242 29.465 0.2046

6 1/4" 6.2500 19.649 30.680 0.2131

6 3/8" 6.3750 20.028 31.919 0.2217

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
6 1/2" 6.5000 20.420 33.183 0.2304
6 5/8" 6.6250 20.813 34.472 0.2394

6 3/4" 6.7500 21.206 35.785 0.2485

6 7/8" 6.8750 21.598 37.122 0.2578
7" 7.0000 21.991 38.485 0.2673
7 1/8" 7.1250 22.384 39.871 0.2769

7 1/4" 7.2500 22.777 41.283 0.2867

7 3/8" 7.3750 23.169 42.718 0.2967
7 1/2" 7.5000 23.562 44.179 0.3068
7 5/8" 7.6250 23.955 45.664 0.3171

7 3/4" 7.7500 24.347 47.173 0.3276

7 7/8" 7.8750 24.740 48.707 0.3382
8" 8.0000 25.133 50.266 0.3491
8 1/8" 8.1250 25.525 51.849 0.3601

8 1/4" 8.2500 25.918 53.456 0.3712

8 3/8" 8.3750 26.301 55.088 0.3826
8 1/2" 8.5000 26.704 56.745 0.3941
8 5/8" 8.6250 27.096 58.426 0.4057

Copyright CRC Press

Appendix A Physical Properties of Materials 701

TABLE A.1 (continued) Areas and Circumferences of Circles and Drill Sizes
Drill Size Diameter (in.) Circumference (in.) Area (in.) Area (ft)

8 3/4" 8.7500 27.489 60.132 0.4176

8 7/8" 8.8750 27.882 61.862 0.4296
9" 9.0000 28.274 63.617 0.4418
9 1/8" 9.1250 28.667 65.397 0.4541

9 1/4" 9.2500 29.060 67.201 0.4667

9 3/8" 9.3750 29.452 69.029 0.4794
9 1/2" 9.5000 29.845 70.882 0.4922
9 5/8" 9.6250 30.238 72.760 0.5053

9 3/4" 9.7500 30.631 74.662 0.5185

9 7/8" 9.8750 31.023 76.589 0.5319
10" 10.0000 31.416 78.540 0.5454
10 1/8" 10.1250 31.809 80.516 0.5591

10 1/4" 10.2500 32.201 82.516 0.5730

10 3/8" 10.3750 32.594 84.541 0.5871
10 1/2" 10.5000 32.987 86.590 0.6013
10 5/8" 10.6250 33.379 88.664 0.6157

10 3/4" 10.7500 33.772 90.763 0.6303

10 7/8" 10.8750 34.165 92.886 0.6450
11" 11.0000 34.558 95.033 0.6600

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
11 1/8" 11.1250 34.950 97.205 0.6750

11 1/4" 11.2500 35.343 99.402 0.6903

11 3/8" 11.3750 35.736 101.6 0.7056
11 1/2" 11.5000 36.128 103.9 0.7213
11 5/8" 11.6250 36.521 106.1 0.7371

11 3/4" 11.7500 36.914 108.4 0.7530

11 7/8" 11.8750 37.306 110.8 0.7691
12" 12.0000 37.699 113.1 0.7854
12 1/4" 12.2500 38.485 117.9 0.819

12 1/2" 12.5000 39.269 122.7 0.851

12 3/4" 12.7500 40.055 127.7 0.886
13" 13.0000 40.841 132.7 0.921
13 1/4" 13.2500 41.626 137.9 0.957

13 1/2" 13.5000 42.412 143.1 0.995

13 3/4" 13.7500 43.197 148.5 1.031
14" 14.0000 43.982 153.9 1.069
14 1/4" 14.2500 44.768 159.5 1.109

14 1/2" 14.5000 45.553 165.1 1.149

14 3/4" 14.7500 46.339 170.9 1.185
15" 15.0000 47.124 176.7 1.228
15 1/4" 15.2500 47.909 182.7 1.269

15 1/2" 15.5000 48.695 188.7 1.309

15 3/4" 15.7500 49.480 194.8 1.352
16" 16.0000 50.266 201.1 1.398
16 1/4" 16.2500 51.051 207.4 1.440

Copyright CRC Press

702 The John Zink Combustion Handbook

TABLE A.1 (continued) Areas and Circumferences of Circles and Drill Sizes
Drill Size Diameter (in.) Circumference (in.) Area (in.) Area (ft)

16 1/2" 16.5000 51.836 213.8 1.485

16 3/4" 16.7500 52.622 220.4 1.531
17" 17.0000 53.407 227.0 1.578
17 1/4" 17.2500 54.193 233.7 1.619

17 1/2" 17.5000 54.978 240.5 1.673

17 3/4" 17.7500 55.763 247.5 1.719
18" 18.0000 56.548 254.5 1.769
18 1/4" 18.2500 57.334 261.6 1.816

18 1/2" 18.5000 58.120 268.8 1.869

18 3/4" 18.7500 58.905 276.1 1.920
19" 19.0000 59.690 283.5 1.969
19 1/4" 19.2500 60.476 291.0 2.022

19 1/2" 19.5000 61.261 298.7 2.075

19 3/4" 19.7500 62.047 306.4 2.125
20" 20.0000 62.832 314.2 2.182
20 1/4" 20.2500 63.617 322.1 2.237

20 1/2" 20.5000 64.403 330.1 2.292

20 3/4" 20.7500 65.188 338.2 2.348
21" 21.0000 65.974 346.4 2.405
21 1/4" 21.2500 66.759 354.7 2.463

21 1/2" 21.5000 67.544 363.1 2.521

21 3/4" 21.7500 68.330 371.5 2.580
22" 22.0000 69.115 380.1 2.640
22 1/4" 22.2500 69.901 388.8 2.700

22 1/2" 22.5000 70.686 397.6 2.761

22 3/4" 22.7500 71.471 406.5 2.823
23" 23.0000 72.257 415.5 2.885
23 1/4" 23.2500 73.042 424.6 2.948

23 1/2" 23.5000 73.828 433.7 3.012

23 3/4" 23.7500 74.613 443.0 3.076
24" 24.0000 75.398 452.4 3.142
24 1/4" 24.2500 76.184 461.9 3.207

24 1/2" 24.5000 76.969 471.4 3.274

24 3/4" 24.7500 77.755 481.1 3.341
25" 25.0000 78.540 490.9 3.409
25 1/4" 25.2500 79.325 500.7 3.477

25 1/2" 25.5000 80.111 510.7 3.547

25 3/4" 25.7500 80.896 520.8 3.616
26" 26.0000 81.682 530.9 3.687
26 1/4" 26.2500 82.467 541.2 3.758

26 1/2" 26.5000 83.252 551.6 3.830

26 3/4" 26.7500 84.038 562.0 3.903
27" 27.0000 84.823 572.6 3.976
27 1/4" 27.2500 85.609 583.2 4.050
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendix A Physical Properties of Materials 703

TABLE A.1 (continued) Areas and Circumferences of Circles and Drill Sizes
Drill Size Diameter (in.) Circumference (in.) Area (in.) Area (ft)

27 1/2" 27.5000 86.394 594.0 4.125

27 3/4" 27.7500 87.179 604.8 4.200
28" 28.0000 87.965 615.8 4.276
28 1/4" 28.2500 88.750 626.8 4.353

28 1/2" 28.5000 89.536 637.9 4.430

28 3/4" 28.7500 90.321 649.2 4.508

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
29" 29.0000 91.106 660.5 4.587
29 1/4" 29.2500 91.892 672.0 4.666

29 1/2" 29.5000 92.677 683.5 4.746

29 3/4" 29.7500 93.463 695.1 4.827
30" 30.0000 94.248 706.9 4.909
31" 31.0000 97.390 754.8 5.241

32" 32.0000 100.53 804.3 5.585

33" 33.0000 103.67 855.3 5.940
34" 34.0000 106.81 907.9 6.305
35" 35.0000 109.96 962.1 6.681

36" 36.0000 113.10 1017.9 7.069

37" 37.0000 116.24 1075.2 7.467
38" 38.0000 119.38 1134.1 7.876
39" 39.0000 122.52 1194.6 8.296

40" 40.0000 125.66 1256.6 8.727

41" 41.0000 128.81 1320.3 9.168
42" 42.0000 131.95 1385.4 9.621
43" 43.0000 135.09 1452.2 10.08

44" 44.0000 138.23 1520.5 10.56

45" 45.0000 141.37 1590.4 11.04
46" 46.0000 144.51 1661.9 11.54
47" 47.0000 147.66 1734.9 12.04

48" 48.0000 150.80 1809.6 12.57

49" 49.0000 153.94 1885.7 13.10
50" 50.0000 157.08 1963.5 13.64

Copyright CRC Press

704 The John Zink Combustion Handbook

TABLE A.2 Physical Properties of Pipe

Wall Sq. Ft. Sq. Ft. Weight Moment
Nominal Schedule Number Thick- Inside Metal outside inside Weight of water of Section Radius
pipe size, ness, Area, Area, surface, surface, per ft, per ft, Inertia, modulus, gyration,
OD, in. a b c in. I.D., in. sq. in. sq. in. per ft per ft lb lb in.4 in.3 in.

… ….. 10S 0.049 0.307 0.0740 0.0548 0.106 0.0804 0.186 0.0321 0.00088 0.00437 0.1271
1/8 40 Std 40S 0.068 0.269 0.0568 0.0720 0.106 0.0705 0.245 0.0246 0.00106 0.00525 0.1215
0.405 80 XS 80S 0.095 0.215 0.0364 0.0925 0.106 0.0563 0.315 0.0157 0.00122 0.00600 0.1146

… ….. 10S 0.065 0.410 0.1320 0.0970 0.141 0.1073 0.330 0.0572 0.00279 0.01032 0.1694
1/4 40 Std 40S 0.088 0.364 0.1041 0.1250 0.141 0.0955 0.425 0.0451 0.00331 0.01230 0.1628
0.540 80 XS 80S 0.119 0.302 0.0716 0.1574 0.141 0.0794 0.535 0.0310 0.00378 0.01395 0.1547

… ….. 10S 0.065 0.545 0.2333 0.1246 0.177 0.1427 0.423 0.1011 0.00586 0.01737 0.2169
3/8 40 Std 40S 0.091 0.493 0.1910 0.1670 0.177 0.1295 0.568 0.0827 0.00730 0.02160 0.2090
0.675 80 XS 80S 0.126 0.423 0.1405 0.2173 0.177 0.1106 0.739 0.0609 0.00862 0.02554 0.1991

… ….. 10S 0.083 0.674 0.3570 0.1974 0.220 0.1765 0.671 0.1547 0.01431 0.0341 0.2692
40 Std 40S 0.109 0.622 0.3040 0.2503 0.220 0.1628 0.851 0.1316 0.01710 0.0407 0.2613
1/2 80 XS 80S 0.147 0.546 0.2340 0.3200 0.220 0.1433 1.088 0.1013 0.02010 0.0478 0.2505
0.840 160 ….. … 0.187 0.466 0.1706 0.3830 0.220 0.1220 1.304 0.0740 0.02213 0.0527 0.2402
… XXS … 0.294 0.252 0.0499 0.5040 0.220 0.0660 1.714 0.0216 0.02425 0.0577 0.2192

… ….. 5S 0.065 0.920 0.6650 0.2011 0.275 0.2409 0.684 0.2882 0.02451 0.0467 0.349

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
… ….. 10S 0.083 0.884 0.6140 0.2521 0.275 0.2314 0.857 0.2661 0.02970 0.0566 0.343
3/4 40 Std 40S 0.113 0.824 0.5330 0.3330 0.275 0.2157 1.131 0.2301 0.0370 0.0706 0.334
1.050 80 XS 80S 0.154 0.742 0.4320 0.4350 0.275 0.1943 1.474 0.1875 0.0448 0.0853 0.321
160 ….. … 0.218 0.614 0.2961 0.5700 0.275 0.1607 1.937 0.1284 0.0527 0.1004 0.304
… XXS … 0.308 0.434 0.1479 0.7180 0.275 0.1137 2.441 0.0641 0.0579 0.1104 0.284

… ….. 5S 0.065 1.185 1.1030 0.2553 0.344 0.3100 0.868 0.478 0.0500 0.0760 0.443
… ….. 10S 0.109 1.097 0.9450 0.4130 0.344 0.2872 1.404 0.409 0.0757 0.1151 0.428
1 40 Std 40S 0.133 1.049 0.8640 0.4940 0.344 0.2746 1.679 0.374 0.0874 0.1329 0.421
1.315 80 XS 80S 0.179 0.957 0.7190 0.6390 0.344 0.2520 2.172 0.311 0.1056 0.1606 0.407
160 ….. … 0.250 0.815 0.5220 0.8360 0.344 0.2134 2.844 0.2261 0.1252 0.1903 0.387
… XXS … 0.358 0.599 0.2818 1.0760 0.344 0.1570 3.659 0.1221 0.1405 0.2137 0.361

… ….. 5S 0.065 1.530 1.839 0.326 0.434 0.401 1.107 0.797 0.1038 0.1250 0.564
… ….. 10S 0.109 1.442 1.633 0.531 0.434 0.378 1.805 0.707 0.1605 0.1934 0.550
1-1/4 40 Std 40S 0.140 1.380 1.496 0.669 0.434 0.361 2.273 0.648 0.1948 0.2346 0.540
1.660 80 XS 80S 0.191 1.278 1.283 0.881 0.434 0.335 2.997 0.555 0.2418 0.2913 0.524
160 ….. … 0.250 1.160 1.057 1.107 0.434 0.304 3.765 0.458 0.2839 0.342 0.506
… XXS … 0.382 0.896 0.631 1.534 0.434 0.2346 5.214 0.2732 0.341 0.411 0.472

… ….. 5S 0.065 1.770 2.461 0.375 0.497 0.463 1.274 1.067 0.1580 0.1663 0.649
… ….. 10S 0.109 1.682 2.222 0.613 0.497 0.440 2.085 0.962 0.2469 0.2599 0.634
1-1/2 40 Std 40S 0.145 1.610 2.036 0.799 0.497 0.421 2.718 0.882 0.310 0.326 0.623
1.900 80 XS 80S 0.200 1.500 1.767 1.068 0.497 0.393 3.631 0.765 0.391 0.412 0.605
160 ….. … 0.281 1.338 1.406 1.429 0.497 0.350 4.859 0.608 0.483 0.508 0.581
… XXS … 0.400 1.100 0.950 1.885 0.497 0.288 6.408 0.412 0.568 0.598 0.549

… ….. 5S 0.065 2.245 3.960 0.472 0.622 0.588 1.604 1.716 0.315 0.2652 0.817
… ….. 10S 0.109 2.157 3.650 0.776 0.622 0.565 2.638 1.582 0.499 0.420 0.802
2 40 Std 40S 0.154 2.067 3.360 1.075 0.622 0.541 3.653 1.455 0.666 0.561 0.787
2.375 80 XS 80S 0.218 1.939 2.953 1.477 0.622 0.508 5.022 1.280 0.868 0.731 0.766
160 ….. … 0.343 1.689 2.240 2.190 0.622 0.442 7.444 0.971 1.163 0.979 0.729
… XXS … 0.436 1.503 1.774 2.656 0.622 0.393 9.029 0.769 1.312 1.104 0.703

Copyright CRC Press

Appendix A Physical Properties of Materials 705

TABLE A.2 (continued) Physical Properties of Pipe

… ….. 5S 0.083 2.709 5.76 0.728 0.753 0.709 2.475 2.499 0.710 0.494 0.988
… ….. 10S 0.120 2.635 5.45 1.039 0.753 0.690 3.531 2.361 0.988 0.687 0.975
2-1/2 40 Std 40S 0.203 2.469 4.79 1.704 0.753 0.646 5.793 2.076 1.530 1.064 0.947
2.875 80 XS 80S 0.276 2.323 4.24 2.254 0.753 0.608 7.661 1.837 0.193 1.339 0.924
160 ….. … 0.375 2.125 3.55 2.945 0.753 0.556 10.01 1.535 2.353 1.637 0.894
… XXS … 0.552 1.771 2.46 4.030 0.753 0.464 13.70 1.067 2.872 1.998 0.844

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
… ….. 5S 0.083 3.334 8.73 0.891 0.916 0.873 3.03 3.78 1.301 0.744 1.208
… ….. 10S 0.120 3.260 8.35 1.274 0.916 0.853 4.33 3.61 1.822 1.041 1.196
3 40 Std 40S 0.216 3.068 7.39 2.228 0.916 0.803 7.58 3.20 3.02 1.724 1.164
3.500 80 XS 80S 0.300 2.900 6.61 3.020 0.916 0.759 10.25 2.864 3.90 2.226 1.136
160 ….. … 0.437 2.626 5.42 4.210 0.916 0.687 14.32 2.348 5.03 2.876 1.094
… XXS … 0.600 2.300 4.15 5.470 0.916 0.602 18.58 1.801 5.99 3.43 1.047

… ….. 5S 0.083 3.834 11.55 1.021 1.047 1.004 3.47 5.01 1.960 0.980 1.385
3-1/2 … ….. 10S 0.120 3.760 11.10 1.463 1.047 0.984 4.97 4.81 2.756 1.378 1.372
4.000 40 Std 40S 0.226 3.548 9.89 2.68 1.047 0.929 9.11 4.28 4.79 2.394 1.337
80 XS 80S 0.318 3.364 8.89 3.68 1.047 0.881 12.51 3.85 6.28 3.14 1.307

… ….. 5S 0.083 4.334 14.75 1.152 1.178 1.135 3.92 6.40 2.811 1.249 1.562
… ….. 10S 0.120 4.260 14.25 1.651 1.178 1.115 5.61 6.17 3.96 1.762 1.549
4 40 Std 40S 0.237 4.026 12.73 3.17 1.178 1.054 10.79 5.51 7.23 3.21 1.510
4.500 80 XS 80S 0.337 3.826 11.50 4.41 1.178 1.002 14.98 4.98 9.61 4.27 1.477
120 ….. … 0.437 3.626 10.33 5.58 1.178 0.949 18.96 4.48 11.65 5.18 1.445
160 ….. … 0.531 3.438 9.28 6.62 1.178 0.900 22.51 4.02 13.27 5.90 1.416
… XXS … 0.674 3.152 7.80 8.10 1.178 0.825 27.54 3.38 15.29 6.79 1.374

… ….. 5S 0.109 5.345 22.44 1.868 1.456 1.399 6.35 9.73 6.95 2.498 1.929
… ….. 10S 0.134 5.295 22.02 2.285 1.456 1.386 7.77 9.53 8.43 3.03 1.920
5 40 Std 40S 0.258 5.047 20.01 4.30 1.456 1.321 14.62 8.66 15.17 5.45 1.878
5.563 80 XS 80S 0.375 4.813 18.19 6.11 1.456 1.260 20.78 7.89 20.68 7.43 1.839
120 ….. … 0.500 4.563 16.35 7.95 1.456 1.195 27.04 7.09 25.74 9.25 1.799
160 ….. … 0.625 4.313 14.61 9.70 1.456 1.129 32.96 6.33 30 10.8 1.760
… XXS … 0.750 4.063 12.97 11.34 1.456 1.064 38.55 5.62 33.6 12.1 1.722

… ….. 5S 0.109 6.407 32.20 2.231 1.734 1.677 5.37 13.98 11.85 3.58 2.304
… ….. 10S 0.134 6.357 31.70 2.733 1.734 1.664 9.29 13.74 14.4 4.35 2.295
6 40 Std 40S 0.280 6.065 28.89 5.58 1.734 1.588 18.97 12.51 28.14 8.5 2.245
6.625 80 XS 80S 0.432 5.761 26.07 8.40 1.734 1.508 28.57 11.29 40.5 12.23 2.195
120 ….. … 0.562 5.501 23.77 10.70 1.734 1.440 36.39 10.30 49.6 14.98 2.153
160 ….. … 0.718 5.189 21.15 13.33 1.734 1.358 45.30 9.16 59 17.81 2.104
… XXS … 0.864 4.897 18.83 15.64 1.734 1.282 53.16 8.17 66.3 20.03 2.060

… ….. 5S 0.109 8.407 55.5 2.916 2.258 2.201 9.91 24.07 26.45 6.13 3.01
… ….. 10S 0.148 8.329 54.5 3.94 2.258 2.180 13.40 23.59 35.4 8.21 3.00
20 ….. … 0.250 8.125 51.8 6.58 2.258 2.127 22.36 22.48 57.7 13.39 2.962
30 ….. … 0.277 8.071 51.2 7.26 2.258 2.113 24.70 22.18 63.4 14.69 2.953
40 Std 40S 0.322 7.981 50.0 8.40 2.258 2.089 28.55 21.69 72.5 16.81 2.938
8 60 ….. … 0.406 7.813 47.9 10.48 2.258 2.045 35.64 20.79 88.8 20.58 2.909
8.625 80 XS 80S 0.500 7.625 45.7 12.76 2.258 1.996 43.39 19.80 105.7 24.52 2.878
100 ….. … 0.593 7.439 43.5 14.96 2.258 1.948 50.87 18.84 121.4 28.14 2.847
120 ….. … 0.718 7.189 40.6 17.84 2.258 1.882 60.63 17.60 140.6 32.6 2.807
140 ….. … 0.812 7.001 38.5 19.93 2.258 1.833 67.76 16.69 153.8 35.7 2.777
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Appendices-Frame Page 706 Tuesday, March 15, 2005 10:00 AM

706 The John Zink Combustion Handbook

TABLE A.2 (continued) Physical Properties of Pipe

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Wall Sq. Ft. Sq. Ft. Weight Moment
Nominal Schedule Number Thick- Inside Metal outside inside Weight of water of Section Radius
pipe size, ness, Area, Area, surface, surface, per ft, per ft, Inertia, modulus, gyration,
OD, in. a b c in. I.D., in. sq. in. sq. in. per ft per ft lb lb in.4 in.3 in.

