+-'

.
c:: CIJ s::: Z
:> .-0 \81
....
ClJ
I.. V c:: ....,
v -lit
""0

c: CIJ to-- ::::s

\81
ra
> -c c:: "'C
0 ~-'
0
",' 0 I.. to Jo..
...., :c
a::
0
::J
ClJ
Co
x LL E .-s:::
s::: I;;
c::
I +-'
",' '" 0
ClJ
""0 I..
....
ClJ to -
ClJ -
\..?
CO
 Preface

The blast furnace can be described from various points of view. The mechanical
engineer will describe the furnace as a strong steel shell with a refractory lining and
a cooling system, which nevertheless is easily attacked by operating challenges.
The chemical engineer will stress the abundance of chemical reactions taking place
within the furnace. The operator will discuss the upsets frequently encountered when
operating a blast furnace. This book reconciles the various points of view. The blast
furnace process, i.e. what is going on within the furnace, is taken as a starting point.

The reason for writing this book was, that in our vision, optimisation of blast furnace
operation is not only based on the experience showing best practices, but also on the
conceptual understanding why something works or does not work. In other words,
operational improvement is not only based on know-how but on know-why as well.
@ Copyright 2004 by Verlag Stahleisen GmbH
The conceptual understanding of the process leads to different strategies to control the
furnace, depending on the results of the analysis.
This work is subject to copyright. All rights are reserved, whether the whole of
part of the materials concerned, specifically those of translation, reprinting,
The authors have been involved in many operational improvement programs, and special
re-use of illustrations, reproduction by photocopying machine or similar means,
blast furnace situations worldwide. They have developed their own understanding of
storage in data-bases, optical data carriers, and microfilm or similar means, and
the process and applied the results of their analysis in various operating furnaces. The
transfer to or transportation in electronic networks.
present book is based on their experience of transferring their understanding to blast
furnace operators. The term "modern" in the title refers to the relatively new application
The use of registered names, trademarks etc. in this publication does not imply,
of coal injection into blast furnaces.
even in the abscence of specific statement, that such names are exempt from the
relevant protective laws and regulations and therefore free for general use.
We are indebted to the people we have worked with. The book would not have been
written without the extensive support of Oanieli Corus BV in promoting blast furnace
Printed in Germany
operational understanding (in particular Mrs. J. Alexander and Messrs. N. Bleijendaal,
Distributors: Verlag Stahleisen GmbH R. van Laar, B. Paramanathan, J. Plooij, E. van Stein Callenfels, P. Zonneveld). We
P.O. Box 105164 thank XLmedia (Mr. E. Engel) for the editing.
0-40042 Dusseldorf
We wish to thank people, who helped us in understanding the blast furnace process.
First of all, we acknowledge the contributions of our home base steel works, Corus
IJmuiden Blast Furnace Department and technologists T Bakker, B. van den Berg,
H. van den Berg, L. Bol, G. Flierman, R. Godijn, G.J. Gravemaker, F. Homminga, T
Huisman, J. van Ikelen, J. Jansen, W. Koen, C. Kolijn, K. van Laar, J. van Langen, R.
Molenaar, J. de pagter, E. Schoone, T Spiering, A. Steeghs and G. Tijhuis. Furthermore,
we are indebted to our many colleagues all over the world:
0 In Argentina, at Siderar Messrs. E. Doro, J.M. Gonzales, F. Giandomenico and O.
Lingiardi. Contents
0 in Australia, at OneSteel Messrs. P Broadbent, A. Laird and J. Tsalapatis.
0 in Italy, at ILVATarantoworks Messrs. S. DeFelice, V. Dimastromatteo, G. Dimaggio
and G. Micci.
0 in Mexico, at AHMSA Messrs. H. Buruato and M. Morales.
0 in the United States, at Ispat Inland Messrs. W. Carter, P Chaubal, M. Ranade, J.
Ricketts and D. luke.
0 in Germany, Mr. H. Lungen.
And many more who have not been mentioned.

IJmuiden, August 2004 Preface i

Maarten Geerdes
Hisko Toxopeus List of Symbols and Abbreviations vi
Cor van der Vliet
Chapter 1 Introduction of the blast furnace process 1
1.1 What is driving the furnace? 3
1.2 The Equipment 5
1.3 Book Overview 9
1.4 Exercises 9

Chapter 2 The blast furnace: contents and gas flow 11

2.1 The generation of gas and permeability of the burden 11
2.2 Furnace Efficiency 15
2.3 Exercises: Gas flow and the contents of a blast furnace 16

Chapter 3 The ore burden: sinter, pellets, lump ore 17

3.1 Introduction 17
3.2 Quality demands for the blast furnace burden 18
3.3 Sinter 21
3.4 Pellets ....
3.5 Lump ore

Chapter 4 Coke 33
4.1 Functions of coke
4.2 Coke quality......
4.3 Coke production
4.4 Coke quality in the blast furnace
Chapter 5 Injection of coal, oil and gas 39 Chapter 11 Special Situations 105
5.1 Coal injection: equipment 40 11.1 Stops and start-ups 105
5.2 Coal specification for PCI.... 11.2 Blow-down 106
5.3 Coal injection in the tuyeres 11.3 Blow-in from new 108
5.4 Process control with Pulverised Coal Injection 43
5.5 Gas flow control................................ Annexes 111
5.6 Circumferential symmetry of injection Annex I Further reading 111
5.7 Other injectants................................. Annex II References 112
Annex III Answers to exercises 113
Chapter 6 Burden calculation and mass balances 47 Annex IV Rules of Thumb 116
6.1 Introduction ...................................
6.2 Burden calculation: starting points Index 117
6.3 An example of a burden calculation 48
6.4 Process calculations: a simplified mass balance 49

Chapter 7 The process: burden descent and gas flow control 55

7.1 Burden descent: where is voidage created? 55
7.2 Burden Descent: System of Vertical Forces 57
7.3 Gas flow in the blast furnace 59
7.4 Fluidisation and channelling 65
7.5 Burden distribution 6E
7.6 Erratic burden descent and gas flow 71
7.7 Blast furnace instrumentation ..............
7.8 Blast Furnace daily operational control

Chapter 8 Blast Furnace Productivity and Efficiency 77

8.1 The raceway................
8.2 Carbon and iron oxides
8.3 Temperature profile .....
8.4 Circumferential symmetry and direct reduction 87

Chapter 9 Hot Metal and Slag 89

9.1 Hot metal and the steel plant 89
9.2 Hot metal composition
9.3 Silicon reduction ........
9.4 Hot metal sulphur 92
9.5 Slag 92
9.6 Hot metal and slag interactions: special situations 95

Chapter 10 Casthouse Operation 97

10.1 Objectives of casthouse operation
10.2 Casthouse layout ..........................
10.3 Dry hearth practice 99
10.4 Opening and plugging the taphole 101
10.5 Hearth liquid level and casthouse operation 102

iv v
List of Symbols and Abbreviations
82,83,84 basicity, ratio of two, three or four components
bar pressure, atmosphere relative
C carbon
cm centimetre
CO carbon monoxide
CO2
CRI
carbondioxide
coke reactivity index Chapter 1
CSR
Fe
GJ
coke strength after reaction
iron
giga joule
Introductionof the
H2
H2O
HGI
hydrogen
water
hard grove index
blastfurnace process
HOSIM hoogovens simulatie (blast furnace simulation)
HV high volatile
IISI International Iron & Steel Institute
ISO International Organisation for Standardization
Two different process routes are available for the production of steel products, namely
JIS Japanese Industrial Standard
the blast furnace with oxygen steelmaking and the electric arc steelmaking route. The
K potassium
routes differ with respect to the type of products that can be made, as well as the
kg kilogram
raw materials used. The blast furnace-oxygen steelmaking route mainly produces
kmole kilomole
flat products, while electric arc steelmaking is more focused on long products. The
LV low volatile
former uses coke and coal as the main reductant sources and sinter and pellets as
m3STP cubic metre at standard temperature and pressure
the iron-bearing component, while the latter uses electric energy to melt scrap. The
mm millimetre
current trend is for electric arc furnaces to be capable of also producing flat products.
Mn manganese Nevertheless, the blast furnace-oxygen steelmaking route remains the primary source
Mt million ton
for worldwide steel production, as shown in Figure 1.
N2 nitrogen
Na sodium
Global finishedsteelconsumption: 934rnillion toq.(144 kg per capita, per year)
O2 oxygen
P phosphorous Resources Processes Finished Products
RAFT raceway adiabatic flame temperature
RR

.. ..
replacement ratio
Iron Ore Oxygen
s second
Si silicon (Sinter, pellet, lump) steelmaking
1,128 mln ton (60%)
Standard Coke coke with 87.5% carbon
t
tHM
ton
ton hot metal / X
Ti
VDEh
VM
titanium
Verein Deutscher Eisenhuttenleute
volatile matter
Scrap
300 mln ton (est.) .. EAF (34 %)
..
Other' (6%)
. = Corex. open hearth. etc.
/
Figure 1: Steelmaking routes and raw materials (I/SI Steel Statistical Yearbook, 2003)

vi
 1 - Introduction of the blast furnace process 1 - Introduction of the blast furnace process

Hot metal is produced in a blast furnace, from where it is transported as liquid As indicated in Figure 3, at any moment, an operating blast furnace contains, from
hot metal to the steel plant where refinement of hot metal to steel takes place by top downwards: :
removing elements such as sulphur, silicon, carbon, manganese and phosphorous. . Layers of ore and coke.
Good performance of the steel plant requires consistent hot metal quality of a given . An area where ore starts to soften and melt, known as the softening-melting zone.
specification. Typically the specification demands silicon content between 0.3% and
0.7%, manganese between 0.2% and 0.4%, phosphorous in the range 0.06-0.08%
. An area where there is only coke and liquidiron and slag, called the "active coke"
or drippingzone.
or 0.1-0.13% and a temperature as high as possible. . The dead man, which describes the coke in the hearth of the furnace.

Inthe blast furnace process iron ore and reducing agents (coke, coal) are A blast furnace has a typical conical shape. The sections from top down are:
transformed to hot metal and slag is formed from the gangue of the ore burden and . Throat, where the burden surface is.
the ash of coke and coal. Hot metal and liquidslag do not mixand remain separate . The shaft, where the ores are heated and reduction starts.
from each other withthe slag floatingon top of the denser iron.The iron can then . The bosh parallel or bellyand
be separated from the slag inthe casthouse, but before that can take place, the . The bosh, where the reduction is completed and the ores are melted down.
transformation to hot metal and slag must be understood. . The hearth, where the molten material is collected and is cast via the taphole.
Let us now consider the contents of a blast furnace at any given moment. Ore
Throat
and coke are charged in discreet layers at the top of the furnace. From studies of
quenched furnaces it was evident that these layers of ore and coke remain until Burden
the temperatures are high enough for softening and melting of the ore to begin.
Quenched furnaces are "frozen in action" with the help of water or nitrogen and Shaft zone Coke
examples of quenched blast furnaces is presented in Figure 2. The quenched 'blast Stack
furnace shows clearly the layer structure of coke and ore. Further analysis reveals
information about the heating and melting of the ore as well of the progress of Cohesive zone
chemical reactions (Bonnekamp et ai, 1984).
Belly/bosh parallel
Active coke zone
Dead Man Bosh
Raceway
-~-=--~-
,~~~ Hearth
*~~"/v;;;:::~ -
Taphole
~~)
~
~"--c'~
~
~;--
~~
~~
ilt~,
~.~ 1.1
Figure 3: The zones in the blast furnace

