SCHLUMBERGER OILFIELD REVIEW

Oilfield Review
Summer 1999

SUMMER 1999

Drilling Risk Management

VOLUME 11 NUMBER 2

Reservoir Model Validation

Real-Time, Wellsite Log Corrections
 Call or Copy and FAX this form.
Multimedia Oilfield Review Archive

Multimedia, graphics and

D
C
Platform: ❏ Macintosh ❏ Windows ❏ UNIX

E
IV
Computer Model

O A 98
H
reprints from
C W 9
ER C
K E -1
V R
Quantity Ordered @$25.00 each = Total $

C I 2
A V 9
B RE : 19
Corporate Express Document and Print Management

Oilfield
D LD LE 3635 Willow Bend, Suite 300

MORA
SI IE B

Houston, Texas 77054

IN F LA

☎1(713)660-7226 • Fax 1(713)660-0362

E OIL AI

Review
E
SE IA AV

E-mail: rmartin1@idt.net
ED OW

Multimedia
IM N

Oilfield Review MORA or Oilfield Review Services Order Form

Archive
Name
LT
U

Company
M

MORA Address
Your electronic journey into the pages of
Oilfield Review begins here. City State Postal Code
This archive provides quick and easy access
Country
Did you know that to Oilfield Review articles published from 1992
to 1998. Select a button to explore articles by
topic, issue or search by keyword.
Telephone

Oilfield Review can Fax E-mail

Method of Payment: ❏ MasterCard ❏ Visa ❏ Check
be found on-line?
About Topics Issues Help Search

Account Number-Please include all digits

Upper sands
Producing Injection Lower sands

Point your browser to

(2 wells)
well wells (3 wells)
M
L N
Production
water Steam K O

www.connect.slb.com to find: J
1.0 1.5
2.0
2.5
P
VAT number (EEC countries only)
0.5

• technical presentations, forums and Steamflooded

upper sands
I A

H B
Please print name as it appears on credit card
publications, including Oilfield Review Waterflooded
lower sands G C

Expiration Date
F D
E

• detailed service information,

applications and case studies A D H 100
90
Order Date Signature

% upper sand production, steamflood

E
80
F L I Before steamflood
B

• log interpretation charts and K

J
70

60
After steamflood

Statement Address (if different from above)

G

nonexclusive seismic maps C 50

40

30
0 ft 300

• an oilfield dictionary, mnemonics Oil from lower sands only

20

10
Oil from upper sands only

database and other oilfield links Commingled oils

0
A B C D E F G H
Well
I J K L

• location and telephone directory

• DataLink* secure data transfer Viewgraph and Slide sets of Oilfield Review Illustrations
• InterACT* real-time data transmission Available from Corporate Express/Ross-Martin Article title No. of sets Subtotal
from wellsites Multimedia (call for prices)
• SuperVision* project monitoring. MORA—the Multimedia Oilfield Review Archive—provides 1.
Connect Schlumberger is the Internet quick and easy access to issues published from 1992 to 1998. 2.
communications resource for Schlumberger MORA preserves the look of the printed magazine in a cross-
clients. Updated daily, this password-protected platform format. Full-color articles can be printed or explored 3.
interface is growing rapidly, providing timely on-screen using key-word searches and cross-document links. 4.
and interactive access to the information, data Specify: ❏ Viewgraphs: $16 each ❏ Slides: $14 each Total $
and tools needed by the E&P industry. Available from Oilfield Review Services:
You can also visit the public Color 35-mm slides and viewgraphs Oilfield Review Back Issues and Articles
Schlumberger Web site at The ideal way to prepare high-quality sales presentations and
www.slb.com. Back-issue price:$15. Call for article reprint price.
lectures—Oilfield Review Services offers color 35-mm slides
* Mark of Schlumberger and viewgraphs of illustrations from Oilfield Review and The Issue date Price Quantity Subtotal
Technical Review. The cost of a slide or viewgraph set ranges 1.
from $120 to $450. Please call for more information. 2.
Article reprints and back issues 3.
Reprints of Oilfield Review articles can be supplied. Minimum 4.
order is 500. Back issues of the Oilfield Review and The
Call for added cost of delivery. Total $
Technical Review can also be ordered, subject to availability.
Specify: ❏ Courier ❏ Air mail
Exhibition posters $
Special display materials such as posters or large transparencies, ORDER TOTAL
incorporating illustrations from Oilfield Review, can be ordered.

Oilfield Review Services also creates tailor-made multimedia

Oilfield Review
Oilfield Review Services
presentations for exhibitions, conferences and seminars, and can
Barbour Square, High Street, Tattenhall,
provide staff and equipment to run the presentations.
Chester CH3 9RF England
☎ (44) 1829-770569 • Fax: (44) 1829-771354
Browse Oilfield Review Services online at www.t-e-s.co.uk E-mail: orservices@t-e-s.co.uk

SMP-6068
Risk Management Improves Drilling Performance

Any context-setting discussion of the oil and gas industry Data acquisition at and ahead of bits—The
must surely begin with commodity prices. Capricious technologies of measurements-while-drilling (MWD)
in nature, oil and gas prices are highly variable and are now mature; logging ahead of drilling bits and
determined by forces beyond our control. In these seismic-while-drilling are being applied and developed
circumstances, attempts to predict future prices are further. Acquiring pore pressures while drilling is still to
futile. Our energies are better spent ensuring that both come, but the strength of constant data updates from
operator and service sector businesses remain profitable, just behind the bit and a rapid forecast ahead of the bit
even at the bottom of a market cycle. are the ways of the future. Even then, having real-time
That’s a tough challenge, particularly when the pursuit data is one thing; being able to synthesize the information
of additional reserves takes the industry into uncharted quickly enough to do something with it is another.
territory. Whether in new geography or new geology, we Drilling simulation—In high-cost wells, mistakes
constantly encounter conditions that test the limits of can be debilitating and expensive. It is better to address
existing know-how and technology; therefore, continuing to potential hazards beforehand in the office, rather than
explore while preserving precious profit margins demands waiting until problems occur on a rig. Although primarily
ongoing improvement in cost and capital efficiency. for training, drilling simulators can be used for testing
A well-known case in point is the deepwater Gulf of alternatives and contingencies by “crashing” a well to
Mexico, which I characterize as drilling in 6000 feet of model catastrophic events for planning purposes.
water, through 6000 feet of salt and unknown stratigraphy Subsurface visualization—The simple capacity to
to a depth of 25,000 feet. This must be accomplished with have all the disciplines responsible for well construction
rigs that are either new and untested or else aging and share a common image of what’s going on in the
stretched to the limits of their capability. The stakes here subsurface cannot be understated. This aspect is more
are huge. Conservatively, there are 5 billion barrels of about facilitating human interactions than about
undeveloped oil-equivalent reserves in Gulf of Mexico advanced information technology.
deep water alone, which implies development spending of Knowledge capture—Information technology (IT)
around $15 billion. Of that amount, about 40% is expected is still in its infancy. Collecting, analyzing and archiving
to be spent on drilling. When multiplied by the number data are key IT elements, but understanding what this
of other provinces that are opening up, the potential for information tells us is even more important. A big
drilling performance improvement swells. prize—order-of-magnitude performance improvement—
Individual wells in the Gulf of Mexico can easily cost rests on integrating real-time data with knowledge
$30 million or more, a number we must reduce. Indeed, systems that suggest what the data mean. For example,
realizing a reduction is directly related to improved “the last time this trend was seen under these
drilling operations, which are as much about managing circumstances, it meant...”
risk as about achieving “breakthrough” performance. Of course, the real power lies in integrating all of
The most damage and highest losses result from big the above in a holistic confluence of technology and
events—chance occurrences that are difficult to recover human interaction for both the planning and execution
from—precipitated by unexpected hazards encountered phases. We may not be there yet, but it’s not far off.
while drilling. The advances to date are building blocks of an exciting
Managing risk, like the approach to oil and gas prices, is future (see “Managing Drilling Risk,” page 2).
mostly about recognizing and then managing uncertainty,
rather than pretending that we can predict future
outcomes. Currently, several factors that govern the
management of drilling risk are coming together to create
new standards for delivering efficient, cost-effective wells. Steve Peacock
Pore-pressure prediction and management—Today, Vice President, Gulf of Mexico Deepwater Exploration
BP Amoco
the approaches to predicting formation pressure are more Houston, Texas, USA
varied than in the past. Integrating these new methods with
future technologies, such as expandable casing and dual- Based in Houston, Texas, USA, Steve Peacock is currently the BP Amoco
Vice President of Exploration for the Gulf of Mexico. In his 21 years with
density mud lift, will allow larger holes to be drilled across BP Amoco, he has had varied assignments around the world. These include
reservoirs and deeper, more efficient setting of casing. serving as Commercial Analyst for the Middle East, Africa, Far East and
Europe, and Exploration Manager for West of Shetlands (UK) and also for
the Southern North Sea gas basin. Steve received a BA degree in natural
sciences from the University of Cambridge in England.
Advisory Panel
Terry Adams Michael Fetkovich
Azerbaijan International Phillips Petroleum Co.
Operating Co., Baku Bartlesville, Oklahoma, USA
Syed A. Ali George King
Chevron Production Co. Amoco
New Orleans, Louisiana, USA Tulsa, Oklahoma
Antongiulio Alborghetti David Patrick Murphy
Agip S.p.A Shell E&P Company
Milan, Italy Houston, Texas, USA
Svend Aage Andersen Richard Woodhouse
Maersk Oil Qatar AS Independent consultant
Doha, State of Qatar Surrey, England