… XXS … 0.875 6.875 37.1 21.30 2.258 1.800 72.42 16.09 162 37.6 2.757
160 …. … 0.906 6.813 36.5 21.97 2.258 1.784 74.69 15.80 165.9 38.5 2.748

… ….. 5S 0.134 10.482 86.3 4.52 2.815 2.744 15.15 37.4 63.7 11.85 3.75
… ….. 10S 0.165 10.420 85.3 5.49 2.815 2.728 18.70 36.9 76.9 14.3 3.74
20 ….. … 0.250 10.250 82.5 8.26 2.815 2.683 28.04 35.8 113.7 21.16 3.71
… ….. … 0.279 10.192 81.6 9.18 2.815 2.668 31.20 35.3 125.9 23.42 3.70
30 ….. … 0.307 10.136 80.7 10.07 2.815 2.654 34.24 35.0 137.5 25.57 3.69
10 40 Std 40S 0.365 10.020 78.9 11.91 2.815 2.623 40.48 34.1 160.8 29.9 3.67
10.750 60 XS 80S 0.500 9.750 74.7 16.10 2.815 2.553 54.74 32.3 212 39.4 3.63
80 ….. … 0.593 9.564 71.8 18.92 2.815 2.504 64.33 31.1 244.9 45.6 3.60
100 ….. … 0.718 9.314 68.1 22.63 2.815 2.438 76.93 29.5 286.2 53.2 3.56
120 ….. … 0.843 9.064 64.5 26.24 2.815 2.373 89.20 28.0 324 60.3 3.52
140 ….. … 1.000 8.750 60.1 30.6 2.815 2.291 104.13 26.1 368 68.4 3.47
160 ….. … 1.125 8.500 56.7 34.0 2.815 2.225 115.65 24.6 399 74.3 3.43

… ….. 5S 0.165 12.420 121.2 6.52 3.34 3.25 19.56 52.5 129.2 20.27 4.45
… ….. 10S 0.180 12.390 120.6 7.11 3.34 3.24 24.20 52.2 140.5 22.03 4.44
20 ….. … 0.250 12.250 117.9 9.84 3.34 3.21 33.38 51.1 191.9 30.1 4.42
30 ….. … 0.330 12.090 114.8 12.88 3.34 3.17 43.77 49.7 248.5 39.0 4.39
… Std 40S 0.375 12.000 113.1 14.58 3.34 3.14 49.56 49.0 279.3 43.8 4.38
12 40 ….. … 0.406 11.938 111.9 15.74 3.34 3.13 53.53 48.5 300 47.1 4.37
12.750 … XS 80S 0.500 11.750 108.4 19.24 3.34 3.08 65.42 47.0 362 56.7 4.33
60 ….. … 0.562 11.626 106.2 21.52 3.34 3.04 73.16 46.0 401 62.8 4.31
80 ….. … 0.687 11.376 101.6 26.04 3.34 2.978 88.51 44.0 475 74.5 4.27
100 ….. … 0.843 11.064 96.1 31.5 3.34 2.897 107.20 41.6 562 88.1 4.22
120 ….. … 1.000 10.750 90.8 36.9 3.34 2.814 125.49 39.3 642 100.7 4.17
140 ….. … 1.125 10.500 86.6 41.1 3.34 2.749 139.68 37.5 701 109.9 4.13
160 ….. … 1.312 10.126 80.5 47.1 3.34 2.651 160.27 34.9 781 122.6 4.07

10 ….. … 0.250 13.500 143.1 10.80 3.67 3.53 36.71 62.1 255.4 36.5 4.86
20 ….. … 0.312 13.376 140.5 13.42 3.67 3.5 45.68 60.9 314 44.9 4.84
30 Std … 0.375 13.250 137.9 16.05 3.67 3.47 54.57 59.7 373 53.3 4.82
40 ….. … 0.437 13.126 135.3 18.62 3.67 3.44 63.37 58.7 429 61.2 4.80
… XS … 0.500 13.000 132.7 21.21 3.67 3.4 72.09 57.5 484 69.1 4.78
… ….. … 0.562 12.876 130.2 23.73 3.67 3.37 80.66 56.5 537 76.7 4.76
14 60 ….. … 0.593 12.814 129.0 24.98 3.67 3.35 84.91 55.9 562 80.3 4.74
14.000 … ….. … 0.625 12.750 127.7 26.26 3.67 3.34 89.28 55.3 589 84.1 4.73
… ….. … 0.687 12.626 125.2 28.73 3.67 3.31 97.68 54.3 638 91.2 4.71
80 ….. … 0.750 12.500 122.7 31.2 3.67 3.27 106.13 53.2 687 98.2 4.69
… ….. … 0.875 12.250 117.9 36.1 3.67 3.21 122.66 51.1 781 111.5 4.65
100 ….. … 0.937 12.126 115.5 38.5 3.67 3.17 130.73 50.0 825 117.8 4.63
120 ….. … 1.093 11.814 109.6 44.3 3.67 3.09 150.67 47.5 930 132.8 4.58
140 ….. … 1.250 11.500 103.9 50.1 3.67 3.01 170.22 45.0 1127 146.8 4.53
160 ….. … 1.406 11.188 98.3 55.6 3.67 2.929 189.12 42.6 1017 159.6 4.48

10 ….. … 0.250 15.500 188.7 12.37 4.19 4.06 42.05 81.8 384 48 5.57
20 ….. … 0.312 15.376 185.7 15.38 4.19 4.03 52.36 80.5 473 59.2 5.55
30 Std … 0.375 15.250 182.6 18.41 4.19 3.99 62.58 79.1 562 70.3 5.53
… ….. … 0.437 15.126 179.7 21.37 4.19 3.96 72.64 77.9 648 80.9 5.50
40 XS … 0.500 15.000 176.7 24.35 4.19 3.93 82.77 76.5 732 91.5 5.48
… ….. … 0.562 14.876 173.8 27.26 4.19 3.89 92.66 75.4 813 106.6 5.46
… ….. … 0.625 14.750 170.9 30.2 4.19 3.86 102.63 74.1 894 112.2 5.44
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Appendices-Frame Page 707 Tuesday, March 15, 2005 10:00 AM

Appendix A Physical Properties of Materials 707

TABLE A.2 (continued) Physical Properties of Pipe

16 60 ….. … 0.656 14.688 169.4 31.6 4.19 3.85 107.50 73.4 933 116.6 5.43
16.000 … ….. … 0.687 14.626 168.0 33.0 4.19 3.83 112.36 72.7 971 121.4 5.42
… … 0.750 14.500 165.1 35.9 4.19 3.8 122.15 71.5 1047 130.9 5.40
80 ….. … 0.842 14.314 160.9 40.1 4.19 3.75 136.46 69.7 1157 144.6 5.37
… ….. … 0.875 14.250 159.5 41.6 4.19 3.73 141.35 69.1 1193 154.1 5.36
100 ….. … 1.031 13.938 152.6 48.5 4.19 3.65 164.83 66.1 1365 170.6 5.30
120 ….. … 1.218 13.564 144.5 56.6 4.19 3.55 192.29 62.6 1556 194.5 5.24
140 ….. … 1.437 13.126 135.3 65.7 4.19 3.44 223.50 58.6 1760 220.0 5.17
160 ….. … 1.593 12.814 129.0 72.1 4.19 3.35 245.11 55.9 1894 236.7 5.12

10 ….. … 0.250 17.500 240.5 13.94 4.71 4.58 47.39 104.3 549 61.0 6.28
20 ….. … 0.312 17.376 237.1 17.34 4.71 4.55 59.03 102.8 678 75.5 6.25

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
… Std … 0.375 17.250 233.7 20.76 4.71 4.52 70.59 101.2 807 89.6 6.23
30 ….. … 0.437 17.126 230.4 24.11 4.71 4.48 82.06 99.9 931 103.4 6.21
… XS … 0.500 17.000 227.0 27.49 4.71 4.45 93.45 98.4 1053 117.0 6.19
40 ….. … 0.562 16.876 223.7 30.8 4.71 4.42 104.75 97.0 1172 130.2 6.17
… ….. … 0.625 16.750 220.5 34.1 4.71 4.39 115.98 95.5 1289 143.3 6.15
18 … ….. … 0.687 16.626 217.1 37.4 4.71 4.35 127.03 94.1 1403 156.3 6.13
18.000 60 ….. … 0.750 16.500 213.8 40.6 4.71 4.32 138.17 92.7 1515 168.3 6.10
… ….. … 0.875 16.250 207.4 47.1 4.71 4.25 160.04 89.9 1731 192.8 6.06
80 ….. … 0.937 16.126 204.2 50.2 4.71 4.22 170.75 88.5 1834 203.8 6.04
100 ….. … 1.156 15.688 193.3 61.2 4.71 4.11 207.96 83.7 2180 242.2 5.97
120 ….. … 1.375 15.250 182.6 71.8 4.71 3.99 244.14 79.2 2499 277.6 5.90
140 ….. … 1.562 14.876 173.8 80.7 4.71 3.89 274.23 75.3 2750 306 5.84
160 ….. … 1.781 14.438 163.7 90.7 4.71 3.78 308.51 71.0 3020 336 5.77

10 ….. … 0.250 19.500 298.6 15.51 5.24 5.11 52.73 129.5 757 75.7 6.98
… ….. … 0.312 19.376 294.9 19.30 5.24 5.07 65.40 128.1 935 93.5 6.96
20 Std … 0.375 19.250 291.0 23.12 5.24 5.04 78.60 126.0 1114 111.4 6.94
… ….. … 0.437 19.126 287.3 26.86 5.24 5.01 91.31 124.6 1286 128.6 6.92
30 XS … 0.500 19.000 283.5 30.6 5.24 4.97 104.13 122.8 1457 145.7 6.90
… ….. … 0.562 18.876 279.8 34.3 5.24 4.94 116.67 121.3 1624 162.4 6.88
20 40 ….. … 0.593 18.814 278.0 36.2 5.24 4.93 122.91 120.4 1704 170.4 6.86
20.000 … ….. … 0.625 18.750 276.1 38.0 5.24 4.91 129.33 119.7 1787 178.7 6.85
… ….. … 0.687 18.626 272.5 41.7 5.24 4.88 141.71 118.1 1946 194.6 6.83
… ….. … 0.750 18.500 268.8 45.4 5.24 4.84 154.20 116.5 2105 210.5 6.81
60 ….. … 0.812 18.376 265.2 48.9 5.24 4.81 166.40 115.0 2257 225.7 6.79
… ….. … 0.875 18.250 261.6 52.6 5.24 4.78 178.73 113.4 2409 240.9 6.77
80 ….. … 1.031 17.938 252.7 61.4 5.24 4.70 208.87 109.4 2772 277.2 6.72
100 ….. … 1.281 17.438 238.8 75.3 5.24 4.57 256.10 103.4 3320 332 6.63
120 ….. … 1.500 17.000 227.0 87.2 5.24 4.45 296.37 98.3 3760 376 6.56
140 ….. … 1.750 16.500 213.8 100.3 5.24 4.32 341.10 92.6 4220 422 6.48
160 ….. … 1.968 16.064 202.7 111.5 5.24 4.21 379.01 87.9 4590 459 6.41

10 ….. … 0.250 23.500 434 18.65 6.28 6.15 63.41 188.0 1316 109.6 8.40
… ….. … 0.312 23.376 430 23.20 6.28 6.12 78.93 186.1 1629 135.8 8.38
20 Std … 0.375 23.250 425 27.83 6.28 6.09 94.62 183.8 1943 161.9 8.35
… ….. … 0.437 23.126 420 32.4 6.28 6.05 109.97 182.1 2246 187.4 8.33
… XS … 0.500 23.000 415 36.9 6.28 6.02 125.49 180.1 2550 212.5 8.31
24 30 ….. … 0.562 22.876 411 41.4 6.28 5.99 140.80 178.1 2840 237.0 8.29
24.000 … ….. … 0.625 22.750 406 45.9 6.28 5.96 156.03 176.2 3140 261.4 8.27
40 ….. … 0.687 22.626 402 50.3 6.28 5.92 171.17 174.3 3420 285.2 8.25
… ….. … 0.750 22.500 398 54.8 6.28 5.89 186.24 172.4 3710 309 8.22
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Appendices-Frame Page 708 Tuesday, March 15, 2005 10:00 AM

708 The John Zink Combustion Handbook

TABLE A.2 (continued) Physical Properties of Pipe

60 ….. … 0.968 22.064 382 70.0 6.28 5.78 238.11 165.8 4650 388 8.15
80 ….. … 1.218 21.564 365 87.2 6.28 5.65 296.36 158.3 5670 473 8.07
100 ….. … 1.531 20.938 344 108.1 6.28 5.48 367.40 149.3 6850 571 7.96
120 ….. … 1.812 20.376 326 126.3 6.28 5.33 429.39 141.4 7830 652 7.87
140 ….. … 2.062 19.876 310 142.1 6.28 5.20 483.13 134.5 8630 719 7.79
160 ….. … 2.343 19.314 293 159.4 6.28 5.06 541.94 127.0 9460 788 7.70

10 ….. … 0.312 29.376 678 29.1 7.85 7.69 98.93 293.8 3210 214 10.50
30 20 ….. … 0.500 29.000 661 46.3 7.85 7.59 157.53 286.3 5040 336 10.43
30.000 30 ….. … 0.625 28.750 649 57.6 7.85 7.53 196.08 281.5 6220 415 10.39

a = ASA B36.10 Steel-pipe schedule numbers

b = ASA B36.10 Steel-pipe nominal wall-thickness designations
c = ASA B36.19 Stainless-steel-pipe schedule numbers

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendix A Physical Properties of Materials 709

TABLE A.3 Commercial Copper Tubing*

The following table gives dimensional data and weights of copper tubing used for automotive, plumbing, refrigeration, and heat
exchanger services. For additional data see the standards handbooks of the Copper Development Association, Inc., the ASTM standards,
and the “SAE Handbook.”
Dimensions in this table are actual specified measurements, subject to accepted tolerances. Trade size designations are usually by actual
OD, except for water and drainage tube (plumbing), which measures 1/8-in. larger OD. A 1/2-in. plumbing tube, for example, measures
5/8-in. OD, and a 2-in. plumbing tube measures 2 1/8-in. OD.

KEY TO GAGE SIZES

Standard-gage wall thicknesses are listed by numerical designation (14 to 21), BWG or Stubs gage. These gage sizes are standard for
tubular heat exchangers. The letter A designates SAE tubing sizes for automotive service. Letter designations K and L are the common sizes
for plumbing services, soft or hard temper.

OTHER MATERIALS
These same dimensional sizes are also common for much of the commercial tubing available in aluminum, mild steel, brass, bronze,
and other alloys. Tube weights in this table are based on copper at 0.323 lb/in3. For other materials the weights should be multiplied by the

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
following approximate factors:

aluminum 0.30 monel 0.96

mild steel 0.87 stainless steel 0.89
brass 0.95

Size, OD Wall Thickness Flow Area Metal Surface Area

Area, Inside, Outside, Weight,
in. mm in. mm gage in.2 mm2 in.2 ft2/ft ft2/ft lb/ft

1/8 3.2 .030 0.76 A 0.003 1.9 0.012 0.017 0.033 0.035
3/16 4.76 .030 0.76 A 0.013 8.4 0.017 0.034 0.049 0.058
1/4 6.4 .030 0.76 A 0.028 18.1 0.021 0.050 0.066 0.080
1/4 6.4 .049 1.24 18 0.018 11.6 0.031 0.038 0.066 0.120
5/16 7.94 .032 0.81 21A 0.048 31.0 0.028 0.065 0.082 0.109

3/8 9.53 .032 0.81 21A 0.076 49.0 0.033 0.081 0.098 0.134
3/8 9.53 .049 1.24 18 0.060 38.7 0.050 0.072 0.098 0.195
1/2 12.7 .032 0.81 21A 0.149 96.1 0.047 0.114 0.131 0.182
1/2 12.7 .035 0.89 20L 0.145 93.6 0.051 0.113 0.131 0.198
1/2 12.7 .049 1.24 18K 0.127 81.9 0.069 0.105 0.131 0.269

1/2 12.7 .065 1.65 16 0.108 69.7 0.089 0.97 0.131 0.344
5/8 15.9 .035 0.89 20A 0.242 156 0.065 0.145 0.164 0.251
5/8 15.9 .040 1.02 L 0.233 150 0.074 0.143 0.164 0.285
5/8 15.9 .049 1.24 18K 0.215 139 0.089 0.138 0.164 0.344
3/4 19.1 .035 0.89 20A 0.363 234 0.079 0.178 0.196 0.305

3/4 19.1 .042 1.07 L 0.348 224 0.103 0.174 0.196 0.362
3/4 19.1 .049 1.24 18K 0.334 215 0.108 0.171 0.196 0.418
3/4 19.1 .065 1.65 16 0.302 195 0.140 0.162 0.196 0.542
3/4 19.1 .083 2.11 14 0.268 173 0.174 0.151 0.196 0.674
7/8 22.2 .045 1.14 L 0.484 312 0.117 0.206 0.229 0.455

7/8 22.2 .065 1.65 16K 0.436 281 0.165 0.195 0.229 0.641
7/8 22.2 .083 2.11 14 0.395 255 0.206 0.186 0.229 0.800
1 25.4 .065 1.65 16 0.594 383 0.181 0.228 0.262 0.740
1 25.4 .083 2.11 14 0.546 352 0.239 0.218 0.262 0.927
1 1/8 28.6 .050 1.27 L 0.825 532 0.176 0.268 0.294 0.655

* Compiled and computed.

From: The CRC Handbook of Mechanical Engineering, CRC Press, Boca Raton, FL, 1998.

Copyright CRC Press

710 The John Zink Combustion Handbook

TABLE A.3 (continued) Commercial Copper Tubing*

Size, OD Wall Thickness Flow Area Metal Surface Area
Area, Inside, Outside, Weight,
in. mm in. mm gage in.2 mm2 in2 ft2/ft ft2/ft lb/ft

1 1/8 28.6 .065 1.65 16K 0.778 502 0.216 0.261 0.294 0.839
1 1/4 31.8 .065 1.65 16 0.985 636 0.242 0.293 0.327 0.938
1 1/4 31.8 .083 2.11 14 0.923 596 0.304 0.284 0.327 1.18
1 3/8 34.9 .055 1.40 L 1.257 811 0.228 0.331 0.360 0.884
1 3/8 34.9 .065 1.65 16K 1.217 785 0.267 0.326 0.360 1.04

1 1/2 38.1 .065 1.65 16 1.474 951 0.294 0.359 0.393 1.14
1 1/2 38.7 .083 2.11 14 1.398 902 0.370 0.349 0.393 1.43
1 5/8 41.3 .060 1.52 L 1.779 1148 0.295 0.394 0.425 1.14
1 5/8 41.3 .072 1.83 K 1.722 1111 0.351 0.388 0.425 1.36
2 50.8 .083 2.11 14 2.642 1705 0.500 0.480 0.628 1.94

2 50.8 .109 2.76 12 2.494 1609 0.620 0.466 0.628 2.51

2 1/8 54.0 .070 1.78 L 3.095 1997 0.449 0.520 0.556 1.75
2 1/8 54.0 .083 2.11 14K 3.016 1946 0.529 0.513 0.556 2.06
2 5/8 66.7 .080 2.03 L 4.77 3078 0.645 0.645 0.687 2.48
2 5/8 66.7 .095 2.41 13K 4.66 3007 0.760 0.637 0.687 2.93

3 1/8 79.4 .090 2.29 L 6.81 4394 0.950 0.771 0.818 3.33
3 1/8 79.4 .109 2.77 12K 6.64 4284 1.034 0.761 0.818 4.00
3 5/8 92.1 .100 2.54 L 9.21 5942 1.154 0.897 0.949 4.29
3 5/8 92.1 .120 3.05 11K 9.00 5807 1.341 0.886 0.949 5.12
4 1/8 104.8 .110 2.79 L 11.92 7691 1.387 1.022 1.080 5.38
4 1/8 104.8 .134 3.40 10K 11.61 7491 1.682 1.009 1.080 6.51
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendix A Physical Properties of Materials 711

TABLE A.4 Standard Grades of Bolts

Part a: SAE Grades for Steel Bolts

From: The CRC Press Handbook of Mechanical Engineering, CRC Press, Boca Raton, FL, 1998.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

712 The John Zink Combustion Handbook

TABLE A.5 Standard Grades of Bolts

Part b: ASTM Grades for Steel Bolts
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

From: The CRC Press Handbook of Mechanical Engineering, CRC Press, Boca Raton, FL, 1998.

Copyright CRC Press

Appendix A Physical Properties of Materials 713

TABLE A.6 Standard Grades of Bolts

Part c: Metric Mechanical Property Classes for Steel Bolts, Screws, and Studs

From: The CRC Press Handbook of Mechanical Engineering, CRC Press, Boca Raton, FL, 1998.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendix B
Properties of Gases and Liquids
TABLE B.1 Combustion Data for Hydrocarbons*

* Based largely on: Gas Engineers’ Handbook, American Gas Association, Inc., Industrial Park, 1967.