~~
~~
What is driving the furnace?

~~~ ~Jff!fJ!
~
--=':.
1.1.1 Process Description
~ ~
The driving forces in the blast furnace are illustrated in Figure 4 (next page)
. burden,
A blast furnace is filled with alternating layers of coke and the iron ore-containing

. Hot blast is blown into the blast furnace via tuyeres. A tuyere is a cooled copper
conical pipe numbering up to 12 in smaller furnaces, and up to 42 in bigger
Figure 2: Dissections of quenched blast furnaces Kakogawa 1 and Tsurumi furnaces through which pre-heated air (up to more than 1,200°C) is blown into the
(Based on Omori et ai, 1987) furnace,
 1 - Introduction of the blast furnace process 1 - Introduction of the blast furnace process

. The hot blast gasifies the reductant components in the furnace, those being coke are between 1,100 and 1,450°C. The temperature differences in the furnace are
as well as auxiliary materials injected via the tuyeres. In this process, the oxygen large. Inthe example the temperature gradients are bigger in the horizontaldirection
in the blast is transformed into gaseous carbon monoxide. The resulting gas has than in the vertical direction, which will be explained in chapter 6.
a high flame temperature of between 2,100 and 2,300°C . Coke in front of the
tuyeres is consumed thus creating voidage.
. The very hot gas ascends through the furnace, carrying out a number of vital
functions.
100°C

. Heats up the coke in the bosh/belly area.

. Melting the iron ore in the burden, creating voidage. 500°C
. Heats up the material in the shaft zone of the furnace.
. Removes oxygen of the ore burden by chemical reactions.
. Upon melting, the iron ore produces hot metal and slag, which drips down through
the coke zone to the hearth, from which it is removed by casting through the
taphole. In the dripping zone the hot metal and slag consume coke, creating
voidage. Additional coke is consumed for final reduction of iron oxide and carbon
dissolves in the hot metal, which is called carburisation.

The blast furnace can be considered as a counter current heat and mass exchanger,
as heat is transferred from the gas to the burden and oxygen from the burden to the
gas. Gas ascends up the furnace while burden and coke descend down through the
furnace. The counter current nature of the reactions makes the overall process an
extremely efficient one.
Figure 5: Temperature profile in a blast furnace (typical example)

Ore consumes coke, 1.2 The Equipment

creates voidage

Coke gasifies in 1.2.1 Equipment overview

front of tuyeres,
creates voidage Burden melts,
An overview of the major equipment is shown in Figure 6 (next page). These include:
creates voidage
. Hot Blast Stoves. Air preheated to temperatures between 1,000 and 1 ,250°C is
produced in the hot blast stoves and is delivered to the furnace via a hot blast
main, bustle pipe and finally through the tuyeres. The hot blast reacts with coke
and injectants, forming the so-called raceway in front of the tuyeres.
Gas
ascent
Burden . Stock house. The burden materials and coke are delivered to a stock house. The
descent materials are screened and then weighted before final delivery into the furnace.
The stock house is operated automatically. Corrections for coke moisture are
generally made automatically. The burden materials and coke are brought to the
top of the furnace via skips or via a conveyor belt, where they are discharged into
the furnace in separate layers of ore and coke.
. Gas cleaning. The top gas leaves the furnace via uptakes and a down-comer.
The top gas will contain many fine particles and so to remove as many of these as
possible the top gas is lead through a dust catcher and wet cleaning system.
Figure 4: The driving force of a blast furnace: the counter current process creates
voidage at the indicated areas causing the burden to descend
. Casthouse. The liquid iron and slag collect in the hearth of the furnace, from where
they are tapped via the taphole into the casthouse and to transport ladles.

A typical example of the temperature profile in the blast furnace is shown in Figure 5.
. Slag granulation. The slag may be quenched with water to form granulated slag,
which is used for cement manufacture.
It is shown that the softening/melting zone is located in an area where temperatures

A
 1 - Introduction of the blast furnace process 1 - Introduction of the blast furnace process

~ Bleeders 1.2.2 Blast furnace construction

There are basically two construction techniques to support blast furnaces. The
classic design utilises a supported ring, or lintel at the bottom of the shaft, upon
Bel Hess Top which the higher levels of the furnace rests. The other technique is a freestanding
construction requiring an independent support for the blast furnace top and the
gas system. The required expansion (thermal as well as from the pressure) for the
installation is below the lintel that is in bosh/belly area for the lintel furnace, while the
Hot Blast Stoves
compensator for expansion in the freestanding furnace is at the top, as indicated in
Figure 8.

Expansion
Figure 6: Blast furnace general arrangement

The top of the blast furnace is closed, as modern blast furnaces tend to operate with
high top pressure. There are two different systems:
. The double bell system, which for burden distribution purposes, requires movable
throat armour. Lintel
. The bell less top, which allows ~asier burden distribution.
Examples of both types are schematically shown in Figure 7.
Expansion

Double-bell top Paul Wurth Top

with movable armour (Bell-Less Top)

Free-standing Furnace Furnace with Support Ring

("German Construction") ("US Construction")
Figure 8: Blast furnace constructions

Distributor Top Bin 1.2.3 Blast furnace development

Small Bell
Blast furnaces have grown considerably in size during the 20th century. In the early
Rotating Chute days of the 20th century, blast furnaces had a hearth diameter of 4 to 5 metres and
Big Bell
were producing around 100,000 tonnes hot metal per year, mostly from lump ore and
Movable armour coke. At the end of the 20th century the biggest blast furnaces had between 14 and
15 m hearth diameter and were producing 3 to 4 MT per year. A typical development
~ of blast furnaces is shown in Table 1 for the situation of the IJmuiden steel works. The
Figure 7: Blast furnace top charging systems ore burden developed, so that presently high performance blast furnaces are fed with
sinter and pellets. The lump ore percentage has generally decreased to 10 to 15%
 --
1 - Introduction of the blast furnace process 1 - Introduction of the blast furnace process

or lower. The reductants used developed as well: from operation with coke only to Presently, very big furnaces reach production levels of 12,000 tld or more. E.g. the
the use of injectant through the tuyeres. Mainly oil injection in the 1960's, while since Oita blast furnace No.2 (NSC) has a hearth diameter of 15.6 meter and a production
the early 1980's coal injection is used extensively. Presently, about 30 to 40% of the capacity of 13,500 t/d. In Europe, the Thyssen-Krupp Schwelgern nr 2 furnace has a
earlier coke requirements have been replaced by injection of coal and sometimes oil hearth diameter of 14.9 m and a daily production of 12,000 t/d.
and natural gas.
1.3 Book Overview
Table 1: Development of the blast furnaces at Corus IJmuiden, The Netherlands
Blast furnace ironmaking can be discussed from 3 different perspectives:
Hearth Diameter m 4.8/5.6 4.8/5.6 5.2/5.9 8.5 8/9 10/11 13/13.8 . The operational approach: discussing the blast furnace with its operational
challenges.
Working Volume

Built
m3 519

1924
519

1926
598

1930
1413

1958
1492

1961
2328

1967
3790

1972
. The chemical technology approach: discussing the process from the perspective of
the technologist who analyses progress of chemical reactions and heat and mass
t/d 280 280 360 3,000 5,000 balances.
Initial Productivity

Most Recent t/d 1,000 1,000 1,100

1,380

3,500
1,700

3,700 7,000 10,500

. The mechanical engineering approach focussing on equipment.
Productivity

2002 1991
The focus of this book is the "operators view", with the aim to understand what is
Last Renovation
going on inside the furnace. To this end the principles of the process are discussed
Demolished 1974 1974 1991 1997 1997
(Chapter 2) followed by the demands on burden quality (Chapter 3) and coke and
auxiliary reductants (Chapters 4 and 5). Simplified calculations of burden and top gas
The size of a blast furnace is often expressed as its hearth diameter or as. its are made (Chapter 6). The control of the process is discussed in Chapter 7: burden
"workingvolume"or "innervolume".The workingvolume is the volume of the blast descent and gas flow control. The issues pertinent to understanding the blast furnace
furnace that is available for the process i.e. the volume between the tuyeres and the productivity and efficiency are presented in Chapter 8. Subsequently, hot metal and
burden level. Definitionsof workingvolume and inner volume are given in Figure 9. slag quality (Chapter 9), casthouse operation (Chapter 10) and special situations like
stops and starts (Chapter 11) are discussed.
Bottom
of bell
1.4 Exercises
Zero level burden
1m below Bottom of 1m below The exercises refer to a large modern blast furnace operated on pulverised coal
bottom movable chute
injection and equipped with high top pressure and high blast temperature. The
of bell armour
working volume is 3,800 m3 and it has four tapholes.
Q) The furnace is producing 10,000 tonnelday with a slag volume of 250 kg/tHM. As
Q) Q) E
E E :J a reductant the furnace uses 300 kg coke per tonne hot metal (tHM) and 200 kg
.2 .2 (5
:> -2
.- >OJ
0 pulverised coal injection (PCI).

-
-
<tS

~
c
~
Q)
c
E
.-
-'"
0 Questions (The answers to this exercise can be found in Annex 111.):
S 1. The International Iron and Steel Institute (IISI) estimates that the worldwide pig
iron production was 575.7 Mt in 2001. How many of these "modern furnaces" are
required to produce this amount? And-if there are three of these furnaces on one
site-how many pig iron producing sites are necessary?
Tuyere level 2. Calculate the annual usage of pellets and coal at such a site. How much finished
steel products are produced? Use the following data:
Taphole level a) From 1 tonne iron 1.1 tonne liquid steel is produced and 0.9 tonne steel finished
Uppermost brick products.
bottom layer b) To produce 1 tonne iron, one requires 1.6 tonne pellets.
c) Coke is produced mainly on local coke plants. 25% of the coal is gasified during
Figure 9: Definitions of working volume and inner volume cokemaking.

8 9
 Chapter 2
The blastfurnace:
contentsand gas flow

2.1 The generation of gas and permeability of the burden

The blast furnace is a counter current reactor (Figure 10). The driving force is the hot
blast consuming coke at the tuyeres. In this chapter the gas flow through the furnace is
analysed in more detail. The charge consists of alternating layers of are burden (sinter,
pellets, lump are) and coke. The burden is charged cold and wet into the top of the
furnace, while at the tuyeres the hot blast gasifies the hot coke. Towards the burden
stockline (22 to 25 m from tuyeres to burden surface) the gas temperature drops from
a flame temperature of 2,200°C to a top gas temperature of 100 to 150 °C.

Gas Burden
ascent descent

Figure 10: The blast furnace as a counter current reactor

11
 2 - The blast furnace: contents and gas flow 2 - The blast furnace: contents and gas flow

The process starts with the hot blast through the tuyeres, which gasifies the coke blast furnace is typically 45 to 55 mm, while the average size of sinter is 18 to 25 mm
and coal in the raceway (Figure 10). The reactions of the coke create hot gas, which and of pellets is 10 to 12 mm. Consequently, the burden layers determine how the
is able to melt the ore burden. Consumption of coke and melting of the ore burden gas flows through the furnace, while the coke layers function as gas distributors.
creates space inside the furnace, which is filled with descending burden and coke.
The oxygen in the blast will gasify the coke to generate carbon monoxide (CO). If gas flows from the bosh upwards, what happens to the gas as it gradually cools
For every molecule of oxygen 2 molecules of carbon monoxide are formed. If blast down?
is enriched from its base level of 21 % to 25% oxygen, then every cubic meter (m3 Firstly, the heat with a temperature in excess of 1 ,400°C, the melting temperature of
STP) oxygen will generate 2 m3 STP of CO. So if the blast has 75% of nitrogen and the slag, is transferred to the layered burden and coke, causing the metallic portion to
25% of oxygen, the bosh gas will consist of 60% (i.e. 75%/(75%+2x25%)) nitrogen melt. In the temperature range from 1,400 to 1,100 °C the burden will soften and stick
and 40% CO gas. In addition a huge amount of heat is generated in the raceway together rather than melt. In the softening and melting zone the remaining oxygen
from the combustion of coke and coal (or oil, natural gas). The heat leads to a high in the ore burden is removed, which generates additional carbon monoxide. This is
flame temperature, which generally is in the range of 2,000 to 2,300°C. Since this referred to as the direct reduction reaction (see section 7.2.1), which only occurs in
temperature is higher than the melting temperature of iron and slag, the heat in the the lower furnace.
hot gas can be used to melt the burden. Flame temperature is discussed in more
detail in section 8.1.3. The gas has now cooled to about 1,1OO°C and additional gas has been generated.
Since the direct reduction reaction costs a lot of energy, the efficiency of the furnace
The hot gas ascends through the ore and coke layers to the top of the furnace. If is largely dependant on the amount of oxygen removed from the burden materials
there was only coke in the blast furnace, the chemical composition of the gas would before reaching this 1,1 OO°C temperature.
remain constant but the temperature of the gas would lower as it comes into contact
In summary:
with the colder coke layers high in the furnace. A presentation of the gas flowing
through a blast furnace filled with coke is presented in Figure 11 . The gas flow has a . Heat is transferred from the gas to the ore burden, which melts and softens (over
prop-flow character. To the experienced blast furnace operator the furnace filled with
coke only may seem a theoretical concept. However, in some practical situations, like . 1,1OO°G).
Residual oxygen in the burden is removed and additional CO is generated. This is
known as the direct reduction reaction.
the blow-in of a new furnace or when taking a furnace out of operation for a long time
(banking) the furnace is almost entirely filled with coke.
Upon further cooling down the gas is capable of removing oxygen from the ore
burden, while producing carbon dioxide (CO2), The more oxygen that is removed,
the more efficient the furnace is. Below temperatures of 1,1 OO°C the following takes
place:
. Heat is transferred from the gas to the burden.
. CO2 gas is generated from CO gas, while reducing the amount of oxygen of
the ore burden. This is called the gas reduction reaction, and in literature it
is sometimes called "indirect reduction" as opposed to "direct reduction". No
~tional gas ~~gleQJJuring th~tion.
A similar reaction takes place with hydrogen. Hydrogen can remove oxygen from
the burden to form water (H2O).

Higher in the furnace, the moisture in the burden and coke evaporates and so is
eliminated from the burden before any chemical reactions take place.

Figure 11: Gas flows in a furnace filled with coke only (left) and in a If we follow the burden and coke on its way down the stack, the burden and coke
furnace filled with alternating layers of coke and are (right). are gradually heated up. Firstly the moisture is evaporated, and at around 500°C the
removal of oxygen begins. A simplified schedule of the removal of oxygen from the
In the normal operational situation the furnace is filled with alternating coke and ore ore burden is shown in Figure 12 (next page).
layers. About 35 to 45 layers of ore separate the coke. It is important to note that the
permeability of coke is much better than the permeability of ore (see also Figure 36).
This is due to the fact that coke is much coarser than sinter and pellets and that the
void fraction within the coke layer is higher. For example, the mean size of coke in a

12 13
 2 - The blast furnace: contents and gas flow 2 - The blast furnace: contents and gas flow

2.2

Hematite (Fe203)
.. ....
8 Iron atom.

......
Oxygen atom

SO0-600°C
~
Furnace Efficiency

process efficiency of the blast furnace, generally considered to be the reductant

.. . -
rate per tonnel1ot metal, is contifiLiOusly-monitored through measurement of th~
- -
Magnetite (Fe304)
..t..,. . chemical composition of the top gas. - -
The efficiency is expressed asthe gas utilisation, that is the percentage of the CO
gas that has been transformed to CO2, as cfefined in the following expression:

Wustite
(FeO).. ..
8. 8... 8. - ~ CO2%

vx> gas utilization = (CO%+CO2%)

8 8 8.
FeOy.
. ... .
Fe8 8 . . . 8
In addition, at modern furnaces the gas composition over the radius is frequently
measured. The latter shows whether or not there is a good balance between the
amount of reduction gas and the amount of ore in the burden. The wall zone !§.
especially important and so the coke percentage in the wall area should not be too
Figure 12: Schematic presentation of reduction of iron oxides and temperature low,-Ihe wall area isthe most difficult place to melt the bura8n as that is wherethe
burden thickness is at it's highest across the radius, and also because the gas at the
The first step is the reduction of the so-called hematite (Fe2O3) to magnetite (Fe304)' wall loses much of its temperature to cooling lOsses. - - ,
The reduction reaction generates energy, so it helps to increase the temperature of
the burden. In addition, the reduction reaction creates tension in the crystal structure
The top gas analysis gives a reasonably accurate indication of the efficiency of the
of the burden material, which may cause the crystal structure to break into' smaller furnace. When comparing different furnaces one should realise that the hydrogen
particles. This property is called low-temperature disinteqration. Several tests are
also takes part in the reduction process (paragraph 7.2.4).
available to quantify the effects (see chapter 3). Further down in the furnace the
temperature of the burden increases gradually until the burden starts to soften and to
The gas utilisation also depends on the amount of oxygen that must be removed.
melt in the cohesive zone. The molten iron and slag are collected in the hearth.
Since pellets have about 1.5 atoms of oxygen per atom of Fe (Fe2O3) and sinter has
about 1.45 (mix of Fe2O3 and Fe3O4), the top gas utilisation will be lower when using
We now consider the interaction b~tween the gas and the ore burden. The more sinter. It can be calculated as about 2.5% difference of the top gas utilisation, when
the gas removes oxygen from the ore burden, the more efficient the blast furnace comparing an all pellet burden with an all sinter burden.
process is. Consequently, intimate contact between the gas and the ore burden is
very important. To optimise this contact the permeability of the ore burden must be as
high as possible. The ratio of the gas flowIng through the ore burden and the amount
5>foxygen to be removed from th~burde~ must also be in balance. -

Experience has shown that many problems in the blast furnace are the consequence
of low permeability ore layers. Therefore, the permeability of the ore layers across the
diameter of the furnace is a major issue. The permeability of an ore layer is largely
determined by the amount of fines (under 5 mm) in the layer. Generally, the majority
of the fines are generated by sinter, if it is present in the charged burden or from lump
ores.
The problem with fines in the furnace is that they tend to concentrate in rings in the
furnace. As fines are charged to the furnace they concentrate at the point of impact
where the burden is charged. They are also generated by low temperature reduction-
disintegration. Thus, it is important to screen the burden materials well, normally with
5 or 6 mm screens in the stock house, and to control the low temperature reduction-
_disintegration cl!aracteristics of the burden.. - - -

14 15
 ---

2 - The blast furnace: contents and gas flow

2.3 Exercises: Gas flow and the contents of a blast furnace

The contents of a blast furnace can be derived from operational results. How long do
the burden and gas reside within the furnace?
Consider an example of a large, high productivity blast furnace with a 14 metre
hearth diameter. It has a daily production of 10,000 t hot metal (tHM) at a coke rate
of 300 kg/tHM and a coal injection rate of 200 kg/t. Moisture in blast and yield losses
are neglected. Additional data is given in Table 2.

Table 2: Data for calculation of blast furnace contents

Chapter 3
The ore burden:
are burden 1,580 kg/tHM 1,800 kg/m3
Coke
Coal
300
200
kg/tHM
kg/tHM
470 kg/m3 87
78
%
%
sinter, pellets, lump ore
Blast Volume 6,500 m3STP/min 1.3 kg/m3STP
Top Gas 1.43 kg/m3STP

O2 in blast 25.6 %
3.1 Introduction
Working volume 3,800 m3 (500 m3used for active coke zone)
Throat diameter 10 m
In the early days of ironmaking, blast furnaces were often located close to are mines.
A charge contains 94.8 t are burden 18 t coke In those days, blast furnaces were using local are and charcoal, later replaced by coke.
A ton hot metal contains 945 45 kg carbon
In the most industrial areas of the 19th century, many blast furnaces were operating
kg Fe
in Germany, England and the US. After the application of the steam engine for ships
Voidage in shaft 30 %
and transportation, the centre of industrial activity changed from the ore sources to the
I
1tonnehotmetalcontains945kg Fe=945/55.6='17,0kmole major rivers, like the river Rhine, and later from the river banks to the coast. This trend
may appear clear at present, but has a short history. As an example: in 1960 there
Questions (answers in Annex III): were 60 operating blast furnaces in Belgium and Luxemburg. Presently (2004), only
1. How much blast oxygen is used per tonne hot metal? six are operating, of which two have the favourable coastal location.
2. How often are the furnace contents replaced?
3. How many layers of are are in the furnace at any moment? The trend towards fewer but larger furnaces has made the option for a rich iron burden
4. What happens with the carbon in the coke and coal? a more attractive one. A rich iron burden translates into a high Fe content and as
5. How much top gas do we get? fine ores are too impermeable to gas, the choice is narrowed down to sinter, pellets
6. Estimate how long the gas remains in the furnace and lump ores. Sinter and pellets are both formed by agglomerating iron ore fines
7. If you get so much top gas, is there a strong wind in the furnace? from the ore mines and have normally undergone an enrichment process, which is not
described here. The quality demands for the blast furnace burden are discussed and
the extent to which sinter, pellets and lump ore meet these demands.

A good blast furnace burden consists for the major part of sinter and/or pellets (Figure
13, next page). Sinter burdens are prominent in Europe and Asia, while pellet burdens
are used in Northern America and Scandinavia. Many companies use sinter as well as
pellets although the ratios vary widely.
 ---

3 - The ore burden: sinter, pellets, lump ore 3 - The ore burden: sinter, pellets, lump ore

0.3

Sinter Pellets Lump c

90% < 25 mm 6-25 mm .Q
11 mm (:t 2 mm) t5
~
LL 0.2
Figure 13: Burden materials ""C

g
Lump ores are becoming increasingly scarce and generally have poorer properties
for the blast furnace burden. For this reason it is used mainly as a cheap alternative
0.1
for pellets. For high productivity low coke rate blast furnace operation the maximum
1 0.5 0
lump ore rate is in the range of 10% to 15%. The achievable rate depends on lump ore
VI
quality and the successful use of higher percentages is known.
Size Distribution V1+Vs

3.2 Quality demands for the blast furnace burden Figure 14: Permeability for gas flow depends on void fraction, which depends on the
ratio of smaller and larger particles. Example of two types of spherical particles,
which are blend in the ratio Volume small (Vs) and Volume large (VI)
3.2.1 Qualitative description
During the first reduction step from hematite to magnetite the structure of the burden
The demands for the blast furnace burden extend to:. materials weakens and fines are generated. Sinter and lump ore are especially prone
. The chemical composition of the burden. After the reduction and melting processes to this effect, known as reduction-disintegration.
A major requirement for the blast furnace burden is to limit the level of fines within the
the correct iron and slag compositions must be made and this will be determined by
the chemical composition of th~ materials charged in the furnace. furnace to as low as possible. This can be achieved by;
. The permeability for gas flow. Good resistance against degradation and no swelling . Proper screening of burden materials before charging. Screens with 5 to 6 mm
holes are normal operational practice.
of the material upon heating.
. The softening and melting properties. Fast transition from solid to liquid state. . Good reduction-disintegration properties.

The reducibility of the burden is controlled by the contact between gas and the burden During charging, fines in the burden material tend to concentrate at the point of
particles as a whole, as well as the gas diffusion into the particles. Whether or not impact on the burden surface. The level of reduction-disintegration increases in
good reduction is obtained in the blast furnace is governed by the layer structure of areas where the material is heated and reduced slowly. A charged ring of burden with
the burden, the permeability of the layers and the blast furnace internal gas flow. The a high concentration of fines will impede gas flow, experience the slower warm-up
reducibility of the burden components will be of less importance if the gas flow within and so result in a higher level of reduction-disintegration.
the furnace does now allow sufficient contact for the reactions to take place.
As soon as burden material starts softening and melting, the permeability for gas is
In the shaft zone the permeability of the burden is determined by the amount of fines, greatly reduced. Therefore, the burden materials should start melting at relatively
(see Figure 14, next page). Fines may be defined as the fraction of the material high temperatures. So that they do not impede gas flow while they are still high
under 5 mm. If there are too many fines, the void fraction used for the transport of the up the stack. Melting properties of burden materials are determined by the slag
reduction gas will diminish and will affect the bulk gas flow through the burden (Hartig composition.
et ai, 2000). There are two sources for fines, those that are directly charged into the Melting of pellets and lump ore starts at temperatures of 1,000 to 1,1 OooG, while
furnace, and those that are generated in the shaft by the process. basic sinter generally starts melting at higher temperatures.

1A 1Q
 3 - The ore burden: sinter, pellets, lump ore 3 - The ore burden: sinter, pellets, lump ore

3.2.2 Ore burden quality tests 3.2.2.2 Tests for reduction-disintegration

Ore burden material is characterised by the following. The reduction-disintegration tests are carried out by heating a sample of the burden
. Chemical composition.
. furnace.
Size distribution, which is important for the permeability of ore burden layers in the
to at least 500°C and reducing the sample with gas containing CO (and sometimes
Hz). After the test the sample is cooled, tumbled and the amount of fines are
measured. The quoted result is the percentage below 3.15 mm.
. Metallurgical properties with respect to: The HOSIM test (blast furnace simulation test) is a test where the sample is reduced
. Cold strength, which is used to characterise the degradation of ore burden to the endpoint of gas-reduction in a furnace. The sample is then tumbled after the
materials during transport and handling. test. The results are the reducibility defined by the time required to reduce the sample