Executive Editor Contributing Editors

Denny O’Brien Rana Rottenberg
Senior Production Editor Illustration
Mark E. Teel Tom McNeff
Senior Editor Mike Messinger
Lisa Stewart George Stewart
Editors Design
Russel C. Hertzog Herring Design
Gretchen M. Gillis Printing
David E. Bergt Wetmore Printing Company, USA

Oilfield Review is published quarterly by Schlumberger to communicate

technical advances in finding and producing hydrocarbons to oilfield
professionals. Oilfield Review is distributed by Schlumberger to its
employees and clients.

Contributors listed with only geographic location are employees of

Schlumberger or its affiliates.

© 1999 Schlumberger. All rights reserved. No part of this publication

may be reproduced, stored in a retrieval system or transmitted in any
form or by any means, electronic, mechanical, photocopying, recording
or otherwise without the prior written permission of the publisher.

Address editorial correspondence to:

Oilfield Review
225 Schlumberger Drive
Sugar Land, Texas 77478 USA

(1) 281-285-8424
Fax: (1) 281-285-8519
E-mail: OilfieldReview@sugar-land.oilfield.slb.com

Address distribution inquiries to:

Mark E. Teel
(1) 281-285-8434
Fax: (1) 281-285-8519
E-mail: teel@sugar-land.oilfield.slb.com

Oilfield Review subscriptions are available from:

Oilfield Review Services

Barbour Square, High Street
Tattenhall, Chester CH3 9RF England

(44) 1829-770569
Fax: (44) 1829-771354
E-mail: orservices@t-e-s.co.uk

Annual subscriptions, including postage, are

160.00 US dollars, subject to exchange rate fluctuations.

An asterisk (*) is used to denote a mark of Schlumberger.

Schlumberger

Summer 1999

Oilfield Review Volume 11

Number 2

2 Managing Drilling Risk

A new initiative to enhance drilling performance significantly reduces
operational costs and nonproductive time. Through close cooperation
between operator and service company, this approach helps provide
planning and real-time drilling solutions that avoid stuck pipe, reduce
drillstring failures, optimize drilling efficiency, and monitor as well as
control borehole stability. This article shows how a risk-management
and loss-control framework is used to combine operator knowledge and
experience with the technical expertise and drilling measurements
provided by Schlumberger to develop surprise-free drilling procedures.

20 Validating Reservoir Models to Improve Recovery

Operating companies report increases in predicted ultimate recovery
on the order of 10% through data integration, reservoir simulation
and proper development strategies. Reservoir models that draw on
all available information must honor data with different scales from
many sources to validate interpretations, decrease uncertainty and
reduce risk. The best models use independent data sets and validation
advances to confirm and constrain multidisciplinary interpretations
as much as possible, improving simulation reliability and extracting
more value from hard-won formation evaluation and production data.

36 Real-Time Openhole Evaluation

A new generation of wireline technology is revolutionizing fundamental
formation evaluation. In this article, we unravel the mysteries of using
real-time environmental corrections and inversions based on forward
models to provide high-quality log measurements and interpretations
at the wellsite. Field examples illustrate recent innovations such as
the HRLA* High-Resolution Laterolog Array tool, which is designed
to provide more accurate resistivity measurements in saline borehole
environments, and post-logging processing techniques to interpret logs
obtained under the most extreme subsurface environmental conditions.

58 Contributors
60 New Books
62 Coming in Oilfield Review
Oilfield Review Services and MORA Order Form (inside back cover)

1
Differential Sticking Geopressure Unconsolidated Zone Fractured or Faulted Zone Undergauge Hole Key Seating

Managing Drilling Risk

Everyone loves a surprise. Everyone, that is, except a driller. Avoiding drilling
surprises means more than being prepared for problems when they occur; it means

Walt Aldred averting them in the first place. New risk management tools help foretell well
Dick Plumb
behavior with enough advance notice to allow drilling teams to calmly make
Sugar Land, Texas, USA
technically sound operational decisions that lead to optimal drilling performance.
Ian Bradford
John Cook
Vidhya Gholkar
Cambridge, England

Liam Cousins
Reginald Minton Oil and gas companies spend about $20 billion No well is drilled without problems. Managing
BP Amoco plc annually on drilling. Unfortunately, not all of that drilling risk means not letting small problems
Aberdeen, Scotland money is well spent. A significant portion, around become big ones. Knowing what the risks are and
15%, is attributed to losses. These include loss of when they are likely to occur keeps surprises to a
John Fuller material, such as drilling equipment and fluids, minimum. Most of the time spent drilling, and
Gatwick, England
and loss of drilling process continuity, called non- most of the cost, is encountered not in the reser-
productive time (NPT). These losses are incurred voir, but in getting to it.
Shuja Goraya
Cabinda, Angola while searching for and implementing remedies Numerous problems taunt the driller, and
to drilling problems. Avoiding drilling problems solutions may be expensive if not impossible in
Dean Tucker cuts finding and development costs and allows some cases (above and next page). Drillpipe can
Aberdeen, Scotland billions of dollars now spent on losses to be better become stuck against the borehole wall by dif-
spent—building and replacing reserves. ferential pressures or lodged in borehole irregu-
For help in preparation of this article, thanks to Laurence larities, requiring skill and force to free it.1 When
Cahuzac and Chin Yuin Hui, Sedco Forex, Montrouge, this fails, sometimes the only solution is to aban-
France; Richard Carossino and Dave Ede, Anadrill, Aberdeen,
Scotland; Charles Cosad, Camco, Houston, Texas, USA; don the stuck portion and drill a sidetrack around
Edward Habgood, GeoQuest, Gatwick, England; and William it, changing the drilling program completely and
Standifird, Anadrill, Youngsville, Louisiana, USA.
APWD (Annular Pressure While Drilling), ARC5 (Array
Resistivity Compensated), DrilCast, DrilMap, DrilTrack,
IDEAL (Integrated Drilling Evaluation and Logging), MDT
(Modular Formation Dynamics Tester), PowerPak, RFT
(Repeat Formation Tester), Schlumberger PERFORM and
SPIN Doctor are marks of Schlumberger.