REFERENCES

American Institute of Physics Handbook, 2nd ed., D.E. Gray, Ed., McGraw-Hill Book Company, NY, 1963.
Chemical Engineer’s Handbook, 4th ed., R.H. Perry, C.H. Chilton, and S.D. Kirkpatrick, Eds., McGraw-Hill Book Company,
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

NY, 1963.
Handbook of Chemistry and Physics, 53rd ed., R.C. Weast, Ed., The Chemical Rubber Company, Cleveland, OH, 1972;
gives the heat of combustion of 500 organic compounds.
Handbook of Laboratory Safety, 2nd ed., N.V. Steere, Ed., The Chemical Rubber Company, Cleveland, OH, 1971.
Physical Measurements in Gas Dynamics and Combustion, Princeton University Press, 1954.

Note: For heating value in J/kg, multiply the value in Btu/lbm by 2324. For flame speed in m/s, multiply the value in ft/s by 0.3048.
From: The CRC Press Handbook of Mechanical Engineering, CRC Press, Boca Raton, FL, 1998.

715
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Appendices-Frame Page 716 Tuesday, March 15, 2005 10:00 AM

716 The John Zink Combustion Handbook

TABLE B.2 Enthalpy of Formation, Gibbs Function of Formation,

and Absolute Entropy of Various Substances at 298 K and 1 atm
o
h f and g of (kJ/kmol), s o (kJ/kmol•K)
o
Substance Formula hf g of so

Carbon C(s) 0 0 5.74

Hydrogen H2(g) 0 0 130.57
Nitrogen N2(g) 0 0 191.50
Oxygen O2(g) 0 0 205.03
Carbon monoxide CO(g) –110,530 –137,150 197.54
Carbon dioxide CO2(g) –393,520 –394,380 213.69
Water H2O(g) –241,820 –228,590 188.72
H2O(l) –285,830 –237,180 69.95
Hydrogen peroxide H2O2(g) –136,310 –105,600 232.63
Ammonia NH3(g) –46,190 –16,590 192.33
Oxygen O(g) 249,170 231,770 160.95
Hydrogen H(g) 218,000 203,290 114.61
Nitrogen N(g) 472,680 455,510 153.19
Hydroxyl OH(g) 39,460 34,280 183.75
Methane CH4(g) –74,850 –50,790 186.16
Acetylene C2H2(g) 226,730 209,170 200.85
Ethylene C2H4(g) 52,280 68,120 219.83
Ethane C2H6(g) –84,680 –32,890 229.49
Propylene C3H6(g) 20,410 62,720 266.94
Propane C3H8(g) –103,850 –23,490 269.91
Butane C4H10(g) –126,150 –15,710 310.03
Pentane C5H12(g) –146,440 –8200 348.40
Octane C8H18(g) –208,450 17,320 463.67
C8H18(l) –249,910 6610 360.79
Benzene C6H6(g) 82,930 129,660 269.20
Methyl alcohol CH3OH(g) –200,890 –162,140 239.70
CH3OH(l) –238,810 –166,290 126.80
Ethyl alcohol C2H5OH(g) –235,310 –168,570 282.59
C2H5OH(l) –277,690 174,890 160.70

Source: Adapted from Wark, K. 1983. Thermodynamics, 4th ed. McGraw-Hill, New
York, as based on JANAF Thermochemical Tables, NSRDS-NBS-37, 1971;
Selected Values of Chemical Thermodynamic Properties, NBS Tech. Note 270-3,
1968; and API Research Project 44, Carnegie Press, 1953.
From: The CRC Handbook of Thermal Engineering CRC Press, Boca Raton, FL, 2000.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendix B Properties of Gases and Liquids 717

TABLE B.3 Properties of Dry Air at Atmospheric Pressure

Symbols and Units:
K = absolute temperature, degrees Kelvin
deg C = temperature, degrees Celsius
deg F = temperature, degrees Fahrenheit
ρ = density, kg/m3
cp = specific heat capacity, kJ/kg·K
cp/cv = specific heat capacity ratio, dimensionless
µ = viscosity, N·s/m2 × 106 (For N·s/m2 (= kg/m·s) multiply tabulated values by 10–6)
k = thermal conductivity, W/m·k × 103 (For W/m·K multiply tabulated values by 10–3)
Pr = Prandtl number, dimensionless
h = enthalpy, kJ/kg
Vs = sound velocity, m/s

*Condensed and computed from: Tables of Thermal Properties of Gases, National Bureau of Standards Circular 564, U.S. Government
Printing Office, November 1955.
From: The CRC Press Handbook of Thermal Engineering, CRC Press, Boca Raton, FL, 2000.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

718 The John Zink Combustion Handbook

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
TABLE B.3 (continued) Properties of Dry Air at Atmospheric Pressure

Copyright CRC Press

Appendix B Properties of Gases and Liquids 719

TABLE B.4 Chemical, Physical, and Thermal Properties of Gases: Gases and Vapors, Including
Fuels and Refrigerants, English and Metric Units

Note: The properties of pure gases are given at 25°C (77°F, 298 K) and atmospheric pressure (except as stated).
From: The CRC Press Handbook of Thermal Engineering, CRC Press, Boca Raton, FL, 2000.

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

720 The John Zink Combustion Handbook

TABLE B.4 (continued) Chemical, Physical, and Thermal Properties of Gases: Gases and Vapors,
Including Fuels and Refrigerants, English and Metric Units

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendix B Properties of Gases and Liquids 721

TABLE B.4 (continued) Chemical, Physical, and Thermal Properties of Gases: Gases and Vapors,
Including Fuels and Refrigerants, English and Metric Units

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

722 The John Zink Combustion Handbook

TABLE B.4 (continued) Chemical, Physical, and Thermal Properties of Gases: Gases and Vapors,
Including Fuels and Refrigerants, English and Metric Units
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendix B Properties of Gases and Liquids 723

TABLE B.4 (continued) Chemical, Physical, and Thermal Properties of Gases: Gases and Vapors,
Including Fuels and Refrigerants, English and Metric Units

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

724 The John Zink Combustion Handbook

TABLE B.4 (continued) Chemical, Physical, and Thermal Properties of Gases: Gases and Vapors,
Including Fuels and Refrigerants, English and Metric Units

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Appendix C
Common Conversions
1 Btu = 252.0 cal 1 in. = 2.540 cm
1055 J 25.40 mm
1 Btu/ft3 = 0.00890 cal/cm3 1J= 0.000948 Btu
0.0373 MJ/m3 0.239 cal
1 Btu/hr = 0.0003931 hp 1 W/sec
0.2520 kcal/hr 1 kcal = 3.968 Btu
0.2931 W 1000 cal
1,000,000 Btu/hr = 0.293 MW 4187 J
1 Btu/hr-ft2 = 0.003153 kW/m2 1 kcal/hr = 3.968 Btu/hr
1 Btu/hr-ft-°F = 1.730 W/m-K 1.162 J/sec
1 Btu/hr-ft2-°F = 5.67 W/m2-K 1 kcal/m3 = 0.1124 Btu/ft3
1 Btu/lb = 0.5556 cal/g 4187 J/m3
2326 J/kg 1 kg = 2.205 lb
1 Btu/lb-°F = 1 cal/g-°C 1 kg/hr-m = 0.00278 g/sec-cm
4187 J/kg-K 0.672 lb/hr-ft
1 cal = 0.003968 Btu 1 kg/m3 = 0.06243 lb/ft3
4.187 J 1 kW = 3413 Btu/hr
1 cal/cm2-sec = 3.687 Btu/ft2-sec 1.341 hp
41.87 kW/m2 660.6 kcal/hr
1 cal/cm-sec-°C = 241.9 Btu/ft-hr-°F 1 kW/m2 = 317.2 Btu/hr-ft2
418.7 W/m-K 1 kW/m2-°C = 176.2 Btu/hr-ft2-°F
1 cal/g = 1.80 Btu/lb 1 lb = 0.4536 kg
4187 J/kg 1 lb/ft3 = 0.0160 g/cm3
1 cal/g-°C = 1 Btu/lb-°F 16.02 kg/m3
4187 J/kg-K 1 lbm/hr-ft = 0.413 centipoise
1 centipoise = 2.421 lbm/hr-ft 1 m= 3.281 ft
1 cm2/sec = 100 centistokes 1 mm = 0.03937 in.
3.874 ft2/hr 1 m2/sec = 10.76 ft2/sec
1 ft = 0.3048 m 1 mton = 1000 kg
1 ft2/sec = 0.0929 m2/sec 2205 lb
1 g/cm3 = 1000 kg/m3 1 MW = 3,413,000 Btu/hr
62.43 lb/ft3 1000 kW
0.03613 lb/in.3 1 therm = 100,000 Btu
1 hp = 33,000 ft-lb/min 1W= 1 J/sec
550 ft-lb/sec 1 W/m-K = 0.5778 Btu/ft-hr-°F
641.4 kcal/hr
745.7 W

TEMPERATURE CONVERSIONS

°C = 5/9 (°F – 32) °F = 9/5°C + 32

K = °C + 273.15 °R = °F + 459.67
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

From: Baukal, C.E., Heat Transfer in Industrial Combustion, CRC Press, Boca Raton, FL, 2000.

725
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Appendices-Frame Page 726 Tuesday, March 15, 2005 10:00 AM

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Glossary
3 “T”s: Time, Temperature, Turbulence calorie: the amount of energy required to raise the tem-
5 “M”s: Meter, Mix, Maintain, Mold, Minimize perature of 1 gram of water by 1°C. The kilocalorie
absolute pressure: the pressure measured relative to a (kcal) is a typical unit of measure in the process
perfect vacuum. Absolute pressures are always pos- industry, 1 kcal = 1000 calories.
itive. The British units for absolute pressure are psia. combustion: the rapid reaction of fuel and oxidant (usu-
atmospheric pressure: the force exerted per unit area by ally oxygen in air) to produce light, heat, and noise.
the atmospheric gases at the earth’s surface. Atmo- Major products of combustion for hydrocarbon fuels

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
spheric pressure varies with altitude. At sea level it (e.g., natural gas, refinery gas, fuel oils) are carbon
is about 14.7 pounds per square inch. At one mile dioxide (CO2) and water vapor (H2O). Trace products
above sea level, it is about 12.2 pounds per square include carbon monoxide (CO) and nitrogen oxides
inch. (NO and NO2), which are pollutants.
atomization: the process whereby a volume of liquid is combustion efficiency: the fraction of carbon in the fuel
converted into a multiplicity of small drops. The that is converted to CO2 in the flue gas, customarily
principal goal is to produce a high surface area to expressed as a percent.
mass ratio so that the liquid will vaporize quickly conduction: the transfer of heat by molecular collision.
and thus be susceptible to combustion. This process is more efficient in metals and other
atomizer: part of an oil gun which breaks up the fuel oil thermal “conductors” and poorer in fluids and insu-
flow into tiny particles by both mechanical means lators such as refractories.
and the use of an atomizing medium. The oil and
convection: the transfer of heat or mass by large scale
atomizing medium mix together in the atomizer and
fluid movements. When the process occurs due to
then flow to the oil tip to be discharged into the
density and temperature differences, it is termed
furnace.
natural convection. When the process occurs due to
audible sound: vibrations in a gas, liquid, or solid with
external devices (such as fans), it is termed forced
components falling in the frequency range of 16 Hz
convection.
to 20 kHz.
convection section: the part of a furnace between the
beta ratio (β): for a single orifice the beta ratio is the ratio
radiant section and the stack. The area is filled with
of the orifice bore diameter to that of the upstream
tubes or pipes which carry a process stream and
pipe diameter. However, since in burner designs typ-
which absorb heat via convection heat transfer from
ically there is more than one orifice at a riser pipe
the hot gases passing through the area on their way
exit, the beta ratio is equal to the square root of the
out the stack. The convection section forms an
ratio between the total area of the fuel ports to that
obstacle to the combustion gas flow and can greatly
of the upstream pipe area.
affect furnace draft in the radiant section of the
Btu (British Thermal Unit): standard measure of energy
furnace.
in the British unit system. 1 Btu is the amount of
heat required to raise the temperature of liquid water dB(A): “A” weighted average of the sound pressure levels
by 1°F. over the entire frequency band. Intended to be a
burner: a device which combines fuel and air in proper more accurate representation of how a human hears
proportions for combustion and which enables the the sound.
fuel–air mixture to burn stably to give a specified Decibel: unit of sound pressure or power. Abbreviation is
flame size and shape. ‘dB’. 1W of sound power is equal to 120 dB. A
burner block: also called “burner tile,” “muffler block,” Log10 scale relates the unit of Watts and dB. Con-
or “quarl.” The specially formed refractory pieces sequently, an increase or decrease of 10 W equates
which mount around the burner opening inside the to a 10 dB difference, while a change of 100W
furnace. The burner block forms the burner’s airflow equates to a 20 dB difference.
opening and helps stabilize the flame. decibel (dB): unit of measure for sound pressure level.
burner capacity: amount of heat release a burner can Developed by Bell Laboratories.
deliver (i.e., amount of fuel which can be completely diffusion (raw gas) flame: combustion state controlled by
burned through a burner) at a given set of operating mixing phenomena. Fuel and air diffuse into one
conditions. another until a flammable mixture ratio is achieved.
727
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Appendices-Frame Page 728 Tuesday, March 15, 2005 10:00 AM

728 The John Zink Combustion Handbook

emissivity: the efficiency with which a material radiates thermal heat liberation: amount of heat released during combustion of
energy, expressed as a fraction between 0 and 1. fuels. One of the criteria for determining what burner to
excess air: the amount of air needed by a burner which is in use in an application. (Also called heat release.)
excess of the amount required for perfect or stoichiometric Higher Heating Value (HHV): the theoretical heat the com-
combustion. Some amount of excess air, depending on the bustion process can release if the fuel and oxidant are
available fuel/air mixing energy, is required to assure thor- converted with 100% efficiency to CO2 and liquid H2O.
ough mixing of the fuel and air for complete combustion. ignition temperature: the temperature required to initiate
flame speed: the rate at which a flame can propagate in a com- combustion.
bustible mixture. If the flame speed is lower than the speed laminar flow: very smooth flow in which all the molecules
of the reacting flow, the flame may lift off the burner. If are traveling in generally the same direction. For internal
the flame speed is higher than the speed of the reacting flows, it occurs at Reynolds numbers less than 2000.
flow the flame may flash back into the burner. lift-off: this condition occurs when the fuel or fuel/air mixture
flammability limits (upper and lower): the upper and lower velocity is too high, thus allowing the fuel to exit the
bounds of the fuel/air mixture which will support combus- stabilizing zone before it has achieved its ignition
tion. The upper flammability limit indicates the maximum temperature.
fuel concentration in air that will support combustion. The Lower Heating Value (LHV): the theoretical heat the combus-
lower flammability limit indicates the minimum fuel con- tion process can release if the fuel and oxidant are con-

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
centration in air that will support combustion. Outside these verted with 100% efficiency to CO2 and H2O vapor. The
bounds the mixture does not burn. process industries generally prefer to use LHV.
flashback: phenomenon occurring only in pre-mix gas burners mixer: the part of a pre-mix burner (also “Gas–Air Mixer”)
when the flame speed overcomes the gas–air mixture flow which uses the kinetic energy of the high velocity fuel gas
velocity exiting the gas tip. The flame rushes back to the stream to draw in part or all of the air required by the
gas orifice and can make an explosive sound when flash- burner for combustion.
back occurs. Flashback is most common when hydrogen noise: any undesirable sound.
is present in fuel gas. normal cubic meter (Nm3): the quantity of a gas that is present
flashing: the process whereby a drop in pressure or increase in in 1 m3 at the thermodynamic conditions of 1 atm and 0°C.
temperature causes vaporization. For an ideal gas there are 22.41 Nm3 in 1 kmol.
fuel NOx: NOx that is formed from nitrogen that is organically NOx: any combination of nitrogen and oxygen in a compound
bound to the fuel molecule. Fuel NOx is most often a prob- form. The most common in terms of environmental con-
lem with liquid fuel or coal firing. Once the nitrogen has siderations is NO, which constitutes 90% of combustion
been cracked from the fuel molecule, the mechanism follows NOx emissions, and NO2. All NO is eventually converted
basically the same path as the prompt NOx mechanism. to NO2 in the atmosphere. Hence, most regulations are
furnace arch: uppermost part of a radiant furnace (also called written to assume that the NOx which is emitted is in the
the “Bridgewall,” a term which came from the original form of NO2. NOx emissions are influenced by many
furnace designs and has remained in use). The last area in factors, including furnace temperature, flame temperature,
an upflow furnace before the convection section. burner design, combustion air temperature, nitrogen con-
furnace draft: the negative air pressure generated by buoyancy tent of liquid fuels, ammonia content of gas fuels, and
of hot gases inside a furnace. The temperature difference other factors.
between gases within the furnace and in the atmosphere oil block: usually a monolithic block located at the center of a
along with furnace and stack height basically determine burner assembly. The oil block acts to stabilize the oil
the amount of draft generated by a furnace. Draft is flame. (Also call the “Oil Tile”)
generally measured in negative inches of water column oil gun: the assembly of parts in a burner which provides atom-
(“-w.c.”; 27.7 inches w.c. = 1 psig). ized fuel oil mixture to the furnace for burning.
furnace or fired heater: a piece of process equipment which is oil tip: part of the oil gun which discharges the atomized fuel
used to heat any of the various process streams in refineries oil mixture into the furnace through multiple openings.
and chemical plants. Furnaces most commonly utilize The hole pattern in the tip has a great effect on the flame
direct combustion of fuels to generate the required heat. size and shape.
gage pressure: the pressure measured relative to the local atmo- orifice discharge coefficient (Cd): the ratio of the actual flow
spheric pressure. Gage pressure may be negative. A neg- through an orifice to that of the theoretical or isentropic
ative gage pressure is known as a suction or vacuum. flow through the orifice. Basically, this parameter is a
gas tip: the part of a burner which discharges the gas fuel via measure of the orifice efficiency. Values are dimensionless
one or more openings into the furnace. The size, arrange- and range from 0.61 for a thin-plate orifice to 0.85 for
ment, and angular disposition of the openings in the tip thick-plate square-edged orifices, and up to 0.90–0.95 for
have a major effect on the size and shape of the flame. tapered orifices.
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Appendices-Frame Page 729 Tuesday, March 15, 2005 10:00 AM

Glossary 729

PAH (polycyclic aromatic hydrocarbons): the carcinogenic sound frequency: the number of pressure waves that pass by
byproducts of some very sub-stoichiometric combustion in a given time. Hertz is the unit of frequency. One wave
processes. Usually absent in process burner flames. per second is one Hertz.
particulates: the residue left over from coal and fuel oil sound power level: the intensity of the sound given off in all
combustion. the directions at the source.
Pascals: a unit of pressure. One Pascal (Pa) is equal to a force sound pressure level: a measure of the acoustical “disturbance”
of one Newton per square meter. produced at a point removed from the source.
pre-mixed flame: before ignition, the fuel and air are intimately sound wave: a wave moving at the speed of sound in a given
mixed. The combustion process is controlled by heat con- medium.
duction and diffusion of radicals.
SSU (seconds, Saybolt Universal): units of kinematic viscosity.
pressure, gas: the force exerted per unit area on a surface cre-
ated by the collision of gas molecules with that surface. stack: the “chimney” or “flue” of a furnace. The stack contains
pressure, static: the pressure of a gas measured at a point where the damper which controls furnace draft.
the gas velocity is zero. stack loss: the fraction of total heat which exits with the flue
pressure, total: the sum of the static pressure and the velocity gas through the stack. The quantity is customarily
pressure of a gas. expressed as a percent of the total heat input. The stack
pressure, velocity or dynamic: the pressure of a flowing gas loss is directly proportional to the stack exit temperature;
attributed to the impact of the gas molecules resulting the higher the temperature, the greater the stack loss.
from the velocity of the gas flow. PV = ρV2/2gc, where ρ staged air: NOx reduction technique predominantly used for
is the density of the flowing gas, V is the velocity, and gc fuel oil firing. The fuel is injected into a fuel-rich primary
is the gravitational constant. zone. This stoichiometry helps to control the fuel NOx
prompt NOx: NOx formed at the initial stages of combustion mechanism. When firing gas, staged air produces higher
that can not be explained by either the thermal mechanism NOx emissions than staged fuel.
or the fuel NOx mechanism. The prompt NOx mechanism staged fuel: NOx reduction technique whereby a small portion
requires the CH radical as an intermediate, so the fuel must of the fuel is injected in a lean primary combustion zone.
have carbon present to create prompt NOx (see Chapter 6). The flue products from this region flow to the secondary
radiant section: the part of a process heater into which the combustion zone where the remainder of the fuel is burned
burners fire. Tubes mounted in this area of the furnace out. The lengthening of the flame creates cooler flame
receive heat principally via direct radiation from both temperatures, thus lowering the thermal NOx.
burner flames and furnace refractory. Physical volume standard cubic foot (SCF): the quantity of a gas that is present
arrangement of the radiant section has a great effect on in 1 ft3 at the thermodynamic conditions of 14.696 psia
burner choice and required flame patterns. and 60°F. For an ideal gas there are 379.7 SCF in 1 lb-mol.
radiation: all warm bodies emit light (electromagnetic radiation steam quality: the fraction of saturated steam that is in the vapor
– mostly infrared). When this radiation is absorbed or state.
emitted by a body, heat is transferred and termed “heat
transfer by radiation.” Such heat transfer requires a line of theoretical flame temperature (adiabatic flame temperature):
sight (view factor) and is proportional to the fourth power the temperature the flame can achieve if it transfers no
of the absolute temperature difference between bodies and heat to its surroundings.
the emissivity of the bodies (see Chapter 3). thermal conductivity: the ability of a material to conduct heat,
ratio of specific heats (k): also known as the isentropic coeffi- expressed as thermal power conducted per unit tempera-
cient. Is equal to the quotient of the heat capacity at con- ture and thickness. Metals and other thermal “conductors”
stant pressure and the heat capacity at constant volume have a large thermal conductivity. Refractories and other
(Cp/Cv). This parameter is tabulated for many pure com- thermal “insulators” have a low thermal conductivity.
ponents at standard conditions, but is technically depen- thermal efficiency: the fraction of total heat input absorbed by
dent on the gas composition and temperature. The values the material being heated. The quantity is customarily
are dimensionless and range from 1.0 to 1.6. expressed as a percent.
sonic flow: when the flow velocity is equal to the speed of sound. thermal NOx: NOx formed via the Zeldovich mechanism. The
The point at which the flow turns sonic is called the critical rate limiting step in this mechanism is the formation of
pressure. This transition occurs at about 12.2 psig for the O radical. This occurs only at high temperatures (above
natural gas at 60°F.
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

about 2400°F). Hence the term thermal NOx, since it is

sound: an alteration of density in an elastic medium that prop- NOx produced in the highest temperature regions of the
agates through the medium (see Chapter 7). flame (see Chapter 6).
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Appendices-Frame Page 730 Tuesday, March 15, 2005 10:00 AM

730 The John Zink Combustion Handbook

thermo acoustic efficiency: equal to the sound power level/heat velocity thermocouple (suction pyrometer): a device for mea-
release. A value used to characterize the amount of com- suring furnace gas temperature. It is comprised of a
bustion noise emitted from a flame. Defined as the ratio thermocouple which has been recessed into an insulating

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
of the acoustical power emitted from the flame to the total shroud, and a suction device such as an eductor which
heat release of the flame. Approximately equal to 1 × 10–6 aspirates large volumes of furnace gas through the shroud
for premixed and turbulent flames and equal to 1 × 10–9 and past the thermocouple. The high velocity of gas
for diffusion and laminar flames. ensures good convective heat transfer to the thermocou-
tramp air: any air which enters (infiltrates) the furnace through ple. The shroud blocks radiant exchange between the ther-
leaks. This air may be measured by the O2 analyzer and mocouple and the surrounding furnace. The velocity
often contributes to the burning of the fuel. thermocouple represents the most accurate means to mea-
turbulent flow: characteristically random flow patterns that sure flue gas temperature. Bare thermocouples are unac-
form eddies from large to small scales. For internal flows, ceptable for this purpose, being in error often by more
it occurs at Reynolds numbers greater than 4000. Turbu- than 100°F due to radiation losses.
lence is integral to the mixing process between the fuel view factor: the fraction of one surface that is visible to another
and air for combustion. (see Chapter 3).
UHC: any unburned hydrocarbon that is emitted in a combustion Watt: unit of measure for power, equal to 1 Joule of energy per
process. Also termed VOC (volatile organic compound). second.