. Reduction-disintegration, which characterises the effect of the first reduction to the endpoint of gas reduction, and the reduction-disintegration represented by the
step and is relevant in the stack zone of the furnace.
. Softening and melting properties, which are important for the formation of the
percentage of fines (under 3.15 mm) after tumbling. Both tests simulate the upper
part of the blast furnace process. The more advanced HOSIM test gives a more
cohesive or melting zone in the furnace. realistic description of the effects in the blast furnace.

It is important for permeability to have a narrow size range and have minimal fines 3.3 Sinter
(less than 5% below 5 mm, after screening in the stockhouse). Measurement of the
percentage of fines after screening in the stockhouse, although cumbersome, might
give indications whether or not excessive fines are charged into the furnace.
3.3.1 Production
A short description of tests used for characterisation of materials is given below with
the objective being to understand the terminology. In many situations tumbler tests Sinter making began as an efficient way to reuse the plant revert materials in the
are used, where a sample of material is tumbled in a rotating drum for a fixed number blast furnace. The sinter making process uses heat to fuse separate iron ore fines
of rotations. The size distribution after tumbling is determined and used as a quality into larger particles that are suitable for charging to the blast furnace. The strength of
indicator (Figure 15). the cohesive force depends on the amount and type of molten material between the
individual particles.
Dwight and Lloyd constructed the first continuous sinter plant in 1906. A schematic

Principle of tumble test:

Sample is tumbled at fixed
~ II:>
presentation is shown in Figure 16. Sinter quality has improved progressively and
in a number of countries (Europe, Japan, Brazil, Korea) sinter is the predominant
blast furnace iron source. Plant reverts like dust from the blast furnace dust catcher,
metallic fines from the blast furnace screens, mill scale and other materials are
number of rotations. Size
reused in sinter making. Modern, large sinter strands are 5 metres wide and an
distribution determined
effective sinter area of 400 m2. Productivity is typically 30 to 45 t/m2/day.
after tumbling. Weight
percentages over or below
certain screen sizes are
used as a quality parameter.

Sinter Hearth
Layer Layer
Travelling Directioll.,
Figure 15: Principle of tumbler test

3.2.2.1 Tests for cold strength

Cold strength is mostly characterised by a tumbler test. For this test, an amount of Stack
material is tumbled in a rotating drum for a specified time interval. Afterwards the Sinter
amount of fines are measured. Breaker

For pellets the force needed to crack the pellets, referred to as the cold compression
Suction Fan
strength, is determined.
Figure 16: Sinter plant

20 21
 ---

3 - The ore burden: sinter, pellets, lump ore 3 - The ore burden: sinter, pellets, lump ore

After delivery of the ore fines to the steel works the sinter fines are stored and formed in the presence of FeO. The heat content of the flame front is lower with
blended in pre-determined ratios. The blending can be done by mixing appropriate fluxed sinter than with acid sinter, which results in lower thermal exposure of the
quantities on a conveyor belt before arrival at the sinter plant or with a blending pile. sinter strand and less wear on the plant.
These piles are constructed by depositing the different iron ores and fluxes in layers
The advantage of adding limestone outside the blast furnace instead of in the
in a longitudinal direction and then perpendicularly reclaiming them. This method of
layering and reclaiming is to ensure a homogeneous mix. furnace is, that the decarbonisation reaction of the limestone (CaCO3 CaO + CO2,
cost 117 kJ/mole) takes place outside the blast furnace and the required heat is not
-
The sinter feed also contains limestone, return sinter and coke breeze. The blend taken from the blast furnace gas flow.
is mixed in a rotating drum, where water is added up to 5 to 7% for primary bonding
between ore particles. It is important that the blend has a good permeability, so Basic points influencing sinter strand productivity, costs and operation are:
that air can be sucked through the blend and the coke breeze (and magnetite) can . The selection of the correct blend of sinter feed is important from a cost point
generate the heat required for the sintering process. of view. Most companies use "value-in-use" models attributing premiums and
penalties to all aspects of the sinter blend. Not only its Fe-value, but also the fuel
The blend is deposited in layers of 35 to 65 cm height on the sinter strand. The requirements and the value of other elements like phosphorous, which will have a
bottom layer is always made of a small layer of sinter (typical 15 to 25 mm size
range) to prevent the flame front from reaching the sinter cars. The sintering process . penalty.
The correct blend includes the size distribution of the various materials. Very fine
starts with ignition of the top layer with fans sucking the heated air down through the materials cannot easily be sintered because of the negative impact on the sinter
sinter bed. The fuels in the bed generate the heat to melt and fuse the ore particles bed permeability. The coke breeze size and amount are important for the sinter
together. At the end of the sinter strand the flame front has passed all the way quality.
through the blend to the bottom. The sinter cake then drops off the end of the strand . The chemical composition of the required sinter is important. At higher basicity,
into a collecting bin. From there it is broken up into manageable chunks, cooled and sinter productivity increases. MgO content must be over 1.4% for good break down
screened into appropriate fractions. The sinter fines (under 5 mm) are reused in the properties during reduction (see section 3.3.3).
sinter blend, some of the material (15 to 25 mm) is recycled for the bottom strand . Since heat is required to decarbonise the limestone, use of burnt lime increases
layer, and the remaining sinter is sent to the blast furnace bins. The process on the the sinter strand productivity, also by increasing the permeability.
sinter strand is shown in Figure 17. . Environmental concerns have major local impact on sinter plant operation. Not
only from the perspective of the use of "waste" materials, but also the emissions of
a sinter plant may be a concern.

3.3.2 Sinter quality

As mentioned above, sinter is made in three different types: acid sinter, self-fluxing
and super-fluxed sinter. The self-fluxing sinter is most common. Since sinter
properties vary with the blend type and chemical composition, only some qualitative
remarks can be made.

The sinter quality is defined by:

0 min
Figure 17: Sinter production progress: the temperature
16 min
profile of the blend during sintering
. Size distribution: sinter mean size ranges from 15 to 25 mm as measured after
the sinter plant. The more basic the sinter, the smaller the average size. Sinter
degrades during transport and handling so sinter has to be re-screened at the
From a chemical point of view it is important to note that the lime (CaCO3, GaG)
blast furnaces to remove all of the fines. Sinter from the stockyard may have very
added to the sinter is important as a flux for the blast furnace. In order to distinguish
different properties from freshly produced sinter directly from the sinter plant. If
between the levels of flux in the sinter, it is categorised into acid sinter, self-fluxing
stock sinter must be used in the blast furnace, it should be charged in a controlled
sinter and super fluxed sinter. Self-fluxing sinter brings the lime required to flux its
fashion, and diluted with as mush fresh sinter as is possible, e.g. by using a
acid components (SiO2). Super-fluxed sinter brings extra CaO to the blast furnace. dedicated bin in the stockhouse to stock sinter.

For self-fluxing and super-fluxed sinter, the lime reduces the melting temperature
. Cold strength: normally measured with a tumble test. The more fuel that is used,
the stronger the sinter. The cold strength influences the sinter plant productivity
of the blend and at relatively low temperatures (1,100 to 1,300°C) strong bonds are
because a low cold strength results in a high fines rate.

22 23
 --
3 - The ore burden: sinter, pellets, lump ore 3 - The ore burden: sinter, pellets, lump ore

. Reduction-disintegration properties. The reduction from hematite to magnetite 3.3.3 Background of sinter properties
generates internal stresses within a sinter particle. The stronger the sinter, the
better the resistance to these stresses. The reduction-disintegration properties Sinter is a very heterogeneous type of material. Research of various type of sinter in
improve with denser sinter structure, i.e. when the sinter is made with more coke a cooled furnace has made clear, that various phases are simultaneously present.
breeze. As a consequence of the higher coke breeze usage, the FeO content of Figure 18. The most important phases present are:
the sinter will increase. From experimental correlations it is well known that for a
given sinter type reduction-disintegration improves with FeO content.
. Primary and secondary magnetite (FesO4). Secondary magnetite is formed during
sintering in the high temperature, reducing areas at the sinter strand, i.e. close to
. The melting of sinter is determined by the chemical composition, that is, the local coke.
chemical composition. Most important is the basicity, the presence of FeO and . Primary and secondary hematite (Fe2Os)' Secondary hematite is formed on the
SiO2. The latter two functioning as components that lower the melting temperature. sinter strand during the cooling down of the sinter in the presence of air (oxygen).
At temperatures of 1,200 to 1,250 DC sinter starts softening and melting. Very
basic parts (CaO/SiO2> 2) melt at higher temperatures, but will still have melting
. Calcium ferrites are structures formed from (burnt) lime (CaO) and iron oxides.

temperatures around 1300 DC in the presence of sufficient FeO. If FeO is low, then It is clear from Figure 18, that at increasing basicity an increased fraction of calcium
melting temperatures exceeding 1,500 DC can be observed. However, melting ferrites can be found. This has major consequences, for the sintering process as well
in a blast furnace differs from melting of "pure" burden materials, since strong as for the use of sinter in the blast furnace.
interactions between different burden components have been observed.
100
A summary of acceptability ranges for sinter quality parameters is given in Table 3.

Table 3: Characterisation of sinter and lump are

80

Mean size Size distribution Average size mm 15--25 mm ISO 4701

< 3.15 mm < 2%

~ 60
"E
CD
Cold strength Size distributio~ after > 6.3 mm > 70-80% ISO 3271 "E
tumbling 0
()
Japanese tumble > 10mm > 52% JIS CD
E
=>
40
Reduction- Size distribution < 3.15 mm < 28% ISO 4696-1 ~
disintegration after reduction and
tumbling (500°C,
RDI-1, static)

02

Reducibility Weight decrease %/min > 0.8% ISO 4695

during reduction
Decrepitation Thermal shock effect < 6.3 mm <8% ISO 8371 0
(lump ore) (700°C) 3.5
0 0.5 1.0 1.5 2.0 2.5 3.0
.. CaO+MgO
BasIcity: SiO2
Figure 18: Phase composition of sinters (after Grebe et ai, 1980)
 3 - The ore burden: sinter, pellets, lump ore 3 - The ore burden: sinter, pellets, lump ore

First, let us consider the liquidus temperatures of sinter-type materials. The acid Second, we consider the reduction-disintegration properties of the sinter. The driving
types of materials have much higher liquidus temperature than basic sinter. This is force of low temperature reduction-disintegration of sinter is the changeover of the
due to the fact, that calcium ferrite type structures have liquidus temperatures as crystal structure from hematite to magnetite, which causes internal stress in the iron
low as 1,200°C (Figure 19), while the acid sinter have liquidus temperatures well oxide crystal structure. So, reduction-disintegration of sinter is related to the fraction
above 1AOO°C. It means also, that sintering of fluxed or superfluxed sinter can of hematite in the sinter. As shown in Figure 18, there are primary and secondary
be accomplished at lower temperatures than sintering of a more acid sinter blend. hematites in the sinter. Especially the latter causes reduction-disintegration, since it
Because of this, acid sinter is generally coarser and has a higher cold strength than is more easily reduced in the upper part of the furnace than primary hematite.
basic sinter.
The higher the hematite percentage in the sinter, the more the sinter is prone to
2 CaO'AI203"SiO2 reduction-disintegration. Or vice versa: there is strong relationship between the
1590°C
FeO content of the sinter and the reduction-disintegration: the higher the FeO
content, the less reduction disintegration will take place. The FeO content of sinter
can be increased by adding more fuel (coke breeze) to the sinter blend. However,
the precise relationship between the FeO content of the sinter and the sinter quality
depends on the ore blend used and is plant-specific. The reduction-disintegration
properties depend on the type of FeO present in the crystal structure. As an example:
a high fraction of magnetite in the sinter blend also gives a sinter with a high
(primary) magnetite fraction. Moreover, in the presence of sufficient Si02 so-called
fayalite structures (2FeO.Si02) can be formed. These structures are very stable and
are difficult to reduce. They are reduced at high temperatures i.e. by the so-called
direct reduction reactions (see section 8.2.1). On the other hand, in the presence
of MgO, spinel structures containing large amounts of FeO can be formed (spinel:
MgO.AI2O3 with FeO). The spinel structures are relatively easy to reduce. Finally,
sinter that has been formed at high temperatures (acid sinter), will contain glass-like
structures and the FeO is relatively difficult to reduce.

It is possible to suppress the formation of secondary hematites by cooling the sinter

Wustite with air with a low O2 percentage (12 to 14%). This is done at Corus IJmuiden
with a gas recycling system. As a result the FeO content of the sinter is relatively
, , high, because less secondary hematite is formed. This has a major benefit for the
, , FeO reduction-disintegration properties of this type of sinter. In addition, the calorific value
1380°C of the blast furnace top gas increases, since less oxygen has to be removed from the
\ ,
\ Liquidus temperature at varying FeO% are burden, which gives a clear financial advantage as well.
\
1500 along line A-A (CaO/SiO2= 1)
During the sintering process there is a major difference between the use of CaO
and MgO as fluxes. Both materials are normally added as the carbonate (limestone:
~ 1300 CaC03 or dolomite CaC03.MgC03). The carbonates are decomposed on the sinter
strand which costs a lot of energy. However, the melts containing substantial amounts
of lime CaO have low liquidus temperatures, i.e. as low as 1,1 OO°C for mixtures of 20
1100 to 27% CaO and iron oxides. For the melts containing MgO - the spinel structures
0% 10% 20% 30% 40% mentioned above- the melting temperatures are much higher. Therefore, it is easier
FeO content (%) to form slag-bonds in the sinter with the help of CaO than with MgO. And generally:
making sinter with CaO can be done at lower temperature. But sinter with high MgO
Figure 19 Example of liquidus temperatures of sinter type materials
is more resistant against reduction-disintegration.
 3 - The ore burden: sinter, pellets, lump ore 3 - The ore burden: sinter, pellets, lump ore

For the final result of the produced sinter, it is important to note, that the sinter blend Pellets
3.4
prior to sintering is rather inhomogeneous. It contains various types of material and
locally there are widely varying compositions present. Ore particles can be as large
as 5 mm, coke breeze up to 3 mm and limestone and dolomite up to 2.5 mm. All 3.4.1 Production of pellets
types of chemical compositions are present on the micro-scale, where the sintering
takes place. Type of materials used, size distribution of the various materials, the Production of pellets is a relatively new development and was driven by the increase
blending of the sinter mix, the amount of slag-bonds forming materials in the blend in very fine ore fractions as a consequence of the ore enrichment methods used.
as well as the amount of fuel used for the sintering all have there own draw-back on Most pellet plants are located close to mines and a schematic of one such plant, a
sinter quality. It makes optimisation of sinter-quality a plant-specific technological travelling grate pellet plant, is shown in Figure 20 (left). Another type of pellet plant
challenge. uses rotary kiln.

Pellet production may be divided into four distinct stages in the travelling grate pellet
In the above sections we stress the importance of reduction-disintegration of sinter.
The lower the reduction-disintegration, the poorer the reducibility of the sinter. plant.
Needle-like structures of Ca-ferrites have a relatively open structure and are easily
accessible for reduction gas in the blast furnace. In cold conditions the sinter is 3.4.1.1 Stage 1-Formation of green pellets
strong (i.e. good tumbler test results), the degradation during transportation is also
good, but the relatively fast reduction in the blast furnace makes the sinter very prone A blend of very fine ores is made, to which fine coke breeze (maximum 1%) may be
to reduction-disintegration. More solid structures in the sinter have better properties added. The blend is pre-pellitised into 'green pellets' on rotating disks or in rotating
in this respect. Reduction-disintegration leads to poorer permeability of the ore- drums. During this step, water and a binder are added to the blend. The adhesion of
burden layers in the furnace and impedes proper further reduction of the iron oxides the particles is accomplished with the thin layer of moisture that is present on each
in the blast furnace. ore particle forming a bridge with the thin layer of moisture on another particle.
When the particle is rotated, more and more particles can adhere and the appropriate
size fraction is selected by screening. The fines are re-circulated in the rotating
drums or disks.

3.4.1.2 Stage 2-Drying

The green pellets are charged via a screen to a travelling grate. The green pellets
have to be dried and the water bond must be replaced with a chemical bond. To
accomplish this, the pellets are gradually heated to 300 to 350°C, which will allow the
binding agent present in the mix, such as bentonite or an organic equivalent, to react.

3.4.1 .3 Stage 3-Firing

The dry pellets are heated to a temperature of 1,250 to 1,350° C and the ore fines
are fused together in a reaction known as induration. The required heat comes from
burners as well as from coke breeze added to the original blend.

3.4.1.4 Stage 4-Cooling

The fired pellets have to cool down in a controlled way to prevent degradation of the
Figure 20: Pellet Plant pellets.

For the heat treatment of the pellets various alternatives are available. A moving bed,
a shaft furnace, or a grate kiln system can be used.
 3 - The ore burden: sinter, pellets, lump ore 3 - The ore burden: sinter, pellets, lump ore

3.4.2 Pellet quality Basic and fluxed pellets have good metallurgical properties for blast furnace
operation. By adding limestone to the pellet blend, the energy requirement of the
With correct chemical composition and induration, pellets remain generally intact in firing/induration increases because of the decarbonisation reaction. For this reason
the blast furnace. Therefore, when judging pellet quality the main issues are: production capacity of a pellet plant can sometimes be 10 t015 percent lower when
. Cold strength, measured as compression strength and the fines generated through
tumbling. A low figure for compression indicates bad or lean firing.
producing basic pellets.

. Swelling properties. With incorrect slag composition pellets tend to have extreme Olivine pellets have MgO instead of GaG, which is added to the blend as olivine sand
swelling properties. Since the phenomenon is well known, it normally does not or serpentine. The pellets are somewhat weaker when tested for cold compression
happen with commercially available pellets. strength.
. The reduction-disintegration properties. These properties are less of a concern
with pellets than with sinter and lump ore. The difference in compression strength might seem large. However, in the blast
. Softening and melting. Pellets tend to melt at lower temperatures than sinter. furnace the pellets are reduced and the difference diminishes during reduction. After
. Moisture level. Moisture in the pellets can interfere with the process. Excessive the first reduction step to Fe304, the cold compression strength drops to 45 to 50 kg
moisture can arise during stocking of pellets. for acid pellets and to 35 to 45 kg for olivine pellets and fluxed pellets. Therefore, the
lower compression strength has no drawback for the blast furnace process.
The slag volume, composition and bonding forces mainly determine the quality of

..
pellets. The three main pellet types are:
Acid pellets.
Basic and fluxed pellets.
The compression strength, especially the fraction with a low compression strength
(under 60 kg/p), is a good indicator for the pelletising process: the more pellets

. Olivine fluxed pellets.

collapse at low compression strength, the poorer the pellets have been fired.
Therefore, pellet quality can be influenced by the production rate: the slower the
grate is moving, the stronger the firing can be, so the induration period increases and
Typical properties of the three types of pellets are shown in Tables 4 and 5. the pellets are stronger.

Table 4: Overview pellet properties A final conclusion on the "optimum" pellet properties has not yet been reached.
A summary of acceptability ranges for pellet quality parameters is given in Table 6.
Acid ++
Table 6: Characterisation of pellets
Basic/ + + +
fluxed
Olivine + + +
Mean size Size distribution 8-16 mm < > 90% ISO 4701
6.3 mm <5%

Table 5: Typical analysis and compression of various types of pellets Cold strength Compression Average kg/p *) > 150 **) ISO 4700
strength % < 60 kg/p <4%
Tumbler strength > 6.3 mm > 95% ISO 3271 or
<0.5 mm <5% ASTM

Acid pellets < 0.15 64%-67% 2-5% <0.2% 250-300 Low Size distribution Strength> 6.3 mm > 80% ISO 13930
< 0.15 64%--67% 2.5--5% 1.3-1.8% temperature after reduction Abrasion < 0.5 mm < 15%
Olivine pellets 150-200
disintegration and tumbling
(dynamic)
Basic pellets - 0.8-1.0 60%--64% 3.5%-5.5% 1.3% 200-250 Reducibility Weight decrease %/min (dR/dt)40 > 0.8% ISO 4695
-1.1-1.3 60%-63% 3.5-5.5%
during reduction
Fluxed pellets 1.5% 200-250
*) kg/p corresponds with metric unit daN/p: decaNewton per pellet (=0.98 kg/p)
**) depends on pellet type. See Table 6.
Acid pellets are strong, but have moderate metallurgical properties. They have good
compression strength (over 250 kg/pellet), but relatively poor reducibility. In addition
the acid pellets are very sensitive to the CaO content. At CaO/SiO2 > 0.3 the pellets
have a strong tendency to swell, which may jeopardise blast furnace operation.

30 31
 --
3 - The ore burden: sinter, pellets, lump ore

3.5 Lump ore

Lump ores are naturally iron-rich materials that are used directly from the mines. The
mines generally produce lump ore as well as (sinter) fines. Major lump ore deposits
are present in Australia (Hamersley, Mount Newman), South America (Carajas, Robe
River) and South Africa (Sishen). In many other places limited amounts of lump ores
are produced.

Lump ores are cheaper than pellets and so in many blast furnaces high amounts of Chapter 4
lump ore are being considered. The lower cost of the lump ore compared with pellets
is offset by the poorer metallurgical properties. Generally speaking, in comparison
with pellets, lump ore:
Coke
. generates more fines during transport and handling.
.. has poorer reduction-disintegration
has a lower melting temperature.
properties.

The fines sometimes tend to accumulate in socalled "piggy-back-fines".

Lump ore is used in an appropriate size fraction, e.g. 8 to 30 mm and the quality
The blast furnace is charged with alternating layers of coke and ore burden. The coke
parameters for lump are similar to those for sinter. .
forms a structure through which the gas can distribute itself and penetrate into the ore
For blast furnace operation at high productivity and high coal injection levels, lump
layers.
ore is a poorer type of burden material, however, since lump ore is a natural material,
Since the 18th century, coke has been the most important source of carbon for the
properties can differ from type to type. Siderar blast furnace in Argentina, for
blast furnace process. From 1960 to now more and more auxiliary reductants are used,
example, operates successfully with a Brazilian lump up to 40% of the burden at high
such as oil, tar, coal and natural gas. The auxiliary reductants are injected through the
furnace productivity.
tuyeres and in doing so the flame temperature drops, so to compensate for this oxygen
enrichment of the blast is often used simultaneously with injection in the tuyeres.

1200 'Ore beneficiation

Input of overseas rich ores
Blast temperature> 1200°C
1000 Oxygen enrichment
~ Top pressure
Burden distribution
~ 800 Gas flow control
0, Improvement of Fe burden
6 Improvement of coke
c Gas flow control
0
600 Small coke in
'5. Fe burden
E
::>
(/)
c 400
0
()
200

0
1950 1960 1970 1980 1990 2000

Figure 21: Coke rate decrease (Courtesy VDEh)

Over the last 50 years, the coke rate of the blast furnace process has considerably
decreased, as illustrated in Figure 21. This is due to the application of auxiliary reductant
injection, and also factors such as improved burden quality, higher blast temperatures,

32 33
 4 - Coke 4 - Coke

bigger furnaces and improved process control. Presently, the reductant rate of larger The various test results to characterise coke and the relevance for the process are
blast furnaces is well below 500 kg/tHM coke equivalent, of which 30 to 40% is injected indicated in Table 7 and in Figure 22.
as coal. Coke rates slightly above or below 300 kg/tHM have been reached and can be
considered normal operational practice. Table 7: Acceptability range for coke quality parameters

4.1 Functions of coke

Mean size Size distribution Average size mm 45-60
The most important functions of coke are: % < 40 mm <25%
% < 10 mm <2%
0 Coke is consumed and generates the heat to melt the burden.
0 Coke provides the structure through which gas can ascend and be distributed Cold Size distribution after 140% > 40 mm > 45 55 Irsid test
strength tumbling 120% > 20 mm >78
through the burden. Coke is a solid and permeable material up to very high
110%<lOmm <19 16
temperatures (over 2,000°C), which is of particular importance in the hearth and M40 % > 40 mm > 80 87 Micum
melting and softening zone. In the area below the melting zone coke is the only M10 %< 10mm <7 5.5 test
solid material remaining. The weight of the blast furnace contents is supported by %>1" > 58
Stability at wharf %>1" > 60 ASTM test
the coke structure. The coke bed has to be permeable, so that iron and slag can > 70
Stability at stockhouse % > 1/4"
accumulate in the hearth and flow to the tap hole. Hardness
0 Coke provides the carbon used for carburisation of the hot metal. % > 10 mm
Strength CSR: reaction(CO2J+ > 58 70
after tumbling
4.2 Coke quality reaction

Reactivity CRI: % weight loss during % <29% 24

The coke quality can be described in two broad categories of its composition and its test
mechanical cold and hot strength. The relevant aspects of the composition are the IRSID, MICUM and ASTM tests differ in their sample preparation. Irsid test takes coke sample> 20mm.
ash and moisture content. Both should be as low as possible. Ash contents generally MICUM test takes coke sample> 63 mm. ASTM test takes coke sample 2"-3"
Reference: after Kolijn (2002), Vander et al (1996)
are 8-11 %, but depend on the local situation. Moisture is a consequence of the
quenching and subsequent handling of the coke. Further aspects of the chemical
composition of the coke are the sulphur and alkali content, which is discussed later.
The physical quality of the coke is characterised by:
0 Size and size distribution. Average mean size of metallurgical coke is typically in
- Size Distribution after Stabilisation
0140,M40, Stability
the range of 45 to 55 mm. The size distribution should be tight to allow for a good Abrasion Resistance
permeability of the coke. A large fraction of coke over 80 mm is an indication of 0110, M10
poor control of the coking process.
0 The resistance to physical degradation due to transport and other mechanical - CokeHotStrength
stress. The parameters used for this property are: 140, M40, and stability. These 0 CSR, Coke Strength after Reaction
parameters indicate the size distribution after stabilisation of coke. (percentage under 10 mm
after reaction at 11OO°C,
0 The resistance against abrasion, which is characterised by a test result called 110
CO2, 2 hours and tumbling)
and M10 and hardness.
0 CRI, Coke Reactivity Index
0 The coke reactivity. The coke can react with CO2,which attacks the coke matrix (percentage weight loss,
and leads to higher coke rate in the blast furnace. Coke with a lower reactivity after aforementioned reaction)
(CRI, coke reactivity index) and higher strength after reaction (CSR, coke strength
after reaction) has higher mechanical strength in the lower part of the furnace.
0 The day-to-day consistency of coke quality, especially size and size distribution,
is often a neglected area in blast furnace operation. Inconsistency in coke quality Figure 22: Coke tests and the blast furnace process
leads to differences in burden distribution and may influence blast furnace internal
gas flow.

CIA CI<::
 ...........-
4 - Coke 4 - Coke

The coke quality is mainly determined by the coal blend used, although, coal
preparation, carburisation time, equipment and quenching also influence coke quality.
Volatiles.,
Usually a blend of different types of coal is used. Coking coals are coal types suitable
for cokemaking and are priced higher than non-coking coals. Optimisation of the coal
blend is an art in itself with a variety of important factors to be taken into account.
These include the volatile matter of the coal and the gas pressure generated during
carbonisation. The final coke should shrink sufficiently so that it can be pushed easily
from the coke chamber, and should have low enough gas pressure not to damage
the walls of the coke chamber. The carbonisation rate depends on the temperature
of the coke battery, the faster the carbonisation rate the more fissures occur and the
smaller the resulting coke. Generally, carbonisation times for slot-oven coke plants
... ... ... ...
are in the range of 16-24 hours. Coal Coal Charging Heating Pushing Quenching
Blending Grinding in oven (16-24 hrs, (water)
4.3 Coke production (4--8 types) 1200°C)
Figure 24: Coke production schedule
The coal blend is heated to high temperatures (1 ,200°C) in an air-free environment
(Figure 23). As the temperature increases, volatile matter of the coal escapes and 4.4 Coke quality in the blast furnace
a solid carbon matrix is formed, termed the carbonisation of coal. A simplified coke
production process is shown in Figure 24 (next page). The coal is heated from the Coke is subjected to mechanical and chemical attack during its descent from
wall to the centre, forming a so-called plasticity zone. As soon as the pla.sticity zones stockline through the stack and down into the hearth. The average size of coke
reach the centre, the coking process is completed. decreases during its descent, as presented in a typical example of coke degradation
in Figure 25. It was found that coke gets smaller over the route from coke plant to
tuyeres while measurement after stabilisation (140) remained constant. The larger the
coke charged to the furnace, the larger it arrives at the tuyeres.

100

80 Strong Coke
E (140= 60)
E
0
"<t
A 60
j
~
OJ
OJ
ro

~
~
40

u u uuu If
00000
L{) 0 000 20
~ 0 co"<t~
~ ~

0
Wharf Screens Screens In Front
Figure23: A coke chamber
Coke Plant BlastFurnace of Tuyeres

Figure 25: Coke degradationfrom wharftotuyeres(after

Van der Velden,2001)
 ...........--
4 - Coke

The coke quality in the hearth has been the subject of much debate (e.g. Geerdes
et ai, 2001). The liquids must be cast from the hearth and permeability of the coke
bed must be good. However, the reaction rate in the hearth is slow since coke is
consumed only by direct reduction reactions and iron carburisation. For this reason
the coke residence time in the hearth is long.

Wear of the hearth refractory by exposure to ring currents of iron along the hearth
wall may be affected by coke quality - the smaller the void fraction, the stronger the
ring currents. Due to the internal configuration of the furnace it is considered likely
that the coke charged into the centre of the furnace is the coke that reaches the
Chapter 5
hearth. Some companies have put extra effort in the charging of large, strong coke
into the centre of the furnace in an effort to maintain good dead man permeability. Injectionof coal, oil and gas
Coke can react with carbondioxide (C + CO2 2 CO). ~

The temperature, where this reaction starts, is in normal situations 1,100 to 1150 °C.
However, the temperature depends especially on the presence of catalysts such as
alkalis (Na, K) and on the ash composition of the coke: the more Fe in the coke ash,
the higher the reactivity and the lower the temperature at which coke gasification by
CO2 starts. Higher coke reactivity leads to higher coke consumption per tonne hot
metal. Use of injection of pulverised (or granular) coal, oil and natural gas can lower the cost
of hot metal. The auxiliary reductants are mainly coal, oil and natural gas, but tar and
other materials can also be used. The precise financial balance depends very much
on local situations.
Up until the early 1980's oil injection was a commonly used, however the changes in
relative prices between coal and oil has resulted in coal becoming the more widely
used injectant. Note, that the preparation of coal for injection involves a rather high
investment cost. The pay-back of the investment heavily depends on the hot metal
production level. Most major sites have been equipped with coal injection. When coke
is scarce and expensive, the feasibility of coal injection for smaller sites increases. The
most important arguments for the injection of coal (or natural gas) in a blast furnace
are;
. Cost savings by lower coke rates. Cost of coke is substantially higher than that of
coal, moreover, the use of an injectant allows higher blast temperatures to be used,
which also leads to a lower coke rate.
. Increased productivity from using oxygen enriched blast.

The reason for the apparent versatility of the blast furnace in consuming all types of
carbon containing materials is that at the tuyeres the flame temperatures are so high
that all injected materials are converted to simple molecules like H2 and CO and behind
the raceway the furnace "does not know" what type of injectant was used.

Coal injection was applied in the blast furnace Amanda of ARMCO (Ashland, Kentucky)
in the 60's. In the early days of coal injection, injection levels of 60-100 kg coal per
tonne hot metal were common. Presently, the industrial standard is to reach a coke
rate of 300 kg/t with injection levels of 200 kg coal per tonne hot metal (Toxopeus et
al,2001).

38 39
 4 - Coke
-
The coke quality in the hearth has been the subject of much debate (e.g. Geerdes
et ai, 2001). The liquids must be cast from the hearth and permeability of the coke
bed must be good. However, the reaction rate in the hearth is slow since coke is
consumed only by direct reduction reactions and iron carburisation. For this reason
the coke residence time in the hearth is long.

Wear of the hearth refractory by exposure to ring currents of iron along the hearth
wall may be affected by coke quality - the smaller the void fraction, the stronger the
ring currents. Due to the internal configuration of the furnace it is considered likely
that the coke charged into the centre of the furnace is the coke that reaches the
Chapter 5
hearth. Some companies have put extra effort in the charging of large, strong coke
into the centre of the furnace in an effort to maintain good dead man permeability. Injectionof coal,oil and gas
Coke can react with carbondioxide (C + CO2 2 CO). -
The temperature, where this reaction starts, is in normal situations 1,100 to 1150 ac.
However, the temperature depends especially on the presence of catalysts such as
alkalis (Na, K) and on the ash composition of the coke: the more Fe in the coke ash,
the higher the reactivity and the lower the temperature at which coke gasification by
CO2 starts. Higher coke reactivity leads to higher coke consumption per tonne hot
metal. Use of injection of pulverised (or granular) coal, oil and natural gas can lower the cost
of hot metal. The auxiliary reductants are mainly coal, oil and natural gas, but tar and
other materials can also be used. The precise financial balance depends very much
on local situations.
Up until the early 1980's oil injection was a commonly used, however the changes in
relative prices between coal and oil has resulted in coal becoming the more widely
used injectant. Note, that the preparation of coal for injection involves a rather high
investment cost. The pay-back of the investment heavily depends on the hot metal
production level. Most major sites have been equipped with coal injection. When coke
is scarce and expensive, the feasibility of coal injection for smaller sites increases. The
most important arguments for the injection of coal (or natural gas) in a blast furnace
are;
. Cost savings by lower coke rates. Cost of coke is substantially higher than that of
coal, moreover, the use of an injectant allows higher blast temperatures to be used,
which also leads to a lower coke rate.
. Increased productivity from using oxygen enriched blast.

The reason for the apparent versatility of the blast furnace in consuming all types of
carbon containing materials is that at the tuyeres the flame temperatures are so high
that all injected materials are converted to simple molecules like H2 and CO and behind
the raceway the furnace "does not know" what type of injectant was used.

Coal injection was applied in the blast furnace Amanda of ARMCO (Ashland, Kentucky)
in the 60's. In the early days of coal injection, injection levels of 60-100 kg coal per
tonne hot metal were common. Presently, the industrial standard is to reach a coke
rate of 300 kg/t with injection levels of 200 kg coal per tonne hot metal (Toxopeus et
ai, 2001).

38 39
 ...........-

5 - Injection of coal, oil and gas 5- Injection of coal, oil and gas

5.1 Coal injection: equipment 5.2 Coal specification for PCI

The basic design for coal injection installations requires the following functions to be Coal types are discriminated according to their volatile matter content. The volatile
carried out: matter is determined by weighing coal before and after heating for three minutes
. Grinding of the coal. Coal has to be ground to very small sizes. Most commonly at 900DC. Coals that have between 6 and 12% volatile mater are classified as low
volatile, those between 12 and 30% are mid volatile and anything over 30% are high
used is pulverised coal: around 60% of the coal is under 75 JJm. Granular coal is
somewhat coarser with sizes up to 1 to 2 mm. volatile. All types of coal have successfully been used.
. Drying of the coal. Coal contains substantial amounts of moisture, 8% to more than The most important properties of the injection coals are:
. High replacement ratio (RR) of coke. The composition and moisture content of the
10%. Since injection of moisture increases the reductant rate, moisture should be
removed as much as possible. coals determine the amount of coke replaced by a certain type of coal. A simple
. Transportation of the coal through the pipelines. If the coal is too small the pneumatic formulafor the replacement ratio (compared withcoke with87.5% carbon) is:
transport will be hampered. It may result in formation of minor scabs on the walls
and also lead to coal leakage from the transportation pipes. RR= 2x C%(coal)+ 2.5x H%(coal)- 2x moisture%(coal)-86 + 0.9x ash%(coal)
. Injection of the pulverised coal: Coal has to be injected in equal amounts through
. Composition: high sulphur and high phosphorous are likely to increase costs in
all the tuyeres. Particularly at low coke rate and high productivity the circumferential
symmetry of the injection should be maintained. the steel plant. These elements should be evaluated prior to the purchase of a
certain type of coal. Young coals (high oxygen content) are known to be more
There are various suppliers available for pulverised coal injection (PCI) installations, susceptible to self-heating and ignition in atmospheres containing oxygen. This is
which undertake the functions mentioned above in a specific way. The reliability of the also an important factor that must be considered with regard to the limitations of
the ground coal handling system.
equipment is of utmost importance, since a blast furnace has to be stopped within one
hour, if the coal injection stops. . Volatile matter: high volatile coals are easily gasified in the raceway, but have

. lower replacement ratio in the process.

Hardness. The hardness of the coal, characterised by the Hardgrove Grindability
Index (HGI) must correspond to the specifications of the grinding equipment. The
Dust Filters
resulting size of the ground coal is also strongly dependent on this parameter and

. must correspond to the limits of the coal handling and injection system.
Moisture content. The moisture content of the raw coal as well as the surface
moisture in the ground coal must be considered. Surface moisture in the ground
Cyclone Distributor coal will lead to sticking and handling problems.
(optional)
From Coal Field - / Potential injection coals can be evaluated on the basis of "value in use", where all
effects on cost are taken into account. It is often possible to use blends of two or
three types of injection coals, so that unfavourable properties can be diluted.
Coal Storage

5.3 Coal injection in the tuyeres

Coals are injected via lances into the blowpipes, ignited and gasified. The coal is in
the raceway area only for a very short time (5 milliseconds) and so the characteristics
Grinding Miils of the gasification reaction are very important for the effectiveness of a PCI system.
Coal gasification consists of several steps as outlined in Figure 27 (next page).
Gas inlet - Firstly the coal is heated and the moisture evaporates. Gasification of the volatile
components then occurs after further heating. The volatile components then ignite
Transport Air/N,
and are gasified, which causes an increase in the temperature, and then finally the
Figure 26: Example of PCI instal/ation coal char is gasified. All of these steps occur sequentially with some overlap.

A.
 5 - Injection of coal, oil and gas
- 5 - Injection of coal, oil and gas

The gasification of coal also depends on the percentage of volatile matter (VM). If
low volatile coals are used, a relatively high percentage of the coal is not gasified
in the raceway and is transported with the gas to the active coke zone. This "char"
@ Ignition/Oxidation of Volatiles
::> will normally be used in the process, but might affect the gas distribution. The high
'@ volatile (HV, over 30% VM) and ultra high volatile coal (over 40% VM) produces
Q) Gasification of Volatiles
0..
E a large quantity of gas in the raceway and a small quantity of char. If the gas
~ combustion is not complete, soot can be formed. Blending a variety of injection coals,
especially high volatile and low volatile coals, gives the advantage of being able to
control these effects.

It has been found that the coke at the border between raceway and dead man
~
Time (msec) contains more coke fines when working at (high) injection rates. This region has been
termed the "bird's nest".
Figure 27: Coal gasification

5.4 Process control with Pulverised Coal Injection

The effects of lance design, extra oxygen and coal type on the coal combustion
have been analysed. Originally, the coal lances were straight steel lances that were
At high Pulverised Coal Injection operation about 40% of the reductant is injected
positioned at or close to the tuyere/blowpipe interface as indicated in Figure 28.
via the tuyeres. Therefore, it is important to control the amount of coal per tonne hot
Generally, the simple arrangement works well, although, occasionally "char" (unburnt
metal as accurate as the coke rate is controlled. The feed tanks of the coal injection
degasified coal) or "soot" (very fine carbon formed from gas) are detected as they
are weighed continuously and the flow rate of the coal is controlled. It can be done
leave the furnace through the top. To avoid this problem, especially at high injection
with nitrogen pressure in the feed tanks or a screw or rotating valve dosing system.
rates, companies have installed different types of injection systems at the tuyeres,
such as: In order to calculate a proper flow rate of coal (in kg/minute) the hot metal production
. Coaxial lances with oxygen flow and coal flow. has to be known. There are several ways to calculate the production. The production

.. Use of two lances per tuyere.

Specially designed lances with a special tip to get more turbulence at the lance tip. rate can be derived from the amount o~ material charged into the furnace. Short-term
corrections can be made by calculating the oxygen consumption per tonne hot
. Coal preheating. metal from the blast parameters in a stable period and then calculating the actual
production from blast data. Systematic errors and/or the requirement for extra coal
can be put in the control model.
When using PCI, deposits of coal ash are occasionally found at the lance tip or
within the tuyere. The deposits can be removed by periodic purging of the lance by
The heat requirement of the lower furnace is a special topic when using PCI. Coal is
switching off the coal while maintaining air (or nitrogen) flow.
not only used for producing the reduction gases, but use of coal has an effect on the
heat balance in the lower furnace. The heat of the bosh gas has to be sufficient to
melt the burden: define the "melting heat" as the heat needed to melt the burden.
The heat requirement of the burden is determined by the "pre-reduction degree",
or how much oxygen has still to be removed from the burden when melting. The
removal of this oxygen requires a lot of energy. The "melting capacity" of the gas is
defined as the heat available with the bosh gas at a temperature over 1 ,400°C. The
melting capacity of the gas depends on:
. The quantity of tuyere gas available per tonne hot metal. Especially when using
Figure 28: Coal injection in the tuyeres and lance positioning high volatile coal there is a high amount of H2 in the bosh gas.
. The flame temperature in the raceway.

.
The speed of gasification increases
The volatility of the coals increases.
as;
The flame temperature in itself is determined by coal rate, coal type, blast
.. The size of the coal particles decreases.
The blast and coal are mixed better.
temperature, blast moisture and oxygen enrichment.

Moreover, as the injection level increases, the amount of coal that leaves the raceway
without being gasified increases.

L1? AC>
 5 - Injection of coal, oil and gas
- 5 - Injection of coal, oil and gas

From the above, the oxygen percentage in the blast can be used to balance the heat 5.5 Gas flow control
requirements of the upper and lower furnace. The balance is dependent on the local
situation. It depends e.g. on burden and coke quality and coal type used. For the For optimised blast furnace process the control of the blast furnace internal gas flow
balance there are some technical and technological limitations, which are presented is critical. A detailed discussion is presented in Chapter 6. Generally speaking, it
as an example in Figure 29. For higher injection rates more oxygen is required. The has been found that the balance between central gas flow and wall gas flow has to
limitations are given by:
. Too low top gas temperature. If top gas temperature becomes too low it takes too
be maintained. In the Corus IJmuiden works the presence of an ore-free centre is
used in order to distribute the bosh gas through the coke slits (Exter et ai, 1997). As
long for the burden to dry and the effective height of the blast furnace shortens.
. Too high flame temperature. If flame temperature becomes too high burden
descent can become erratic.
a consequence the root of the melting zone can descend quite close to the tuyeres.
The associated risk of a low root of the melting zone is damage to the tuyeres

. Too low flame temperature. Low flame temperature will hamper coal gasification
(leakage and/or tipping).
High PCI requires a better circumferential symmetry of the process, not only from the
and melting of the ore burden.
. Technical limitations to the allowed or available oxygen enrichment.
burdening, but also from the blast and coal injection distribution through the tuyeres.

41 5.6 Circumferential symmetry of injection

If every tuyere in a blast furnace is considered as part of the blast furnace pie and is
37 responsible for the process to the stock-line, it is self evident that the circumferential
symmetry of the process has to be assured to reach good, high performance. The
~
~ various systems in use for PCI have different methods to ensure a good distribution.
(j) 33
co
]5
.!:
c
~
>. 29
><
0

25

21
Normal Operation One Tuyere Off
100 150 200 250 300
Figure 30: Schematic presentation of the effect of no PCI on one tuyere
Coal Injection (kg/thm, r.r. = 0.87)

Figure 29: Limiting factors affecting raceway conditions with Pulverised Coal Injection Table 8: Coke use per tuyere in case a single tuyere receives no coal
(RAFT= Raceway Adiabatic Fiame Temperature)
Pc:1 at all tuyeres PCI at one tuyere off

Note especially the effect of the use of extra coal injection for recovery of a cooling . Coke rate 300 kg coke/t . Tuyere uses 3,300 kg/hr, coke
furnace. By putting extra coal on the furnace the production rate decreases by about .. PCI 200 (R.R. = 0.85)
Total
170 kg coke/t
.
only
Productionincrease at
470 kg coke/t
2.5% for every 10 kg extra coal per tonne hot metal. Simultaneously, the flame
temperature drops by about 32°C. Therefore, depending on the specific situation, the
. Production 10 t/hr per this tuyere without PCI of
tuyere 3,300/1,600 = 205%
melting capacity per tonne hot metal may even decrease. If the flame temperature' . Carbon balance per tuyere:
drops from 2,050 to 2,018°C, the melting capacity decreases by 5% (32/2,050-1,400) .. Coke
Coal
3,000 kg/hr
1,700kg/hr
and the production decreases by 2.5% resulting in a decrease of melting heat per
. Total 4,700kg/ht
tonne hot metal by 2.5%. If the chilling furnace has insufficient melting capacity of . Ironcarburisation -500 kg/hr
the gas, extra PCI may worsen the situation. In such a situation the efficiency of the . To direct reduction' -900 kg/hr
process must be improved, i.e. by a lower production rate and lower blast volume. . Used at tuyeres 3,300 kg/hr
. of whichcoal 1,700 kg/hr
. of whichcoke 1,600 kg/hr

44 45
 ............-
5 - Injection of coal, oil and gas

However, the largest deviation from circumferential symmetry occurs when no coal
is injected in a particular tuyere. If no injection is applied, the production rate at that
particular tuyere doubles. Consequentially, the blast furnace operator has to take
care that all tuyeres are injecting coal. In particular, where two tuyeres next to each
other are not injecting coal the equalising effects between the tuyeres are challenged.
Especially if the furnace is operating at high PCI rates, the situation is rather serious
and short-term actions have to be taken to correct the situation.

This point can be illustrated from Table 8 and Figure 30 (previous page). The
calculation shows, how much coke is consumed in front of a tuyere, where coal
Chapter 6
injection is switched off. At high injection rates, the production can increase twofold.
Note, that this is an example, since in such a situation neighbouring tuyeres will tend Burdencalculation
to contribute. Moreover, the calculation does not take the oxygen of the coal itself into
account.
and mass balances
5.7 Other injectants

As stated earlier, all types of (hydrocarbon) injectants can be used. A comparison

of replacement ratio, typical chemical composition and effect on flame temperature
iaregiven in Table 9. 6.1 Introduction

Table 9: Typical data for injectants The blast furnace is charged with pellets, sinter, lump ore and coke, while additional
reductant might be injected through the tuyeres. The steel plant requires a defined
quality of hot metal and the slag has to be chosen for optimum properties with respect
to fluidity, desulphurising capacity and so on. Therefore, the blast furnace operator
Coal 0.80 78-82 4-5.5 1--4 -32 has to make calculations to select the blast furnace burden. The present chapter
Oil 1.17 87 11 2 -37
first indicates the conditions for a burden calculation, which is then illustrated with a
practical example. Later in the chapter the burden calculation is taken a step further to
Natural 1.05 57 19 -45
indicate the process results. To this end a simple one-stage mass balance is used.
gas

Tar 1.0 87 6 2 -25 6.2 Burden calculation: starting points

*) Compared with standard coke with 87.5% C
**) When injecting additiona/10 kg/tHM

Starting points for burden calculations are the hot metal and slag quality.
. Hot metal quality: silicon, typically 0.4 to 0.5%. Low sulphur (under 0.03%) and
defined phosphorous levels, which vary due to variation in burden materials from
0.05% to 0.13%.
. Slag quality: generally the lower the slag volume the better. Typically the four major
constituents of slag contain about 96% of the total volume: AI2Os (8 to 20%), MgO (6
to 12%), SiO2 (28 to 38%) and CaO (34 to 42%). For slag design, see chapter 10.

The burden has to fulfil requirements with respect to:

. Maximum phosphorous input, since phosphorous leaves the furnace with the iron.
. Maximum alkali input, especially potassium, which can attack the refractory and
affect the process. Typically a limit of 1 to 1.5 kg/tHM is used.
. Maximum zinc input: zinc can condense in the furnace and can, similar to alkali, lead
to a Zn cycle. Typically, limits for zinc input are 100-150 g/tHM. With high central gas
temperatures, zinc and alkali are partly removed with the top gas.

46 47
 JIll""""
Burden calculation and mass balances Burden calculation and mass balances

6.3 An example of a burden calculation 6.4 Process calculations: a simplified mass balance

The burden calculation uses the chemical composition (on a dry basis) and the weights The calculations of the previous section can be extended to include the blast into
of the various materials in a charge as input parameters. A charge consists of a layer of the furnace. In doing so the output of the furnace can be calculated: not only the hot
burden material and coke with its auxiliary reductants as injected through the tuyeres. metal and slag composition and the reductant rate, but the composition of the top gas
In order to be able to do the calculation, the yield losses when charging the furnace are as well. Calculation of the top gas composition is done in a stepwise manner in which
also taken into account. The present example is restricted to the components required the balances of the gas components (nitrogen, hydrogen, oxygen, CO and CO2) and
to calculate the slag composition. The four main components (SiO2, CaD, MgO and iron and carbon are made. For the calculations the example of a 10,000 t/d furnace is
A12O3)represent 96% of the total slag volume. The other 4% consist of MnO, S, K20, P used. The stepwise approach indicated in Table 11.
and many more. The losses from the materials charged through the top into the blast
furnace are taken into account and are normally based on samples of material from the Table 11: Stepwise approach for a simplified mass balance
dust catcher and scrubber systems. The calculation is presented in Table 10. Input Element Nitrogen (N2) Hydrogen Iron (Fe) Carbon (C) Oxygen (°2)
(Hz)
Table 10: Simplified burden calculation Main Sources Blast Injection Burden Coke Burden
Blast Injection (52%)
Che"} analysis Moisture Blast (48%)
Ash Moisture Fe CaO MgQ AI203
What to know N2% in blast H%in %Fe ore %C in coke %°2 wind
Coke 9% 5% 2% 0.5% 5.0% 3.0% reductant burden and injectant