2 Oilfield Review
Reactive Formation Mobile Formation Collapsed Casing Junk Cement-Related Drillstring Vibration

Wellbore Geometry Poor Hole Cleaning

> Common drilling problems.

potentially adding millions of dollars to the well Making drilling decisions to correct these 1. Bailey L, Jones T, Belaskie J, Houwen O, Jardine S,
McCann D, Orban J and Sheppard M: “Causes, Detec-
cost. Drilling at a high rate of penetration can problems is a complex process because many tion and Prevention,” Oilfield Review 3, no. 4 (October
save time and money, but when accompanied by factors have to be considered. For example, 1991): 13-26.
too low a drillstring rotation rate or mud flow rate increasing mud weight to control wellbore sta- Adelung D, Askew W, Bernardini J, Campbell AT Jr.,
Chaffin M, Congras G, Hensley R, Kirton B, Reese R and
that fails to lift rock cuttings to surface, the result bility in one interval in the well may cause Sparling D: “Techniques for Breaking Free,” Oilfield
is stuck pipe. Faults and fractures that the well- fracturing elsewhere. Solutions are often well- or Review 3, no. 4 (October 1991): 27-35.
Cline M, Granger G, Hache J-M and Lands J: “Backoff
bore encounters open conduits for loss of drilling field-specific. Basics,” Oilfield Review 3, no. 4 (October 1991): 48-51.
fluid to the formation.2 Excessively high mud Successful drilling hinges on developing a 2. Addis T, Last N, Boulter D, Roca-Ramisa L and
pressure can fracture the formation and cause sound plan, continually updating it in light of new Plumb D: “The Quest for Borehole Stability in the
Cusiana Field, Colombia,” Oilfield Review 5, no. 2/3
lost circulation. Too low, and the mud pressure information and keeping the involved personnel (April/July 1993): 33-43.
fails to keep high-pressure formations under con- informed on a timely basis. The plan must include
trol, resulting in gas kicks or worse, blowouts. procedures to follow under normal circumstances
Drillstring vibrations can weaken and destroy and methods for dealing with the most likely and
pipe and equipment as well as seriously damage most severe problems that may be encountered.
the wellbore. And some of these problems, even With the proper training, a well-defined drilling
if they don’t completely suspend the drilling pro- process, sufficient data and tools for interpre-
cess, jeopardize subsequent logging, completion tation, successfully drilling a well should be a
and production. routine process.

Summer 1999 3
 Advances in Drilling Technology
Multisensor Schlumberger introduced Completely
Partially mechanized IDEAL system with mechanized
MWD 1 pipe handling
introduced instrumented motor pipe-handling
Dual/Tri-Act
Baker Hughes derrick
Spinning chain Teleco Sperry-Sun introduced
replaced by introduced introduced integrated services First sonic tool
pipe spinner simple MWD LWD 2 MHz bit + motor + MWD introduced

71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
High-rate
mud telemetry Eastman & Smith PowerPak Schlumberger Rotary
introduced motors introduced introduced systems
steerable motors ARC5 tool
Offshore high-pressure
high-temperature Step change
(HPHT) drilling Topdrive in QHSE

> Time line from 1971 to 1999 showing recent advances in drilling technology.

Background do operating companies acknowledge that the Second, wells are becoming more complex.
During the last twenty years, the industry has drilling process still needs to improve? The phys- Extended-reach and horizontal wells react differ-
celebrated innovations in drilling practices from ical forces acting on the borehole haven’t ently to earth stresses than do vertical or low-
the introduction of measurements-while-drilling changed. What has happened? angle wells. Drilling multilateral wells requires
(MWD) and steerable motors to computerized Two things have changed. First, exploration extraordinary accuracy and control. Deepwater
rigsite displays and high-resolution while-drilling and production (E&P) companies have altered and high-pressure, high-temperature wells offer
logs (above). In the early 1990s, different operator their internal structures and reduced their work additional challenges. Wells are being drilled in
and service companies applied the power of forces. Many senior, experienced hands have left tectonically active and remote areas where the
maturing while-drilling measurements to adopt the industry. Companies are operating with a infrastructure may be less well developed and
new methods of stuck-pipe avoidance and other bare minimum of personnel. Experienced people communication problematic.
drilling training programs.3 Why, ten years later, who remain may be specialized, and hence not
suited for the integrative role required.

Asset Office Rigsite

Onshore Drilling Team Schlumberger PERFORM Engineer

Earth Model Well Plan

Drill

Revise Interpret Detect

> Integrated drilling process. The phases of a drilling project require joint effort by the asset office and the rig, and encompass
construction of the earth model and well plan, the actual drilling, detection and interpretation of information obtained while
drilling, and ultimately, revision of the model.

4 Oilfield Review
 Seismic Data Drilling Data Log Data Calibration Data

2D cross sections Exploration well reports Deep resistivity Microfrac, XLOT

3D velocity cubes Directional surveys Gamma ray Kicks and losses
Structural interpretation Bit records Oriented multi-arm calipers Gas and flow checks
Seismic attribute maps Time-depth curves Sonic P,S Cavings
Time and depth relations Mud weights Bulk density Cores
Mud logs, drilling fluids Borehole images RFT and MDT pressure
NPT—kicks, losses, stuck pipe
Correlation with geology

Mechanical Earth Model

Faults Elastic Rock strength Pore pressure Stress profiles Stress

Formation tops parameters profile profile Sv, Sh, SH direction

> A partial list of the types of data that Well Plan and Performance Prognosis
contribute to a complex mechanical earth model.

A New Approach engineer and geologist balance the requirements the life of a field, but at its core remain the three
To drill successfully amid these changes and of target location, cost and drillability. Many principal phases that govern the very existence of
challenges requires a new approach to the more factors must be incorporated into a com- a well: developing the proper plan, executing it,
drilling process. In recent years, oil companies plete well plan. These include casing design, and learning from the ongoing process.
and service companies have developed more completion requirements, life-of-field issues, rig The earth model can be simple or complex,
cooperative relationships that make it easier for size and selection, personnel considerations, depending on the information available and the
both to achieve their objectives. The way of costs, cement design, liners, drillstring and BHA requirements of the well. Creating a complex
doing business together has evolved from one of design, and availability of equipment. earth model can require dozens of input and data
managed opposition to one of aligned objectives, The best drilling plan optimizes well location integration steps. In short, every pertinent data
with oil and service companies cooperating to and trajectory, but also minimizes the risk of source is used, from drilling reports, logs and
face the uncertainty and risks of the subsurface. wellbore instability and stuck pipe, improves well tests in offset wells to seismic sections, velocity
The approach taken by the Schlumberger com- productivity and accelerates the drilling learning cubes and structural interpretations (above).
panies to provide technical and decision support process. The plan should flag intervals in which 3. Bradley WB, Jarman D, Auflick RA, Plott RS, Wood RD,
to operators has reduced drilling costs by as much geologic risks such as pore pressure, fracture Schofield TR and Cocking D: “Task Force Reduced
Stuck-Pipe Costs,” Oil & Gas Journal 89, no. 21 (May 27,
as 50% in a wide variety of drilling environments. pressure and other wellbore instabilities can 1991): 84, 86, 88-89.
The complete process integrates the efforts of oil threaten wellbore integrity. To achieve this, the Nordt DP and Stone MS: “Professional Development of
company and service company personnel at the plan must be evaluated to identify all risks before New Rig Supervisors a Must,” Oil & Gas Journal 90,
no. 43 (October 26, 1992): 77-80, 83-84.
office and on the rig, during all stages of well any action takes place.
planning and drilling and through every phase of On the rig, the well is drilled according to the
a drilling project (previous page, bottom). drilling plan. During drilling, information is col-
Simply put, the process begins in the office lected, interpreted and fed back to the drilling pro-
with construction of an earth model. The model is cess, to the well plan, or to the earth model itself.
then used as part of the well planning process to Through modification and updating, the well plan
create the best drilling plan. This is a multidisci- becomes a living document rather than a static
plinary optimization process in which the drilling one. Drilling risks are also continually reevaluated.
The process is valid for wells drilled throughout

Summer 1999 5
 Elastic Strength Earth Stress and Pore Pressure
Young’s
0 modulus 100 0 Friction angle 70
Unconfined
Poisson’s compressive strength Stress Stress direction Sh
ratio
Stratigraphy 0 10 20 kPa 400 0 MPa 200 W N E
Structure and Stratigraphy
1.0

Grain
support
facies

Fault?
Clay
support
facies

Pp Sh SH SV Regional
trend
> Earth model example. The earth model houses all information on rock properties and behavior and is used
during all phases of the life of the well, including trajectory and wellbore stability planning, bit and rate of
penetration (ROP) selection, pore-pressure prediction, casing design, sand control and reservoir stimulation.