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Index

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Index
A heating, 524 ANOVA, see Analysis of variance
infiltration, 601 ANSI, see American National Standards
Absolute viscosity, 119, 122
leaks, 494 Institute
Accident, factors contributing to, 328
low excess, 200 APH, see Air preheater
ACERC, see Advanced Combustion
metering, 356 API, see American Petroleum Institute
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Engineering Research Center

mixer assembly, 355 Aromatic hydrocarbons, 215
Acetaldehyde, 216
-only systems, 533 Aromatics, 177
Acetone, 216
pollution, environmental regulations combustion data for, 45
Achimedes’ principle, 129
limiting, 362 properties of, 713
Acid gas(es), 639
preheat, 62 Arrestors, 629
composition of used in CFD study, 309
effects, 55 Asphalt, 5
removal, 657, 663, 665
reduction, 200 Asphyxiation, 214
systems, 682
temperature, 195 Atmospheric attenuation, 231, 246
Acrylonitrile, 667 preheater (APH), 438, 561, 562
Actuator, 385 Atmospheric distillation, 5
pressurized, 584 Atmospheric pressure, 148, 715–716
Additives, 199 properties of dry, 715
Adiabatic equilibrium reaction process, 54 Atomic weights, list of for common
-quality regulations, 198 elements, 36–37
Adiabatic flame temperature (AFT), 46, 184, registers, 458, 464
185, 213 Atomizer(s)
staging, 517 eternal mix, 586
Adjustable-plug venturi quench, 656 temperature, combustion, 476
Adsorption, with solid desiccants, 158 fitted onto oil guns, 585
valve internal mix, 584
Advanced Combustion Engineering butterfly-type combustion, 391
Research Center (ACERC), 291 spill-return, 585
characterizer, 392, 393
Aeration rate, 274 Atomizing medium block valve, 486
VOC-laden, 280
Aerospace applications, industrial Autoignition
Alcohols, 216
combustion and, 8 ethylene oxide plant explosion caused
Alkanes
AFT, see Adiabatic flame temperature by, 343
combustion data for, 45
Afterburn, 472 temperature, of methane, 56
non-methane, 404
Agency approvals and safety, 379 Automatic steam control, 627
properties of, 713
Air Autooxidation, 343
Alkenes
ambient atmospheric, 352 Available heat, 64, 65
combustion data for, 45
-assisted flare Averaged velocity, 140
properties of, 713
burner, 608 Aviation gasolines, 175
Alkylation, 5, 13
horizontal settling drum at base of, Avogadro’s number, 35
Ambient air firing, 533
618 Axial flow burners, 565
Ambient atmospheric air, 352
augmenting, 533 American National Standards Institute
blast, 586 (ANSI), 540
control, 358, 456 American Petroleum Institute (API), 190,
control device 339, 358, 597
B
picture of, 480 Ammonia injection, use of to control NOx, Back surfaces, radiant exchanges between,
schematic, 480 413 90
dampers, 464 Analog controllers, 394 Baghouse, 657, 658
density, combustion, 264 Analog control systems, 376 Balanced-draft heater, logic diagram for
excess, 471, 555, 556 Analog devices, 385 tuning, 492
flare(s), 272 Analog loop, 378 Benzene, 216
annular, 611 Analysis of variance (ANOVA), 409, 411 Bernoulli equation, 130, 131
comparison of to prediction, 276 for factorial design, 414 Bessel functions, 80
flow with separate model effects, 409 Beta ratio, 153
balancing of by windbox model, 567 table, testing separately for bias in, 413 Bias, 412
control, 480 Analytical instruments, types of, 389 Blackbody(ies)
distribution, to burners, 564 Analyzer(s) emissive power graph, 103
rate, air valve position vs., 392 in situ extractive, 472 radiation, 88, 91, 95
rate, O2 trim of, 395 oxygen, 473 source, 90
/fuel mixture, bigger explosions of, 235 portable, 472 Blockage, 532
-to-fuel ratio, 38, 257 Annular air flare, 611 Blocking generator, 418
/gas mixtures, 605 Annunciators, 383 Blood-borne pathogens, 346

733
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 734 Tuesday, March 15, 2005 10:02 AM

734 The John Zink Combustion Handbook

Blower(s), 670 BOOS, see Burner out-of-service grid

centrifugal, relative characteristics of, 673 Boussinesq hypothesis, 292 gas flame from 531
inlet, 674 Box heater, 17 heat distribution by, 532
location, 671 Breathing, 502 heat input, average flame length as function
speed control, 677 Bridgewall temperature, 496 of, 551
Bluff body, 67 Bulirsch-Stoer technique, 282 high pressure drop, 643
Boiler(s), 648 Buoyancy, 129 horizontal floor-fired, 14
cleanliness, 561 Burnback, 67 ignition
design, impacts of on NOx emissions Burner(s), 15–28 attempt, 336
correlations, 553 air-assisted flare, 608 ledge, 482
firetube, 648, 650 air dampers, 457 improper installation of, 451
fluidized bed, 524 air registers, 457 improvement of mass flow distribution to,
load, influence of on NOx, 558 arrangements, 20 563
low-NOx retrofit, 569 axial flow, 565 in-duct, 524
multi-burner, 559 block(s) inline, 530
municipal solid waste, 413 for floor-mounted service, 367 installation, 450, 456
Nebraska, 584 material, 370 light-off sequence, simplified flow diagram
performance, effects of burner retrofits on, valves, 485, 495 of standard, 377
553 boiler, see Boiler burners low-NOx, 552, 575
single-burner industrial, 549 bolt circle, on new heater, 451 low pressure drop, 642
-specific burner requirements, 546 capacity curves burner manufacturers use for maintenance activities, 459
tube, 218, 532 sizing, 262 management system, 540

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
utility, 549, 572 CFD simulation of, 306 manufacturers, 111, 262
watertube, 649 combination, 418, 482, 643 materials, 369
Boiler burners, 547–586 combustion medium pressure drop, 643
atomizers for boiler burners, 584–586 mounted on top of heater, 452
instability noise, 237
boiler design impacts on NOx emissions mounting, 451
noise, 236
correlations, 553–563 natural draft, 4, 357, 358, 480
roar, 237
boiler design, 553–554 noise
competing priorities, 15–16
boiler load influence on NOx, 558–561 abatement techniques, 242
condition, 495
boiler/system condition impacts on curve, 231
configuration, by heater type, 12
combustion and NOx formation, example, 244
connection of to heater, 455
561–563 NOx, 202
controller, 374
excess air, 555–558 example, 402
conventional, round flame, 367
boiler-specific burner requirements, response from, 404
COOL TECHNOLOGY, 518, 519
548–553 oil
crude unit, 19
conventional burner technology for common, 479
damper position, 495
boilers, 548–550 -fired, experiencing flame lift-off, 513
design factors, 16–21 needing service, 28
design features of low NOx burners,
development of for petrochemical troubleshooting for, 520
552–553
applications, 291 operation, ultra-low emissions, 582
effects of burner retrofits on boiler
performance, 553 diffusion, 25, 67, 481, 482 out-of-service (BOOS), 202, 573
low-NOx burner technology for boilers, downfired, 369 partially premixed, 25
550–551 drawing, 450 performance, 459
staged burner design philosophy, duct, 524, 534 physical model of, 534
551–552 arrangement, 533 pilot, 465, 483, 514
current state-of-the-art concepts for multi- drilled pipe, 536 piping, 453
burner boilers, 563–575 fluidized bed startup, 529 plenum, viewing oil flame through, 174
combustion optimization, 563–568 low emission, 536 ports, 509
methods to reduce NOx emissions, eduction processes in premixed, 257 potential problems, 27–28
568–575 elevation, 470 pre-installation work, 450
low-NOx burners for packaged industrial firing heavy oil, 180 premix, 21, 22, 480, 481
boilers, 575–580 flame lift-off from, 513 pressure drop, 252, 260
ultra-low emission gas burners, 580–584 flare, 605 problem, 568
background, 580 flat flame, 368 radiant wall, 103, 260, 369
implementation of NOx formation forced draft, 358, 437, 439, 480 raw gas, 206, 357, 364, 366, 481, 482, 515
theory for ultra-low-NOx emissions, free-standing flat flame, 368 refractory, 494, 506
580–581 fuel flows of, 67 requirements, boiler-specific, 546
ultra-low emissions burner design, fuel-staged, 405, 407 retrofits, effects of on boiler performance,
581–582 furnace combination, 111 553
ultra-low emissions burner operation, gas, 23 role of
582–584 troubleshooting for, 519 in furnace, 112
Boiling points, general fraction, 177 typical excess air values for, 490 in heat transfer, 110
Bolts, standard grades of, 709, 710, 711 general burner types, 21–27 sound pressure level, 237
Boltzmann constant, 88 geometry, 310, 405 staged
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 735 Tuesday, March 15, 2005 10:02 AM

Index 735

air, 25, 517 casing and refractory, 487–488 Canadian Standards Association (CSA), 540
design philosophy, 551 combination burners, 482 Capacity curves, 262, 263, 265
fuel, 26, 517, 518 diffusion or raw gas burner, 481–482 CARB, see California Air Resources Board
systems, ZTOF, 623 flame patterns, 483–485 Carbon dioxide, 99, 218
technology, 526 pilot burners, 483 Carbon monoxide (CO), 439, 538
boiler, 548 premix burner, 480–481 analyzer, 389
NOx, 518 measurements, 470–479 formation, 60
test setup, 433 combustion air temperature, 476 /opacity level, 567
throat, 354, 455 draft, 470 oxidation reaction, 314
tile excess air or excess oxygen, 471–473 Carrier gas, ammonia in, 669
broken, 494 flue gas temperatures, 476–477 Cartesian coordinates, 76, 134
ledge in, 361 fuel flow, 473 Cartesian differential equation set, 294
picture of showing multiple tile pieces, fuel pressure, 473–476 Castable refractory, 646, 684, 689
454 fuel temperature, 476 Catalyst configuration, 671
support for, 367 process fluid parameters, 479 Catalytic cracking, 5
tip, plots of contours of streamfunction with process tube temperature, 477–479 Catalytic hydrocracking, 5
increasing backpressure at, 322 operational considerations, 488–499 Catalytic hydrotreating/hydroprocessing, 5
Todd Combustion low NOx, 578, 579 developing startup and shutdown Catalytic oxidizers, 647

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
turndown, 425 procedures for fired heaters, 497–498 Catalytic reforming, 5, 13
types, 21, 362 developing emergency procedures for Categorical factor, 418
ultra-low emissions, 580, 583 fired heaters, 499 CEMs, see Continuous emissions monitors
ultra low-NOx, 517 heater operation data, 496–497 CEN, see European Committee for
unmodified, 308 heater turndown operation, 492–493 Standardization
utility, NOx vs. load with firing natural gas inspection and observations inside CENELEC, see European Committee for
on, 559 heater, 493–494 Electrotechnical Standardization
venturi-style, 552, 558
inspection and observations outside Center fire gas (CFG), 576
wall-fired, 15, 368
heater, 495–496 Central-composite designs, 419
zone heat release (BZHR), 551, 554, 575
target draft level, 488 Central tube wall process heater, 16
Burner design, 263, 351–370, 432
target excess air level, 489–491 Centrifugal blowers, relative characteristics of,
air control, 358–359
Burner testing, 431–447 673
air metering, 356–358
burner test setup, 433–437 Ceramic temperature, 281
forced draft, 358
application, 434 CFD, see Computational fluid dynamics
natural draft, 357
test furnace selection criteria, 434–437 CFG, see Center fire gas
burner types, 362–367
instrumentation and measurements, Chemical
oil or liquid firing, 364–367
437–442 heat release, 137
premix and partial premix gas, 363–364
emissions analysis, 438–441 industry, major fired heater applications in,
raw gas or nozzle mix, 364
flame dimensions, 441 14
combustion, 352–353
fuel flow rate metering, 441 manufacturing process, by-products of, 63
final, 312
furnace gas temperature measurement, process industry (CPI), 288, 361
fuel metering, 353–356
gas fuel, 353–354 438 reaction(s)
liquid fuel, 354–356 heat flux, 442 global, 204
ignition, 360–361 measuring air-side pressure and modeling of, 252
initial, 312 temperature, 437–438 reactors, idealized, 258
materials selection, 369–370 test matrix, 442–444 Chemical Manufacturers Association (CMA),
mixing fuel/air, 359–360 definition of data to be collected, 444 630
co-flow, 360 heater operation specifications, 442 Chemiluminescent analyzer, 197
cross-flow, 360 performance guarantee specifications, CHEMKIN, 125, 258
entrainment, 360 442–443 Chlorinated hydrocarbons, 309
flow stream disruption, 360 Burning Chloroform, 216
mounting and direction of firing, 367–369 points, division of incoming gas stream into, Choked flow, 147, 275
conventional burner, round flame, 609 mass flux, 275
367–368 smokeless, 603 test rig, 147
downfired, 369 velocity, 296 Circles, areas and circumferences of, 693–701
flat flame burner, 368 Butane lighter, shadow photograph of burning, Clean Air Act Amendments, 631
radiant wall, 369 236 CMA, see Chemical Manufacturers Association
patterned and controlled flame shape, Butylene dehydrogenation, 14 CMC, see Conditional moment closure
361–362 BZHR, see Burner zone heat release CO, see Carbon monoxide
problems, 27 Coal, 72
pollutants, 362 Coking, 11, 13, 462
ultra-low emissions, 581 Cold flow furnace, 264
Burner/heater operations, 469–499 C Colebrook formula, 151
heater and appurtenances, 479–488 Cabin heater, 16, 18, 28 Collecting surface area, 659
air flow control, 480 Calcium chloride, dehydration with, 158 Combination burners, 418, 482, 643
burner block valves, 485–487 California Air Resources Board (CARB), 197 Combustible waste gas streams, 160
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 736 Tuesday, March 15, 2005 10:02 AM

736 The John Zink Combustion Handbook

Combustion Combustion equipment, experimental design Compressibility factor, 127

air, 19 for, 401–429 Compressible flow, 144, 145
blower inlet preheat, 524 accounting for fuel mixtures, 420–424 Computational fluid dynamics (CFD), 82, 110,
density, 264 building 2n-level factorials from two- 142, 265, 288
preheating of, 63 level factorials, 424 -based combustion submodels, 296
temperature, 26, 476 combining mixture and factorial designs, burner simulation, 306
basics, 639 423–424 code, 293
chemistry experimental designs for mixtures, engineer, 289
of hydrocarbons, 34 421–423 model, 302, 534
models, 299 orthogonal mixture designs, 423 background, 291
as controlled process, 352 combining domain knowledge with SED, for design optimization, 534
control strategies, 303, see also Combustion 424–427 elements in, 290
controls practical considerations, 425 ethylene pyrolysis furnace, 305
data, for hydrocarbons, 45 semi-empirical models, 425–426 rendered view of, 303
efficiency, 630 sequential experimental strategies, specialist, 289
equipment, see also Combustion equipment, 426–427 study, acid gas used in, 309
experimental design for important statistics, 408–412 vendors, 303
designers, 224 analysis of variance, 409 Computational fluid dynamics based
noise, 243 ANOVA with separate model effects, combustion modeling, 287–325
gases, polynomial expression for, 125 409–411 case studies, 305–319
-generated particles, 217 F-distribution, 409 ethylene pyrolysis furnace, 305–306
heat of, 38 pooling insignificant effects, 411–412 incineration of chlorinated
instability noise, 234 standard errors of effects, 412 hydrocarbons, 309–319
kinetics, 60 linear algebra primer, 427–429 sulfur recovery reaction furnace,
modeling, 258 identity, inverse, and transpose, 428–429 307–309
modification, 194, 199, 201 matrix addition, 429 venturi eductor optimization, 319
non-premixed, 299 matrix multiplication, 428 xylene reboiler, 306
optimization, 563 Taylor and MacLaurin series combustion submodels, 296–302
overall sound pressure level from, 235 approximations, 427–428
combustion chemistry models, 299–301
premixed, 301 linear transformations, 407–408
pollutant chemistry models, 301
process(es), 58 matrix solution, 406–407
radiation models, 297–299
approximations involved in modeling, method of least squares, 405–406
solution algorithms, 297
296 power of SED, 402–404
turbulence models, 301–302
modification, 667 second-order designs, 419–420
model background, 291
product(s) central-composite designs, 419–420
simulation model, 291–296
flare, 629 practical considerations, 420
convergence criteria, 295
temperature, 59 statistical experimental design principles,
modeling basis, 296
reactions, molar ratios for, 37 404–405
model validation, 295–296
roar, flare, 234 two-level factorial designs, 412–419
solution technique, 295
safety, see Combustion safety foldover, 416–417
transport equations, 292
staged, 200, 573 including categorical factors, 418–419
submodels, CFD-based, 296 interactions, 412 turbulence equations, 292–294
substoichiometric, 46, 47 method of steepest ascent, 416 solution methodology, 302–305
sustainable, for methane, 353 orthogonal blocking, 417–418 analysis of results, 303–305
systems, scaling of from laboratory scale to pure error and bias, 412–414 problem setup, 302–303
industrial scale, 291 screening designs, 415–416 solution convergence, 303
tetrahedron, 330 serial correlation and lurking factors, Conditional moment closure (CMC), 300, 301
turbulent, regimes of, 299 416 Conduction
waste, 670 two-level fractional factorials, 414–415 heat transfer, 71, 137
Combustion controls, 373–399 Combustion safety, 327–349 one-dimensional steady-state, 73
in radial coordinate systems, 76
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

controllers, 394–398 design engineering, 339–346

control schemes, 389–393 fire extinguishment, 344 steady-state, 77
fully metered cross limiting, 392–393 flammability characteristics, 339–342 transient, 77, 80
parallel positioning, 389–392 ignition control, 342–344 unsteady-state, 78
fundamentals, 374–383 safety documentation and operator Conductivity analyzer, 389
agency approvals and safety, 379–380 training, 344–346 Configuration factor, 92
analog control systems, 376–378 overview, 329–339 Conservation
control platforms, 374–376 codes and standards, 338–339 energy, 134, 137, 275, 289
discrete control systems, 376 combustion tetrahedron, 330–331 mass, 35, 133, 134, 274, 289
failure modes, 378–379 definitions, 329 momentum, 133, 134, 136, 253, 275
pipe racks and control panels, 381–383 explosion hazards, 334–337 species, 134, 138
primary measurement, 383–389 fire hazards, 331–334 Constant surface heat flux, 80
analog devices, 385–389 process hazard analysis, 337–338 Constant velocity, 134
discrete devices, 383–385 sources of information, 346–347 Contact resistance, definition of, 75
tuning, 398–399 Composite wall, 74 Continuous emissions monitors (CEMs), 207
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 737 Tuesday, March 15, 2005 10:02 AM