Coal 6% 1% 0% 0.2% 3.0% 1.5% Main output Top gas Top gas Hot metal Top gas (85%) Top Gas
via Hot metal .CO (32%)
Sinter
Pellets
1%
1%
1%
1%
58%
65%
4.0%
3.5%
8.3% 1.4%
1.3%
0.6%
0.8%
(15%) ..
CO2 (64%)
H2O (4%)

Lump 3% 1% 61% 4.0% 1.0% What to know N2 % in top H2 efficiency Hot metal Rates per
gas composition tonne
Weight After losses Composition
Burden kgltHM' kg/tHM Input in Calculation of Top gas H2 % in top Oxygen input Top gas
volume gas via burden composition
Coke 300 ,294 1 15 0 0 9
CO & CO2 %
Coal 200 200 0 6 0 0 3

Sinter 1,000 990 575 40 82 14 6

The approach is as follows:
Pellets 500 495 322 17 0 6 4 Step 1: nitrogen balance: from the nitrogen balance the total top gas volume is
Lump 80 79 49 3 0 0 1 estimated.
Total 1580 947 81 82 20 23 Step 2: hydrogen balance: from hydrogen input and a hydrogen utilisation of 40%
the top gas hydrogen can be estimated. In practice hydrogen utilisations of
Correction HM silicon 0.46%= 10 kg Si02/tHM -10 38-42% are found.
Slag 71 82 20 23 Step 3: iron and carbon balance: the carbon use per tonne is known from the hot
Results metal chemical composition and coke and coal use per tonne.
Step 4: oxygen balance: the burden composition gives the amount of oxygen per
Slag volume *) kgltHM 204 Si02 CaO MgO AI2O3 tonne hot metal input at the top, while also the amount of oxygen with the
Slag composition 35% 40% 10% 11% blast is also known per tonne hot metal.
Basicity CaO/Si02 1.16 Step 5: the balances can be combined to calculate the top gas composition.
(CaO+MgO)/Si02 1.45
The calculations are based on basic chemical calculations. Starting points for the
(CaO+MgO)/(Si02+AI203) 1.10 calculations are, that:
11% . 12 kilogram of carbon is a defined number of carbon atoms defined as a kilomole.
AI2O3
Ore/coke ratio 5.3 . Every mole of an element or compound has a certain weight defined by the
periodic table of the elements.
*) 1.04x(SiO2+CaO+MgO+AIP:JY

48 49
 Burden calculation and mass balances
---,....---
Burden calculation and mass balances

. 1 kmole of a gas at atmospheric pressure and O°C occupies 22.4 m3 STP. Table 14: The nitrogen balance and top gas volume
The properties of the various components used for the calculations are indicated Nitrogen from (1-0.256)x 4836 m'STP/min
in Table 12. The present balance is used for educational purposes figures and blast 6500
compositions are rounded numbers. Effects of moisture in pulverised coal and the From coal 16 m' STP/min
argon in the blast are neglected.
From coke 23 m' STP/min

Table 12: Properties of materials used for mass balance calculations Total input 4875 m' STP/min
Top gas nitrogen 48.5 %
Top gas volume 10051 m' STP/min
N2 28 kg/kmole CO 28 kg/kmole

°2 32 kg/kmole CO2 44 kg/kmole 6.4.2 The hydrogen balance

H2 2 kg/kmole
C 12 kg/kmole Moisture in the blast and coal reacts to H2 and CO according to:
Fe 55.6 kg/kmole
H2O+ C ~ H2+ CO
Si 28 kg/kmole
1 kmol gas (Nz. O2. etc) ~ 22.4 m' STP All hydrogen in coal and coke are converted to H2 in the furnace. In the furnace the
1 tonne hot metal contains 945 kg Fe~ 945/55.6 ~ 17.0 kmoie
H2 is reacting to H2O; part of the hydrogen is utilised again.
Since the top gas volume is known as well as the hydrogen input, the top gas
6.4.1 The Nitrogen Balance hydrogen can be calculated, if a utilisation of 40% is assumed. There are ways to
check the hydrogen utilisation, but it is beyond the scope of the present exercise.
Nitrogen does not react in the blast furnace, so it escapes unchanged via the top gas. Table 15 shows the input and calculates the top gas hydrogen.
At steady state the input equals the output and the top gas volume can be calculated
Table 15: The hydrogen balance
with a nitrogen balance given the nitrogen input and the nitrogen concentration in the
top gas. The input data for a simplified model are shown in Table 13 and the top gas
volume is calculated in Table} 4.
From blast 7
Table 13: Mass balance input
From coal 56
Blast volume 6500 m' STP/min
From coke 4
Oxygen in blast 25.6 %
Total input 67 750
Moisture 10 g/m' STP
Utilisation 40%. so
Production 6.9 tHM/min
Left in top gas 60% 450
Coal rate 200 kg/tHM
Top gas volume 10051
Coke rate 300 kg/tHM
H2 in top gas 4.5%

CO2 22 vol%
6.4.3 The iron and carbon balance
CO 25 vol%

H2 4.5 vol%
Hot metal contains 945 kg Fe per tonne. The balance is taken by carbon (45 kg),
N2 48.5 vol% silicon, manganese, sulphur, phosphorous, titanium and so on. The precise Fe
C% H% 0% content of hot metal depends slightly on the thermal state of the furnace and quality
Coal 78
of the input. For the balance we use 945 kg Fe/tHM. This amounts to 17 kmole
4.5 7 1.4
(947/55.6).
Coke 87 0.2 1.4

50 51
 .,.....
Burden calculation and mass balances Burden calculation and mass balances

The carbon balance is more complicated. The carbon is consumed in front of the From the combination of the carbon balance and the oxygen balance we can now
tuyeres and is used during the direct reduction reaction (see section 8.2.1). The derive the top gas utilisation, as shown in Table 18.
carbon leaves the furnace via the top gas and with the iron. The carbon balance is
made per tonne hot metal. Table 16 shows the results. The carbon via the top gas is Table 18: Calculation of top gas utilisation
also given in katom per tonne hot metal.

Table 16: Carbon balance Carbon via top gas 31.0

Oxygen via top gas 47.6

Carbon from 261 Oxygen bound to hydrogen -1.9

coke
Oxygen as CO and CO2 45.7
Carbon from coal 156

Total carbon use 417 45.7

Oxygen balance: CO+ 2x CO2
Carbon via iron -45 31.0
Carbon balance: CO + CO2
Carbon via top 372 31.0 14.7
CO2
gas
CO 16.3

6.4.4 The oxygen balance Utilisation CO2/ 47.3%

(CO+CO2)
CO2volume 2283 m3 STP/min CO2% 22.7%
The oxygen balance is even more complicated. Oxygen is brought into the furnace
with the blast, with PCI, with moisture and with the burden. It leaves the furnace CO 2539 m3 STP/min CO% 25.3%
through the top. The burden with sinter contains not only Fe2O3 (a/Fe ratio 1.5) but
some Fe3O4 (a/Fe ratio 1.33) as well. The a/Fe ratio used here is 1.46, which means
The calculations can be used to check the correct input data. More advanced models
that for every atom of Fe there is 1.46 atom O. On a weight basis it means, that for
are available, which take into account the heat balance of the chemical reactions as
every tonne hot metal, which contains 945 kg Fe atoms there is 397 kg a-atoms. All
well (e.g. Rist and Meysson, 1966). The models are useful for analysis, especially
this oxygen leaves the furnace with the topgas. The balance is given in Table 17.
! questions like "are we producing efficiently?" and for prediction: what if PCI is
increased? hot blast temperature is increased? and so on.
Table 17: The oxygen balance

Input From blast 240 342

From blast moisture 8
From coal 14
From burden 397

Total input 762

Output via top gas 762 47.6

6.4.5 Calculation of top gas analysis

The oxygen in the top gas is leaving the furnace in three different states:
. Bound to the hydrogen. The quantity is known since we know how much hydrogen

. has been
Bound to
converted to process water.
carbon as CO.
. Bound to carbon as CO2,

52 53
 Chapter 7
The process: burden descent
and gas flowcontrol

7.1 Burden descent: where is voidage created?

The burden descends in the blast furnace from top to bottom. Figure 32 shows a
representation of the burden descent. It is indicated with stock rods, which are resting
on the burden surface and descending with the burden between charging. The burden
surface descends with a speed of 8 to 15 em/minute.

Zero level burden

1.30 m (Stockline)

1. Stock rod descending

2. Stock rod on burden
3. Stock rod descending
with burden
4. Stock rod ascending
for charging

6m
24/0l/0303""", 21/01/""""'" "/0110"',"""

Figure 31: Stable burden descent

 ~
7 - The process: burden descent and gas flow control 7 - The process: burden descent and gas flow control

In order for the burden to descend, voidage has to be created somewhere in the must be found in the configuration of the melting zone. The materials "glue" together
furnace. Where is this voidage created? See Figure 32. and can form internal bridges within the furnace. Poor burden descent arises at the
. Firstly, coke is gasified in front of the tuyeres, thus creating voidage at the tuyeres. cohesive zone. The effect of a slip is, that the layer structure within the furnace is
. Secondly, the hot gas ascends up the furnace and melts the burden material. So disrupted and the permeability for gas flow deteriorates (See Figure 48).
the burden volume is disappearing into the melting zone.
. Thirdly, the dripping hot metal consumes carbon. It is used for carburisation of the Zero level burden
~'i !
~I

iron as well as for the direct reduction reactions, so below the melting zone coke is ~I "
consumed.
It is possible to indicate how much each of the three mechanisms contributes to the
amount of voidage created. A large part of the voidage is created at the melting zone. Hanging
In a typical blast furnace on high PCI, only about 25% of the voidage is created at the
tuyeres.
Slow Descent

Ore burden consumes coke

by solution loss,
creates voidage Slipping
Coke gasifies in
front of tuyeres,
creates voidage Burden melts,
creates voidage

Gas Burden
ascent 6m ~OOl
descent 1'/01/03 20,55,25 11/01/03 20,59,16 18/01/0302,55,25

Figure 33: Irregular Burden Descent

7.2 Burden Descent: System of Vertical Forces

The burden descends because the downward forces of the burden exceed
counteracting upward forces. The most important downward force is the weight of
Figure 32: Creation of voidage in the Blast Furnace
the burden; the most important upward force is the pressure difference between the
blast and top pressure. The cohesive zone is the area with the highest resistance to
This implies that the mass flow of material is strengthened towards the ring where the
gas flow, which leads to a high pressure drop over the cohesive zone and to a large
highest amount of ore is charged into the furnace. Therefore, at low coke rates high
upward force. If this pressure difference becomes too high, the burden descent can
ore concentration at any ring in the circumference, especially in the wall area, has to
be avoided. be disturbed (Figure 34, next page). This happens for instance, when a blast furnace
is driven to its limits and exceeds the maximal allowable pressure difference over the
burden.
Sometimes the burden descent of a blast furnace is erratic. How is the mechanism'?
Ore burden materials and coke flow rather easily through bins, as can be observed
In addition to the upward force arising from the blast pressure, friction forces from the
in the stock house of a blast furnace. Hence in the area in the blast furnace where
descending burden are impacting on the burden descent: the coke and burden are
the material is solid, the ore burden and coke flow with similar ease to the void areas.
pushed outward over a cone of stationary or slowly descending central coke. Also
Nevertheless, blast furnace operators are familiar with poorly descending burden the wall area exerts friction forces on the burden. These friction forces can become
(Figure 33, next page, for an example). Also the phenomenon of "hanging" (no
rather large.
burden descent) and "slips" (fast uncontrolled burden descent) are familiar. From the
analysis in this section it follows that, in general, the cause of poor burden descent

56 57
 - The process:
- - The process:
JIll"""'"
7 burden descent and gas flow control 7 burden descent and gas flow control

The coke submerged in hot metal also exerts a high upward force on the burden due 7.3 Gas flow in the blast furnace
to buoyancy forces (Figure 35) as long as the coke is free to move upwards and does
not adhere to the bottom.
The gas generated at the tuyeres and at the melting zone has a short residence time
of 6 to 8 seconds in the blast furnace (section 2.4). During this time the gas cools
Weight of burden above 1 3850 t down from the flame temperature to the top gas temperature, from 2,000 to 2,200oG
softening/melting zone down to 100 to 150oG, while simultaneously removing oxygen from the burden. The
vertical distance between tuyeres and stockline is around 22 metres. Therefore, the
Friction against wall
gas velocity in the furnace is rather limited, in a vertical direction about 2 to 5 mis,
which is comparable with a wind speed of 2 to 3 Beaufort. During the 6 to 8 seconds
Pressure difference melting l' 2450 t the chemical reactions take place. How is the gas distributed through the furnace?
zone - furnace top
First consider the difference between the coke layers and the ore burden. It is
Dead man coke
important to note, as indicated in Figure 36, that ore burden has a higher resistance
to gas flow than coke. The resistance profile of the furnace determines how gas flows
Result without friction 1400 t through the furnace. The gas flow along the wall can be derived from heat losses or
Figure 34: System of vertical forces in the Blast Furnace hot face temperatures as the gas will heat the wall as it travels past as described in
the next section.

Resulting upward force: 6.2 ton/m3

boP
r r r r r r r 1

Weight of coke: 1 ton/m3 Voidage Diamater

Ore Burden low small 80%

Weight of hot metal: 7.2 ton/m3
Coke high large

Figure 35: Upward force from hearth liquids

20%
In operational practice poor burden descent is often an indicator of a poor blast
furnace process. The reasons can be: .
. The upward force is too high. Experienced operators are well aware of the
maximum pressure difference over the burden that allows smooth operation. If the
maximum allowable pressure difference is exceeded (generally 1.6 to 1.9 bar),
r r r r r r r 1
the process is pushed beyond its capabilities: burden descent will become erratic,
resulting in frequent hanging, slipping and chills. Figure 36: Pressure loss through coke and are

. A hot furnace is also known to have poorer burden descent. This is because the
As soon as the ore burden starts to soften and melt at about 1,1 OooG, the burden
downward force decreases due to the smaller weight of burden above the melting
zone. In addition, there is more slag hold-up above the tuyeres, because of the layer collapses and becomes (nearly) impermeable for gas. If this happens in the
longer distance and the (primary) slag properties. centre of the furnace the central gas flow is blocked.
. Burden descent can be very sensitive to casthouse operation because of the
above-mentioned upward force on the submerged coke.
. The friction forces become high due to softening and melting material in the
melting zone leading to bridge formation.

58 59
 "........-
7 - The process: burden descent and gas flow control 7 - The process: burden descent and gas flow control

7.3.1 Observation of heat fluxes through the wall therefore, the efficiency of a blast furnace is determined by the progress of the
chemical reactions and thus by the gas flow through the furnace.
Figure 37 shows the temperature at the hot face of the furnace wall. It has been
observed in many furnaces, that suddenly the temperature rises in minutes and Two basic types of gas distribution can be discriminated: the "central working"
decreases over the next hour(s). This is often attributed to the loss of scabs (build- furnace and the "wall working" furnace. The typology has been developed to explain
up) on the furnace wall. differences in operation. Intermediate patterns can also be observed.
The explanation put forward in this book is that such temperature excursions are due
to "short-circuiting" of gas along the furnace wall. These "short-circuits" are due to In the "central working" furnace the gas flow is directed towards the centre. In this
the formation of gaps along the furnace wall creating a very permeable area where case the centre of the furnace contains only coke and coarse burden materials and
the hot gasses preferentially flow. Low CO2 concentrations in the wall area during is the most permeable area in the furnace. The melting zone takes on an "inverted V
such excursions have been observed and confirm the "short-circuiting". The basic shape".
premise of the present book is that heat losses through the wall are caused by gas In a "wall working" furnace the gas flow through the centre is impeded, e.g. by
flow along the walls. The gas is more or less directly coming from the raceway. softening and melting burden material. The gas flows preferentially through the zone
with highest permeability, i.e. the wall zone. In this case the melting zone takes the
800 form of "W shape". Figure 38 shows both types.

Both types of gas flow can be used to operate a blast furnace, but have their own
drawbacks. The gas flow control is achieved with burden distribution.
600

~Q)
~
~Q) 400
0..
E
~
200

0
0 3 6
Time of day (hrs)
Figure 37: Temperatureat hot face

Why does the gas flow along the wall? Gas takes the route with the lowest Figure 38: Two types of melting zone,
resistance. The resistance for gas flow in a filled blast furnace is located in the ore Central working (left) and Wall-working (right)
layers, since its initial permeability is 4 to 5 times less than the permeability of coke
layers. There are two areas in the blast furnace that have the highest permeability: 7.3.3 Central working furnace
the centre of the furnace if it contains sufficient coke and the wall area. At the wall
there can be gaps between the descending burden and the wall. In the centre of the, The two types of gas flow through a furnace can be achieved with the help of the
furnace there can be a high percentage of coke and there can be relatively coarse
burden distribution. In Figure 39 (next page) the ore to coke ratio over the radius
ore burden due to segregation.
is shown for a central working furnace. In the figure the centre of the furnace only
contains coke. Therefore, in the centre of the furnace no melting zone can be formed
7.3.2 Two basic types of melting zone and the gas is distributed via the coke slits from the centre towards outside radius of
the furnace. The melting zone gets an inverted V or even U shape. The central coke
The efficiency of the furnace is determined by the amount of energy used in the column not only serves as a gas distributor, but as well as a type of pressure valve: it
process. Heat losses to the wall and excess top gas temperature are examples of functions to stabilise the blast pressure.
energy losses. The top gas contains CO and H2' which have a high calorific value,

60 61
 --,........
7 - The process: burden descent and gas flow control 7 - The process: burden descent and gas flow control

Burden distribution at throat level The central working furnace can give very good, stable process results with respect
100 to productivity, hot metal quality and reductant rate. It also leads to long campaign
length for the furnace above the tuyeres. However, the process is very sensitive for
deviations in burden materials, especially the size distribution, casthouse operation,
~
~
circumferential symmetry of injection and stops /starts of the furnace.
OJ
E
..::! 50 7.3.4 Wall working furnace
0
>
OJ
-""
0 In Figure 40 the wall working furnace is presented. Melting ore burden blocks the
0
centre of the furnace and the gas flow is directed towards the wall area.

Burden distribution at throat level

0
- Wall Centre 100
Figure 39: Central working furnace

It depends on the type of burden distribution equipment how the coke can be brought ~
~
OJ
to the centre. With a bell-less top the most inward positions of the chute can be E
used. With a double bell system the coke has to be brought to the centre by coke ..::!
0 50
>
push (see below) and by choosing the right ore layer thickness in order to prevent the OJ
-""
flooding of the centre with ore burden materials. 0
0
In the central working furnace there is a relatively small amount of hot gas at the
furnace wall: hence low heat losses. As a result the melting of the burden in the wall
area takes place close to the tuyeres, so the root of the melting zone is low in the 0
furnace. The risk of this type of process is that ore burden is not melted completely Wall Centre
before it passes the tuyeres. This could lead to the observation of lumps of softened
Figure 40: Wall working furnace
ore burden through the tuyere peep sites. This can lead from slight chilling of the
furnace (by increased direct reduction) and irregular hot metal quality to severe chills The gas flow causes high heat losses in the area of the furnace where a gap can
and damage of the tuyeres.
be formed between burden and wall i.e. in lower and middle shaft. The melting
zone gets a W shape or even the shape of a disk. In this situation the root of the
Limiting the risk of a low melting zone root can be done with gas and burden melting zone is higher above the tuyeres, which makes the process less sensitive
distribution. Operational measures include the following.
. Maintain a sufficiently high coke percentage at the wall. Using nut coke in the
for inconsistencies. The process can be operated rather efficiently, however, due to
the high heat losses the wear of the refractory in the shaft is much more pronounced
wall area can also do this. Note that an ore layer of 55 cm at the throat needs than with the central working furnace. The gas passing along the wall can also cool
about 20 to 22 cm of coke for the carburisation and direct reduction. So if the coke
down rapidly and in doing so lose its reduction capabilities. Moreover the fluctuations
percentage at the wall is under 27%, a continuous ore burden column is made at
in the pressure difference over the burden are more pronounced, which leads to
the wall.
. Ensure a minimum gas flow along the wall in bosh and belly, which can be
limitations in productivity.

monitored from heat loss measurements and/or temperature readings. If the gas 7.3.5
flow along the wall becomes too small, it can be increased by means of burden
Gas distribution to ore layers
distribution (more coke to the wall or less central gas flow) or by increasing the gas
Gas produced in the raceway is distributed through the coke layers in the cohesive
volume per tonne hot metal (by decreasing oxygen).
. Control the central gas flow. Note that the gas flow through the centre leaves the
Zone and into the granular coke and ore layers, as shown in Figure 41.

furnace at a high percentage of CO and H2 and a high temperature. The energy

content of the central gas is not efficiently used in the process and thus the central
gas flow should be kept to a minimum.

62 63
 JIll"""'"
7 - The process: burden descent and gas flow control 7 - The process: burden descent and gas flow control

100%

Pellets over 10 mm

Sinter over 10 mm 50%

Sinter and pellets

under 10 mm

Sinter and pellets

under 5 mm
0
Wall Centre

Figure 41.. Schematic presentation of gas distribution through coke layers Figure 42: Distribution of fines over the radius, double bell simulation (after Geerdes et ai, 1991)

The ore burden layers account initially for about 80% of the resistance to gas flow. In summary:
. The permeability of the ore burden is determined by the amount of fines.
The reduction process takes place within these layers.
. The amount of fines is determined by:
.The screening efficiency in the stock house.
What determines the contact between the gas and the ore burden layers?
The most important factor determining the permeability to gas flow is the voidage .. The method of burden distribution used.
The physical degradation during transport and charging.
between particles. As shown in F~gure 14 (Section 3.2.1), the voidage between
particles depends heavily on the ratio of coarse to small particles. The wider the size .The low temperature degradation properties of the burden.
distribution, the lower the voidage. Moreover, the finer the materials, the lower the These effects cause a ring of burden material with poor permeability in many
permeability (Chapter 3). operating blast furnaces. This ring of material in particular is often difficult to reduce
In practical operations the permeability of ore burden material is determined by the and to melt down. Sometimes, unmolten ore burden materials are visible as scabs
amount of fines (percentage under 5mm). Fines are very unevenly distributed over through the peepsites of the tuyeres. These scabs can cause operational upsets
the radius of the furnace, as is indicated by the typical example shown in Figure like chilling the furnace or tuyere failures. It is a misunderstanding to think that these
scabs consist of accretions fallen from the wall.
42 (next page). Fines are concentrated along the wall especially under the point of
impact of the new charge with the stockline.
7.4 Fluidisation and channelling
If a bell-less top is used, the points of impact can be distributed over the radius.
With a double bell charging system the fines are concentrated in a narrow ring at the The average gas speed above the burden is rather low, as shown in chapter 2.
burden surface and close to the wall. When the burden is descending the coarser However, in a central working furnace the gas speed might locally reach 10 m/s
materials in the burden follow the wall, while the fines fill the holes between the or more especially in the centre of the furnace. This is well above theoretical gas
larger particles and do not follow the wall to the same extent as the coarser particles. velocities at which fluidisation can be observed.
Therefore, upon descent the fines in the burden tend to concentrate even more. Coke fluidises much more easily than ore burden because of its lower density. It is
believed that the ore burden secure the coke particles in the centre, nevertheless, if
Moreover, sinter and lump ore can break down during the first reduction step (from local gas speeds become too high, fluidisation may occur. Fluidisation of coke has
hematite to magnetite). This effect is stronger if the material is heated more slowly. been observed in operating furnaces as well as models of the furnace. It leads to
Thus, the slower the material is heated the more fines are generated, the extra fines a relatively open structure of coke. It has even been observed, that pellets on the
impede the gas flow even more, giving rise to even slower heating. border of fluidising coke "dives" into the coke layers.

64 65
 .,.....
7 - The process: burden descent and gas flow control 7 - The process: burden descent and gas flow control

15 Conditions in Fines concentrate at the point of impact and the coarse particles flow "downhill"
furnace center while the fine particles remain below the point of impact. This mechanism, known as
~ 10 segregation, is also illustrated in Figure 44.
E-
~
"(3
0
Q)

~
ro
5
0 4
3
10 20 30 40 60

Particle diameter (mm)

Figure 43: Gas velocities for fluidisation of are burden and coke. Area indicated is range from
800°C to 300°C at 1 atmosphere pressure (after Biswas, 1981)

If the fluidisation stretches itself into the lower furnace, channelling can take place, {{{{{{{{{{{{{{{
short-circuiting the lower furnace (or even the raceway) with the top. These are open Figure 45: Coke push effect with gas flow
channels without coke or ore burden in it. Channelling is observed as a consequence
of operational problems, for example, delayed casts can create higher local gas When burden is charged into the furnace, it pushes the coarse coke particles on the
speeds, resulting in channelling. During channelling, the gas might escape through top of the coke layer towards the centre. This effect is called coke push and is more
the top with a very high temperature and low utilisation, since the gas was not in pronounced when the furnace is on blast. It is illustrated in Figure 45.
good contact with the burden.
7.5.2 The charging equipment
7.5 Burden distribution
The type of charging mechanism used has a major impact on the distribution of fines.
Burden distribution can be used tIDcontrol the blast furnace gas flow. The conceptual Figure 46 shows the bell-less top and double bell systems.
framework of the use of burden distribution is rather complex, since the burden
distribution is the consequence of the interaction of properties of the burden materials
with the charging equipment.

7.5.1 Properties of burden materials

Figure 44 shows the angles of repose of the various materials used in a blast
furnace. Coke has the steepest angle of repose, pellets have the lowest angle of
repose and sinter is in between. Hence, in a pellet charged furnace the pellets have
the tendency to slide to the centre.

Coke: 35-38° Sinter: 29-33°