N
0 10
20
30
40
50

60

70

80

W 90 E
> Which way to drill in a South American
field. With rock mechanics data such as 100
expected stress state, pore pressure and
rock failure parameters from a variety of 110
sources, a drilling risk profile can be plotted.
Red signifies risky, difficult drilling and 120
blue is less risky and easier. The numbers
around the arc represent azimuth; traveling 130
along a radius is the same as taking a path 140
of constant azimuth. Distance from the 150
center depicts inclination from vertical. The 160
180 170
center of the circle represents a vertical S
wellbore, and the outer edge represents all
possible horizontal wellbores. This plot
indicates that it is easier to drill a horizontal
well than a vertical well given the particular 0.9 0.95 1.0
stress state. Drilling difficulty

6 Oilfield Review
 Schlumberger PERFORM Workflow
DrilMap

DrilCast Prepare risk assessment

for each hole section
DrilTrak

Develop forward plan Drilling and geological

and contingencies for each program roles and
hole section with drilling team responsibilities

Summary and detailed Prepare daily risk assessment Current rig

risk report for next 24 hr earth state

Observations
Develop forward plan Data collection
24-hr activity forecast
and contingencies for next Interpretation
Roles and responsibilities
24 hr with drilling team Analysis

To drilling team Compliance Yes

with plan?

No
No Yes
> The Schlumberger PERFORM workflow. Diagnose
Diagnose Contingencies?
Report to
Responsibilities extend from risk assessment company
and contingency planning to data collection and representative
analysis, then to reporting, well plan updating Develop plan
Develop plan
and activity forecasting. The colors in the upper
left key refer to display, reporting or analysis Loss No loss
tools described in subsequent figures.
Event Near-miss
report report

(A full treatment of the rock mechanics involved tories could be drilled (previous page, bottom). Once the best plan has been formulated, fol-
is beyond the scope of this article.4) The result- Drilling a horizontal well at a 90° azimuth was lowing it through at the rig can be a surprisingly
ing mechanical earth model consists of forma- predicted to be the least risky: wells at other challenging feat. To accomplish this, the
tion tops, faults, elastic parameters, stress inclinations and azimuths would be prone to Performance through Risk Management effort, or
directions and variations with depth, and rock borehole collapse. Schlumberger PERFORM initiative for short, has
strength and pore-pressure profiles (previous The best plan according to any earth model been launched within Anadrill. Schlumberger
page, top). must be reconciled with trajectory goals of that PERFORM efforts have already reduced NPT by
Once a target has been selected, it can be well to optimize the process as a whole. For as much as 40%, saving as much as $300,000 per
reached from many directions. Selecting the path example, in one well, the preferred trajectory may well. The concept is simple and most of the steps
with the least risk requires an understanding of have a 62°-inclination in one section, but hydrau- are almost intuitive, but a structured approach is
the stress state and the rock parameters, and lics analysis may indicate that hole-cleaning required for success. The approach comprises a
how the drilling process will interact with them. problems at this inclination could endanger well workflow, software tools and engineer to ensure
An example of the information that can be integrity. Two or more sections drilled at safer that the technical solutions derived in the plan-
extracted from an accurate mechanical earth angles, though seemingly more time-consuming, ning stage become operationally effective solu-
model comes from a South American field. For could optimize the overall drilling process. tions to aid decisions that help avoid drilling
this field, a risk profile was created that color- problems (above).
coded the difficulty with which particular trajec- 4. For more: Fjaer E, Holt R, Horsrud P, Raaen A and Risnes
R: Petroleum Related Rock Mechanics. New York, New
York, USA: Elsevier Science Publishing Company, 1992.
Alsen J, Charlez P, Harkness R, Last N, McLean M and
Plumb R: “An Integrated Approach to Evaluating and
Managing Wellbore Instability in the Cusiana Field
Colombia, S. America,” paper SPE 30464, presented at
the SPE Annual Technical Conference and Exhibition,
Dallas, Texas, USA, October 22-25, 1995.

Summer 1999 7
 DrilMap

PP
500
N
1000 Stuck pipe at 1100 ft
DrilBase
1500
BHA packoff at 1700 ft
2000

2500
Depth, ft

50-bbl kick at 3200 ft

3000

3500 Hole collapse

at 3700 ft
4000

Pipe stuck
4500
at 4700 ft

5000
Time
8 lbm/gal 14 lbm/gal

> Mapping out the drilling plan. The DrilMap screen displays the planned well trajectory, expected
pore pressures (PP), and two drilling time-versus-depth curves—one optimal (blue) and the other
taking into account potential hazards (red). Hazards are identified with specific depths and tied to
the DrilBase database containing previous drilling and near-miss reports and contingency plans.

The goal of the Schlumberger PERFORM engi- Because each well can host a distinct set of In the planning stage of a drilling project, the
neer is to work with operators to significantly these problems, a specially trained engineer is Schlumberger PERFORM engineer works with the
reduce cost and nonproductive time through inte- assigned to each job. The quality of the person- operator staff to identify potential hazards,
gration of planning and real-time drilling solutions. nel can make or break the process. As general develops methods for detecting them, and finally
A risk-management and loss-control framework qualifications, the engineer must have good prob- with the drilling team, formulates contingencies
combines Schlumberger technical expertise and lem-solving, data-integration and communication to complete the drilling plan. The engineer delivers
measurements with operator knowledge and skills, a solid technical background in petroleum a DrilMap display that links well geometry, geo-
experience to develop operational solutions. or drilling engineering, ample seniority and expe- logical and hazard information with contingency
Communications and teamwork are essential in rience with operator organizations. Technical plans to form a complete process map for the
implementing these solutions. training includes Schlumberger courses on well (above).
The process concentrates on the follow- drilling mechanics, wellbore stability, pore-pres- During drilling, the engineer evaluates the
ing areas: sure analysis, bit performance and drilling fluids. well condition to identify any new hazards that
• wellbore stability and fluid loss Operational problem-solving techniques and may have developed and at every tour provides
• pore-pressure analysis communication skills are sharpened through an updated risk assessment and 24-hr forecast
• stuck pipe and pipe lost in hole problem-simulation exercises. Additional training (next page). The DrilCast report enumerates the
• drillstring failure prevention includes industry-standard courses in stuck-pipe conditions and potential hazards ahead and
• drilling efficiency, rate of penetration prevention and well control. explains how to detect and manage them.
and bit optimization. Detailed planning before a potential hazard is
encountered and accurate identification of the
hazards reduce the risks of losses and signifi-
cantly improve performance.
(continued on page 11)

8 Oilfield Review
 DrilCast

PP
6:00 am 2/3/99

DrilBase

11,000 ft

DrilCast Summary Report

11,500 ft

6:00 pm 2/3/99
> Forecasting drilling activity. The DrilCast display is a graphical
daily report of what should be observed and what might be
13 lbm/gal 14 lbm/gal encountered in the next 24 hours. Each hazard is linked with a
method for its detection and a contingency plan for mitigating
actions. A summary report is distributed to the drilling team at
the morning meeting. Detailed reports, including roles and
responsibilities, are given to each drilling team member.

Summer 1999 9
 PERFORM Daily Report

Client IPM Client Representative

Well Deepwater location. Randall Anderson
Section 14 3/4 X 17 1/2 Drilling Assembly
Perform Engineer
Date 2/24/99 6:00 William B. Standifird

24-Hour Summary For Time Period... 2/23/99 6:00 To 2/24/99 6:00

RIG OPERATIONS
Start Time Rig Operation Comments
2/23/99 6:00 Under-reaming Under-reamed to 7038 ft
2/23/99 7:30 Short trip Hole stable, 500u B/U gas
2/23/99 15:30 Drilling ahead Drilling new formation
TRENDS
Time Period Variable Noteworthy Behavior
24 hr MWD SHOCKS Transverse shocks increase while reaming sands.