Index 737

Controlled variable (CV), 394, 395 D Drier off-gas, 352

Controller(s), 394 Drilled pipe duct burner, 536
Damköhler number, 291
analog, 394 Drilling and exploration techniques, 167
Damper(s), 465, 488
modes of operation of, 397 Drill sizes, areas and circumferences of,
air, 464
output mode, 393 693–701
burner air, 457
pass flow, 498 Droplet size
fan, 457
comparison, between standard and newer oil
Control panel(s), 381 plenum, 489, 490
gun, 279
inside of, 382 radial inlet, 676
distribution, 278
large, 382 stack, 457, 495
Dry air, properties of at atmospheric pressure,
small, 382 Data, definition of to be collected, 444
715–716
Control platforms, 374 DCS, see Distributed control system
DSP, see Digital signal processing
Control schemes, 389 Decibel, 226
Duct burner(s), 523–544
Defining contrast, 415
Control signal, fuel flow rate versus, 392 applications, 524–526
Degrees of freedom (DF), 408
Control system(s) air heating, 524–526
Dehydration, with calcium chloride, 158
analog, 376 cogeneration, 524
Delayed coking, 5
discrete, 375 fume incineration, 526
Demountable derrick, 626
distributed, 375 stack gas reheat, 526
Density, definition of, 118
instrumentation and, 675 arrangement, 533
Derivative gain, 396
Control valve, 397 drilled pipe, 536
Derrick
body, 385 flame, 7
demountable, 626
-supported flare, 626 fluidized bed startup, 529
bypass, 619
Design engineering, 339 in large duct, 7
characteristics, 385
Destruction efficiency, 630, 681–682 low emission, 536
typical staging, 628
Destruction and removal efficiency (DRE), 280, technology, 526–544
Control volumes, 132, 133 accessories, 540
Convection, 82, 511 644
DF, see Degrees of freedom design considerations, 528–539
coefficient, 79, 84 design guidelines and codes, 540–544
Diesel engine exhaust, 352
correlation, for flow in circular tube, 86 grid configuration (gas firing), 527
Differential formulation, 134
forced, 104 Diffuser grid configuration (liquid firing), 527
heat transfer condition, 494 in-duct or inline configuration, 526–527
in banks of tubes, 87 cone, 464, 510 maintenance, 539
coefficient, 78 Diffuse surfaces, view factors for, 92 Dynamic pressure, 131
laminar flow, 83 Diffusion
mixed, 82 burner, 25, 67, 481
Convective heat transfer, 137 flame(s), 299
Conventional burner, round flame, 367 burners, 482 E
painting of, 24 Early flame zone, 317
Converging–diverging nozzle, 255
-mixed flames, 21 Eddy
Conversions, common, 723
Digital signal processing (DSP), 228 -breakup model, 299
Cooler fireboxes, 478
Dioxins, 219 diffusivity, definition of, 293
Cooling, Newton’s law of, 82, 281 formation, 360
Direct numerical simulations (DNS), 297
COOL TECHNOLOGY burner, 518, 519 Eduction processes, 252
Direct spray contact quench, 645
Copper tubing, commercial, 707–708 Discharge coefficient, 148, 152, 153 in pilots, 256
Coriolis flow meter, 388 Discrete control systems, 375 in premixed burners, 257
Correction factor, 95 Discrete devices, 383 Eductor(s)
CPI, see Chemical process industry Dispersion, 632, 633 modeling of, 252
Cracking furnaces, 34 Distillation, 180 optimization, venturi, 319
Crocco similarity, 296 Distributed control system (DCS), 375, 540 throat, re-circulation region in, 320
Cross fluctuating velocity terms, 142 DNS, see Direct numerical simulations EF, see Exposure factor
Crude oil, 173, 175, 181 Domain knowledge, combining of with SED, Electrical heating, 202
Crude unit burners, 19 424 Electromagnetic radiation, spectrum of, 89
Double-block-and-bleed, for fuel supply, 379 Electromechanical relays, 540
CSA, see Canadian Standards Association
Downcomer, 656 Electronic ignitor
Current-to-pressure transducer, 385
Downfired burner, 369 ST-1S pilot tip without, 466
CV, see Controlled variable
Down-fired tests, test furnace for simulation of, ST-1SE pilot tip with, 466
Cycloalkanes, 177 Electronic linkage, 390
435
Cyclone separator, 618, 655 Draft, 130 Electrostatic precipitator (ESP), 659, 660
Cylinder, temperature distribution in, 76 expression of, 470 Ellipsoidal radiometer schematic, 443
Cylindrical coordinate system, 77 level, 488 Emergency flow rate, maximum, 594
Cylindrical differential equation set, 294 loss, 470 Emergency shutdown (ESD), 338
Cylindrical furnace, radiation heat transfer in, targets, 496 Emission(s), 631
107 type, 26
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`--- analysis, 438
Cylindrical heater, 17 DRE, see Destruction and removal efficiency guarantees, 443
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 738 Tuesday, March 15, 2005 10:02 AM

738 The John Zink Combustion Handbook

noise, 443 Excess air, 471, 555 excess oxygen level within, 504
NOx, high, 515 empirical evidence of effect of on NOx, 556 good flame within, 486
particulate, 442 level, 489 temperature, effect of on NOx, 515
pollutant, 432 species concentration vs., 39 very bad flame pattern in, 486
regulations, 111 stoichiometric ratio and, 38 Fired heater(s), 9–15
Emissivity(ies) Excess oxygen, 471 applications, in chemical industry, 14
of carbon dioxide, 99 Expansion joints, 455 design, 70
for various surfaces, 100 Experimental fluid dynamics, 288 developing emergency procedures for, 499
of water vapor, 99 Explosion(s) developing startup and shutdown procedures
Enclosed flares, 241, 622 ethylene oxide plant, 343 for, 497
Endwall fired burner arrangement, 20 furnace, 336 major refinery processes requiring, 13
Energy hazards, 334 process heaters, 11–15
conservation of, 134, 137, 275, 289 piping, 335 reformers, 10–11
conversion, 608 potential dangers caused by, 328 tube rupture in, 332
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

flux, 73 source of in furnaces, 337 Firetube boiler, 648, 650

kinetic, 276, 293 stack, 335 Firewood principle, 609
minimum ignition, 342 tank, 335 Firing
thermal unburned fuel accumulations causing, 393 combination, 482
turbulent, 294 Exposure factor (EF), 232 mounting and direction of, 367
Energy Information Administration, 168 Fixed effects, 417
Enthalpy of formation, 714 Fixed heater size distribution, 12
Environmental requirements, 597 Flame
Equation(s) F burst, 105
Bernoulli, 130, 131 Factorial(s) color observed in, 267
Favre-averaged, 292, 301 diffusion, 21, 24, 299
design(s)
ideal gas, 127 dimensions, 432, 441, 442, 443
ANOVA for, 414
Le Chatelier, 339 duct burner, 7
combining mixture and, 423
momentum, 142 envelope
two-level, 412
Navier–Stokes, 145, 292 of oil flame, 485
saturated, 415
partial differential, 295 with x-y-x axes, 485
two-level, 424
radiation transport, 297 equivalence ratio, 193
Factor space, 416
radiative transfer, 97 fireflies leaving, 507
Factory Mutual (FM), 540, 675
Redlich–Kwong, 128 -to-flame similarity of appearance, 566
Failure modes, 378
Reynolds-averaged conservation, 301 flat, 363, 368
and effect analysis (FMEA), 337
of state, 126 -front generator (FFG), 615
safety concerns and, 390
steady-state conduction, 77 self-inspirating, 615, 616
Fan(s), 670
transport, 292 slipstream, 615, 616
dampers, 457
turbulence, 292 gas, 104, 105, 531
wall to wall, 288 noise, 237, 243 gaseous fuel, 165
Equilibrium wheel designs, 672 holder, 67, 360, 361
chemistry assumption, 300 Fanno flow, 148 oxidation of, 539
reaction process, adiabatic, 54 Fault trees, 337 warped, 539
thermodynamics and, 54 Favre-averaged conservation equations, 301 ignition, forced-draft, 612
ESD, see Emergency shutdown Favre-averaged transport equations, 292 impingement, 9, 333, 493
ESP, see Electrostatic precipitator F-distribution, 409, 410–411 possible cause of on tubes, 504
Esters, 216 Feedback, 378 on tubes, 28, 503
Eternal mix atomizers, 586 Feedforward, 378, 379 infrared thermal image of in furnace, 98
Ethers manufacture, 5 FFG, see Flame-front generator inside thermal oxidizer, 317
Ethylbenzene dehydrogenation, 14 FFT, see Fourier transform analysis leaning, 509
Ethylene FGR, see Flue gas recirculation lift-off, 513
cracking furnace, 111 Fiberglass-reinforced plastic (FRP), 654 long, narrow, 484
furnaces, test furnace for simulation of, 434 Fieldbus, 375 long smoky, 507
hydration, 14 Finite-rate chemistry, 296 luminous, photographic view of, 101
oxide (ETO), 343, 619 Fire(s) nonluminous, photographic view of from
plant, multiple ZTOF installation in, 593 common ignition source of, 343 John Zink gas burner, 102
/polyethylene gases, 186 diluted fuel, 571 oil, flame envelope of, 482
pyrolysis furnace extinguishment, 344 pattern, 483, 493
CFD model of, 305 hazards, 331 in firebox, 486
rendered view inside, 304 ignition sources of major, 342 irregular, 505, 506
ETO, see Ethylene oxide plenum, 507 pilot, stable ignition of, 336
European Committee for Electrotechnical potential dangers caused by, 328 presence of soot in, 298
Standardization (CENELEC), 540 tetrahedron, 330 process heater oil, 276
European Committee for Standardization Firebox(es) properties, 61
(CEN), 339 consequence of too much draft in, 488 pulled toward wall, 28
Event trees, 337 cooler, 478 pulsating, 502
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 739 Tuesday, March 15, 2005 10:02 AM

Index 739

quality, 476 factors influencing flare design, 594–598 Fanno, 148

radiation, 101, 105, 107, 598 environmental requirements, 597–598 fluid, kinetic energy in, 258
round, 362, 367 flow rate, 594 fuel, 473
scanners, 384 gas composition, 594–595 incompressible, 136
short, bushy, 484 gas pressure available, 596 inviscid, 145, 290
simulated reforming gas, 169 gas temperature, 595–596 laminar, 138, 141
smoking tendencies, predicting, 266 safety requirements, 597 meter(s), 387
sparklers leaving, 507 social requirements, 598 Coriolis, 388
speeds, 67 utility costs and availability, 596–597 magnetic, 388
stability, 493, 564 flame, thermogram of, 602 orifice, 388
stabilizer, 360, 361, 463 forced-draft Dragon liquid, 613 positive displacement, 388
cones, 369 guy wire-supported, 625 turbine, 388
design of, 534 high-pressure, 244, 245 vortex shedder, 387
flow patterns around, 537 horizontal, 592 pass, 496
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

temperature, 61, 101 modeling of air-assisted, 270 patterns, around flame stabilizer, 537
adiabatic, 46 multi-point, 592 pulsating, 620
reduction, 559 noise quasi-one-dimensional isentropic, 145
transfer of heat from, 58 abatement techniques, 239 rate, FGR, 571
unstable, 494 effect of distance on, 246 reversal, 658
visible, 604 level, engineer measuring, 235 stream disruption, 360
Flameholder, 534 non-smoking, 267 switches, 384
Flamelet, 296, 299, 300 offshore oil rig, 7 turbulent, 138
Flamesheet, 296 pilot, 614, 616 types of, 138
Flammability pressure relief vessel venting to, 131 viscous, 290
characteristics, 339, 340 self-supported, 624 volume equivalent of, 185
limits, for gas mixtures, 66 Flowmeter, ultrasonic, 388
single-point, 591, 593
Flare(s), 5, 589–634 Fluctuating velocity, 140
smokeless
air-assisted smokeless, 610
operation, 266 Flue gas(es), 72, 471
annular air, 275, 611
rates, modeling of, 252 condition of, 653
applications for throughout hydrocarbon
steam, 272 cooled, 652, 668
industry, 590
smoking, 267 counter-current flow of, 664
available gas pressure at, 596
stack, flaming liquid engulfing, 600 drawn through U.S. EPA sampling train, 440
burner(s), 605
steam, 237, 254, 268 flow
RIMFIRE®, 611, 613
control valve on, 626 minimizing, 653
state-of-the-art Steamizer™, 610
with steam eductor, 254 NOx reduction vs., 570
capacity, smokeless, 266
support structures, 623 hot, 511, 512
combustion
systems, 590–594 measured pollutant concentration, 190
instability noise, 236
applications, 591 oxygen concentration in, 425
roar, 234
flare system types, 591–593 processing methods, 647
combustion products, 629–633
dispersion, 632–633 major system components, 593–594 recirculation (FGR), 20, 22, 550
emissions, 631–632 objective of flaring, 590 flow distribution, improvement of to
reaction efficiency, 630–631 purpose, 590 burners, 565
controls, 624 tip(s) flow level, 559
derrick-supported, 626 exit velocity, 599 flow rate, 571
design considerations, 598–605 Zink double refractory severe service, hot, 568
air/gas mixtures, 605 606 implementation, NOx reduction by, 568
air infiltration, 601 vendors, 273, 622 recycle, 668
flame radiation, 601–603 Flarestack explosion, due to improper purging, shearing effect of high-velocity, 661
hydraulics, 599–600 336 ST, 130
liquid removal, 600–601 Flaring temperature, 476
noise/visible flame, 604 event, 591 Fluid(s)
reliable burning, 599 objective of, 590 absolute viscosity vs. temperature for, 122
smoke suppression, 603–604 Flashback, 67, 505 coking/flexicoking, 5
early model smokeless, 591 Flash point, of liquid, 179 density, 118, 139
enclosed, 241, 622 Flat flame burner, 368 dynamics, theoretical, 288
equipment, 605–629 Flat-shaped flame, 363 flow
enclosed flares, 622–623 Flexicoking waste gas, 162 conservation of mass for, 135
flare burners, 605–614 Flow(s) kinetic energy of, 258
flare controls, 624–629 air, balancing of by windbox model, 567 Newtonian, 136, 292
flare support structures, 623–624 choked, 275 packet, velocity of, 130
knockout drums, 618–619 compressible, 144 properties, 118
liquid seals, 619–620 configuration, 644 velocity profile of, flowing along solid
pilots, ignitors, and monitors, 614–618 control, method of, 674 surface, 119
purge reduction seals, 620–622 controller gains, 396 viscosity, 139
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 740 Tuesday, March 15, 2005 10:02 AM

740 The John Zink Combustion Handbook

Fluid dynamics, fundamentals of, 117–154 CO and unburned, 214 high, 510
flow types, 138–148 combustion, stagewise, 572 pretreatment, 199
compressible flow, 144–148 composition, 195, 205, 437, 496 rapid oxidation of, 641
turbulent and laminar flow, 138–144 dilution, 571 refinery/chemical plant, 528
fluid properties, 118–128 distribution, liquid, 279 selection of test, 434
density, 118 fires, diluted, 571 source, 344
equations of state, 126–128 flow, 473 spray, 277
specific heat, 125–126 balancing techniques, 567 -staged burner, 405, 407
viscosity, 119–125 rate, control signal vs., 392 staging strategies, 291
fundamental concepts, 128–138 rate metering, 441 storage tanks, test, 436
Bernoulli equation, 130–132 free jet of, 363 supply, double-block-and-bleed for, 379
control volumes, 132–134 gas, 353 switching, 198
differential formulation, 134–138 firing ports, 513 system preparation, 497
hydrostatics, 128–130 piping system, 459 temperature, 476
pressure drop fundamentals, 148–153 pressure measurement, 475 train, 540
basic pressure concepts, 148–149 tips, 209, 462 oil, 543
discharge coefficient, 152–153 gaseous, 18, 158–165 typical main gas, 541
loss coefficient, 151–152 combustible waste gas streams, 160–165 typical flared gas compositions, 185–186
roughness, 149–151 liquefied petroleum gas, 159 ethylene/polyethylene gases, 186
Fluidized bed molecular weights and stoichiometric oil field/production plant gases, 186
boilers, 524 coefficients for common, 37 other special cases, 186
startup duct burner, 529 natural gas, 158–159 refinery gases, 186
Fluidized catalytic cracking, 13 photographs of gaseous fuel flames, 165 valves, 380
FM, see Factory Mutual physical properties of, 165 Fully metered cross limiting scheme, 392
FMEA, see Failure modes and effect analysis refinery gases, 159–160 Fume incineration, 526
Foldover, 416 gas property calculations, 183–185 Fundamentals, 33–67
Forced convection, 104 derived quantities, 185 combustion kinetics, 60–61
Forced draft flammability limits, 184–185 fuel-bound NOx mechanism, 61
air preheater, 438 lower and higher heating values, 184 prompt-bound NOx mechanism, 61
burners, 358, 437, 439, 480 molecular weight, 183–184 reaction rate, 60–61
ignition flame, 612 specific heat capacity, 184 thermal NOx formation, 60
Formaldehyde, 216 viscosity, 185 conservation of mass, 35
FORTRAN, 282 high inert composition, 359 equilibrium and thermodynamics, 47
calculation and standardized storage of hot burning, 536 flame properties, 61–67
chemical kinetic data, 258 hydrocarbon-based gaseous, 528 available heat, 64
conversion of thermodynamic information injection nozzle parameters, 516 flame speeds, 67
into NASA polynomials using, 125 injector spud, 530 flame temperature, 61–64
Forward tip blade operating curve, 674 introduction, 640 flammability limits for gas mixtures,
Fourier number, 78 leaks, 344 66–67
Fourier transform analysis (FFT), 228 liquid, 165–183 minimum ignition energy, 64–66
Fourier transform infrared (FTIR), 440 high viscosity, 365 general discussion, 54–59
gas analyzer, 440 history, 165–172 air preheat effects, 55–57
system, 440 liquid naphtha, 179 fuel blend effects, 57–59
Fractional factorials, 412, 414 oil recovery, 173–175 ideal gas law, 35–38
Free jet, 142 oils, 178–179 net combustion chemistry of hydrocarbons,
entrainment, 143 physical properties of liquid fuels, 34–35
interaction with surrounding fluids, 143 179–183 overview of combustion equipment and heat
structure, 143 production, refining, and chemistry, transfer, 34
Freestream velocity, 87 175–178 stoichiometric ratio and excess air, 38–46
Friction vaporization of, 354 adiabatic flame temperature, 46
coefficient, 86 metering, 353, 354 heat of combustion, 38–45
factor, 84, 85, 150 mixtures, accounting for, 420 substoichiometric combustion, 46, 47–53
FRP, see Fiberglass-reinforced plastic nonluminous, 18 uses for combustion, 34
FTIR, see Fourier transform infrared nozzles, 266 Furans, 219
Fuel(s), 157–187 oils, 72 Furnace(s)
/air piping system, heavy, 460 air in-leakage influence, 562
mixing of, 359 piping system, light, 461 burner combination, 111
ratio, horizontal imbalances of, 573 temperature, 563 cleanliness, effect of on NOx emissions, 561
unbalance, vertical, 573 viscosity of, 183 cold flow, 264
at ambient temperature, 61 orifice-to-eductor throat diameter, 319 cracking, 34
blend piping design, 458 cylindrical, radiation heat transfer in, 107
composition, 205 preheat temperature, 56 design, terrace wall, 111
effects, 57 pressure, 213, 473, 496 ethylene
-bound NOx mechanism, 61 drop, 443 cracking, 111
chemical properties of, 717–722 graph of heat release vs., 475
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
pyrolysis, 304, 305
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 741 Tuesday, March 15, 2005 10:02 AM

Index 741

test furnace for simulation of, 434 shearing effect of high-velocity, 661 properties of dry air at atmospheric pressure,
explosions in, 336 temperature, 476 715–716
gas, 290, 296 fuel, 353 Gasolines, aviation, 175
flow patterns, 109 capacity curve, 355 Gauss constant, 143
radiation, 105 train, typical main, 541 Gibbs function of formation, 714
recirculation, schematic, 201 geometries, mean beam lengths for, 100 GLC, see Ground-level concentration
temperature measurement, 438 ignition characteristics of, 340 Glossary, 727–728
heat transfer, 104 injectors, 583 Gravitational body force, 136, 137
hot oil, 14 jet Gravitational constant, 130
hydrogen reforming, 111 Mach number of, 238 Gray/diffuse surfaces, radiant exchange
infrared thermal image of flame in, 98 mixing noise, 238 between, 91
low temperature, 109 noise, 238, 241, 242 Grease, 5
manufacturers, 111 kinetic theory of, 122, 126 Grid burner(s)
process, heat transfer in, 102 lighter-than-air, 130 gas flame from, 531
reforming, 34 lines, pilot, 514 heat distribution by, 532
role of burner in, 112 mass flow rate of secondary, 253 Ground-level concentration (GLC), 632, 683
source of explosions in, 337 mixtures Guy wire-supported flare, 625
sulfur recovery reaction, 307 flammability limits for, 66
temperature, 191 multi-component, 339
test viscosity of, 120
selection criteria, 434 oil field/production plant, 186 H
for simulation of down-fired tests, 435 oxygen-containing, 601 Halogenated hydrocarbon systems, 683
for simulation of up-fired tests, 435 partial premix, 363 Hard refractory, 645
/thermal oxidizer/incinerator/combustion premix, 363 Hastelloy, 685
chamber, 644 pressure available, 596 Hazardous waste incinerators, 291