Figure 44: Segregation and angles of repose
Pellets: 25-26°
- Coarse Material - Fine Material

Figure 46: Bell-less top charging (left) and double bell charging (right):
comparison of the segregation of fines on the stockline

66 67
 7 - The process: burden descent and gas flow control
~ 7 - The process: burden descent and gas flow control
----
In a bell-less top the possibility exists to distribute the fines in the burden over The gas flow is closely monitored in order to control it. Instrumentation of the blast
various points of impact by moving the chute to different vertical positions. Coke can furnace is described in the next section.
be brought to the centre by programming of the charging cycle. With a double bell
charging system there is less possibility to vary the points of impact and fines will be The most important parameters to define the actual gas flow are:
concentrated in narrower rings. Modern blast furnaces with a double bell charging 0 Burden descent (stock rods, pressure taps) and pressure difference over the
system are mostly equipped with movable armour, which give certain flexibility with burden.
respect to distribution of fines and the are to coke ratio over the diameter. However, 0 The wall heat losses or temperatures at the wall.
its flexibility is inferior to the more versatile bell-less system. 0 Stockline gas composition and temperature profile.

7.5.3 Mixed layer formation Gas flow control and optimised burden distribution are found on a trial and error
basis, and have to be developed for every furnace individually. Some general
remarks can be made.
The model of thinking applied up to here takes clean are and coke layers as a
1. Gas flow is mainly controlled with coke to are ratio over the radius. An example of
starting point. However, since the average diameter of coke 45 to 55 mm is much a calculated burden distribution is shown in Figure 47. Note the are free centre.
larger than that of pellets and sinter (typically under 15 mm and 25 mm respectively)
2. The centre of the furnace should be permeable and no or minimal (coarse) are
burden components dumped on a coke layer will tend to form a mixed layer. This burden should be present.
mixed layer will have permeability comparable with the are layer. The formation of 3. The coke percentage at the wall should not be too low. Note that 55 cm of are in
mixed layers is also produced by protruding or recessed parts of the wall: such as throat consumes about 20 cm of coke for direct reduction and carburisation. A
protruding cooling plates, missing arm our plates, wear of refractory at the throat continuous vertical column of burden material should be prevented. A coke slit
and so on. The mixed layers have a different permeability and can give rise to should be maintained between all are layers.
circumferential process asymmetry. 4. Concentration of fines near the wall should be prevented.
5. The central gas flow is governed by the amount of are burden reaching the centre.
7.5.4 Gas flow control The amount of are reaching the centre heavily depends on the are layer thickness
and the amount of coarse coke lumps. To reach a stable gas flow the central gas
The optimised gas flow in a modern furnace operated at high productivity and low flow should be kept as consistent as possible and consequently, when changes
coke rate has the inverted V shaped melting zone type as described above. However, in are to coke ratio are required, the are layer should be kept constant. This is
the gas escaping through the (ore-free) centre leaves the furnace with a low especially important when changing the coal injection level as this will result in big
utilisation. This loss of "unused" Igas should be minimised. changes in the relative layer thickness of ore and coke are made.
If the central gas flow is too high, there is a too small gas flow along the wall for 6. The coke layer thickness at the throat is typically in the range of 45 to 55 cm. In
heating, reduction and melting of the are burden and consequently the root of the our example in section 2.3 it is 49 cm. The diameter of the belly is more than 45%
melting zone comes close to the tuyeres. In this situation the reductant rate will higher than the diameter of the throat. Hence, the surface more than doubles
be high and there is a high chance of tuyere damage. It is essential that the gas during burden descent and the layer thickness is reduced to less than half the
flowing though the centre distributes itself through the coke slits to the burden layers. layer thickness at the throat. Japanese rules of thumb indicate that the layer
Therefore, the permeability of the central coke column must not be too high, which thickness at the belly should not be less than 18 cm. The authors have, however,
means that the diameter of the central coke column must not be too wide. successfully worked with a layer thickness of coke at the belly of 14 cm.
If the central gas flow is (partially) blocked, a relatively large part of the gas escapes
along the wall and is cooled down. The result is the part of the gas is cooled down In the practical situation small changes in are layer thickness can strongly influence
low in the furnace and the reduction reactions slow down. In this situation the central
central gas flow. This effect is generally stronger in double bell- movable arm our
gas flow is small and heat losses are high. furnaces than in furnaces equipped with a bell-less top.
Experience has shown that wall gas flow and central gas flow are strongly correlated.
Gas flow control is based on keeping the balance between central and wall gas flow An example for a burden distribution control scheme is given in Table 19 (next page).
to the optimum. If more central gas flow is required then Coke 3 replaces schedule Coke 2. Replacing
Coke 2 with Coke 1 reduces central gas flow.
The difficulty with gas flow control is that the gas flow is influenced by many changes
in burden components, process parameters and installation specifics. The variation in
the percentage of fines near (but not at) the wall and the low temperature breakdown
properties of the burden are especially important.

69
68
 7 - The process: burden descent and gas flow control
l--- 7 - The process: burden descent and gas flow control

Table 19: Bell-/ess top charging schedules with varying central gas flow
Position 11 10 9 8 7 6 5 4 3 2
Wall Centre
Coke 1 More 14% 14% 16% 14% 14% 14% 6% 8%
central
Coke 2 Normal 14% 14% 14% 14% 14% 14% 6% 10%
Coke 3 Less central 14% 14% 12% 14% 14% 14% 6% 12%
Ore 16% 16% 16% 12% 10% 10% 10% 10%

Similar schedules can be developed for a double bell charging system. With a double
bell system, the use of ore layer thickness can also be applied: a smaller ore layer
gives higher central gas flow and vice versa. wall
If a major change in coke rate is required, the operator has the choice either to Figure 47: Example of burden distribution with an ore-free centre
change the ore base and keep the coke base constant, or change the coke base and are burden penetration in coke layer
and keep the ore base constant. Both philosophies have been successfully applied.
The operators keeping the coke base constant point to the essential role of coke for 7.6 Erratic burden descent and gas flow
maintaining blast furnace permeability, especially the coke slits.
The burden descent sometimes becomes erratic (see Figure 33). What happens
The authors, however, favour a system in which the ore base is kept constant. The in the furnace if it hangs and slips? The mechanism of hanging and slipping is
gas distribution is governed by the resistance pattern of the ore burden layers and illustrated in Figure 48 on the next page.
- as mentioned above - by the amount of ore burden that reaches the centre. The
latter can change substantially when changing the ore base, especially in furnaces First, the furnace hangs because the net downward force is too low. This can be
equipped with double bell charging. caused by high friction forces. Note, that granular coke, pellets and sinter flow easily
An illustrative example showing a change in coke rate from 350 kg/tHM to 300 kg/ downwards, as can be observed in the stockhouse. Therefore, high friction forces
tHM is presented in Table 20. The ore base is kept constant and coke base reduced. arise at the cohesive zone, where bridges of melting ore burden are formed.
Experience has shown that relatively minor changes in burden distribution will be
required for optimisation of the central gas flow (i.e. coke distribution). The burden Second, while the furnace hangs, the process continues: coke is consumed and ore
distribution adjustments can be applied as a second step if required. burden melts. Therefore, voidage arises in or below the melting zone.

Table 20: Coke base change when PCI rate changes Third, when this voidage becomes too big, it collapses: the furnace slips. The layer
Old situation New Situation structure is completely disrupted and the gas flow through these layers is impeded.
Coke rate 350 kg/tHM
This leads again to areas in the furnace where ore burden is insufficiently reduced
300 kg/tHM
and remains in a cohesive state for too long. These areas will form the bridges for
Coke base 21 t 18 t
next time the furnace hangs.
Ore base 90t 90t
Burden distribution No change until required
The problem can only be solved by re-establishing the layer structure within the
furnace, which means, that the complete content of the furnace has to be refreshed:
the furnace has to be operated on reduced blast volume for five to ten hours.
Burden distribution changes should be based on an analysis of the causes of
changes in gas flow. The gas flow can also be influenced by operational problems,
such as a low burden level or problems in the casthouse. In this situation adjustments
in the burden distribution will not give satisfactory results. Heat losses through the
wall are very closely related to burden descent. Therefore, the cause of high heat
loads should be analysed together with other process data. An example of a burden
distribution is shown in Figure 47 (next page).

70 71
 ",......-
7 - The process: burden descent and gas flow control 7 - The process: burden descent and gas flow control

7.7 Blast furnace instrumentation

An overview of blast furnace instrumentation as discussed in various parts of the text

is given in Figure 49.

Profilemeter

Top gas: CO, CO2, H2' temperature (total)

Probes: CO, CO2, H2' temperature (radial)
Burden level

Skin flow temperatures

In-burden probes: CO, CO2, H2' temperature

(retractable and/or fixed probes)
Cooling losses: 6x8 panels in total
Shaft pressure: 2 columns, 6 positions
Blast: volume, pressure, O2, temperature
A - Normal Operation B - Upset: Furnace Hangs! Hearth: refractory, shell, cooling water temperature
Rings of melting ore burden and coke No burden descent.
Slag: composition and flow
slits alternating in cohesive zone.
Iron: temperature, composition and flow

Figure 49: Overview of Blast Furnace instrumentation

7.8 Blast Furnace daily operational control

In this section the blast furnace daily operational control is discussed. The better
the consistency of the blast furnace input, the lower the need for adjustments in the
process. Ideally, a good consistency of the input allows the operator to "wait and
see". The need for daily operational control is a consequence of the variability of the
input and - sometimes- the equipment.

The process must be controlled continuously, which may require changes to be made
on a daily or even shift basis. The changes are aimed towards:
. Correct iron and slag composition. The burden and coke are adjusted to get
the correct chemical composition of the iron and slag. For the latter especially
the basicity of the slag is important because of its effect on hot metal sulphur.
Correct iron and slag composition also implies control of thermal level, since the
hot metal silicon is correlated with the hot metal temperature. So, there are daily
requirements for burden calculations with updated chemical analysis of the burden
C - Cause 0 - Burden Slips
components and actual burden, and frequent adjustments of the thermal level of
Cohesive material forms bridge Voidage collapses, severe disruption
the furnace. Adjusting the coke rate or the auxiliary reductant injection through the
structure. Since process continues, of layer structure. Gas flow impeded
tuyeres can achieve the latter.
voidage is created and increases. and gas jets along furnace wall.
Figure 48: Hanging and slipping: a mechanism of erratic burden descent
. Stable process control. Burden descent (as measured by the stock rods,
Figure 31, or pressure taps, Figure 50), blast furnace productivity and efficiency
are evaluated on the basis of hourly data. Raceway conditions (e.g. flame
temperature) are monitored or calculated. The total process overview gives an
indication whether or not adjustments are required. Pressure taps indicate whether

72 73
 ,...-
7 - The process: burden descent and gas flow control 7 - The process: burden descent and gas flow control

or not "short circuiting" of gas flow along the wall takes place. In stable periods the
layers of coke and ore can be followedwhen passing the pressure taps.
. Gas flow control. The subject of gas flow control is discussed in more detail below.

Figure 50: Pressure taps indicating the stability of the process, 24hr graphs.
The example shows stable (left) and unstable (right) operation, with short-circuiting
of gas flow encircled in red. (Courtesy: Siderar, Argentina)

Measurements and data required for daily gas flow control are shown i(1Figure
52 (next page). The gas flow through the furnace can be monitored with the help
of global top gas composition, top gas composition across the radius, heat losses
at the wall and gas flow along the wall. The latter can be measured with the short Figure 51: Presentation of process data in an operational furnace. The weekly graph gives an
in-burden probes: the probes measure the temperature about three metres below the overview from the stability and development of the process. From top downwards:
burden level up to 50 cm into the burden. If temperatures are low (under 100°G) the Tope - CO utilisation (%), skin flow temperature roC)and top temperature roC);
Flujo T - Total heat loss and sum offields (GJ/hr);
burden is not yet dry and more gas flow in wall area is required to increase the drying
Arrabio - Hot metal temperature roC) and silicon (%);
capacity at the wall. I
Viento - Blast volume (Nm3/min) and top pressure (bar)

If the furnace seems in need of an adjustment of the gas flow, a change to the
burden distribution can be considered. However, a thorough analysis of the actual
situation has to be made. For example, consider the situation whereby high central
temperatures are observed. If these high central temperatures are observed together
with low heat losses and low gas utilisation, then the central gas flow can be
considered to be too high. The appropriate action in this case would be to consider
changes to the burden distribution to decrease the central gas flow. If, on the other
hand, the high central temperatures are combined with a good gas utilisation and
good wall gas flow, then there is no need to change the layers of ore and coke.
The appropriate action in this scenario would be to consider working with lower gas
volume per tonne HM Le. with higher oxygen enrichment.

Note also, that the heat losses are very sensitive to the burden descent. Irregular
burden descent leads to gaps at the wall and high heat losses. So, if a furnace
is showing high heat losses, again, the cause should be investigated in detail
before adjusting burden distribution. For example, if a blast furnace is pushed to its
production limits and burden descent suffers due to the high pressure difference over Figure 52: Example of gas flow control. The radial gas distribution is measured with above
the burden, the solution of the high heat losses is to reduce production level (or gas burden probes, expressed as CO utilisation (7 day graph). The decreasing gas utilisation in the
volume) and not to adjust burden distribution. centre of the furnace (point 1 and 2, yellow and dark green) shows increased central working.

74 75
 ~

Chapter 8
Blast Furnace Productivity
and Efficiency

The production rate of a blast furnace is directly related to the amount of coke used in
front of the tuyeres in a stable situation. This is due to every charge of coke at the top
of the furnace bringing with it an amount of ore burden materials. In a stable situation
the hot metal is produced as soon as the coke is consumed. The productivity of a
blast furnace increases as less reductant is used per tonne hot metal. In the present
chapter the basics behind blast furnace productivity, the chemical reactions and
efficiency are discussed (see also Hartig et ai, 2000).

8.1 The raceway

8.1.1 Production rate

In the raceway hot gas is formed which melts the burden material and is used to
drive the chemical reactions in the furnace. Given a certain amount of coke and
coal used per tonne hot metal, the production rate of a blast furnace is determined
by the amount of oxygen blown through the tuyeres. The more oxygen that is blown
into the furnace, the more coke and coal are consumed and form carbon monoxide
(CO), and the higher the production rate becomes. In addition, the lower the
reductant requirement per tonne of hot metal (tHM), the higher the production rate. A
quantitative example is indicated below. Coke (and coal) are not only gasified in front
of the tuyeres, but are also used for carburisation of iron (hot metal contains 4.5%
C) and for direct reduction reactions (section 8.2). The coke rate is expressed as
standard coke, i.e. coke with a carbon content of 87.5%.

In an operating blast furnace the use of the reductants can be as follows:

Coke rate 300 kg/tHM 300 kg/tHM
Coal injection 200 kg/tHM replacement ratio 0.85 170 kg/tHM
Total (as standard coke) 470 kg/tHM

1.- 77
 r- 8 - Blast Furnace Productivity and Efficiency
8 - Blast Furnace Productivity and Efficiency

Total (as standard coke) 470 kg/tHM 8.1.3 Raceway flame temperature
Required for carburisation 50 kg/tHM
Required for direct reduction 100 kg/tHM The flame temperature in the raceway is the temperature that the raceway gas
Gasifies in front of the tuyeres 320 kg/tHM reaches as soon as all carbon, oxygen and water have been converted to CO and H2.
The flame temperature is a theoretical concept, since not all reactions are completed
The 320 kg/tHM standard coke which is used in front of the tuyeres consists of 170 in the raceway. From a theoretical point of view it should be calculated from a heat
kg/tHM coke equivalent injected as coal and so per tonne hot metal, 150 kg coke balance calculation over the raceway. For practical purposes linear formulas have
(320-170 kg) is gasified at the tuyeres. been derived. Flame temperature is normally in the range of 2,000 to 2,300°C and is
Note the issue of efficiency: if the same amount of oxygen is blown into the furnace, influenced by the raceway conditions. The flame temperature increases if:
thus maintaining same blast volume and blast conditions, while the reductant rate is . Hot blast temperature increases.
10 kg/tHM lower, the production rate will increase. At a 10 kg/tHM lower reductant
rate the production will increase by 3% (320/310-100%)! Conversely, if extra coal
. Oxygen percentage in blast increases.

is put on the furnace for thermal control, the production rate will decrease if blast The flame temperature decreases, if:
conditions are maintained. This is a simplified approach. Secondary effects, like . Moisture increases in the blast.
the effect on gas flow throughput, the effect on flame temperature and the oxygen
content of the coal, have been neglected.
. Reductant injection rate increases, since cold reductants are gasified instead of
hot coke. The precise effect depends also on auxiliary reductant composition.

8.1.2 Bosh gas composition Table 21 gives some basic rules with respect to flame temperature effects.

Table 21.. Flame temperature effects, rules of thumb

The heat of the blast and the heat generated by the reactions of coke (an.d coal/
auxiliary reductants) in the raceway are used to melt the burden. The heat available Unit Cl1ange flame temp. Top temp.
I!
(°C) (°C)
to melt the burden depends on the amount of gas produced and on the flame
temperature, known as the "raceway adiabatic flame temperature" (RAFT). Blast 'c + 100 + 65 -15
temp.
The amount and composition of the raceway gas can be calculated using the Coal + 10 30 +9
kglt

--
following reactions that take place in the raceway:
2C+ O2 2CO
H2O + C CO + H2
I
Oxygen
Moisture
%

g/m3 STP
+
+
1
10
+ 45
50
-15
+9

In and directly after the raceway all oxygen is converted to carbon monoxide and all The top gas temperature is governed by the amount of gas needed in the process;
water is converted to hydrogen and carbon monoxide. the less gas is used, the lower the top gas temperature and vice versa. Less gas per
ton hot metal results in less gas for heating and drying the burden.
Consider the following example; the blast furnace in section 2.3 has a blast volume
of 6,500 m3 STP wind with 25.6% oxygen. Ignoring the effects of moisture in the blast 8.2 Carbon and iron oxides
and the coal injection, what would be the raceway gas volume and composition?
In the preceding section the formation of gas in the raceway has been described.
Blast into the furnace (per minute):
.. Oxygen:
Nitrogen: 4,836 m3 STP/min
1,664m3STP/min
((1-Q.256)x6,500)
(0.256x6,500)
What happens with the gas when it ascends through the furnace and cools down?
First consider what happens with the carbon monoxide.

The oxygen generates two molecules of CO for every O2 molecule, so the gas
volume is 8,164 m3 STP/min (4,836+2x1 ,664). The gas consists of 59% nitrogen
. C + 1/2 O2 -
Carbon can give two types of oxides:
CO + heat (111 kJ/mole).
This reaction takes place in the raceway
(4,836/8,164) and 41% CO (2x1,664/8,164).
. C + O2 - CO2 + heat (391 kJ/mole).
This reaction does not take place in the raceway and is more typical in an area
The calculation can be extended to include the moisture in the blast and the injection
such as a power plant.
of coal (or other reductants). This is done in section 6.4.

78 79
 .,.....
8 - Blast Furnace Productivityand Efficiency 8 - Blast Furnace Productivity and Efficiency

Note that in the second step much more heat is generated than in the first step,

---
The reactions can be indicated as below
therefore, it is worthwhile to convert CO to CO2 as much as possible in the process. 2 FeOo.5+ CO 2 Fe + CO2
The ratio CO2/ (CO+CO2) is called the gas utilisation or gas efficiency and is used + C~+C 2CO
extensively in blast furnace operation.
Total 2 FeOo.5+ C 2 Fe + CO
In Figure 53, the equilibrium between CO .. C + CO2 is presented for various The direct reduction reaction requires an enormous amount of heat, which is also
temperatures. The line indicates the equilibrium of the "Boudouard" reactions.
provided by the specific heat contained in the hot raceway gas.
At temperatures above 1,1 OO°C all CO2 is converted to CO, if in contact with coke.
Therefore, at the high temperatures in the bosh and melting zone of the blast furnace The direct reduction reaction is very important for understanding the process. In a
only carbon monoxide is present.
modern blast furnace the direct reduction removes about a third of the oxygen from
At temperatures below 500°C all CO has the tendency to decompose into C+CO2. the burden, leaving the remaining two-thirds to be removed by the gas reduction
The carbon formed in this way is very fine and is called "Boudouard" carbon. reaction. The amount of oxygen to be removed at high temperatures, as soon as
the burden starts to melt, is very much dependent on the efficiency of the reduction
In operational practice the carbon monoxide decomposition can be observed in
processes in the shaft. See section 8.2.2.
refractory material, where there is a CO containing atmosphere in the correct
temperature region. This generally is a very slow process. Note two important observations:

%CO %CO2 CO2

..Direct reduction uses carbon (coke) and generates extra CO-gas.
Direct reduction costs a lot of energy.
in gas in gas CO+CO2
0 50 100 In operational practice the direct reduction can be monitored. In many blast furnaces
the direct reduction rate (the percentage of the oxygen removed from the burden by
10 40 direct reduction) or the solution loss (the amount of coke used for the reaction) are
80 calculated on line. Experienced operators are well aware that as soon as the direct
reduction rate or the solution loss increases, the blast furnace starts to descend
20 30 faster, the cohesive zone will come down as the coke below it is consumed. And the
60
CO furnace will chill. When properly observed, chilling can be prevented, for example by
CP2
30 20 using extra coal injection.
40
8.2.2 Gas reduction or "indirect" reduction
40 10 20
As soon as temperatures of the gas reduce, the CO2 becomes stable and reduction
50 0
400 600 800 1000 1200
0
reactions can take place, such as:
Hematite: 6 Fe2O3+ 2 CO --
4 Fe3O4+ 2 CO2 Generates heat 8 kj/mole Fe

Figure 53: Boudouard

Temperature (°G)

reaction: the drawn line indicates equilibrium

Magnetite: 4 Fe304 + 4 CO
Wustite: 6 FeO + 3 CO -
12 FeO + 4 CO2 Requires heat 5 kj/mole Fe
6 FeOo.5+ 3 CO2 Requires heat 6.5 kj/mole Fe

The reduction is called "gas reduction" because the oxygen is removed from the
8.2.1 Direct reduction burden materials with CO gas. H2 reacts in a similar way. In literature it is also often
called "indirect" reduction, since carbon is only indirectly involved in this reaction. The
As the hot reducing gases produced in the raceway ascend through the lower reduction of the FeOo5 takes place via the direct reduction.
furnace, they transfer heat to the ore burden to the extent that it becomes molten at
the lower levels of the melting zone. They also remove oxygen from the iron oxides, Following the burden descent from the stockline, the reduction from hematite to
i.e. they reduce the iron oxides, which contain approximately one oxygen for every magnetite starts around 500°C. The reduction from magnetite to wustite takes place
two iron atoms. The CO2 produced from the reaction immediately reacts with the in the temperature zone from 600 to 900°C, while the reduction from wustite to iron
carbon in the coke to produce CO. The total reaction is known as direct reduction, takes place in the temperature region between 900 and 1,1 OO°C. At the start of
because carbon is directly consumed. melting (1,100 to 1150°C) FeOo.5 is normally reached. Here FeO is used as a symbol
for wustite, however the most stable composition is FeOo.95'

80
81
 8 - Blast Furnace Productivity and Efficiency
~ 8 - Blast Furnace Productivity and Efficiency

The reactions are shown in Figure 54. percentage CO. Gas utilisation should be below 30%. If CO2 is higher, wustite is no
longer converted to iron. From these measurements it is clear that the reduction from

+
.. .. .."". 0 0 wustite to iron comes close to equilibrium.

.t. ...
Carbon Monoxide e"" Hematite (Fe203)
I Gas The progress of the reduction reactions in a blast furnace can be detected in two
. 0 e different ways:
.
I +...' ""I"" I
Reduction

Gas
Magnetite (Fe304)
Burden: from quenched furnaces an overview of the progress of the reduction can
be derived. An example is shown in Figure 56

+ e. e"" e"" . Reduction 0 . Gas: by sending gas sampling devices down into the furnace, the progress of

CarboilPioxit!e
... . . . . . . I Gas Wustite (FeO)
temperature/gas composition can be derived. Figure 57 (next page) shows typical
results from a gas sampling exercise. The data can be depicted in the graph of
. Reduction
+. e .. .e . e. 0 e the equilibrium between gas and iron oxides. The gas normally shows a "thermal
reserve zone", that is, a zone in which the temperature does not change rapidly
Carbon e I Direct FeOy, as well as a "chemical reserve zone", a zone in which the chemical composition of

e . Reduction
+
. . .
e e 0
Fe
the gas does not change. The thermal reserve zone decreases and can disappear
when the furnace is pushed to high productivities.

Figure 54: Overview of the reduction of iron oxides

%CO %CO2 CO2.

in gas in gas CO+CO2
0 50 100

10 40 Magnetite 80

20 30 60
Wustite

30 20 40

40 10 Iron 20

50 0 0
400 600 800 1000 1200
Temperature (°C)
- 0.75
- 0.50
Figure 55: Schematic presentation of the relation between temperatures, CO/CO2 gas