24 hr WOB,TQA,ECD,SPP,TFLOW,TRPM ECD and TQA spiking when annulus loads above under-reamer.
EVENTS
When? What? How? Why?

2/24/99 6:00 Drilled sand lobe at 7228 ft

24-Hour Forecast For Time Period... 2/24/99 6:00 To 2/25/99 6:00

This is a tough section. Depleted zone at 7375 ft is next major hazard.

RIG OPERATIONS
ITEM Operation When? Possible Hazards Severity Probability

1 Drilling Cutting sands First depleted sand at 7375 ft. High High
Stuck pipe, lost circulation.
2 Drilling Cutting sands MWD shock high when U/R hits sands. High High
MWD shock > 22 can damage BHA quickly.
3 Drilling Pumping ECD will spike as U/R packs off. High High
Back-reaming, U/R Stuck-pipe situation.
4 Short Trip Pulling up Swab formation into wellbore. Med High
Back-reaming Gas or fluid entering wellbore.
HAZARD DETECTION METHODS
ITEM
1 Identify sand locations and verify stability. 7215, 7375, 7565 and 7745. Use offset e-logs/mud logs

2 Monitor MWD shocks on Anadrill display.

3 Monitor ECD closely. Spikes are rapid and must be addressed quickly.

4 Monitor trip speeds (swab-surge), gain/loss and gas.

PROPOSED HAZARD PREVENTION ACTION

ITEM
1 Prepare LCM and other LC systems. Keep the pipe moving. Survey at 7200, 7350 and 7800. DO NOT survey
if formation is unstable. Stuck pipe is more expensive than a GYRO in casing.
Torque and Slump is first action to differential sticking or coming off slips.
2 Rotate during connections.
Notify PERFORM Engineer. May need to adjust RPM/WOB to control vibrations and avoid BHA damage.

3 Consider picking up and back-reaming until ECD stabilized. First move is in opposite direction of resistance.

4 Back-ream or pump out of hole. Circulate gas out of hole if encountered. Work tight spots and keep
pulling speeds minimal.
Please contact the Schlumberger PERFORM Engineer if there are any questions or transmission errors: Call Ext. 158 (rig) 3460 (town)

> Daily report of past and future drilling activity.

10 Oilfield Review
 DrilTrak

lbm/gal
PP
500
N
1000 GO

1500

2000
BHA packoff at 1700 ft

2500
Depth, ft

GO
3000

Hole collapse
3500
at 3700 ft

4000
GO

4500

5000
Near Miss Time
Loss

> Tracking drilling progress. The DrilTrak plot updates the drilling map while drilling. Changes in the trajectory are recorded along with the response of
the well and the effectiveness of the drilling plan. Hazards that were avoided with no material or process loss are reported as near misses (green arrows).
Losses are reported as events (red arrows).

In this example from a deepwater well coordi- sands, under-reaming or pumping), underway Suites of data evaluation and problem diagno-
nated by the Schlumberger Integrated Project when the hazard is met; the type of hazard sis tools have been developed to support these
Management (IPM) group, the daily report and its consequences; the severity; and the prob- drilling displays. Diagnostic tools, such as the
includes a summary of rig operations, trends and ability. Methods for detecting each hazard are SPIN Doctor stuck drillstring prevention software,
events of the past 24 hours along with the fore- listed, as are actions to prevent an event from zero in on the most probable cause for each prob-
cast for the next day (previous page). The look- causing loss. lem by asking the user a series of questions. The
ahead portion lists four possible hazards that may A member of the drilling team monitors well SPIN Doctor application also contains links to
be encountered in the upcoming hole section. The conditions continuously to determine if the well electronic documents such as the Schlumberger
section is ranked as a tough one, with a depleted is behaving as planned (above). If the well is not Stuck Pipe Handbook for more in-depth investi-
zone ahead posing the next major hazard. The proceeding as expected, the appropriate contin- gation into unforeseen problems, and can be
hazards are identified according to several fac- gency is identified. The driller can then follow the custom-hyperlinked to any desired electronic
tors: the operation (drilling, back-reaming or plan for that contingency. If none of the planned resource, including proprietary drilling process
tripping), and the specific procedure (cutting contingencies is appropriate, the problem is ana- manuals and help files (next page).
lyzed, and a new action plan is developed with
the drilling team.

Summer 1999 11
 SPIN DOCTOR In addition to connecting to analysis and
DIAGNOSIS
diagnostic programs, the DrilTrak system incor-
Differential Sticking porates drilling alarms that analyze while-drilling
Welcome to Poor Hole Cleaning measurements in real time to alert drillers to
Unconsolidated Formations
severe problems. These alarms warn of high
Fractured/Faulted Formations
friction factors, bearing failures, low drilling effi-
Junk in Hole
Stuck drillstring prevention software.
Cement Blocks ciency and bit performance, washouts and kicks.
Version 3.4 Reactive Formations
Geopressured Formations Understanding Risk
Formation Ledges The Schlumberger PERFORM approach is built
Key Seating
on a foundation of risk management and loss-
Wellbore Geometry
Undergauge Hole
control methodologies. Controlling loss requires
Click on a diagnosis name for a handbook date.
Mobile Formations an understanding of event causation, or the
To proceed, is this:
Collapsed Casing act or process of causing events, problems or
a real stuck pipe incident? just a training exercise?
String Component Failure accidents that lead to loss. A model of event
Bit Failure causation catalogs the stages in the evolution
Restart Back Next Notes Produce log Help of an event from its original controlled state
Schlumberger Drillers Stuck Pipe Handbook (next page). In the earliest stage preceding an
event, inadequacies in the system, or in stan-
SPIN DOCTOR
dards or compliance generate the potential for
DIAGNOSIS
Differential Sticking
an event. In the case of drilling, the system is the
Poor Hole Cleaning basis of design for the well; the standard is the
Overpull in New Hole
Unconsolidated Formations drilling program; and compliance is making sure
Fractured/Faulted Formations the well is behaving as anticipated. Underlying
Junk in Hole problems that can be traced back to this first
Is the overpull in the new hole section?
Cement Blocks
stage might be inappropriate casing or drilling
Reactive Formations
Geopressured Formations
fluids design or a drilling rig unsuitable for the
Definitely yes
Formation Ledges particular drilling program. In and of themselves,
Probably yes Key Seating these do not cause a drilling mishap, but trying to
Possibly yes Wellbore Geometry adjust daily drilling activities to work around
Indeterminate Undergauge Hole these fundamental flaws requires human energy
Mobile Formations
Possibly no that could be better spent following the drilling
Collapsed Casing
Probably no String Component Failure routine. This first stage is the one in which the
Definitely no Bit Failure longest decision time—months in most cases—
Restart Back Next Notes Produce log Help is available to avert a problem, and the most
Schlumberger Drillers Stuck Pipe Handbook brain-power, in terms of numbers of highly
trained personnel, can focus effort on a solution.
In the second stage, basic causes of an event
SPIN DOCTOR
can be attributed to personal factors and job or
DIAGNOSIS
Differential Sticking
system factors. Examples in drilling could be
Restricted Circulation Poor Hole Cleaning inferior or insufficient training, delaying a bit
Unconsolidated Formations change in anticipation of the end of the work
Fractured/Faulted Formations shift or not putting a cover on the hole when the
Junk in Hole drillstring is pulled out. Taken individually or
Is circulation restricted?
Cement Blocks
even together, these factors do not cause a prob-
Reactive Formations
Geopressured Formations
lem, but may allow problems to develop. Actions
Definitely yes
Formation Ledges at this stage typically are based on decisions
Probably yes Key Seating made days to hours before an event, by one per-
Possibly yes Wellbore Geometry son or a few on the rig.
Undergauge Hole
Indeterminate The third stage describes immediate causes
Mobile Formations
Possibly no of an event, such as substandard conditions,
Collapsed Casing
Probably no String Component Failure practices or acts—letting equipment fall into
Definitely no Bit Failure disrepair, accidentally dropping a small hand tool
Restart Back Next Notes Produce log Help down the hole, or improperly interpreting a mea-
Schlumberger Drillers Stuck Pipe Handbook surement. Decisions—not to report the faulty
> Three panels from the SPIN Doctor stuck drillstring prevention software. As the user answers ques-
hardware or lost screwdriver or not to mention
tions about a drilling problem and the accompanying well conditions and drilling activity, the system what the shale shaker is accumulating—are
rules out some mechanisms and highlights increased probability for others. In this case, the final diag- made days to minutes before the event, by some-
nosis is poor hole cleaning. one on the rig, often acting under stress.