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
tube, 8 property(ies) Hearing
vertical tube, 104 calculations, 183 threshold of, 226
wall of selected, 72 typical range of human, 225
cross section of, 108 purge, 597, 601 Heat, see also Heat transfer
heat transfer through, 108 radiation capacity, 125, 182
absorption and emission in, 93 of combustion, 38
properties, 298 conduction, 71
recirculation, 20, 200 damage, 331
G refinery, 159, 186 exchangers, 70, 652
Gage comparison of to test blend, 437 flux, 104, 442
pressure, 149 example, 437 recovery
sizes, 707 simulated refinery, 39, 48 area (HRA), 562
Gas, see also Gases and liquids, properties of; specification sheet, test procedure, 445–446 cooling by, 648
Natural gas stack, reheat, 526 cooling without, 652
acid, 639, 657, 663 temperature, 438, 595 efficiencies (HREs), 282
analyzer, FTIR, 440 tip(s), 456, 462, 641 thermal oxidation system, 680
behavior of, 127 cleaning, 462 regenerator performance, modeling of, 252
burner(s), 23 corroded, 504 release, 185, 602
troubleshooting for, 519 trapped purge, 621 chemical, 137
typical excess air values for, 490 treatment, 5 graph of fuel pressure vs., 475
typical premixed, 23 turbine exhaust, 352, 531 graphical representation of, 66
center fire, 576 valve data, 392 Heater(s)
chemical properties of, 717–722 waste, 599 balanced-draft, 492
combustion wet fuel, 160 box, 17
data comparing single- and three-stage, Gaseous fuel(s), 158 burner mounted on top of, 452
574 flames, photographs of, 165 cabin, 16, 18, 28, 34
polynomial expression for, 125 mixtures collapse, 499
composition, 54, 55, 594 physical constants of typical, 163 connection of burner to, 455
emissivity, 96 volumetric analysis of, 163 cutout, 451
ethylene/polyethylene, 186 molecular weights and stoichiometric cylindrical, 17
-fired furnaces, 290, 296 coefficients for common, 37 existing, 450
firing, grid configuration, 527 physical properties of, 165 fired, 9
flame(s), 104, 105, 531 Gases and liquids, properties of, 713–722 developing emergency procedures for,
flammability characteristics of, 340 chemical, physical, and thermal properties 499
flexicoking waste, 162 of gases, 717–722 developing startup and shutdown
flow, diverting, 619 combustion data for hydrocarbons, 713 procedures for, 497
flue, 72, 471 enthalpy of formation, Gibbs function of major refinery processes requiring, 13
hot, 511, 512 formation, and absolute entropy of tube rupture in, 332
recirculation, 550 various substances, 714 natural-draft, logic diagram for tuning, 491
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 742 Tuesday, March 15, 2005 10:02 AM

742 The John Zink Combustion Handbook

new, 450 Horizontal floor-fired burners, 14 ports, 513

operation Hot flue gas, 130 sources, of major fires, 342
data, 496 Hot oil furnace, 14 spontaneous, 342
specifications, 442 Hottel’s assumption, 95 transformers, 384
preparation, 450 HPI, see Hydrocarbon processing industry zone, in natural draft burners, 360
process, 11, 12 HRA, see Heat recovery area Ignitors, 614
refinery, 11 HREs, see Heat recovery efficiencies Incendiarism, 342
refractory, 453 Human ear, cross section of, 225 Incineration, upstream, 664
shell, 495 Human error, 338 Inclined manometer, 129
shutdown of, 498 Hybrid systems, 375 Incomplete combustion products (ICPs), 548,
turndown operation, 492 Hydraulics, 139, 599 550
warped steel on shell of, 451 Hydrocarbon(s) Incompressible flow, 136
Heat transfer, 69–114 -based gaseous fuels, 528 In-duct burners, 524
conductive, 71, 137 chemical degradation of, 479 Industrial equipment, noise-producing, 224
convection, 82–87 chlorinated, 309 Industrial insurance carriers, 339
laminar flow convection, 83–84 coking of, 462 Industrial noise pollution, 231
Newton’s law of cooling, 82 combustion data for, 45 Industrial Risk Insurers (IRI), 339, 675
turbulent external flow, 85–87 industry, 4, 590 Infrared pyrometer, 478
turbulent internal flow, 84–85 long-chain, 538 Infrared temperature measurement, 92
heat transfer in process furnaces, 102–112 net combustion chemistry of, 34 Inline burner, 530
analysis of radiation heat transfer, paraffinic, 269 In-scattering, 97
106–107 processing industry (HPI), 361 Insignificant effects, pooling of, 411
flame radiation, 105 properties of, 713 In situ extractive analyzers, 472
furnace gas flow patterns, 109–110 reforming, 14 Installation and maintenance, 449–466
furnace gas radiation, 105–106 systems installation, 450–459
heat transfer in process tube, 109 halogenated, 683 air control, 456–458
heat transfer through wall of furnace, sulfur-bearing, 682 burner installation inspection, 456
108–109 burner mounting, 451–453
temperature vs. viscosity for, 123
refractory surface radiation, 106
unburned, 433, 439, 538, 548, 597 burner pre-installation work, 450–451
role of burner in heat transfer, 110–112

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
vapor pressures for light, 341 connection of burner to heater, 455–456
mechanisms of, 70
viscosity vs. temperature for range of, 477 fuel piping design, 458–459
in packed bed, 281
Hydrodesulfurization, 13 preparation of heater, 450
in process furnaces, 102
Hydrogen tile installation, 453–455
in process tube, 109
bubble technique, 150 maintenance, 459–465
radiation, 87–102
-to-carbon weight ratio, 165 air registers and dampers, 464–465
blackbody radiation/Planck distribution,
cyanide, 667 flame stabilizer, 463–464
88–90
enrichment, 631 gas tip and orifice cleaning, 462
equation of radiative transfer, 97–101
production, 8 oil tip and atomizer cleaning, 463
infrared temperature measurement,
purification, 161 pilot burners, 465
92–93
reforming furnace, 111 tile, 463
mean-beam-length method, 95–97
Hydrostatics, 128 Internal flue gas recirculation (IFGR), 576, 578
radiant exchange between black
surfaces, 90–91 Hydrotreating, 13 Internal mix atomizers, 584
radiant exchange between gray/diffuse Internal mix twin fluid atomizer, 364, 365
surfaces, 91 Internal tube fouling, 493
radiation in International Electrochemical Commission
absorbing/emitting/scattering media,
I (IEC), 540
93–95 ICPs, see Incomplete combustion products Inviscid flows, 145, 290
radiation emitted by flame, 101–102 Ideal gas IRI, see Industrial Risk Insurers
view factors for diffuse surfaces, 92 equation, 127 Isomerization, 5
role of burner in, 110 law, 35, 126, 263 Isotropic turbulence, 142
thermal conductivity, 71–82 Idealized chemical reactors, 258
one-dimensional steady-state Identity matrix, 428
conduction, 73–77 IEC, see International Electrochemical
transient conduction, 77–82 Commission J
Heavy fuel oil piping system, 460 IFGR, see Internal flue gas recirculation Jet(s), 142
Heavy oils, 178, 180 Ignition, 360 engine, flow around air intake of, 144
Helium balloon, attached to ground, 129 characteristics, of liquids and gases, 340 high-pressure, 607
HHV, see Higher heating value control, 342
HIGH box, 490 energy, minimum, 64
Higher heating value (HHV), 184 flame, forced-draft, 612
High pressure drop burner, 643 fuel pressure, unsuccessful, 336 K
High-pressure flare, 244, 245 graphical representation of, 66 Karlovitz number, 299
High viscosity liquid fuels, 365 ledge, burner, 482 KE, see Kinetic energy
Horizontal flares, 592 pilot, 615, 640 Ketones, 216
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 743 Tuesday, March 15, 2005 10:02 AM

Index 743

K factors, 132 Long-chain hydrocarbons, 538 regenerative thermal oxidizer performance,

Kiln off-gas, 352 Loop seals, 629 280–282
Kinematic viscosity, 120 Loss coefficient(s), 151 Matrix
Kinetic energy (KE), 276 definition of, 152 addition, 429
definition of turbulent, 293 for various fittings, 152 arithmetic, rules of, 406
in fluid flow, 258 Lower flammability limit (LFL), 184, 339 multiplication, 428
Kinetic theory of gases, 126 Lower heating value (LHV), 184, 436, 437, 630 Maximum continuous rating (MCR) conditions,
Knockout drums, types of, 618 Low pressure drop burners, 642 567
Low-viscosity liquids, 366 Maxwell relation, 144, 145, 146
LPG, see Liquefied petroleum gas MCR conditions, see Maximum continuous
Lube oil, 5 rating conditions
L Luminous flame, photographic view of, 101 Mean beam length, 95, 96, 100
Laminar flow, 138 Mean droplet size, 277, 278
Lumped capacitance, 77, 78, 79
convection, 83 Mean square residual (MSR), 408
Lurking variables, 416
of smoke, 141 Mean squares (MS), 409
Laminar regime, friction factor for flow in, 150 Mechanical linkage, 389
Laplacian operator, 73 Mechanical stops, 386
Medium pressure drop burner, 643
Large control panel, 382 M Metering
Large eddy simulations (LES), 297 Mach number, 144, 146, 276
Laser Doppler velocimetry (LDV), 295 air, 356
flows, 292 fuel, 353
LDV, see Laser Doppler velocimetry gas jet, 238
Le Chatelier equation, 339 Methane
maximum, 255 auto-ignition temperature of, 56
LES, see Large eddy simulations
relation, 147 concentration, predicted centerline profiles
Level controls, 628
MacLaurin series approximation, 427, 428 for excess air, 316
LFL, see Lower flammability limit
Magnetic flow meter, 388 consumption, 314, 317
LHV, see Lower heating value
Maintenance procedures, refinery damaged due graph of sustainable combustion for, 353
Liftoff, 67
to improper, 346 oxygen–nitrogen system, 258
Light fuel oil piping system, 461
Manipulated variable (MV), 394 stoichiometric O2:CH4 ratio for, 55
Lightning, 342
Manometer Methanol, 216
Light oils, 178
inclined, 129 Method of least squares, 405
Linear algebra, 427
U-tube, 128, 129 Method of steepest ascent, 416, 417
Linear burner elements, 530
Manometry, 128 Middle ear, bones in, 224
Linear transformations, 407
Manufacturing tolerances, 153 MIE, see Minimum ignition energy
Line sizing, 458
Mass Minimum ignition energy (MIE), 64, 342
Liquefied petroleum gas (LPG), 338
conservation, 133, 135, 274, 289 Minutes per repeat (MPR), 396
Liquid(s), see also Gases and liquids,
flow distribution, improvement of to Mixed convection, 82
properties of
burners, 563 Mixed-is-burnt assumption, 296
autoignition temperature, 343
selective detector (MSD), 440 Mixtures, experimental designs for, 421
firing, 364, 527 Model
flammability characteristics of, 340 Mathematical modeling, of combustion systems,
251–284 aerodynamic simulation, 566
flash point of, 179 burner, 516, 534
fuel(s), 165 burner pressure drop, 260–266
CFD, 302, 305
atomization system, 354 education processes, 252–258
combustion chemistry, 299
atomizer/spray tip configurations, 356 education processes in pilots, 256–257
development, regenerative thermal oxidizer,
capacity curve, 357 education processes in premixed
280
distribution, 279 burners, 257–258
eddy breakup, 299
firing, typical excess air values for, 490 steam flare education modeling,
flamelet, 300
high viscosity, 365 254–256
plug flow reactor, 259
physical properties of, 179 flare smokeless operation, 266–273
pollutant chemistry, 301
vaporization of, 354 application to steam flares, 268–270
radiation, 297
ignition characteristics of, 340 modeling air-assisted flares, 270–273 second-order, 428
low-viscosity, 366 predicting flame smoking tendencies, semi-empirical, 425
naphtha, 179 266–268 smokeless nozzle scaling, 270
oxidizers, 330 idealized chemical reactors and combustion turbulence, 292, 301
pour point of, 179 modeling, 258–260 validation, 295
removal, 600 perfectly stirred reactor, 259 weighted-sum-of-gray-gases, 298
seals, 619, 621 plug flow reactor, 258–259 windbox, air flow balanced by, 567
tips, 641 systems of reactors, 259–260 Modeling
viscosity for multiple constituent, 120 oil gun capacities, 273–277 air-assisted flares, 270
Liquified petroleum gas (LPG), 159 results, 277 burner pressure drop, 252
Literature review, 7–9 two-phase flow analytical development, chemical reactions, 252
combustion, 8 274–277 computational fluid dynamic, 534, see also
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

combustion in process industries, 9 oil gun development, 277–280 Computational fluid dynamics based
process industries, 9 overview, 252 combustion modeling
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 744 Tuesday, March 15, 2005 10:02 AM

744 The John Zink Combustion Handbook

eductors, 252 Newton’s law of cooling, 82, 281 fan noise abatement techniques, 243
flare smokeless rates, 252 Newton’s second law, 137, 289 flare noise abatement techniques,
heat regenerator performance, 252 NFPA, see National Fire Protection Association 239–242
oil gun performance and improvement, 252 Nitric oxide (NO), 438 valve and piping noise abatement
physical, 534 Nitrobenzene, 667 techniques, 243
sample results of simplified, 261 Nitrogen oxides (NOx), 38, 191 analysis of combustion equipment noise,
steam flare eduction, 254 analyzer, 389 243–246
Molar ratios, for combustion reactions, 37 boiler load influence on, 558 high-pressure flare, 244–246
Molecular weight, 183 burner(s) multiple burner interaction, 243–244
Molten salt system, 687 low, 202 burner combustion, 236, 237
Molten substances, 342 technology, 518 combustion
Momentum control equipment, 243
conservation, 133, 253, 275 methods, 666 instability, 234
equation, 142 technique, reduction efficiencies for, 198 contributions, based on mathematical model,
flux normal to control surface, 255 technologies, in process heaters, 199 245
Monitors, 614 degree of power functional, 559 emissions, 443
Monte Carlo method, 297 effect of boiler design on, 554 exposures, OSHA permissible, 234
Moody diagram, 151 effect of bound nitrogen in liquid fuel on, fan, 237
Motorboating, when firing oil, 512 516 flare combustion instability, 236
MPR, see Minutes per repeat effect of burner model on, 516 fundamentals of sound, 224–231
MS, see Mean squares effect of firebox temperature on, 515 basics of sound, 224–228
MSD, see Mass selective detector emission(s), 61 measurements, 228–231
MSR, see Mean square residual correlations, boiler design impacts on, gas jet, 238, 242
Muffler elbow, 265 553 glossary, 246–248
Multi-burner boilers, 559 effect of furnace cleanliness of, 561 industrial noise pollution, 231–234
Multiple burner interaction, 243 factors, for typical process heaters, 193 international requirements, 232–233
Multi-point flares, 592 fuel composition effects on, 211 noise sources and environment
Municipal solid waste boiler, 413 high, 515 interaction, 234
MV, see Manipulated variable methods to reduce, 568 OSHA requirements, 232
test data, 577 mechanisms of industrial combustion
theory for ultra-low-, 580 equipment noise, 234–239
from ultra-low emissions burner, 584 combustion roar and combustion
N empirical evidence of effect of excess air on, instability noise, 234–237
Naphtha 556 fan noise, 237–238
distillation curve, 181 formation, 60 gas jet noise, 238–239
liquid, 179 excess O2 influence on, 556 valve and piping noise, 239
NASA fundamentals, 581 meter, schematic of, 228
equilibrium code, 126 prompt-, 61 pollution, 231, 234
nozzle performance data, 253 thermal, 60 radiating from valve, 240
polynomials, 125 fuel-bound nitrogen, 537 reduction plenum, 451
National Electrical Code (NEC), 338, 379, 540 as function of burner geometry, 405 screech, 245
National Fire Protection Association (NFPA), generation, with firing natural gas, 555 shock-associated, 238
338, 379, 675 important factors affecting, 192 /visible flame, 604
Natural draft influence of oxygen on, 408 Nonluminous flame, photographic view of from
burner, 4, 357, 358, 4, 480 mechanism, fuel-bound, 61 John Zink gas burner, 102
gas burner, 27 minimization, 688 Non-smoking flares, 267
heater, logic diagram for tuning, 491 municipal solid waste boiler using ammonia Normal forces, 136, 137
Natural gas, 18, 72, 158, 528 injection to control, 413 NOx, see Nitrogen oxides
components, commercial, 159 predictions, improving, 319 Nozzle mix burner, 364
example pipeline quality, 158 reduction, 565 Null hypothesis, 409
flame produced by burning, 595 data, 570 Nusselt correlation, 282
NOx generation with firing, 555 by FGR implementation, 568 Nusselt number, 84, 281
NOx vs. heat load with firing, 577 philosophy, 561
production, 638 regulations for, 196
reforming, 14 relative steam flow vs., 557
species concentration response from burner, potential factors O
vs. excess air for, 39 affecting, 404 OASPL, see Overall sound pressure level
vs. stoichiometric ratio for, 48 strategies for reducing, 198 Occupational Safety and Health Act (OSHA),
Tulsa, 170, 171, 172 thermal, 537 232
Navier–Stokes equations, 145, 292 NO, see Nitric oxide, 438 HAZWOPER, 346
Nebraska boiler, 584 Noise, 223–249 permissible noise exposures, 234
NEC, see National Electrical Code abatement techniques, 239–243 Octave bands, 230
Newtonian fluid, 136, 292 burner noise abatement techniques, Off-gas, 352
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
Newton iteration, 259 242–243 Offshore oil rig flare, 7
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 745 Tuesday, March 15, 2005 10:02 AM

Index 745

Oil(s), 178 Out-scattering, 97 PCDF, see Polychlorinated dibenzofuran

atomizer cleaning, 463 Overall heat transfer coefficient, 76 Polychlorinated dibenzo-p-dioxin
burner(s) Overall sound pressure level (OASPL), 246 (PCDD), 219
common, 479 Oxidation reactions, 639 PDEs, see Partial differential equations
needing service, 28 Oxidizer, 19 PDPA, see Phase Doppler Particle Analyzer
tile, 455, 463 catalytic, 647 Perfectly stirred reactor (PSR), 259
troubleshooting for, 520 liquid, 330 Performance guarantee specifications, 442
crude, 173, 175 switching, 199 Petrochemical
deposits, found in United States, 166 industry, 4, 590
thermal, see Thermal oxidizers
derrick, 174 manufacturing, 638
Oxygen
discovery, successes of American, 166 process heaters, 291
analyzer, 389, 473
exploration, 169 Petroleum
concentration, in flue gas, 425
field/production plant gases, 186 industry, quantitative listing of products
-fired burner, experiencing flame lift-off, 513 condition made by U.S., 175
firing, 364 excess, 309 refinery, typical, 4
flame of insufficient, 503 refining processes, 5
flame envelope of, 482 stoichiometric, 315 PFDs, see Process flow diagrams
viewing of through burner plenum, 174 -containing gases, 601 PFR, see Plug flow reactor
fuel, 72 content PGI, see Power generation industry
train, 543 excess air indication by, 471 PHA, see Process hazard analysis
viscosity of, 183 TEG, 538 pH analyzer, 389
gun(s), 551 -enriches streams, 352 Phase Doppler Particle Analyzer (PDPA), 279
atomizers fitted onto, 585 excess, 471 Physical modeling, 534
capacities, 273 fuel chemical compound rate of reaction Physical properties of materials, 693–711
development, 277 with, 364 areas and circumferences of circles and drill
droplet size comparison between -to-fuel ratio, 38 sizes, 693–701
standard and newer, 279 levels, cost of operating wit higher excess, commercial copper tubing, 707–708
modeling, 283 474 physical properties of pipe, 702–706
oil–steam spray discharging from, 508 location for measuring excess, 472 standard grades of bolts, 709, 710, 711
performance and improvement, mass fractions, 311 PICs, see Products of incomplete combustion
modeling of, 252 P&IDs, see Piping and instrumentation
sensors, 336
heavy, 178, 180 diagrams
sources, 393
light, 178 Pilot(s), 614
trim, 393
motorboating when firing, 512 burners, 465, 483, 514
Ozone, 438
recovery, 173 eduction processes in, 256
reserves, United States, 168 flames, stable ignition of, 336
residual, 178 gas
sour crude, 176 air mixture, 514
spillage, main cause of, 507
P lines, 514
steam spray, discharge of from oil gun, 508 Packed column, 663, 665 train, 542
sweet crude, 176 Paraffinic hydrocarbons, 269 ignition, 615, 640
tip(s), 456, 484 Paraffins monitor, SoundProof acoustic, 617
cleaning, 463 combustion data for, 45 oil train, 544
drillings, 483 properties of, 713 parts, 614
train, typical pilot, 544 Parallel positioning, 389 tip, 456, 465
viscosity of mid-continent, 124 electronically linked, 391 VYD burner gas tip in diffuser with, 457
well, capping of burning, 175 mechanically linked, 390 Pipe
Olefins, 404 variation of, 391 physical properties of, 702–706
combustion data for, 45 Parallel stress, 136 racks, 381
properties of, 713 Partial differential equations (PDEs), 295 Pipeline transmission network, 158
OPEC, see Organization of Petroleum Exporting Partial premix gas, 363 Piping
Countries Particle burner, 453
Operator training and documentation, 345 explosions in, 335
carryover, 218
Optical sensing, 617 noise abatement techniques, 243
entrainment, 217
Organic heat-transfer fluid heat exchangers, 652 upstream, 604
removal efficiencies, 661
Organization of Petroleum Exporting Countries Piping and instrumentation diagrams (P&IDs),
Particulate(s), 217, 440, 640
(OPEC), 166 345
Orifice flow meter, 388 /acid gas removal, 657 Planck distribution, 88, 89
Orthogonal blocking, 416, 417 emissions, 442 Plane wall, 73, 74
Orthogonal mixture designs, 423 matter (PM), 539 Plant schematic, typical, 525
Orthogonal subspace, 423 removal, 660 Plastic
Osborn Reynolds’ experimental apparatus, 140 Pasquill–Gifford–Turner classification, 632 fiberglass-reinforced, 654
Oscillation, 398 Pass flow, 496, 498 refractory, 645, 684
OSHA, see Occupational Safety and Health Act PCB, see Polychlorinated biphenyls Platinum, use of in resistance temperature
Outlet damper flow control, 676 PCDD, see Polychlorinated dibenzo-p-dioxin detectors, 387
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