composition and iron oxides. The drawn lines indicate equilibrium.

The equilibrium between the various iron oxides and the gas is shown in Figure 55.
~
The figure shows at what level of temperatures and gas compositions further gas Figure 56: Reduction progress in a quenched furnace (Hirohata, after Omori, 1987, p. 8)
reduction of the burden is no longer possible.
The reduction of wustite to iron requires the highest CO concentration, as can be
seen in Figure 55, the reduction of wustite requires a gas with a relatively high

82 83
 ........-
8 - Blast Furnace Productivity and Efficiency 8 - Blast Furnace Productivity and Efficiency

1500
There is a need of 10 atoms carbon per 3 atoms of Fe. So the carbon requirement is
'" 57 kmole (10/3x 17), which corresponds to 684 kg carbon (57x12). Again, the extra
:; 1000
45 kg carbon in iron has to be added giving a carbon rate of 729 kg/t and a coke rate
'"
"§

of 833 kg coke per tonne hot metal (729/0.875). This reaction has a poor coke rate
E
~
500
100
and a high heat excess.
0 80 Magnetite
The conclusion of the considerations above is, that the counter-current character of
the blast furnace works efficiently to reduce the reductant rate by combining direct
60 60
and gas reduction reaction. Approximately 60-70% of the oxygen is removed by gas
co
and the remaining oxygen is removed by direct reduction.
40 40
~w

8.2.4 Hydrogen
20 20 Iron

Hydrogen is formed from moisture in the blast and injectants in the raceway.
0 0
0 100 200 400 600 800 1000 1200
Hydrogen can act as a reducing agent to remove oxygen and form water. The
300
Time Temperature (OC)
reaction is comparable with that for carbon monoxide:
H2+ FeO ~ Fe + H2O.
Figure 57: Gas composition in operating furnace. CO, CO2, H2 and temperature were measured The major differences with the reactions for hydrogen and carbon monoxide are as
with descending probes (Chaigneau et aI, 2001). Typical measurements from various furnaces
follows:
are shaded. After McMaste~ 2002.
. Figure 57 shows the equilibrium of the iron oxides and hydrogen. Hydrogen
8.2.3 is more effective for the reduction at temperatures above 900°C. From
Gas reduction and direct reduction
measurements in the blast furnace it has been derived, that hydrogen reactions
are already completed at this temperature.
The direct reduction and gas reduction reaction combine to create a very efficient
process. Suppose that all oxygen is removed by direct reduction. Then, the following
. Hydrogen utilisation as measured from the top gas is normally around 40% while
CO utilisation is close to 50%.
reaction takes place:
Fe2O3 + 3C 2Fe + 3 CO (requires 2.6 MJ per tonne HM).
~

cO2 H2Oprocess
%CO %CO2
in gas in gas CO+CO2 or (H2+H2Oprocess)
Iron contains about 945 kg Fe per tonne hot metal. Coke contains about 87.5%
carbon. Atomic weights of Fe and Care 55.6 and 12 respectively. A tonne of iron 0 50 100

contains 17 kmole (945/55.6). For every atom of iron we need 1.5 atoms of carbon,
so the carbon requirement is 25.5 kmole (1.5x17), which is 306 kg carbon (25.5x12). 10 40 Fe304 80
In addition, about 45 kg carbon is dissolved in iron. In total, 351 kg carbon is used per
tonne hot metal, which corresponds to 401 kg of coke. This is a very low equivalent
coke rate and a blast furnace will not work, because the heat generated in this 20 30 60
reaction is too low.

30 20 40
Now consider that all reduction reactions are done via the gas reduction, what coke
rate is required in this situation? It is assumed that coke combustion generates the
CO required. The reaction is: . 40 10 20
3 FeO + 3 CO 3Fe + 3 CO2,
~

We only consider the reduction of wustite since the resulting gas is powerful enough 50 0 0
to reduce magnetite and hematite. We know from the above (Figure 55) that for gas 400 600 800 1000 1200
reduction the maximum gas utilisation is 30%. To get 30% gas utilisation more CO is
needed and the reaction becomes: Temperature (°C)
3 FeO + 10 CO 3 Fe + 3 CO2 + 7 CO
~ (gas utilisation = 3/(3+7) = 30%) Figure 58: Equilibrium iron oxides with hydrogen
So the coke requirement is calculated as above: every tonne iron contains 17 kmole.

84 85
 ..,.....
.
8 - Blast Furnace Productivity and Efficiency 8 - Blast Furnace Productivity and Efficiency

Note that the hydrogen utilisation cannot be measured. The H2O formed in the 8.4 Circumferential symmetry and direct reduction
process cannot be discriminated from the water put in the furnace with coke and
burden moisture.
High performance operation of a blast furnace requires that the complete
The hydrogen utilisation of the top gas is defined as H2/(H2+H2Oprocess)'
circumference of the furnace contributes equally to the process. A furnace can be
When working at high hydrogen input (via moisture, natural gas, coal), the
divided into sectors in which every tuyere forms one sector. See Figure 30 for an
competition between the reduction reactions will lead to lower top gas CO2 utilisation.
example.
The simple reasoning is, that H2 competes with CO. All oxygen taken by H2 is not
If all sectors do not contribute equally to the process, asymmetry in the melting zone
taken by CO2 and thus CO increases and CO2 decreases. 1% extra H2 in topgas will
will arise, as shown in Figure 60. Local heat shortages will drive the melting zone
lead to 0.6% extra H2Oprocess in top gas and thus to a 0.6% lower CO2 and a 0.6%
downwards in certain sectors and upwards in other sectors. This can result in an
higher CO percentage. 1% extra topgas hydrogen leads to a decrease in topgas
increase in direct reduction in some sectors. Increasing the thermal level of the entire
CO-utilisation of 1.3%, e.g from 49% to 47.7%.
furnace affecting its overall efficiency can only compensate for the effect and not
If a more advanced model is used and the efficiency of the furnace is kept constant resolve it.
at the so called FeO level, a 1% increase in topgas hydrogen leads to a decrease of
0.8% in topgas CO-utilisation.

8.3 Temperature profile

The temperature profile and the chemical reactions in a blast furnace are closely
related. It is summarised in Figure 59. The reduction of the oxides to wustite takes
place at temperatures between 800 and 900 °C. Thereafter, in the temperature range
of 900 to 1,1 OO°C, the wustite can be further reduced indirectly without interference
from the Boudouard reaction. This chemical preparation zone can take up to 50 to
60% of the height of the furnace and has a relatively constant temperature. This
region is called the thermal reserve zone.

Hematite
(Fe203)*.
-. -.*. -...
Magnetite
(Fe304)*'
- .-* -"-
*,.
Wustite(FeO) -'II..

. . - .. I ~900to
1100oe
Figure 60: Asymmetric melting zone

FeOY2
- . -. / Asymmetry in the process can arise from various sources.
// ..
-.-. -.
From uneven coal injection. Especially tuyeres without PCI (section 5.6).
lover
1100oe Blast distribution: if the blast speed is too low (under 100 m/s), tuyeres will not
Fe
efficiently function as blast distributors. This can be observed especially at
the tuyeres opposite the inlet between hot blast main and bustle main. Blast
Figure 59: Progress of the reduction reactions and temperature of the burden distribution can also be effected by plugged tuyeres (above a taphole or refractory
hot spots) and slag deposits in the tuyere.
. By asymmetry of the charging. With a bell-less top this can be prevented by
alternating the coke and ore top bins and by changing the rotational direction of the
chute. With a double bell system it is possible to alternate the last skip in a dump.
Note that the changes have to be made on a time scale smaller than the blast
furnace process i.e. more frequent than every six hours.

86 87
 """"""
8 - Blast Furnace Productivity and Efficiency

. Worn refractory or armouring plates at the top of the furnace.

. Deviation of furnace centre line from vertical line. This is especially a concern in
older furnaces.

Measures to correct for deviations of circumferential symmetry are available, such as

removing PCI injection from specific tuyeres. However, it is preferred to eliminate the
causes of the circumferential asymmetry instead of correcting for it.

Asymmetry in the gas flow can be derived from the radial heat loss distribution. In the
figure below, the heat losses are measured in eight segments of the furnace over four
Chapter 9
vertical sections. Extended asymmetry can be investigated with the help of this type
of data and graphs. Hot Metal and Slag

Typical hot metal and slag compositions are given in Table 22. Hot metal leaves the
furnace with a temperature typically in the range between 1,480 and 1,520 DC.

Table 22: Typical hot metal and slag composition

J+J6frnefal

Iron Fe 94.5% CaD 40% 34--42%

Carbon C 4.5% MgO 10% 6-12%

Silicon Si 0.40% Si02 36% 28-38%

AI2O3 10% 8-20%

Manganese Mn 0.30%

Sulphur S 0.03% Sum 96%

Phosphorous P 0.07% Sulphur 1%

9.1 Hot metal and the steel plant

Figure 61: 24 hrs heat loss distribution (blue). Note a slight process asymmetry
One day graph of eight sections, four levels. Hot metal is used for the production of steel. In a steel plant the hot metal is refined
so that the (chemical) composition can be adjusted to the metallurgical requirements.
The refining process is usually achieved in two steps:
. Removal of sulphur from the hot metal by means of desulphurisation. In most
cases the sulphur is removed with carbide and lime (stone) or magnesium,
according to:
. 2 CaO + 2 [S] +CaC2~ 2 (CaS) + CO (gas) or Mg + [S] ~ (MgS).
(Square brackets, i.e. [S], show that material is dissolved in the hot metal. Round
brackets, i.e. (CaS), show material dissolved in slag.)

88 89
 .."....-
9 - Hot Metal and Slag 9 - Hot Metal and Slag

. Removal of carbon, silicon, manganese and phosphorous. These elements react 9.3 Silicon reduction
with the oxygen blown into the converter. The "affinity" for oxygen decreases in the
sequence Si>Mn>C>P>Fe. In this sequence material is refined in the converter Silicon, manganese and phosphorous oxides are reduced via the direct reduction
process. At the end of the refining process iron can be reoxidised, which is reaction. Out of these three, the silicon reactions are of particular interest. The hot
sometimes required to heat up the steel before casting. Si, Mn, P and FeO are metal silicon is a sensitive indicator of the thermal state of the furnace, and the silicon
removed with the slag phase, the C as CO or CO2 in the gas phase. variation can be used to analyse the consistency of the process. For these reasons
the silicon reactions are discussed in more detail. The reduction of silicon takes place
The important considerations for a steel plant are: via three steps (Figure 62):
. Consistent quality: the control of the converter process incorporates "learning", . Formation of gaseous SiO in the raceway. The first reduction step takes place at
which adjustments to the process settings are necessary on the basis of expected the very high flame temperatures of the raceway. The silicon comes from the ash
outcome versus the actual outcome. The more consistent the iron quality, the of the coke (and coal). The higher the coke ash, the higher the silicon in hot metal.

.
better the results in the steel plant.
Hot metal silicon, manganese, titanium and temperature are important energy
. contact with the iron can be reduced as follows: SiO + [C]
Further reduction by means of direct reduction with the iron. The SiO gas in
~ [Si] + CO (square
sources for the converter process and effect the slag formation. brackets indicate solution in iron).
. Hot metal phosphorous has a major influence on steel production process. In the . The more intimate the contact between iron and gas, the higher the hot metal
blast furnace 97 to 98% of the phosphorous leaves the furnace with the hot metal. silicon content. The higher the height that the iron drips down, the greater is the
. Hot metal sulphur is a problem because sulphur is difficult to remove in the contact between the hot gasses and the liquid metal, leading to higher hot metal
converter process. For high grades of steel a maximum sulphur level of 0.008% is temperatures. The longer contact allows more SiO gas to react with the carbon in
required, while the blast furnace produces hot metal with a content of 0.030% and the hot metal, leading to higher hot metal silicon content. Therefore, a high-located
higher. Therefore, an external desulphurisation step is often required. melting zone corresponds with high hot metal temperature and high hot metal
silicon.
9.2 Hot metal composition

The final hot metal composition is the result of a complex process of iron-slag
interactions as the various elements are divided ov~r the slag and iron phases. The
dispersion of an element over the two phases depends on the slag and hot metal
composition as well as temperatura, as discussed below. As an illustration the typical
percentages of elements entering the slag and iron phases are indicated in Table 23.

The following points should be noted:

. Silicon, titanium and sulphur are concentrated in the slag.
. Manganese is concentrated in the hot metal.
. Some of the potassium is discharged from the top.
. Nearly all the phosphorous goes to the hot metal.

Table 23: Typical distributions of selected elements over iron and slag

%
FinalSi in iron
Silicon 46 5 11% 41 89% I
6 4.5 75% 1.5 25%
Figure 62: Reactions of silicon in the blast furnace
Manganese
Titanium 3 0.7 23% 2.3 77% . The hot metal silicon is in equilibrium with the slag. Important aspects are:
Sulphur 3 0.3 10% 2.7 90% . When iron droplets descend and pass through the slag layer, the silicon can be
Phosphorous 0.5 0.48 96% 0 0% reoxidised if FeO is present in the slag, according to:
Potassium 0.15 0 0% 0.11 73%
. [Si] + 2 (FeO) + 2 [C] ~ (SiO2)+ 2[Fe] + 2 CO.
The more basic the slag (less SiO2 in slag), the lower the hot metal silicon.

90 91
 .." ~

9 - Hot Metal and Slag 9 - Hot Metal and Slag

. The hot metal formed in the centre has high silicon, while the hot metal formed
at the wall has low hot metal silicon. The cast result is an average value.
Table 25: Typical slag compositions

Typical Range
CaD 40% 34-42%
Hot metal silicon and manganese are both indicators of the thermal state of the 6-12%
MgO 10%
furnace. Manganese shows a quicker response on process changes due to the fact
Si02 36% 28-38%
that the equilibrium with the remaining slag in the furnace is faster for manganese
due to the smaller fraction of manganese in the slag. AI2O3 10% 8-20%

Total 96% 96%

9.4 Hot metal sulphur

The hot metal sulphur is governed by the input of sulphur, the slag composition and 9.5.2 Slag properties
the thermal state of the furnace. The most important parameters are:
. 8ulphur input: the sulphur input is typically 2.5 to 3.5 kg/tHM. The main sources Slag has much higher melting temperatures than iron. In practice it is more correct
to think in temperature ranges than in melting points, as composite slags have a
being coke and the auxiliary reductant such as coal or oil.
. The division of sulphur between iron and slag, indicated by the (8)/[8] ratio. This melting trajectory rather than a melting point. At the solidus temperature the ore
burden starts melting.The liquidus temperature is the temperature at which the slag
ratio is very sensitive to the slag basicity and the thermal level of the furnace (hot
metal temperature or hot metal silicon). is completely molten. At temperatures below the liquidus temperature solid crystals
. The slag volume: the lower the slag volume per tonne hot metal, the higher the hot are present. These solid crystals increase the viscosity of the slag. In our experience
metal sulphur at the same (8)/[8]. the behaviour of slag can be well understood on the basis of its liquidus temperature.
Liquidus temperatures are presented in ternary diagrams as shown in Figure 63.
Most companies have their own correlations between (8)/[8] and the slag basicity These diagrams have been developed for pure components and in practice the
and thermal level. The correlations are derived on the basis of historical data for a liquidus temperatures are somewhat lower. 8ince in the ternary diagrams only three

..
blast furnace. As a basic guide: to reduce hot metal sulphur by 5%:
reduce input by 5%.
Increase basicitiy by 0,02 (basicity defined as CaO+MgO/8iO2) or
components can be indicated, one of the major slag components is taken as fixed.
i.e. AI2O3 content is 10%. Diagrams at different AI2O3 percentages are presented
in Figure 64 (next page). The typical slag composition for a blast furnace slag is
. Increase hot metal silicon by 0.06%.
I
also indicated (Table 24). Note that the liquidus temperature is above 1,400 °C and
that the liquidus temperature increases when CaO increases (i.e. when the basicity
9.5 Slag increases).
510,

9.5.1 Slag composition and basicity

10'" "¥'>
81ag is formed from the gangue material of the burden and the ash of the coke
and auxiliary reductants. During the process primary slag develops to a final
slag. Composition ranges are presented in Table 25 on the next page. Four major
components make up about 96% of the slag, these being 8iO2' MgO, CaO and
A12O3'The balance is made up of components such as manganese (MnO), sulphur
(8), titanium (TiO2) , potassium (K2O), sodium (Na2O) and phosphorous (P). These
components have a tendency to lower the liquidus temperature of the slag. The
definitions of basicity are given in Table 24.

Table 24: Definitions of basicity (weight percentage)

82 CaO/Si02
83 CaO+MgO/Si02 CoO ..,---,0 '" ~ ~ - "'"

84 (CaO+MgO)/ Figure 63: Phase diagram of liquidus temperatures of blast furnace slag system for 10% AI203'
(Si02+AI2O3) The slag composition CaO=40%, MgO= 10% and 8i02= 36% is also indicated. To this end the
components have to be recalculated from 96% to 100% of the slag. (After slag atlas, 1981.)

92 93
 9 - Hot Metal and Slag 9 - Hot Metal and Slag

SiO2
SiO2 1300°C .
I 5% AI2O3 110%A12031
- BF Slag

1
- - - 1400°C CD Olivine fluxed pellets
1500°C @ Typical basic pellets
1600°C @) Superfluxed sinter
1800°C @ Fluxed sinter

CaO 10 20 30 40 50 60 70 80 MgO CaO 10 20 30 40 50 60 70 80MgO

CaO 10 20 30 40 50 60 70 80 MgO
115% AI2O3 I
- 1300°C
Figure 65: The slag composition of typical pellets and sinter qualities
- - . 1400°C
-- - 1500°C
- -- 1600°C
1800°C

Primary slag fluxed by FeO

CaO 10 20 30 40 50 60 70 MgO
Figure 64: of slag liquidus temperatures at various AI;Oslevels. (After slag atlas, 1981.)
Final slag fluxed by SiO2

In Figure 65 (next page), the composition of the slag resulting from a burden of
self fluxed sinter and pellets is indicated. The liquidus temperatures of the "pure"
components give high liquidus temperatures for the slag, well above 1,500 DC. How
is it possible that the material melts in the cohesive zone?
Figure 66: Slag formation
The secret behind the melting of sinter and pellets is, that the ore burden contains a
lot of FeO, which lowers the melting temperature or, as mentioned earlier, lowers the 9.6 Hot metal and slag interactions: special situations
liquidus temperature and solidus temperature. The primary slag, i.e. the slag formed
during melting process and prior to solution of the coke ash components into the During special blast furnace situations like a blow-in or a very hot furnace the hot
slag, is made liquid due to dissolved FeO (Figure 19). The final slag is made liquid metal silicon can rise to very high values. Since the silicon in the hot metal is taken
through the solution of SiO2 as indicated in Figure 66 (next page). from the SiO2 in the slag, the consequence is that the basicity increases. This leads
to high slag liquidus temperature. In a situation with very high basicity the final slag is
no longer liquid in the furnace and cannot be cast. It will remain in the furnace where
it can form a ring of slag, particularly in the bosh region. Burden descent and casting
will be disrupted. Therefore, for special situations where hot metal silicon is expected
too be high, the slag should be designed to handle the high hot metal silicon. To this
end, extra SiO2 has to be brought into the furnace and the recommended method is
the use of siliceous lump ore.

94 95
 ........
9 - Hot Metal and Slag

Slag Volume 250 kg/t 240 kg/t 230 kg/t

Chapter 10
Casthouse Operation
SiO2

basicity (B3) 1.37 1.53 1.73

10.1 Objectives of casthouse operation
liquidus CC) 1420 1460 1528

Figure 67: Slag properties under special situations Excellent operational results from a blast furnace can only be reached if three
conditions are met:
Some companies use quartzite, which is suitable to correct the basicity in normal . Good burden and coke quality.
operation however, it is not suitable for chilled situations, since the liquidus
temperature of quartzite itself is very high (1 ,700°e). The effect of the use of a
..Good gas flow control.
Good casting of the furnace.
siliceous ore can also be shown in the ternary diagram in Figure 68: by working at a The present chapter deals with good casting practice.
lower basicity, the liquidus temperature decreases along the indicated line.
The liquid iron and slag collect in the furnace hearth well below the tuyeres. Iron and
SiO2 - 1300°C slag do not mix: slag has a lower specific weight (2.4 t/m3) than iron (7.2 t/m3) and
- - - 1400°C "floats" on the iron. The implication of this is that droplets of iron pass through a layer
of slag. Iron and slag come close to thermal and chemical equilibrium. A schematic
- --- 1500°C
presentation of a hearth and a taphole are presented in Figure 69 (next page). The
-- 1600°C taphole is a refractory construction and on the inside a refractory "mushroom" consists
- -- 1800°C of solidified taphole clay.
. BF Slag
A blast furnace is tapped 8 to 14 times per day through a taphole. The average
- LowBasicity
Burden duration of a cast is 90 to 180 minutes. In this time the furnace produces a
considerable part of its working volume. As shown in chapter 2, the residence time
of the burden is 5 to 6 hours. Therefore, for a two hour cast about one third of the
contents of the furnace is transformed into molten iron and slag.

Spraying of the liquids due to gas from the raceway escaping from the taphole
indicates the end of a cast.
CaO 10 20 30 40 50 60 70 80 MgO
Figure 68: Effect of low basicity burden on slag liquidus termperatures

96 97
 .
10 - Casthouse Operation 10 - Casthouse Operation

through, but diverts the slag to the slag runner. The slag is then usually granulated by
water or dumped into slag pits. The iron is collected into torpedoes. Two torpedoes
are located at each iron runner and can be filled using a tilting runner, which allows
the operator to exchange a torpedo during a cast.

BOF
Bird's Nest BOF

Slag
Granulation

Dead Man

Mushroom

Figure 69: Hearth BOF W0r Slag

Granulation

The liquid level in the hearth has two major effects on the blast furnace process: Figure 70: Example of a casthouse layout (three tapholes)
. The liquids in the hearth effect the burden descent: the higher the liquid level, the
stronger the upward force from the submerged coke (Figure 35).
. If slag reaches the level of the tuyeres and cannot be drained out through the
Skimmer
taphole, the gas flow is severely affected. This can result in poor reduction of the
burden and therefore a chilling furnace.
. Slag can be blown high up in the dead man, impeding normal gas distribution.
In order to prevent the effects, the hearth liquid level has to be kept under control and
preferably at a constant and low level (see De Pagter and Molenaar, 2001).
The objective of casthouse operation is to get the liquids from the furnace without
interfering with the blast furnace process. Trough Slag Runner Iron Runner
Note that the storage capacity of the hearth depends on the void fraction of the coke Figure 71: Iron/slag separation
in the hearth. The void fraction can be very low; with values of 25% void fraction
possible. In static areas of coke in the hearth the void fraction can be even lower. 10.3 Dry hearth practice
10.2 Casthouse layout Most high productivity furnaces cast through alternating tapholes, Le., when one
taphole is plugged, the other taphole is opened. With this practise the gap time
A modern blast furnace has at least two tapholes, with furnaces as big as 14 metres between the casts can be reduced to as low as zero, where continuous tapping
hearth diameter equipped with up to four tapholes. A schematic layout of a casthouse becomes possible. The advantage of the method is that the liquid level in the hearth
is shown in Figure 70, including a layout of the runner system. The iron is cast into can be kept consistently at a low level and interference with the process is avoided.
the main iron runner, or trough. Iron and slag can be separated easily because Be it by incident or by purpose, sometimes only one taphole is available and the
they do not mix due to their difference in specific weight. Figure 71 shows the iron liquids are cast via one taphole. In this situation there will be a gap time between
and slag flows through the main trough to a skimmer, which allows the iron to flow the casts of 20 to 45 minutes. This time is needed to clean the taphole area and for

98 99
 10 - Casthouse Operation 10 - Casthouse Operation

curing of the taphole clay. As a consequence of 1-taphole operation the liquid level 10.4 Opening and plugging the taphole
within the furnace will vary much more (section 10.5.1). Control of taphole length in
this situation is more difficult. There appears to be a maximum iron production that
can be cast with 1-taphole operation. It is estimated to be 6,000 t/d for an 11 metre 10.4.1 Iron and slag flow
furnace and 8000 t/d for a 14 metre furnace.
The iron and slag flow from a taphole are determined by the characteristics of the
The "ideal" casthouse operation for a big, high productivity furnace is continuous, taphole i.e.:
alternating casts with very similar cast times (10 casts/day) and almost continuous . The length of the taphole, which is affected by the way of plugging and clay quality.
slag flow. Slag coverage should be more than 95% of the time the iron is cast. See . The diameter of the taphole and especially the wear of the taphole over time.
Figure 72. . The roughness of the inner surface of the taphole.
. pressure.
The pressure inside, consisting of the blast pressure and the "hydrostatic"

Taphole1
Initially, during a cast, the liquid flow is lower than the production rate, thus the
Taphole 2 liquid level in the hearth moves upwards. As soon as the slag is tapped, the taphole
starts to wear out and the liquid flow will be higher than the production rate. At
Taphole 3 the beginning of a cast only iron might be cast. Even with good casting there is a
variation in the hearth liquid level of up to one metre inside the furnace. The taphole
Taphole 4 clay quality determines the resistance to attack of slag and appropriate choice of clay
has to be made. This is not discussed here.
8 10 14 16 18 20 22 0 . 2
Time of day (hrs) 10.4.2 Opening the taphole
Figure 72: Casting presentation for a 4 taphole furance, where tapholes 1 and 3 are operated
alternately without gaps Two methods for opening the taphole are applied: drilling a taphole open and soaking
bar practice. With soaking bar practice a bar is put into the clay just after plugging
and is pulled out 20 to 45 minutes later in order to open the taphole. The advantage
is that opening the taphole is very fast and thought to be reliable. It is especially
suitable for single taphole operations. Many companies have tried the soaking bar
technique, but the majority uses the drilling method. The soaking bar technique has a
high impact on the taphole bricks and blocks.

Good drilling has to fulfil the following requirements:

(DID 3
>-
0) 0
-£
. The drill position should be reproducible and at the same position: the centre of the
taphole. To this end the drill should be automatically secured on the furnace.
~.§-2
gO)
- >
. Drill diameter: a selection of drill diameters should be available (40-60 mm). The
diameter of the hole should be constant over its complete length. To this end the
:§.8 1 drill should be guided and the forward force, the rotating speed and hammering
~ ~
l! E should be properly controlled. Wear of the drill bit should be minimised, which can
~ 0
0 60 120 180 be achieved by using nitrogen or nitrogen/water for purging and cooling during
240 300 360
drilling.
minutes . The length of the taphole: drilling directly into the hot metal bath should open the