12 Oilfield Review
 Causation Model Making known to all rig personnel the techni-
cal reasons behind contingency actions is
Control Basic Immediate Incident Loss
causes causes another area in which good communication plays
an important role. As in most situations influ-
Inadequate Personal Substandard Event Unintended
factors acts or harm or enced by habit, the easiest thing for a driller to
practices damage do is what’s been done before. But if, when the
•System
•Standard Job or system Substandard time comes, it’s important to do something dif-
•Compliance factors conditions
ferent, the driller is much more likely to react cor-
rectly if the reason is understood. The case study
Threshold in the next section demonstrates how communi-
limit
cation, risk analysis, proper measurements and a
> Evolution of an event. Inadequacies in the basic system, personal factors and
substandard practice all contribute in more or less identifiable ways to a drilling
team approach help drill wells where success
problem. Most events are not even reported until past the critical state of loss. previously had been elusive.

In the fourth stage, the event, or incident, The same elements make an effective Controlling Instability
occurs. Drillpipe gets stuck or the well takes a approach to dealing with drilling incidents, and Experts estimate that wellbore instability costs
kick. There may be only minutes to make the right several of these have been incorporated into this the industry more than $1 billion per year. The
decision. The person making the decision that new strategy for drilling. Better communication in industry average cost of nonproductive time—
might free the pipe or prevent disaster is acting the form of near-miss reporting, documentation of often due to wellbore instability—works out to
under tremendous stress, and so with reduced process compliance, increased awareness of team about $1.5 million per well, and in extreme cases
ability. Experts in the management of crises, such goals and understanding of the technical reason- can reach $16 million for a single well.
as wars and natural disasters, report that under ing behind contingency actions is the most impor- Wellbore instability occurs when earth
comparable levels of stress, decision-makers uti- tant factor in applying these risk management and forces or interactions between the formation
lize only one-fourth of the information available. cause analysis methods to drilling operations. and the drilling fluid act to squeeze, stretch,
The final stage, the actual loss, results in Near-miss reporting is considered standard constrict or otherwise deform the borehole.
unintended loss or damage to property and the HSE practice for successfully reducing the fre- Consequences of wellbore instability are stuck
drilling process. The bottomhole assembly (BHA) quency of workplace errors and accidents, but pipe and BHAs, excessive trip and reaming time,
and a section of drillpipe are lost in the hole, or a before the introduction of the Schlumberger mud losses, fishing or loss of equipment, side-
kick advances to a situation that can be con- PERFORM methodology, it had not been applied tracks, inability to land casing, and poor logging
trolled only by killing the well. Afterwards, the to drilling. In the past, when a well was com- and cementing conditions.
incident is finally reported. pleted on schedule without major problems, Drilling plans include stability studies based
These risk management and cause analysis everyone involved congratulated each other on a on information from neighboring wells so that
concepts have their origins in health, safety and job well done, but little thought was given to optimal drilling trajectories, mud programs and
environment (HSE) awareness initiatives. Most analyzing the process that produced the success- drilling practices can be established in advance.
companies in the E&P industry have comprehen- ful result. The well may have been drilled without However, the earth doesn’t always behave as
sive, effective HSE awareness and training pro- major problems, but it almost certainly was not predicted and sometimes the forces act contrary
grams. Maintaining an active training program is drilled without any problems at all. That it to expectations.
recognized as being as important as any other appeared to be so was because each of the small Wellbore instability often can be managed if
aspect of doing business. HSE training programs difficulties encountered along the way had been it can be detected in time. Control mechanisms
are based on the understanding that most inci- dealt with successfully. The story behind each of include changing mud chemistry, mud weight and
dents that lead to loss are caused by human error, the forgotten small problems and its solution is flow rate to exert more or less pressure on the
error that could be prevented with proper care. the secret of the well’s success. formation or changing rate of penetration (ROP)
In the E&P industry, operators have examined Identifying drilling predicaments and report- or drillstring revolutions per minute (rpm) to facil-
occurrences of drilling problems and report that ing them as soon as possible increases the likeli- itate hole cleaning.
most unscheduled events can be attributed to hood that a small problem will be recognized and In an effort to develop a capability for
human error. In one published report, 65% of solved at an early causation stage, before it real-time detection and control of wellbore
stuck-pipe incidents could be directly related to becomes unmanageable. Documenting the steps stability—while the well is being drilled—a
inadequate planning; 68% of incidents occurred taken to solve the problem produces two addi- partnership was formed in 1996 between Amoco,
within two hours of a tour change.5 tional benefits: the first is a report of the drilling The Netherlands Institute of Applied Geoscience,
Most of the techniques used in HSE training history, complete with a record of how personnel GeoQuest and Schlumberger Cambridge Research,
courses are designed to combat human nature— responded to problems. This record shows how England. Partial funding was supplied by the
to slow down speedy driving, do away with lazy successfully workers comply with procedures. European Union THERMIE program.
waste-disposal habits or avoid distraction during The second is an archive of problems and solu- 5. Watson B and Smith R: “Training Reduces Stuck Pipe
machine operation. Managers understand the tions that can be tapped in the future, whether in Costs and Incidents,” Oil & Gas Journal, 92, no. 38
(September 19, 1994): 44-47.
need for constant vigilance and annual retrain- deeper sections of the same hole, or in other
ing, and employees are required to keep their wells or other fields.
training records up to date. Near-miss reporting
helps employees become more aware of situa-
tions and conditions that could lead to accidents.

Summer 1999 13
 The methodology was tested in the Valhall
field, a major chalk reservoir discovered in 1975
and operated currently by BP Amoco Norge, with
partners Elf, Amerada Hess and Enterprise. The
field contains 600 million bbl [95 million m3] oil
N reserves, with a centralized production complex
0
in 70 m [230 ft] of water. Reservoir depth is
2500 m [8200 ft]. Overall development objectives
-1000
W are to increase the value of Valhall assets to
1 billion barrels [160 million m3], partly through
-2000 section extended-reach drilling into downflank reserves.
Problem
Earlier drilling problems on Valhall were
-3000 numerous and typically included packoffs and
4000 stuck pipe, tools lost in hole, mud losses, side-
2000 tracking and inability to land casing or drill out of
-1000 -2000
0 casing. As a consequence, there is a high risk
2000 1000
0 3000 that wells will be suspended or abandoned
4000
before reaching the target.
The field test of the methodology, which was
> Problem section predicted in Valhall trajectory. Borehole inclination, earth stresses developed at Schlumberger Cambridge Research,
and formation characteristics combine to make this inclined section of the borehole called for an integrated approach to wellbore
prone to cavings that could lead to stuck pipe if not properly managed.
instability control. The design stage comprised
data gathering, mechanical earth model
construction, well stability strategizing and for-
0
mulation of a drilling plan. Execution included
Eldfisk
1000
drilling monitoring, data acquisition and instabil-
Breakouts
ity detection. Evaluation consisted of interpreta-
Horizontal Valhall
tion of observations, updating the model and
2000
stress recommending future actions.
Pore pressure
No