746 The John Zink Combustion Handbook

PLC, see Programmable logic controller static, 130, 131 Radiant wall burner, 369
Plenum swing adsorption (PSA), 161 photographic view of, 103
burner mounted in common, 452 switches, 383 picture of, 260
damper, 489, 490 total, 130, 131 Radiation, 87
fire, 507 transmitters, 387 in absorbing/emitting/scattering media, 93
noise reduction, 451 units of, 149 absorption, in gases, 93
Plug flow reactor (PFR), 258 upstream, 147 calculation methods, 604
Plugged gas ports, 539 velocity, 132 electromagnetic, 89
Plugging, with polymers, 462 Problem burner, 568 exchange rate, net, 96
Plume heights, estimating, 633 Process flame, 105
PM, see Particulate matter flow diagrams (PFDs), 345 furnace gas, 105
Pneumatic control valve, 385 fluid parameters, 479 heat transfer, 70
Pollutant(s) furnaces, heat transfer in, 102 analysis of, 106
accurate measurements of, 197 hazard analysis (PHA), 337, 344, 345 correction factor, for mixtures of water
chemistry models, 301 heater(s), 11 vapor and carbon dioxide, 100
Pollutant emissions, 189–220, 432 examples of, 12, 16 in cylindrical furnace, 107
carbon dioxide, 218 heat balance, 21 intensity of, 94
combustibles, 214–217 NOx control technologies in, 199 models, 297
CO and unburned fuel, 214–215 oil flame, 276 refractory surface, 106
volatile organic compounds, 215–217 heat transfer tube, cross section of, 109 thermal, 88
conversions, 190–191 industries, 4–7, 9 transport equation (RTE), 297
dioxins and furans, 219 hydrocarbon and petrochemical wavelength of, 88
emission in hydrocarbon and petrochemical industries, 4–5 Radiative exchange, network representation of
industries, 190 power generation industry, 5–7 between surfaces, 92
nitrogen oxides, 191–214 thermal oxidation, 7 Radiative heat transfer, 137
abatement strategies, 198–204 tubes, 477, 493 Radiometer schematic, ellipsoidal, 443
field results, 204–214 variable (PV), 395, 396, 397–398 Radiosity, 89, 91
measurement techniques, 197–198 Products of incomplete combustion (PICs), 309 Random effects, 417
regulations, 196–197 Random error, 408
Profibus, 375
theory, 192–196 Raw gas burner(s), 206, 364, 366, 481, 482
Programmable logic controller (PLC), 374, 375,
particulates, 217–218 effect of excess oxygen on NOx in, 515
380, 540
sources, 217–218 typical throat of, 357
Prompt-NOx formation, 61
treatment techniques, 218 Reaction
Propane, 23, 170, 595
SOx, 219 efficiency, 630
Propylene, 23, 595
Polychlorinated biphenyls (PCB), 683 rate, 60
PSA, see Pressure swing adsorption
Polychlorinated dibenzofuran (PCDF), 219 Reactor(s)
PSR, see Perfectly stirred reactor
Polymerization, 5 plug flow, 258
Pulsating flame, 502
Polymers, plugging with, 462 systems of, 259
Pulse-jet, 658
Polynomial expression, for combustion gases, Reboiler, xylene, 306, 307
Pulverized coal combustor, 291
125 Reburning, 201
Pure error, 412
Portable analyzers, 472 Recuperative preheat exchanger, 649
Pure tone, 224
Port mix twin fluid atomizers, 365 Recycle flue gas, 668
Purge
Positioner, 386 Redlich–Kwong equation, 128
Position switches, 383 controls, 629 Reed wall, 504
Positive displacement flow meters, 388 gas, 597, 601 Refinery
Potential energy, change in, 130 reduction seals, 620 /chemical plant fuels, 528
Pour point, of liquid, 179 PV, see Process variable damage of due to improper maintenance
Power generation industry (PGI), 5, 361 Pyrometer, infrared, 478 procedures, 346
Prandtl–Meyer expansion waves, 147 flow diagram, 6, 176
Premix fuel composition, 446
burner, 21, 22, 480, 481 gas(es), 159, 186
gas, 363
Q comparison of to test blend, 437
metering orifice spud, 355 Quasi-one-dimensional isentropic flow, 145 composition of typical, 160
Pressure Quench/two-stage acid removal system, 666 example, 437
atmospheric, 148 heaters, 11
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

constant, 258 Reformate extraction, 14

definition of, 148 Reformers
drop, 148, 152 R terrace wall, 10
dynamic, 131 Radial blade operating curve, 673 top-fired, 10
energy, change in, 130 Radial coordinate systems, 76, 79 Reforming furnaces, 34
forces, 136, 137 Radial inlet damper, 676 Refractory(ies)
fuel, 213 Radiant exchanges castable, 646, 684, 689
gage, 149, 495 between black surfaces, 90 failure, 487
relief vessel, venting to flare, 131 between gray/diffuse surfaces, 91 hard, 645
stagnation, 148 Radiant intensity, 94 plastic, 645, 684
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 747 Tuesday, March 15, 2005 10:02 AM

Index 747

soft, 645, 646 Seal(s) Soil fuel carryover, 218

surface radiation, 106 liquid, 619 Solenoid valve(s)
Refrigerants, chemical properties of, 717–722 loop, 629 most common types of, 384
Regenerative thermal oxidizer (RTO), 280 purge reduction, 620 three-way, 386
bed retrofit applications, 282 smoke signals from surging liquid, 620 Solid desiccants, adsorption with, 158
model development, 280 Second law of thermodynamics, 71 Solvent deasphalting, 5
burner pressure drops, 283 Second-order designs, 419 Soot
Regen tile, 366 Second-order regression model, 428 blowers, online cleaning with, 688
Reheater (RH), 561 Seconds Saybolt Furol (SSF), 182 formation
Relative NOx, 559 Seconds Saybolt Universal (SSU), 182 favorable condition for, 279
Relative roughness, 150 SED, see Statistical experimental design precursors to, 298
Relay system, 374 Selective catalytic reduction (SCR), 203, 550, rates, 267
Renormalization Group (RNG), 302 667 presence of in flame, 298
Selective noncatalytic reduction (SNCR), 203, SOPs, see Standard operating procedures
Repeats per minute (RPM), 396
204, 550, 667 Sound(s)
Residual fuel oil (RFO), 561–562
Self-supported flare, 624 fundamentals of, 224
Residual oils, 178
SEM, see Scanning electron microscopy level(s)
Resistance temperature detectors (RTDs), 387
Semi-empirical models, 425 meter, block diagram of, 229
Reynolds averaging, 139, 292, 301
Semi-Implicit Method for Pressure Linked overall, 229
Reynolds number, 83, 138, 145, 180
Equations (SIMPLE), 295 of various sources, 234
Reynolds stress model (RSM), 302 Semi-infinite solids, transient conduction in, 80 low-frequency, 228
RFO, see Residual fuel oil Sequential experimental strategy, 426 power level, 226
RH, see Reheater Serial correlation, 416 pressure level (SPL), 225, 226, 235
RIMFIRE® flare burner, 611, 613 SG, see Specific gravity aircraft-to-ground propagation in, 231
RNG, see Renormalization Group SH, see Superheater burner, 237
Rockets, industrial combustion and, 8 Shadow photograph, of burning butane lighter, spectrum, of high-pressure flare, 245
Rosin–Rammler relationship, 278 236 spectrum, 233
Rotatability, 420 Shape factor, 92 speed, definition of, 146
Roughness, 149, 150 Shear stress, 119, 136, 137 SoundProof acoustic pilot monitor, 617
Round-shaped flame, 362 Shock-associated noise, 238 Sour crude oil, 176
RPM, see Repeats per minute Shock waves, 245 South Coast Air Quality Management District
RSM, see Reynolds stress model Shooting Runge–Kutta solver, 277 (SCAQMD), 196
RTD, see Resistance temperature detectors Shutdown string, 381 Species, conservation of, 134, 138
RTE, see Radiation transport equation Sidewall fired burner arrangement, 20 Specific collector area (SCA), 659
RTO, see Regenerative thermal oxidizer SIMPLE, see Semi-Implicit Method for Specific collector surface (SCS), 659
Runge–Kutta solver, shooting, 277 Pressure Linked Equations Specific gravity (SG), 182
Run indicators, 384 Simplex-centroid design, 421 Specific heat, 125, 182, 184
Runner pipes, severe sagging of, 539 Simplex designs, 415 Spectral blackbody emissive power, 90
Rust preventatives, 175 Simulated refinery gas Spill-return flow atomizer, 585
species concentration vs. excess air for, 39 Spin diffuser, 464
species concentration vs. stoichiometric SPL, see Sound pressure level
ratio for, 48 Spontaneous ignition, 342
S Simulated reforming gas flame, 169 Square matrix, 429
Saddle data, 281 Simulation model, aerodynamic, 566 SSF, see Seconds Saybolt Furol
Single-point flares, 591, 593 SSOVs, see Safety shutoff valves
Safety
Slipstream flame-front generator, 615, 616 SSU, see Seconds Saybolt Universal
documentation, operator training and, 344
Small control panel, 382 Stabilizer, damaged, 464
requirements, 597
SMD, see Sauter mean diameter Stack(s)
shutoff valves (SSOVs), 379, 384
Smoke damper, 495, 457
Salts/solids systems, 686
emission, from stack, 514 emissions, 496
Sankey diagram, 64
generation, 334 explosions in, 335
Saturated factorial, 415 from incense, 138 gas reheat, 526
Saturation quench section, 653 laminar flow of, 141 smoke emission from, 514
Sauter mean diameter (SMD), 277–278 suppression, 598, 603 temperature, high, 511
SBCR, see Selective noncatalytic reduction Smokeless burning, 603 Staged air burner, 25, 517
SCA, see Specific collector area Smokeless flare Staged burner design philosophy, 551
Scaling functional, 272 capacity, 266 Staged combustion, 200, 573
Scanning electron microscopy (SEM), 441 early model, 591 Staged flare system, general arrangement of,
SCAQMD, see South Coast Air Quality of paraffinic hydrocarbons, 269 598
Management District Smokeless nozzle scaling model, calibration of, Staged fuel burner, 26, 517, 518
SCR, see Selective catalytic reduction 270 Stagnation pressure, 148
Screech noise, 239, 240, 245 Smokeless steam flare, 272 Standard deviation, 409
Screening designs, 415 Smoking flares, 266 Standard errors of effects, 412
Scrubbers, 664, 665, 687 SNCR, 669 Standard operating procedures (SOPs), 345
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
SCS, see Specific collector surface Soft refractory, 645, 646 Stanton number, 84, 85
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 748 Tuesday, March 15, 2005 10:02 AM

748 The John Zink Combustion Handbook

START box, 490 TE, see Thermal energy non-acid gas endothermic waste
Static electricity, 343 TEG, see Turbine exhaust gas gas/waste liquid system, 677–681
Static mixers, 160 Temperature non-acid gas exothermic waste gas/waste
Static pressure, 130, 131, 596 bridgewall, 496 liquid system, 681–682
Static sparks, 342 control valve (TCV), 498 NOx minimization or reduction systems,
Statistical experimental design (SED), 402 distribution, in cylinder, 76 688–689
combining domain knowledge with, 424 drop, due to thermal contact resistance, 75 salts/solids systems, 686–688
contrast of classical experimentation and, firebox, effect of on NOx, 515 system generating steam, 679
403 flame, reduced, 559 Thermal radiation, 88
principles, 404 flue gas, 476 Thermoacoustic efficiency (TAE), 235
Steady-state conduction equation, 77 fuel, 476, 563 Thermocouples, 386, 387, 478
Steam gas, 595 Thermodynamics
-assisted flare, 237, 241 indicating controller (TIC), 389 equilibrium and, 47
control, automatic, 627 process tube, 477 Maxwell relation of classical, 144
flare(s), 268 stack, high, 511 relations, 144
eduction modeling, 254 switches, 383 second law of, 71
with steam eductor, 254 tube, 487, 496 Thermophoresis, 661
tube layout, third-generation, 254 Terrace wall Thermowell, 386
flaring smoking tendencies, predicting, 269 furnace design, 111 Three-component system, 422
-to-hydrocarbon ratios, 269, 271 reformers, 10, 436 Three-way solenoid valve, 386
injection, 201, 607 Test Threshold of pain, 227
manifold, 606 fuel storage tanks, 436 Throat velocity, 147
reforming, composition of, 161 furnace TIC, see Temperature indicating controller
superheater, 14 selection criteria, 434 Tile
trapped, 333 for simulation of up-fired tests, 435 burner refractory, 506
Steamizer™, 610 matrix, 442, 444 height, 455
Stefan–Boltzmann law, 79, 90 procedure installation, 453, 454
Stoichiometric oxygen condition, 315 gas specification sheet, 445–446 oil burner, 455, 463
Stoichiometric ratio, excess air and, 38 example, 447 tolerances, 454
Submerged quench, 655 Tetracontane, 177 Time-averaged emission, 298
Substoichiometric combustion, 46, 47 Theoretical fluid dynamics, 288 TNG, see Tulsa Natural Gas
Suction pyrometer, 207, 386 Thermal conductivity, 71, 81 Todd Combustion low NOx burner, 578, 579
Sulfur Thermal contact resistance, temperature drop Toluene, 216
-bearing hydrocarbon systems, 682 due to, 75 Top-fired reformers, 10
dioxide, 539 Thermal cracking, 13, 14 Total dry products, 38

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---
oxides, 219 Thermal diffusivity, 78 Total pressure, 130, 131
recovery reaction furnace, 307 Thermal energy (TE), 276 Total thermal resistance, 74
Superheater (SH), 34, 561 Thermal NOx, 60, 537 Total wet products, 38
Supersonic eductor performance, analysis of, Thermal oxidizer(s), 8, 637–689 Touchscreen, 376
254 basic system building blocks, 640–670 Transducer, current-to-pressure, 385
Surface burners, 640–644 Transformers, ignition, 384
convection condition, 81 catalytic systems, 646–647 Transient conduction, 77
exchange, 93 flue gas processing methods, 647–670 Transmitters, pressure, 387
Sweet crude oil, 176 furnace/thermal Transport
Sweetening/sulfur removal, 5 oxidizer/incinerator/combustion characteristics of turbulent, 292
Swirler, 361 chamber, 644–645 equations, 292
Swirl vanes, 583 refractory, 645–646 properties, 122
Switches blowers, 670–674 Troubleshooting, 501–520
flow, 384 combustion basics, 639–640 flame impingement on tubes, 503–504
position, 383 carbon monoxide, acid gases, 639–640 cause and effect on operation, 503–504
pressure, 383 material and energy balance, 639 corrective action, 504
temperature, 383 NOx formation, 639 indications of problem, 503
System shutdown oxidizing/reducing combustion flame lift-off, 513
local request required after, 380 processes, 639 corrective action, 513
unsatisfactory parameter, 380 particulate, 640 effect on operation, 513
control systems and instrumentation, indications of problem, 513
675–677 flashback, 505
flame inside, 317 cause and corrective action, 505
T geometric information describing, 314 effect on operation, 505
TAE, see Thermoacoustic efficiency picture of, 260 indications of problem, 505
Tanks, explosions in, 335 sample results of simplified modeling for, high fuel pressure, 510
Taylor series approximation, 427 262 effect on operation, 510
TCV, see Temperature control valve system configurations, 677–689 indications of problem, 510
TDMA algorithm, 295 acid gas systems, 682–685 solution and corrective action, 510
Copyright CRC Press
Provided by IHS Markit under license with CRC Press Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Elsheikh, Baher
No reproduction or networking permitted without license from IHS Not for Resale, 01/15/2018 21:50:21 MST
2337-Index-frame Page 749 Tuesday, March 15, 2005 10:02 AM

Index 749

high NOx emissions, 515–519 Tuning, 398 control, 385, 397

corrective action, 515–519 Turbine gas, data, 392
effect on operation, 515 exhaust gas (TEG), 352, 524, 531 noise radiating from, 240
indication of problem, 515 oxygen content, 538 pneumatic control, 385
high stack temperature, 511 quenching effects of, 529 remote isolation, 333
corrective action, 511 flow meters, 388 safety shutoff, 379, 384
effect on operation, 511 power augmentation, 532 solenoid, 384, 386
indications of problem, 511 Turbulence steam control, 626
irregular flame patterns, 505–507 equations, 292 Van Driest hypothesis, 293
cause and corrective action, 506–507 Leonardo da Vinci’s view of, 140 Vapor(s)
effect on operation, 506 model(s), 292, 301 chemical properties of, 717–722
indications of problem, 505–506 /radiation interaction, 298, 299 pressures, for light hydrocarbons, 341
leaning flames, 509–510 Turbulent combustion regimes of, 299 Vaporization process, 337
causes and corrective action, 509–510 Turbulent energy, 294 Variable loop gain, 398
effect on operation, 509 Turbulent external flow, 85 Variance, 408
indications of problem, 509 Turbulent flow, 138 VC process heater, see Vertical cylindrical
long smoky flames, 507–508 Turbulent internal flow, 84 process heater
corrective action, 508 Turbulent kinetic energy, definition of, 293 Velocity
effect on operation, 508 Turndown operation, heater, 492 boundary layer thickness, 85
indications of problem, 507 Two-level factorials, 424 constant, 134
main burner fails to light-off or extinguishes design, 412 energy, change in, 130
while in service, 508–509 fractional, 414 gradient, mathematical expression of, 119
corrective action, 508–509 Two-phase flow analytical development, 274 pressure, 132
effect on operation, 508 thermocouple, 386, 387, 478
indications of problem, 508 Venturi eductor optimization, 319
motorboating when firing oil, 512–513 Venturi quench, adjustable-plug, 656
effect on operation, 512 U Venturi scrubber, 665, 687
indications of problem, 512 UFL, see Upper flammability limit Venturi-style burners, 558
solution and corrective action, 512–513 UHCs, see Unburned hydrocarbons Venturi-style low NOx burner, 522
oil spillage, 507 UL, see Underwriters Laboratories Venturi tube, burning inside, 505
cause and effect on operation, 507 Ultra-low emissions burner, 580, 581, 582, 583 Vertical cylindrical (VC) process heater, 16, 34
corrective action, 507 Ultra low-NOx burner, 517 Vertical tube furnace, oil and gas firing in, 104
indications of problem, 507 Ultrasonic flow meter, 388 View factor(s)
overheating of convection section, 511–512 Unburned fuel, 214 aligned parallel rectangles, 97
corrective action, 512 Unburned hydrocarbons (UHCs), 433, 434, 439, coaxial parallel disks, 97
effect on operation, 512 538, 548, 597 computation of, 106
indications of problem, 511 Underwriters Laboratories (UL), 540 diffuse surfaces, 92
pilot burner fails to ignite or extinguishes Universal gas constant, 126 perpendicular rectangles with common
while in service, 514 Unsatisfactory parameter system shutdown, 380 edge, 97
corrective action, 514 Unstable flame, 494 two-dimensional geometries, 94
effect on operation, 514 Upfired burner arrangement, 20 Visbreaking, 5, 13
indications of problem, 514 Up-fired tests, test furnace for simulation of, 435 Viscosity, 180, 185
pulsating flame, 502–503 Upper flammability limit (UFL), 184, 339 absolute, 119
cause and effect on operation, 502–503 Upstream pressure, 147 conversion table, 121
corrective action, 503 U.S. EPA, 197 definition of, 119
indications of problem, 502 sampling system schematic as recommended fluid, 139
smoke emission from stack, 514–515 by, 198 fuel oils, 183
corrective action, 514–515 sampling train, flue gas drawn through, 440 kinematic, 120
effect on operation and equipment, 514 Utility mid-continent oils, 124
indications of problem, 514 boilers, 549, 572 temperature vs. for hydrocarbons, 123
Tube(s) burners, NOx vs. load with firing natural gas Viscous flow, 290
bank, constants of equation for, 87 on, 559 Visual Basic, 277
failures, 331, 333 costs, 596 VOCs, see Volatile organic compounds
flame impingement on, 503 U-tube manometer, 128, 129 Volatile organic compounds (VOCs), 214, 215
furnaces, used in hydrogen production, 8 Vortex shedder flow meter, 387
metal temperature, 496
rupture, in fired heater, 332
support color, 494 V
temperatures, 487 Vacuum distillation, 5 W
-to-wall spacing, 110 Valve(s) Wake area showing mixing vortices, 141
venturi, 505 atomizing medium block, 486 Wall
Tulsa Natural Gas (TNG), 170, 207, 208, 256, burner block, 485, 495 composite, 74
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