Figure 73: Hearth liquid level variation (in meter) for alternating casts. furnace. It is counterproductive to try to save a drill bit and rod by stopping drilling
earlier.
Alternating casts and 1-taphole operation give different variations in the liquid level in
the hearth, as shown in Figure 73. In some situations drilling cannot open the taphole. In this case, manual oxygen
lancing is applied. Lancing should be the last resort, since reproducibility is poor and
there are risks for taphole damage and break out.

100 101
 10 - Casthouse Operation 10 - Casthouse Operation

10.4.3 Plugging the taphole and taphole length 10.5.2 Casthouse delays

Clay quality and proper plugging are very important for the length of the taphole and The consequences of casthouse delays can be very severe.
for the flow rates of iron and slag. Plugging has to meet the following requirements: First, if starting the cast is delayed-say delay times of 45 to 75 minutes-the
. Plugging has to be done on the same position as the drill in order to avoid clay burden descent is effected: the higher upward force prevents the burden descending
spillage. Clay spillage may lead to a "secondary" tap stream. (Figure 35). The slower burden descent will be followed by a period with faster
. The speed of the piston and the pressure on the clay has a direct effect on the descent, during which there will be less heat available to melt the burden and the
proper injection of the clay into the hole. consequence is that hot metal temperature and silicon will decrease. From an
. When using tar-bonded clay the curing time should be sufficiently long to ensure operational point of view it is important to be able to keep the casthouse always
under control. It is recommended to decrease production level at an early stage to
proper clay conditioning. To this end the clay gun has to be positioned and
pressurised in front of the taphole for 15 to 30 minutes. prevent more severe effects.
. The length of the taphole is determined by the quantity of clay injected. Therefore,
Second, if the casting is delayed beyond the moment that slag reaches the
more clay is injected than required to fill the taphole. The excess clay is used to
form the mushroom at the inside of the taphole. The bigger the furnace, the longer tuyere level, the slag will be pushed inwards and will block part of the gas flow.
the taphole i.e. a 2.5 metre taphole length for an 11 metre furnace and 3 metres for Consequently, if the furnace was operating with central gas flow, the gas flow pattern
a 14 metre furnace. will be redistributed towards the wall. The moment when slag reaches the tuyeres is
. Gas tightness. After drilling the hole, the taphole clay should be sufficiently gas visible from a gradual increase of blast pressure/burden resistance or decrease of
tight to prevent spraying of the casts. Re-pressurising the taphole clay directly permeability index. In this situation the chilling of the furnace will be more severe than
after plugging might help to improve gas tightness. above.

10.5 Hearth liquid level and casthouse operation The very bad situation can arise where even the iron can reach the tuyere level. If
that is the case, iron will flow through the tuyeres into the blowpipes and a blowpipe
will burn causing an emergency stop of the furnace.
10.5.1 The effect of 1-side casting and alternating casts compared
10.6 Exercise
In Figure 75, the effect of the casthouse operation on hearth liquid level is compared
for alternate casts and 1-side casting. Since at 1-side casting there is a gap time of
Suppose the cast is dry and we cannot open the next cast. Suppose we continue
30 to 40 minutes between the casts, the hearth liquid level rises much more than in
the production at full blast. What is the time needed to fill the hearth from taphole to
a situation with alternate casts. The calculation made for 1-side casting shows, that
tuyeres?
the liquid level might reach 2 to 3 metres above the taphole. This has also an effect
Use the following data:
on the descending speed of the burden: it is not unusual, that at the end of a cast
. The furnace produces iron (6.7 t/min or 400 t/hr) and slag (100 t/hr).
the burden descent and thus charging rate increases. The reason is that the upward
force from the liquids in the hearth decreases, see Figure 35.
. Hot metal weighs 7.2 t/m3, slag weighs 2.4 t/m3.
. Voidage is 20%.
. Hearth diameter: 14 m.
. Distance taphole-tuyeres: 4.2 m.

Wk-v-J0 60 120 180

minutes
240 300 360
:~
0

Figure 75: Liquid level in hearth compared for alternating

60 120 180
minutes
240 300 360

casts (left) and 1-side casting (right)

102 103
 Chapter 11
Special Situations

11.1 Stops and start-ups

When a blast furnace in full operation is stopped, some of the processes continue.
While the blast is stopped, the direct reduction reactions within the furnace continue
as well as heat losses to the wall. The consequence is that the temperature of the
material in the melting zone is reduced to around 1,050 to 1,1 OooG, which is the start
of the carbon solution loss reaction. The decreasing temperature re-solidifies the
melting materials. Therefore, after a stop it takes some time for the burden to start
descending. The burden descent restarts as soon as the "old" melting zone is molten
(Figure 76).

Operating Furnace After Stop

1100°C 1100°C
t t

t
1400°C
Melting zone solidifies:
FeO+ C ~ Fe+ CO-heat

Figure 76: Solidified melting zone as consequence of a stop

105
 .......
11- Special Situations 11 - Special Situations

The heat shortage for a stop of a furnace operating with PCI is even worse: during 40
the stop procedure the coal injection is switched off from the furnace and during the
start-up it takes time to restart the PCI. An additional reductant shortage results. ;g
g",.. 30
c
0
:;:0
In addition, after a stop the hot metal silicon sometimes rises to very high values, '(ij 20
0
especially if during the stop/start procedure the furnace is operated at a low blast Co
E
volume. As shown in Figure 68, the basicity of the slag will be affected by the high 0 10
()
hot metal silicon and might even solidify within the furnace. This results in disturbed
burden descent. Heating up the slag is the only solution, which can be achieved by 0
charging coke blanks into the furnace.
0

So, in order to compensate for the heat losses during a stop and the risk for high hot
metal silicon, the following measures have to be applied:
. Extra reductant into the furnace. Coke as well as auxiliary reductants are possible.
~

.s
-10

Additional reductant is needed for a period that the furnace is not operated on PCI.
. Design slag composition for low basicity at high hot metal silicon. Use of a
siliceous lump ore is recommended.
~ OJ
I -20

Even if a stop is unplanned, taking these measures after the stop is worthwhile.
-30
For a blow-in after a stop major pitfalls are: 7
. Too fast blow-in. The solidified melting zone will take time to melt during the
start-up. If allowed time is insufficient, the pressure difference over the burden can
1 2 3 4
hours
5 6

increase too much, leading to gas escaping along the wall (high heat losses) and Figure 77: Typicalprogress of a blow--down
poor burden descent.
. Too fast restart of the PCI. Since the melting zone is solidified, there is a risk that
Moreover, generally H2 increases as a consequence of the (unavoidable) contact
solid agglomerates will block the hot blast through the tuyere. If this happens, the
of spraying water with the hot coke. At the end of the blow-down, when the level
coal will still be blown into the blowpipe where it can cause blowpipe failure. It is
of the coke is coming close to the tuyeres, the CO2 formed at the tuyeres has
recommended to restart coal injection only when the burden starts descending.
. Too high slag basicity.
insufficient opportunity to be transformed to CO and the CO2 percentage in the top
gas increases. As soon as half of the oxygen is in CO2 (Le. when the CO2 percentage
equals half the CO percentage), the furnace should be isolated from the gas system.
11.2 Blow-down
Normally, a blow-down takes 10 to 12 hours, after a preparatory stop, to reach the
Blowing down a blast furnace requires operating the furnace without simultaneous tuyere level.
charging of the furnace. Therefore, all material charged into the furnace is exposed to
the same temperatures and reduction processes as if the furnace was fully charged. Prior to the blow-down the furnace contains coke in the active coke zone and dead
man, and alternating layers of coke and ore in melting zone and stack zone. Since
However, since the temperature of the shaft gas is not transferred to the cold charge, during the blow down the coke of the active coke zone and dead man will be gasified,
the off-gas temperatures increases and the gas composition changes. Since the there is coke excess in the blast furnace. During the latter stages of the blow down
equipment has not been designed to withstand the high top gas temperatures, the reduction reactions have largely stopped, so any auxiliary reductant injection can be
top gas temperatures are kept under control by spraying water. The water sprayed stopped during the early stages of the blow down. The moment is indicated by the
above the burden should be prevented from reaching the burden surface, either gas analysis: as soon as the CO2 percentage starts to decrease to below 10%, then
directly via descent on top of the burden or indirectly via the wall. Special atomising there is little iron left to reduce.
nozzles are required and the success of the blow-down heavily depends on proper
spraying. The progress of the blow-down process can be measured from the burden The burden level in the furnace is difficult to measure with standard stock rods.
level as well as from the analysis of the top gas composition. Since less and less The stock rods have to be equipped with chain extensions and recalibrated for
oxygen is removed from the ore, the CO2 percentage decreases and CO percentage the purpose. The stock rods should be used only at intervals, since the high
increases (Figure 77, next page).

106 L ,107
 ........
11 - Special Situations 11- Special Situations

temperatures above the burden may cause chain breakage. Radar level indicators 0 In the early stages of a blow-in, blast temperature should be maximised and blast
can be used if reliable. Indications from the level of the burden can also be obtained moisture minimised.
from:
0 Heating up the hearth requires some 7 to 8 hours after the blow-in. Heat is
0 The pressure taps. generated from coke used at the tuyeres.
0 The casthouse operation i.e. the quantity of iron cast.
0 Calculation of the amount of coke consumed in front of the tuyeres. 11.3.2 Starting the reduction processes
The required condition of the furnace after the blow-down depends on the purpose During the early stages of the blow-in while the hearth is heating up, the reduction of
of the blow-down and consequent repair. Generally the walls have to be clean. the iron oxides has not yet begun due to the temperatures being too low. Therefore,
Cleaning of the hearth is another important topic. If solid skulls and scabs are one has to consider the increased amount of direct reduction. The situation may
expected in the hearth and have to be removed prior to the blow-down, the furnace become difficult if the level of direct reduction is too high, (and gas reduction is low).
can to be operated for a prolonged period on a high thermal level, relatively low This situation manifests itself from:
PCI rate and without titanium addition. The effect of these measures is, however, The gas utilisation.
uncertain. 0

The direct reduction, as manifest from CO+CO2 exceeds "normal" values.

11.3 Blow-in from new
The gas utilisation is an indication of the amount of gas reduction taking place, while
the total CO and CO2 percentage is an indication for the direct reduction. Especially
Blowing in a furnace from new can be considered in two phases: the CO2 percentage indicates if gas reduction takes place.

Phase 1: Heating-up the hearth. 11.3.3 Slag formation

Phase 2:Starting the reduction reactions and iron production.
The two phases are discussed separately below. In general, the slag during blow-in has to be designed for high hot metal silicon.
However, with the proposed method the hot metal silicon should be under control.
11.3.1 Heating up the hearth If we continue to follow the "two-phase" blow-in approach mentioned here, during
the first phase of the blow-in about 350 tonne coke is gasified in 8 hours and the slag
Heat is generated by the reaction 9f coke carbon to CO. Coke generates 55 kJ per formed comes only from the coke ash. Taking 10% ash and 30% of the ash as A12O3'
mole carbon, when reacting to CO, which corresponds to 3.9 MJ/ t coke. we get during the first 8 hours 35 tonne of a high AI2O3 slag. This will not cause a
problem in the furnace because of the small volume. The coke ash can be diluted,
The heat requirement in the early stages of the blow-in is for the following: e.g. by using a high siliceous ore in the coke blank. In order to dilute to 20% A12O3'
0 Heat coke in the hearth, dead man and active coke zone to 1500DC. some 30 tonne of a siliceous ore has to be added to the 350 tonne coke blank.
0 Heat required for evaporation of moisture from the coke.
0 Heat required to compensate for moisture in blast dissociating into hydrogen gas 11 .3.4 Hot metal quality during blow-in
(H2O+ C CO + H2).
~

0 Heat to compensate for loss of heat through the wall. As soon as the hearth is heated the hot metal temperature exceeds 1400 DC. As
soon as the top temperature exceeds dew-point, all excess moisture has been
A detailed analysis of the heat requirement to fill the hearth, dead man and active removed from the furnace and the process has started. There is only limited heat
coke zone with coke of 1500DC indicates the following: required for heating -up and drying of refractories, if compared with the heat
0 Moisture in the coke can be neglected. requirements of the process itself. So as soon as hot metal temperature reaches
0 The heat required filling the hearth, dead man and active coke zone with hot coke 1400 DC and top temperature exceeds 90 DC,the process has to be brought back to
of 1500DC requires an amount of coke gasified to CO of about two-thirds of the normal operation conditions.
estimated volume of the hearth/dead man/active coke zone.
0 Additional heat requirement arises from the water dissociation reaction and the However, in this situation the coke rate in the furnace is still very high and the hot
heat losses through the wall. For example, if 300 tonne coke is required to fill metal silicon will rise to 4 to 5%. The hot metal silicon can be reduced by putting
hearth, dead man and active coke zone with coke, a coke blank is required with a a normal coke rate in the furnace. The "normal" coke rate at "all coke" operation
total weight of 600 tonne: 300 tonne to fill hearth, dead man and active coke zone is about 520 kg/tHM. In doing so, however, it takes considerable time to consume
with coke and 300 tonne for the generation of heat to bring the coke to 1500DC. all excess coke, which is present in the furnace. More rapid decrease of hot metal

108 109
 -..........-
11 - Special Situations

silicon can be reached, if a lower coke rate is charged and auxiliary injection is used
as soon as required. The injectant is switched on, as soon as the hot metal silicon
decreases below 1%.

An example of such a rapid blow-in of a furnace is presented in Figure 78. At the

blow-in the furnace was started-up with eight tuyeres (of 36). After opening all
tuyeres, a "heavy" burden (coke rate 450 kg/tHM) was put in the furnace 50 hours
after the blow-in and coal injection was put on the furnace 58 hours after the blow-in.
Hot metal silicon reached the 1% mark 60 hours after the blow-in. The 4th day after Annexes
the blow-in, average hot metal silicon was 0.95% and productivity was 2.1 t/m3
WV/d.

800 5

:'2 700 4
:;!;
C,
-"
c 600 3

500 2

0 400 1
() I .. Annex I Further reading
300 0
0 20 40 60 80 1. Biswas, AK.: Principles of Blast Furnace Iron making, Cootha Publishing House,
Hours after blow-in Brisbane, Australia, 1981.
Figure 78: Charged coke rate and hot metal silicon after blow-in
I 2. Committee on Reaction within Blast Furnace, Omori, Y. (chairman): Blast furnace
phenomena and modelling, Elsevier, London, 1987.
3. IISI website: worldsteel.org.
4. McMaster University: Blast Furnace Iron making Course (every 2 years),
Hamilton, Ontario, Canada, 2002.
5. Meyer, K.: Pelletizing of iron ores, Springer Verlag, Berlin, 1980.
6. Peacy, J.G. and Davenport, WG.: The iron blast furnace, Pergamon Press,
Oxford, UK, 1979.
7. Rist, A and Meysson, N.: A dual graphic representation of the blast-furnace
mass and heat balances, Iron making proceedings (1966), 88-98.
8. Rosenqvist, T: Principles of extractive metallurgy, McGrawHill, Singapore, 1983.
9. Schoppa, H.: Was der Hochofner von seiner arbeit wissen muss, Verlag
Stahleisen, Dusseldorf, Germany, 1992.
10. Turkdogan, ET (1984), Physicochemical aspects of reactions in iron making and
steelmaking processes, Transactions ISIJ, 24, 591-611.
11. Wakelin, D.H.: The making, shaping and treating of steel, 11th edition, AISE
Steel Foundation, 1999.
12. Walker, RD.: Modern Ironmaking Methods, Institute of Metals, London, UK,
1986.

110 111
 Annexes Annexes

Annex II References Annex III Answers to exercises

1. Biswas, A.K.: Principles of Blast Furnace Ironmaking, Cootha Publishing House, 1.4 Raw material flows
Brisbane, Australia, 1981.
2. Bonnekamp, H., Engel, K., Fix, W, Grebe, K. and Winzer, G.: The freezing 10.000 t/d gives about 3.5 million tonne per year (at an availability of 96%). So total
with nitrogen and dissection of Mannesmann's no 5 blast furnace. Ironmaking world production of 575.7 million ton can be produced in 165 furnaces. With three
proceedings, 1984, Chicago, USA, 139-150. of these furnaces per site, there would be a worldwide need for 55 of these large
3. Chaigneau,R., Bakker, 1., Steeghs, A. and Bergstrand, R.: Quality assessment ironmaking sites, each producing 10.5 million tonne.
of ferrous burden: Utopian dream? 60th lronmaking Conference Proceedings,
2000, Baltimore, 689-703. The resulting annual flow of materials would be:
4. Committee for Fundamental Metallurgy of the Verein Deutscher Eisenhuttenleute:
Slag atlas, Verlag Stahleisen, Dusseldorf, Germany, 1981.
..Steel output: 0.9x1 0.5 = 9.45 million tonnes steel products per year.
Pellets required 1.6x1 0.5 = 16.8 million tonnes.
5. Exter, P. den, Steeghs, A., Toxopeus, H. Godijn, R., Van der Vliet, C., Chaigneau, . Coal required:
R. and Timmer, R.: The formation of an ore free blast furnace center by bell Forcokemaking:10.5x0.3/0.75 = 4.2 million tonnes
charging, 56th Ironmaking Conference Proceedings, 1997, Vol 56, 531-537. For PCI: 10.5xO.2= 2.1 million tonnes
6. Geerdes,M., Van der Vliet, C., Driessen,J. and Toxopeus,H.: Control of high
productivity blast furnace by material distribution, 50th Ironmaking Conference 2.3.1 How much blast oxygen is used per tonne hot metal?
Proceedings, 1991, Vol 50, 367-378.
7. Geerdes, M., Toxopeus, H., Tijhuis, G. and Veltman, P.: Blast Furnace Operations
Oxygen from the blast volume amounts to 0.256x 6,500 m3 STP/min = 1,664 m3 STP
for Long Hearth Campaigns, 3rd IAS lronmaking conference, Buenos Aires,
Argentina, 2001, 97-105. oxygen/min. The production rate is 10,000/(24x60) = 6.94 tHM/min.
So the oxygen use is 1,664/6.94 = 240 m3 STP blast oxygen/tHM.
8. Grebe, K., Keddeinis,H. and Stricker, K.: Untersuchungen uber den
Niedrigtemperaturzerfall von Sinter, Stahl und Eisen, 100, (1980), 973-982.
9. Hartig, W, Langner, K., Lungen, H.B. and Stricker, K.P.: Measures for increasing 2.3.2 How often are the furnace contents replaced?
the productivity of blast furnace, 59th Ironmaking Conference Proceedings,
Pittsburgh, USA, 2000, vol 59, 3-16. To produce a tonne of hot metal, the furnace is charged with:
10. Kolijn, C.: International Cokemaking issues, 3rd McMaster Cokemaking Course, . 300 kg coke: 0.64 m3 (300/470)
McMaster University, Hamilton, Canada, 2001. . 1,580 kg sinter/pellets: 0.88 m3 (1,580/1,800)
11. Pagter, J. de and Molenaar, R.: Taphole experience at BF6 and BF7 of Corus . Total per tonne of hot metal: 1.52 m3
Strip Products IJmuiden, McMaster Ironmaking Conference 2001, Hamilton,
Canada. Production is 10,000 tonne per day, which is 10,000x1.52 m3 = 15,200 m3 per day.
12. Schoone, E.E., Toxopeus, H. and Vos, D.: Trials with a 100% pellet burden, 54th This material can be contained in the working volume of the furnace, with exception
Ironmaking Conference Proceedings, Nashville, USA, 1995, vol 54, 465-470. of the volume used for the active coke zone. So the contents of the furnace are
13. Toxopeus,H., Steeghs, A. and Van den Boer, J.: PCI at the start of the 21st refreshed 4.6 times per day (15,200/(3,800-500)). This means the burden charged at
century, 60th lronmaking Conference Proceedings, Baltimore, USA, 2001, vol 60, the top reaches the tuyeres in 5.2 hours.
736-742.
14. Vander, 1., Alvarez, R., Ferraro, M., Fohl, J., Hofherr, K., Huart, J., Mattila, E.,
2.3.3 How many layers of ore are in the furnace at any moment?
Propson, R., Willmers, R. and Van der Velden, B.: Coke quality improvement
possibilities and limitations, Proceedings of 3rd International Cokemaking
The number of ore layers depends on the layer thickness or the weight of one layer
Congress, Gent, Belgium, 1996, vol 3, 28-37. in the burden. It can vary from furnace to furnace and depends on the type of burden
used so there is a large variety of appropriate burden thicknesses. A typical range
is 90-95 tonne of burden per layer. A layer contains 94.8 tonne, so about 60 tonne
hot metal. In 5.2 hours, the furnace produces 2,167 tonne, which corresponds to 36
layers of ore (2,167/60). In our example, taking a throat diameter of 10m, the ore
layer is 67 cm and the coke layer is an average of 49 cm at the throat.

112 L 113
 Annexes Annexes

2.3.4 What happens to the carbon of the coke and coal? It is possible to make the corrections mentioned above. Take an average temperature
of the gas of 900°C and an average pressure of 4 bar, and then the effects are:
One tonne of HM requires:
. 300 kg coke, C content 87%:
. Increase in residence time owing to higher pressure: 4/1 = 4 times longer.
261 kg C . Decrease in residence time owing to higher temperature 273/(273+900)= 0.23
. 200 kg coal, C content 78%: 160kg C times shorter.
. Total carbon: 417 kg C . Decrease in residence time due to extra gas from direct reduction is 8,164/9,987 =
0.82 times shorter.
About 45 kg carbon dissolves in the hot metal. The balance leaves the furnace . In total, the residence time is shorter by a factor of 0.75 (4xO.23xO.82), so the
through the top, which is 421-45 = 372 kg. It leaves the furnace as CO and CO2, corrected residence time is 8xO.75 = 6 seconds.

2.3.5 How much top gas do we get? 2.3.7 If you get so much top gas, is there a strong wind in the
furnace?
Production per minute is 10,000/(24x60) = 6.94 tonne HM/min. For a tonne hot metal
6,500/6.94 = 936 m3 STP blast is used, which weighs 1,218 kg. The balance is No, at the tuyeres there are high wind velocities (over 200 m/sec), but top gas
presented in Table 26. The weight of the top gas is the known inputs minus the output volume is about 9,970 m3 STP/min. Over the diameter of the throat, at a gas
of iron and slag. Total top gas weight is twice the weight of the iron! temperature of 120°C and a top pressure of 2 bar, top gas velocity is 1,0 m/sec: on
the Beaufort scale this corresponds to a wind velocity of 1. Through the voids the
Table 26: Weight balance example velocity is about 3 m/s. Note, that in the centre the velocity can be much higher, so
Input In kg/tHM Output In kg/tHM that even fluidisation limits can be reached (See 7.4).
Blast 1,218 Iron 1,000
Coke 300 240
10.6 Hearth liquid level
Slag
Coal 200 Top gas 2,058 The production of liquid is:
Burden 1,580 . Iron: 6.7 t/min = 400 t/hr = 400/7.2 m3/hr = 89 m3/hr.
3,298 3,298 . Slag: 0.25x6.7 t/min = 100 t/hr = 100/2.4 m3/hr = 42 m3/hr
So in total, the liquid formed per hour (no casting) amounts to 131 m3/hr.

The raceway gas volume is 8,164 m3 STP/min (6,500x(1 +0.256)), and the top gas
volume is 9,987 m3 STP/min (2,058/1.43 = 1,439 m3/tHM). So the gas from direct
reduction reactions is approximately 1,800 m3 STP/min (9,987-8,164).
.
The amount of liquid that can be stored in the hearth:
The hearth surface is 3.14/4 x14x14 = 154 m2.
. Voidage is 20% and height 4.2 meter. Between taphole and tuyeres we can store
0.2x4.2x154 = 129 m3 STP of liquid, which is sufficient for 1 hour of production
2.3.6 Estimate how long the gas remains in the furnace without casting!

The blast volume is 6,500 m3 STP/min with 25.6% oxygen. Since for every
unit of oxygen two units of CO are produced, the raceway gas amounts to
6,500x(1 +0.256)=8,164 m3 STP. This gas has a higher temperature (decreasing from
some 2,200°C to 125°C top gas temperature), the furnace is operated at a higher
pressure (say 4.8 bar, absolute at the tuyeres and 3 bar, absolute at the top) and
extra gas is formed by the direct reduction reaction (see exercise 2.3.5). If all these
effects are neglected, the exercise is straightforward:
Suppose the void fraction in the burden is 30%, then the open volume in the furnace
is (3,800-1 OO)x0.30 = 1,100 m3 STp, through which 8,164 m3 STP gas is blown per
minute. So the residence time of the gas is (1,100/8, 164)x60 = 8 seconds.

114 115
 Annexes

Annex IV Rules of Thumb

Table 27: Rules of thumb for daily operation of the blast furnace process, a typical example
Unit Change Coke
Rate
(kglt)
Si % + 0.1 + 4
Moisture g/m3
STP
+ 10 + 6 Index
Top pressure bar + 0.1 - 1.2
Coal kglt + 10 - 9
Oil kglt + 10 - 11

Oxygen % + 1 + 1

Blast temperature °C +100 - 9

Slag kglt + 10 + 0.5

Cooling losses GJ/hr + 10 + 1.2

Gas Utilization % + 1 - 7 Alkali 47 Coke, production 36,37
Angles of repose 66 Coke, quality 34
Bird's nest 43 Coke, tests 35
Table 28: Rules of thumb for daily operation of the blast furnace process (constant blast volume) Blast furnace construction 7 Cold strength, ore burden 20
Blast furnace temperature profile 86 Compression of pellets 30
U Flam~temp.
CC) Blow-down 106 CRI 35
Blast temp. °C +100 + 65
Blow-in after reline 108 CSR 35
- 15
Blow-in after stop 105 Decrepitation 24
Coal kglt + 10 - 30 + 9
Bosh gas composition 78 Dolomite 27
Oxygen % + 1 + 45 - 15 Boudouard reactions 80 27
Fayalite
Moisture g/m3 STP + 10 - 50 + 9 Burden calculation 48 Fluidisation 66
Burden descent 55,71 Forces, vertical 57
Burden distribution 66 Forces, vertical, from coke in hearth 58
Burden distribution (fines) 65 Gas utilisation, calculation 52
Carburisation 34 Gas utilisation 15
Casthouse, 1 taphole operation 102 Hanging 57,72
Casthouse, dry hearth practice 99 Hardgrove index 41
Casthouse layout 98 Hardness 35
Channelling 65 Heat fluxes, through furnace wall 59
Char 42 Hematite 14,25
Coal injection, coal selection 41 Hot metal desulphurisation 89
Coal injection, equipment 39 Hot metal quality 90
Coal injection, gasification 42 Hot metal silicon 91, 96, 110
Coal injection, lances 42 Hot metal sulphur 92
Coal injection, oxygen enrichment 43 Hot metal/slag elemental distribution 90
Coal injection, replacement ratio 41 110 35
Cohesive zone, types of 61 140 35
Coke, degradation 37 Instrumentation 73

116
L 117
 Index

Limestone decarbonisation 27 Symmetry, circumferential 45,87

Lintel 7 Taphole 101
Liquidus temperature 93 Temperature, flame 79
Lump ore 32 Temperature profile 86
M10 35 Volume, definitions 8
M40 35 Wustite 14,25
Magnetite 14,25 Zinc 47
Mass balance 49
Melting capacity of raceway gas 43
Melting zone, see: cohesive zone
Mixed layer 68
Ore burden quality 24,31
Pellet production 29
Pellet quality 31
Permeability 19,59,64
Pressure taps 74
Production rate 77
Quenched furnaces 2,83
Raceway 77
Reducibility 24,31
Red uction-d isintegration 25
Reduction of iron oxides 14,81
Reduction of iron oxides,
direct and gas reduction 84
Reduction of iron oxides,
direct reduction 80
Reduction of iron oxides,
gas reduction 81
Reduction of iron oxides, hydrogen 85
Residence time, gas 114
Residence time, ore burden 113
Resistance, see: permeability
Rules of thumb for daily operations 116
Sinter, effect basicity on structure 25
Sinter production 21
Sinter quality 24
Sinter types 26
Slag, primary 95
Slag, properties 93
Slag basicity, definition 92
Slag basicity, special situations 95,109
Slipping 57,72
Solidus temperature 93
Solution loss 81
Soot 42
Spinel 27
Stops 105

118