In the planning phase for Well 2/8-A3C, a

Measured depth, m

rw
eg ector

3000
UK

mechanical earth model was generated that

ian

Hod
s

sec

tor
described the state of stress, rock properties and
tor

4000 sec failure mechanisms active in this region of the

ish
Dan
0 km 10 Valhall structure. A mud-weight window was cal-
5000 0 miles 6.2 culated taking into account traditional wellbore
instabilities, and other problems such as fracture
6000 zones—existing natural fractures—were identi-
fied. A problem interval at 4000 m [13,100 ft]
7000 Bergen
measured depth in the 121⁄4-in. hole section was
6 8 10 12 14 16 18 NORWAY
Mud weight, lbm/gal
flagged as a zone where fracture zones could
Stavanger become destabilized (above left).6
> Valhall location (right) and mud-weight window (left).
The pore pressure and minimum horizontal stress curves 0 km 200 Depending on depth, the calculated mud-
are taken from the mechanical earth model. The breakout 0 miles 124 weight window is either extremely narrow or
curve (red) is calculated as the mud weight needed to nonexistent (left). Instability was inevitable. At
ensure that none of the rock around the hole will be DENMARK
too high an effective mud weight, the fracture
stressed beyond failure. Mud weight needs to lie between
the breakout curve and the horizontal stress curve (blue). zone would be driven beyond its precarious
In some sections of the hole, this is not possible. balance and cause irreparable borehole collapse.
But for any mud weight below the fracture
pressure, breakouts would occur. The solution,
therefore, was based on recognition of the
inevitability of formation failure. The only way
to drill the well was to let the instability occur,
then manage it. Breakout problems would be
controlled by good hole cleaning. Fracture
zones, however, are uncontrollable, and must be
kept stable.

14 Oilfield Review
 Typically, drilling in the Valhall Tertiary strata If, in spite of the low mud weight, blocky Angular Caving
started with a mud weight of 14.3 lbm/gal cavings were seen at surface, it would mean
[1.71 g/cm3]. As drilling proceeded and cavings, that the fracture zones were being invaded. This
caused by shear failure of the wellbore wall, would require addition of lost-circulation mate-
were observed, the mud weight would be rial to the mud in order to seal the fractures.
increased steadily, often exceeding 16 lbm/gal A Schlumberger PERFORM engineer was
[1.92 g/cm3]. This caused problems in the lower stationed on the rig to monitor surface and down-
section, as it produced wellbore pressures above hole measurements and advise on stuck-pipe
the fracture gradient. Mud was lost, and large issues: in particular to monitor and analyze cav-
amounts of blocky cavings were produced from ings and act as liaison between the drilling staff
the naturally fractured zones, resulting in pack- on the rig and the wellbore-stability team onshore.
offs. The new strategy proposed that drilling Three aspects of cavings information were
should begin with mud at 14.2 lbm/gal tallied. First, the rate of cavings production at the Tabular Caving
[1.7 g/cm3], barely lower than usual, but that this shale shakers—the coarse solids separators
value should not be increased unless absolutely on any rig—was recorded every 30 minutes by
necessary in response to gas, positive flow measuring the time required to fill a bucket. This
checks or other signs of overpressure. If cavings method may seem crude, but is reliable and
were produced by shear failure, they would be versatile in terms of the number of different rigs
removed by good hole-cleaning practices rather to which it can be applied. More sophisticated
than be suppressed by higher mud weight. solids-measuring devices have been tried, but
The equivalent circulating density (ECD) few have been satisfactory.
would be kept lower than the minimum hori- Second, the dominant shape of the cavings
zontal stress—15.3 lbm/gal [1.83 g/cm3] in the was noted. Initially, the intention was to classify
problem zones, except in extreme circumstances. cavings into three types: angular ones originating
> Tabular caving (bottom) from natural fractures
ECD is the effective mud weight that generates from breakouts, blocky from naturally fractured
and angular caving (top) caused by breakouts.
the downhole pressure observed while pumping, zones, and elongate or splintery cavings from Scale is in mm.
and is generally greater than the mud weight zones of elevated pore pressure. Unfortunately,
measured at surface because of frictional pres- most cavings were just nondescript pieces of One of the tasks was to carefully monitor the
sure drop in the annulus and cuttings loading in broken rock. However, some did indicate they rate of penetration and the ECD. If the latter
the mud. In earlier Valhall wells, ECD was were from breakouts, and some from overpres- crept up to 15.3 lbm/gal, there would be the risk
allowed to exceed 15.3 lbm/gal, with consequent sured zones. Only two cavings were seen during of mud invading fracture zones and causing per-
loss of mud to the fractures in the formation— the entire drilling program that came unam- manent formation damage. If the ECD got too
an expensive problem, but not one previously biguously from fracture zones, attesting to the low, cuttings and cavings could be accumulating
regarded as threatening to wellbore integrity. correctness of the selected drilling strategy. around the bottomhole assembly, eventually pre-
This new drilling strategy made the explicit Third, the geological age of the cavings gives venting fluid flow and sticking the drillstring in
assumption that cavings produced by shear fail- an idea of where they are coming from in the the hole. Rate of penetration is important in con-
ure stemming from low mud weight would occur interval. This required micropaleontological trolling ECD. If too much rock is drilled too
in quantities controllable by hole cleaning, but analysis that was not available immediately. quickly, the suspended cuttings increase the mud
that cavings produced by mud invasion and stim- When the results did arrive, they indicated that density and hence the ECD. While it is clear this
ulation of fracture zones would be uncontrol- all cavings were coming from the upper openhole might lead to problems, one of the traditional
lable. It was clearly important to know whether section that had been exposed to drilling fluids aims of the drilling crew on a rig is to drill as fast
mud invasion was occurring, and so a further part the longest. as possible. Crews assume that high ROP will
of the strategy was to monitor mud volume for Onshore at the BP Amoco drilling team office, help reach target depth more quickly, and some-
losses, and also monitor cavings at surface to real-time data were displayed. The real-time times pay bonuses are tied to beating drilling
identify their source. This would be done by clas- drilling parameters display proved popular, and schedules. In most areas, however, including the
sifying their shapes; shear-induced cavings from gave the onshore drilling and wellbore-stability North Sea, a longer term view must be taken;
breakouts are angular, those from fracture zones staff close contact with drilling operations. The high instantaneous drilling rates can lead to
are tabular and parallel-sided (above right). wellbore-stability team attended morning drilling problems that cost more to solve than is saved in
6. Kristiansen TG, Mandziuch K, Heavey P and Kol H: meetings, advised on stability issues and gave a drilling time.
“Minimizing Drilling Risk in Extended Reach Wells at class on wellbore stability to this group and one
Valhall Using Geomechanics, Geoscience and 3D
Visualization Technology,” paper SPE/IADC 52863, pre- from another platform in the Valhall. The classes
sented at the SPE/IADC Drilling Conference, Amsterdam, focused the attention of the crew on the avoid-
The Netherlands, March 9-11, 1999.
ance of instability problems, rather than the
traditional reactive approach, and allowed the
staff to meet and question the scientists
and engineers who would be influencing their
drilling procedures.