437 butterfly-type combustion air, 391 -fired burner, 15

750 The John Zink Combustion Handbook

-fired flat flame burner, 368 Wave patterns, development of orderly, 238 WSPA, see Western States Petroleum
plane, 73 Weighted-sum-of-gray-gases (WSGG) Association
Reed, 504 method, 306
to wall equations, 288 model, 298
Warped flame holders, 539 Welding, 335, 607
Waste(s) WESPs, see Wet electrostatic precipitators
X
combustion systems, 670 Western States Petroleum Association (WSPA), Xylene
gas 190 isomerization, 14
composition, 599 Wet electrostatic precipitators (WESPs), 660, reboiler, 306, 307
streams, combustible, 160 662
introduction, 641 advantages of, 663
salt-containing, 686 flue gas coming into, 662
Wien’s displacement law, 90
Z
water-based, 687
ZDR severe service flare tips, see Zink double
Water Windbox
refractory severe service flare tips
-based wastes, 687 air flow modeling, 563 Zeldovich mechanism, 192
exiting faucet at low velocity, 139 model, air flow balanced by, 567 Zink double refractory (ZDR) severe service
from faucet showing transition, 141 Wing geometry, 534 flare tips, 606, 607
gas shift reaction, 47 Wobbe index, 436 Zink Thermal Oxidizer Flare (ZTOF), 598, 622
injection, 201, 241 Woofing, 502 burner systems, 623
vapor, emissivity of, 99 Word length, 415 schematic of, 623
Watertube boiler, 649 WSGG, see Weighted-sum-of-gray-gases ZTOF, see Zink Thermal Oxidizer Flare
--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

--`,,`,,``,,``,```````,,`,,,`-`-`,,`,,`,`,,`---

Copyright CRC Press

Combustion Handbook - Fundamentals For Power, Marine & Industrial Applications - The Coen & Hamworthy
100% (8)
Combustion Handbook - Fundamentals For Power, Marine & Industrial Applications - The Coen & Hamworthy
555 pages
Combustion Handbook 2 PDF
100% (4)
Combustion Handbook 2 PDF
471 pages
Combustion Engineering and Gas Utilization by Cornforth, J. R.
100% (3)
Combustion Engineering and Gas Utilization by Cornforth, J. R.
908 pages
Burner Handbook
100% (8)
Burner Handbook
169 pages
Industrial Furnaces, 0471387061
90% (10)
Industrial Furnaces, 0471387061
492 pages
North America Combustion Handbook Vol2 PDF
100% (4)
North America Combustion Handbook Vol2 PDF
471 pages
Burners and Their Controls
100% (2)
Burners and Their Controls
75 pages
(V. Ganapathy) Steam Plant Calculations Manual
91% (11)
(V. Ganapathy) Steam Plant Calculations Manual
226 pages
Practical Guide To Industrial Boiler Systems
100% (20)
Practical Guide To Industrial Boiler Systems
375 pages
PDF
No ratings yet
PDF
5 pages
Burner Calculation
67% (3)
Burner Calculation
29 pages
Boiler Control Systems Engineering
100% (17)
Boiler Control Systems Engineering
79 pages
DARBY - Chemical Engineering Fluid Mechanics 3edition PDF
100% (8)
DARBY - Chemical Engineering Fluid Mechanics 3edition PDF
578 pages
Burners Handbook
50% (2)
Burners Handbook
169 pages
Basic of Boiler Design
100% (6)
Basic of Boiler Design
186 pages
Industrial Steam Systems Fundamentals and Best Design Practices PDF
100% (6)
Industrial Steam Systems Fundamentals and Best Design Practices PDF
236 pages
Fluidized Bed Combustion
100% (7)
Fluidized Bed Combustion
600 pages
John Zink Combustion Handbook
No ratings yet
John Zink Combustion Handbook
148 pages
The John Zink Combustion Handbook Second Edition 3 Volume Set Baukal Download Full Chapters
No ratings yet
The John Zink Combustion Handbook Second Edition 3 Volume Set Baukal Download Full Chapters
167 pages
1386 - FM Front Page Industrial Burner Handbook
0% (1)
1386 - FM Front Page Industrial Burner Handbook
14 pages
(Ebook) Oxygen-Enhanced Combustion by Baukal, Charles E ISBN 9781439862308, 1439862303 Full
No ratings yet
(Ebook) Oxygen-Enhanced Combustion by Baukal, Charles E ISBN 9781439862308, 1439862303 Full
109 pages
(Ebook) Oxygen-Enhanced Combustion by Baukal, Charles E ISBN 9781439862308, 1439862303 Sample
No ratings yet
(Ebook) Oxygen-Enhanced Combustion by Baukal, Charles E ISBN 9781439862308, 1439862303 Sample
166 pages
The John Zink Combustion Handbook Second Edition 3 Volume Set Baukal Sample
100% (1)
The John Zink Combustion Handbook Second Edition 3 Volume Set Baukal Sample
114 pages
The Coen Hamworthy Combustion Handbook Fundamentals For Power Marine Industrial Applications 1st Edition Stephen Londerville Full Digital Chapters
100% (2)
The Coen Hamworthy Combustion Handbook Fundamentals For Power Marine Industrial Applications 1st Edition Stephen Londerville Full Digital Chapters
90 pages
The John Zink Combustion Handbook Second Edition 3 Volume Set Baukal Latest PDF 2025
No ratings yet
The John Zink Combustion Handbook Second Edition 3 Volume Set Baukal Latest PDF 2025
134 pages
Steam Generation Solutions
No ratings yet
Steam Generation Solutions
6 pages
Catalogo Steam Generation Solutions
No ratings yet
Catalogo Steam Generation Solutions
6 pages
The Coen Hamworthy Combustion Handbook Fundamentals For Power Marine Industrial Applications 1st Edition Stephen Londerville PDF Download
No ratings yet
The Coen Hamworthy Combustion Handbook Fundamentals For Power Marine Industrial Applications 1st Edition Stephen Londerville PDF Download
116 pages
Industria1 C o M B U S T I o N Pollution and C o N T R o L PDF
No ratings yet
Industria1 C o M B U S T I o N Pollution and C o N T R o L PDF
890 pages
The Coen Hamworthy Combustion Handbook Fundamentals For Power Marine Industrial Applications 1st Edition Stephen Londerville Online Reading
No ratings yet
The Coen Hamworthy Combustion Handbook Fundamentals For Power Marine Industrial Applications 1st Edition Stephen Londerville Online Reading
94 pages
Baukal 2003 Até PG 90
No ratings yet
Baukal 2003 Até PG 90
91 pages
Combustion Engineering 2nd Edition Kenneth W. Ragland (Author) Sample
100% (3)
Combustion Engineering 2nd Edition Kenneth W. Ragland (Author) Sample
82 pages
COEN Startup Burners
No ratings yet
COEN Startup Burners
8 pages
Baukal CH., Ed. Oxygen-Enhanced Combustion (CRC, 1998) (T) (356s) PDF
100% (2)
Baukal CH., Ed. Oxygen-Enhanced Combustion (CRC, 1998) (T) (356s) PDF
356 pages
Combustion Engineering 2nd Edition Kenneth W. Ragland Full Chapters Instanly
100% (4)
Combustion Engineering 2nd Edition Kenneth W. Ragland Full Chapters Instanly
109 pages
Combustion Engineering 2nd Edition Kenneth W. Ragland (Author) Instant Download
No ratings yet
Combustion Engineering 2nd Edition Kenneth W. Ragland (Author) Instant Download
51 pages
COEN 13116 Duct Burner Solutions LR Web
No ratings yet
COEN 13116 Duct Burner Solutions LR Web
2 pages
Introduction To Heat Transfer
No ratings yet
Introduction To Heat Transfer
15 pages
Synthesis Gas Combustion Fundamentals and Applications 1st Edition Tim Lieuwen Online PDF
100% (4)
Synthesis Gas Combustion Fundamentals and Applications 1st Edition Tim Lieuwen Online PDF
139 pages
Fire Fundamentals and Control First Edition Haessler Download
No ratings yet
Fire Fundamentals and Control First Edition Haessler Download
171 pages
Clinic Planning 2010 Final
No ratings yet
Clinic Planning 2010 Final
8 pages
Handbook of Induction Heating, Second Edition Valery Rudnev Latest PDF 2025
No ratings yet
Handbook of Induction Heating, Second Edition Valery Rudnev Latest PDF 2025
154 pages
Toc
No ratings yet
Toc
11 pages
Jan F. Kreider Et Al. "Introduction To The Buildings Sector" Ed. Jan F. Kreider Boca Raton, CRC Press LLC. 2001
No ratings yet
Jan F. Kreider Et Al. "Introduction To The Buildings Sector" Ed. Jan F. Kreider Boca Raton, CRC Press LLC. 2001
11 pages
Combustion Fossil Power PDF
100% (3)
Combustion Fossil Power PDF
1,114 pages
Handbook of Induction Heating, Second Edition Valery Rudnev Available All Format
No ratings yet
Handbook of Induction Heating, Second Edition Valery Rudnev Available All Format
82 pages
Statistical Quality Control 1st Edition M. Jeya Chandra Online Version
100% (2)
Statistical Quality Control 1st Edition M. Jeya Chandra Online Version
82 pages
John Zink - ATTACH H VCU Catalogs Leaflets (Brochure)
No ratings yet
John Zink - ATTACH H VCU Catalogs Leaflets (Brochure)
8 pages
COE 0003 Industrial Burner Brochure M Web
No ratings yet
COE 0003 Industrial Burner Brochure M Web
8 pages
Catalogo COOLstar Burner
No ratings yet
Catalogo COOLstar Burner
2 pages
BMS Catalog
No ratings yet
BMS Catalog
28 pages
Homogeneous Charge Compression Ignition HCCI and Controlled Auto Ignition CAI Engines For The Automotive Industry 1st Edition Hua Zhao Download
No ratings yet
Homogeneous Charge Compression Ignition HCCI and Controlled Auto Ignition CAI Engines For The Automotive Industry 1st Edition Hua Zhao Download
76 pages
Combustion Engineering 2nd Edition Kenneth W. Ragland Full Digital Chapters
100% (2)
Combustion Engineering 2nd Edition Kenneth W. Ragland Full Digital Chapters
163 pages
Reducing Quench
No ratings yet
Reducing Quench
68 pages
Handbook of Induction Heating, Second Edition Valery Rudnev Newest Digital Edition
No ratings yet
Handbook of Induction Heating, Second Edition Valery Rudnev Newest Digital Edition
162 pages
BACnet Application Guide For Consulting Engineers 125-1984T
No ratings yet
BACnet Application Guide For Consulting Engineers 125-1984T
170 pages
Combustion and Incineration Processes Applications in Environmental Engineering Third Edition Environmental Science Pollution Walter R. Niessen Download
No ratings yet
Combustion and Incineration Processes Applications in Environmental Engineering Third Edition Environmental Science Pollution Walter R. Niessen Download
52 pages
Bij 4 01 2012
No ratings yet
Bij 4 01 2012
24 pages
Boiler C&i
No ratings yet
Boiler C&i
28 pages
BI IDCbookextract REV3
No ratings yet
BI IDCbookextract REV3
28 pages
Introduction To Process Engineering and Design, Second Edition PDF
92% (13)
Introduction To Process Engineering and Design, Second Edition PDF
1,138 pages
Industrial Ventilation A Manual of Recommended Practice For Design 9781607261087 Compress
100% (12)
Industrial Ventilation A Manual of Recommended Practice For Design 9781607261087 Compress
686 pages
Process Plant Design and Simulation Handbook
100% (14)
Process Plant Design and Simulation Handbook
468 pages
Heat Transfer Calculations
83% (12)
Heat Transfer Calculations
727 pages
Heat Exchanger Design Essentials
92% (49)
Heat Exchanger Design Essentials
972 pages
Chemical Engineering Design Principles Sinnott PDF
100% (13)
Chemical Engineering Design Principles Sinnott PDF
1,436 pages
Pump Handbook
100% (15)
Pump Handbook
257 pages
ASME B31.3 Piping Workshop
96% (52)
ASME B31.3 Piping Workshop
382 pages
Chemical Engineering With Computer Applications
93% (15)
Chemical Engineering With Computer Applications
287 pages
Valves, Piping and Pipelines Handbook, Third Edition
98% (54)
Valves, Piping and Pipelines Handbook, Third Edition
696 pages
Advanced Heat Transfer, 3rd Edition
100% (12)
Advanced Heat Transfer, 3rd Edition
578 pages
Stoichiometry and Process Calculations (2017) PDF
100% (10)
Stoichiometry and Process Calculations (2017) PDF
613 pages
Introduction To Process Engineering and Design @ProcessEng
100% (10)
Introduction To Process Engineering and Design @ProcessEng
761 pages
Process Plant Layout and Piping Design
95% (38)
Process Plant Layout and Piping Design
361 pages
Chemical Engineering Design Guide
100% (11)
Chemical Engineering Design Guide
170 pages
Power Plant Engineering - Nodrm
100% (12)
Power Plant Engineering - Nodrm
1,142 pages
Handbook of Mechanical Engineering Terms
96% (51)
Handbook of Mechanical Engineering Terms
366 pages
Modeling and Simulation of Chemical Process Systems (2019) PDF
100% (11)
Modeling and Simulation of Chemical Process Systems (2019) PDF
519 pages
Piping Materials Guide
90% (10)
Piping Materials Guide
356 pages
Thermodynamics For Engineers
100% (14)
Thermodynamics For Engineers
620 pages
Fundamentals of Engineering Thermodynamics 2nbsped 812032790x 9788120327900 Compress
100% (6)
Fundamentals of Engineering Thermodynamics 2nbsped 812032790x 9788120327900 Compress
720 pages
Duct Systems Design Guide
100% (14)
Duct Systems Design Guide
186 pages
Process Engineering Design Guide PDF
90% (10)
Process Engineering Design Guide PDF
90 pages
Fundamentals of Process Safety Engineering
100% (10)
Fundamentals of Process Safety Engineering
499 pages
Heat Exchanger Design Handbook
100% (19)
Heat Exchanger Design Handbook
1,245 pages
Pipe Drafting and Design Third Edition PDF
93% (41)
Pipe Drafting and Design Third Edition PDF
475 pages
Piping
96% (46)
Piping
739 pages
Stress Concentration Vs Stress Singularity 1751747300
No ratings yet
Stress Concentration Vs Stress Singularity 1751747300
15 pages
Static Equipment Awarness Session - 02
100% (5)
Static Equipment Awarness Session - 02
156 pages
Materials Selection and Corrosion Control
No ratings yet
Materials Selection and Corrosion Control
1 page
Nureg 6007
No ratings yet
Nureg 6007
122 pages
Heat Exchangers Training Guide
100% (1)
Heat Exchangers Training Guide
24 pages
Heat Exchangers Course Baher Part04
No ratings yet
Heat Exchangers Course Baher Part04
32 pages
Clamp Design Baher 1651934316
No ratings yet
Clamp Design Baher 1651934316
10 pages
Static Equipment Material Selection
50% (2)
Static Equipment Material Selection
58 pages
Key Changes in API 571-2020 Part 1
100% (6)
Key Changes in API 571-2020 Part 1
66 pages
Mock Up of Tube To Tubesheet Joints in Shell and Tube Heat Exchangers
100% (2)
Mock Up of Tube To Tubesheet Joints in Shell and Tube Heat Exchangers
4 pages
SYN Catalyst Migration Problem of 2000 MTPD Ammonia Plant 1754577147
No ratings yet
SYN Catalyst Migration Problem of 2000 MTPD Ammonia Plant 1754577147
10 pages
Heat Exchangers Course Baher Part03
100% (2)
Heat Exchangers Course Baher Part03
18 pages
Heat Exchangers Course Baher Part02
100% (1)
Heat Exchangers Course Baher Part02
21 pages
Codes and Standards: Petroleum Engineering Technology (Pet) Department Polytechnic Abu Dhabi (Adpoly)
100% (5)
Codes and Standards: Petroleum Engineering Technology (Pet) Department Polytechnic Abu Dhabi (Adpoly)
37 pages
Utility Boiler Condition Assessment
100% (1)
Utility Boiler Condition Assessment
12 pages
FSA Gasket Handbook - June 2017
No ratings yet
FSA Gasket Handbook - June 2017
138 pages
RR 253
No ratings yet
RR 253
47 pages
Flange and Gasket Detail PDF
No ratings yet
Flange and Gasket Detail PDF
74 pages
Supercritical and Ultra-Supercritical Power Plants - SEA's Vision or Reality?
No ratings yet
Supercritical and Ultra-Supercritical Power Plants - SEA's Vision or Reality?
23 pages
Air-Cooled Heat Exchanger Weld Integrity
No ratings yet
Air-Cooled Heat Exchanger Weld Integrity
6 pages
Gaskets
No ratings yet
Gaskets
58 pages
54521
No ratings yet
54521
29 pages
3-34C-5 Tube To Tubesheet Welds Integrity
No ratings yet
3-34C-5 Tube To Tubesheet Welds Integrity
5 pages
4688 Tube To Tube Sheet Weld Scanning Service
No ratings yet
4688 Tube To Tube Sheet Weld Scanning Service
2 pages
5 Shell-And-tube Heat Exchanger - Eckrich
No ratings yet
5 Shell-And-tube Heat Exchanger - Eckrich
22 pages
Repair of Tube-Tubesheet Weld Cracks in A Cracked Gas/Steam Heat Exchanger
No ratings yet
Repair of Tube-Tubesheet Weld Cracks in A Cracked Gas/Steam Heat Exchanger
7 pages
Static Equipment Webinar for Engineers
100% (12)
Static Equipment Webinar for Engineers
95 pages
ISO and EIGA Standards For Cryogenic Vessels and Accessories
100% (1)
ISO and EIGA Standards For Cryogenic Vessels and Accessories
36 pages
Attock Refinery Limited (Annual Report 2007)
100% (1)
Attock Refinery Limited (Annual Report 2007)
124 pages
American Fuel Oil Blend Components.: General
No ratings yet
American Fuel Oil Blend Components.: General
2 pages
Marine Fuel Stability and Compatibility Testing
No ratings yet
Marine Fuel Stability and Compatibility Testing
4 pages
MN-2-011-27 Bunker Pollution Convention PDF
No ratings yet
MN-2-011-27 Bunker Pollution Convention PDF
6 pages
Gas Turbines Fuels Flexibility
No ratings yet
Gas Turbines Fuels Flexibility
35 pages
Coal Documentation
No ratings yet
Coal Documentation
74 pages
Burners PDF
100% (1)
Burners PDF
26 pages
Catalog Motoare MAN 2
100% (2)
Catalog Motoare MAN 2
120 pages
Comisia Europeana 2023 - 07 - 24 - With - Taxes - 2161
No ratings yet
Comisia Europeana 2023 - 07 - 24 - With - Taxes - 2161
8 pages
Europeanoilproductspecs
No ratings yet
Europeanoilproductspecs
71 pages
MCQ Worldscale
No ratings yet
MCQ Worldscale
16 pages
Dromon TP On EEDI EEXI
No ratings yet
Dromon TP On EEDI EEXI
33 pages
Towing Operations (Appendix)
No ratings yet
Towing Operations (Appendix)
3 pages
Philippine Excise Tax Overview
No ratings yet
Philippine Excise Tax Overview
10 pages
Del II Vedlegg F-1 Stanag 3149eed10 Minimum Quality Surveillance For Fuels
0% (1)
Del II Vedlegg F-1 Stanag 3149eed10 Minimum Quality Surveillance For Fuels
135 pages
Lesson 8: Fuels and Grease Flash and Fire Points of Liquid Fuels and Grease
No ratings yet
Lesson 8: Fuels and Grease Flash and Fire Points of Liquid Fuels and Grease
19 pages
HFO Instruction Manual Westfalia
100% (10)
HFO Instruction Manual Westfalia
216 pages
Nordic Hunter - Q88 - 28nov2022
No ratings yet
Nordic Hunter - Q88 - 28nov2022
8 pages
Container & Reefer Ship Specs
No ratings yet
Container & Reefer Ship Specs
5 pages
PDVSA Piping Material Specs Index
No ratings yet
PDVSA Piping Material Specs Index
6 pages
ASTM D7566 23a
No ratings yet
ASTM D7566 23a
14 pages
Shell Operational Guidelines For VLSFO
No ratings yet
Shell Operational Guidelines For VLSFO
2 pages
Boiler Basics for Power Sector Pros
50% (2)
Boiler Basics for Power Sector Pros
44 pages
Crude Oil Distillation and Secondary Processes
No ratings yet
Crude Oil Distillation and Secondary Processes
16 pages
Glove
No ratings yet
Glove
9 pages
Marine Fuel Standards Guide
No ratings yet
Marine Fuel Standards Guide
2 pages
Combustion Fundamentals
No ratings yet
Combustion Fundamentals
47 pages
Industry Code of Practice On Ship Recycling
100% (1)
Industry Code of Practice On Ship Recycling
12 pages
Rotary Pumps Tushaco Catalogue
No ratings yet
Rotary Pumps Tushaco Catalogue
5 pages
2024 Energy Statistics
100% (1)
2024 Energy Statistics
65 pages