Summer 1999 15
 An example of the Schlumberger PERFORM
process in action can be seen in the crew’s reac-
tion to an anticipated problem. During drilling,
gas levels and fluid volumes require continuous
monitoring to ensure that any gas is detected and
there is no risk of a kick developing. When back-
ground gas levels were high in the interval from
2100 to 2200 m [6890 to 7220 ft], the standard
response would have been to increase mud
weight substantially to suppress gas influx into
the borehole. This would have led to the destabi-
lization of the critical fractured zone lower,
between 4100 and 4300 m [13,450 to 14,100 ft].
The driller was advised that mud weight had
to be kept low, and that another way to control
gas leakage was to slow down. The mud-
weight increase was restricted to 14.6 lbm/gal
[1.75 g/cm3] and the rate of penetration was
reduced to below 30 m/hr [98 ft/hr]. The lower
ROP decreased the rate at which crushed rock
released gas into the annulus, and these actions
reduced background gas levels from the 20 to > Structure of the salt dome responsible for the Mungo field accumulation. White curves
35% range down to less than 5%, while avoiding are well trajectories and the yellow lines on the dome are interpreted faults.
problems deeper in the well.
The reservoir was penetrated ahead of sched-
ule, with much lower mud loss to the formation
than usual and negligible activation of fracture
zones. The asset team acknowledges that the
implementation of real-time wellbore-stability
500 Well trajectory
control significantly reduced the risk and drilling
costs to the top of the reservoir, and achieved
optimal well construction technique earlier in the
field development cycle.
1000 Tilt of 27.6 degrees
on horizontal stress
> Wellbore trajectory on a vertical slice through
the stress field modeled around the Mungo field.
Measured depth, m

In an unperturbed earth, the maximum principal

stress is vertical, and the intermediate and
minimum ones, horizontal. With the intrusion 1500
of salt, the stress field has been perturbed and
tilted. This earth model can be interrogated at
any point and the stresses visualized as axes
through a sphere. In this example, the model has
been interrogated at 1500 m: the red crosshairs Outline
2000 of diapir
and circle indicate that the maximum and interme- Sigma1 30.6
diate principal stresses are tilted 27.6 degrees.
Sigma2 24.58
Sigma3 23.54
2500 Tilt 27.6
x coord 1004
z coord 1407

3000
0 500 1000 1500 2000 2500 3000
Distance, m

16 Oilfield Review
 Another field in the North Sea experienced 0
similar gains in drilling efficiency through opti- W E
mized planning and monitoring of wellbore stabil- 30 in.

ity and hole-cleaning practices. Development

wells in the Mungo field in the Eastern Trough
Area Project (ETAP) encountered extreme instabil- 500
ity problems as they neared the flanks of the over- 18 5/8 in.
hanging salt diapir (previous page, top). Tectonic Fracture-bounded
caving
activity associated with the salt emplacement Weak bedding-
had fractured and weakened formations through plane caving
which wells were to pass on their way to the 1000 13 3/8 in.
reservoir. The highly disturbed sediments around
the salt intensified the hole-cleaning problem.
Cavings, whether bounded by fractures or by

Measured depth, m
weakened bedding planes, clogged the wellbore.
Stuck-pipe problems were especially severe in
1500
the long, inclined 60° sections of the S-shaped
trajectories that were necessitated by the cen- Gas-oil contact, 1680 m
trally located platform. Overpressured formations
and high-pressure chalk rafts added further risk to Top chalk
the drilling program. Top reservoir
The four wells in the first phase of Mungo 2000
development drilling had experienced large cost
9 5/8 in.
overruns in the 121⁄4-in. sections. For the subse-
Salt dome
quent phase of development, a mechanical earth
model was constructed for the Mungo structure
and used to plan the second phase of develop- 2500
ment wells. Some of these wells pierced the salt Oil-water contact, 2645 m
for a 133⁄8-in. casing point then followed the salt
downflank to the reservoir sand (previous page,
bottom). As in some sections of the Valhall wells,
0 500 1000 1500 2000
stress profiles indicated cavings would be abun-
dant, so good hole-cleaning practices would be Distance, m

crucial to successful drilling. Downhole monitor- > Cavings shapes predicted and found along the trajectory. The volume around the top
ing of ECD with the APWD Annular Pressure of the salt dome was predicted to be highly fractured and prone to fracture-bounded
While Drilling tool would help the engineer cavings. Deeper along the inclined section, cavings were found to separate along
weaknesses in bedding planes.
detect hole-cleaning problems before they could
cause stuck pipe.7
In Well P2, the first well of the second phase Currently, the Mungo wells team has a lar morning call where everyone is briefed and
of Mungo development, wellbore instability did Schlumberger PERFORM engineer offshore and a made fully aware of any potential problems for the
cause large amounts of cavings to enter the geomechanical expert onshore as part of the next 24 hours. This process worked well on the
borehole (right). However, the combination of drilling team. This engineer and members of the recently drilled P3 well. A situation involving pos-
surface detection of cavings and cuttings, down- onshore team, consisting of the geomechanical sible losses was avoided by keeping the ECD low
hole measurements for hydraulics monitoring expert, drilling engineer, directional planner and while drilling through a fracture. A small volume of
and attentive drilling overcame this problem. The geologist, hold a morning conference call to dis- mud was lost, but drilling continued unabated.
NPT was significantly reduced, with substantial cuss what has occurred over the last 24 hours and
7. For more on the application of the APWD tool in the
cost savings. what can be expected for the upcoming day. The Mungo field and others: Aldred W, Cook J, Bern P,
results of this meeting are presented at the regu- Carpenter B, Hutchinson M, Lovell J, Rezmer-Cooper I
and Leder PC: “Using Downhole Annular Pressure
Measurements to Improve Drilling Performance,” Oilfield
Review 10, no. 4 (Winter 1998): 40-55.

Summer 1999 17
 ultimately lead to higher per-foot costs. Stick-slip and downhole measurements of weight on bit
GHANA NIGERIA
occurs when high friction between the bit and torque, shocks and vibrations provided a clear
CAMEROON the formation actually stops the bit from rotat- guide to controlling stick-slip, shocks and vibra-
ing—the stick phase—even though the drillpipe tions by modifying WOB (below). Surface (SWOB)
is still being turned at a constant rate on surface. and downhole weight on bit (DWOB) were seen to
CONGO After a short delay, slip takes over when torque correlate closely with the occurrence of torsional
GABON
built up in the twisted drillpipe overcomes the vibrations at XX325 ft brought on by stick-slip, so a
friction and the bit turns, but several times faster stick-slip threshold weight was determined, under
than the speed transferred from the rotary table which the WOB would allow smooth drilling. For
Cabinda ZAIRE
or topdrive. Torsional vibration, or oscillation of thresholding purposes, the downhole weight on bit
the drillstring around its rotational axis, is one of measurement was more reliable than that mea-
ANGOLA the three modes of drillstring vibration, the other sured on surface. For example, at XX360 ft, where
two being axial—along the long axis of pipe, and torsional vibrations are low, the DWOB lies below
lateral—from side to side across the pipe.8 the threshold, but the SWOB is above it. This is in
Introduction of the Schlumberger PERFORM contrast to the next lower section in which DWOB
> Cabinda, a northern enclave of Angola on technique produced immediate results. In the first (and SWOB) are above the threshold, and vibra-
the west coast of Africa, having crude oil as
well to use such an engineer, monitoring surface tions are set in motion.
its dominant export.

Optimizing Bits and Drilling Practice Lateral Vibration SHK Width

The Schlumberger PERFORM techniques for opti- 0 G 40 0 µsec 3000
Depth, ft

mizing drilling performance can be applied to SWOB STOR Axial Vibration SHK Peak
0 klbf 40 0 klbf 20 0 G 10 0 G 200
other drilling challenges. In addition to managing
ROP DWOB DTOR Downhole Torsional Vib
wellbore instability and promoting good hole- 0 ft/hr 50 0 klbf 40 0 klbf 10 100 rpm 300 0 radian/sec 3000
cleaning practice, the methodology has been
used to improve drilling efficiency by supporting
bit selection and appropriate drilling practice to
reduce damage to drillstring components.
Chevron is drilling and operating offshore in XX320
the Cabinda enclave of Angola (above). Their cur-
rent efforts concentrate on the South Sanha
fields where the main reservoir, the Pinda forma-
tion, is the deepest and hardest to drill. The
interbedding of hard and soft layers in the Pinda
XX340
formation plays havoc with drilling equipment,
and it is a challenge to prolong the lives of bits
and other BHA components. In one instance,
after drilling just two wells, Sedco Forex had to
scrap about 80 joints of heavy-weight and stan-
dard drillpipe due to eccentric wear. XX360
The main goals for the Schlumberger PERFORM
engineer were to improve ROP and eliminate drill-
string failures. In essence, this meant finding ways
to ensure that all the rig energy imparted through Stick-slip
the rotary table or topdrive to the drillstring and bit threshold
XX380 weight
be used constructively to cut rock rather than to
destroy the bit and drillstring. The difference
between the two situations sometimes can be
small, and the best way to avoid the latter is by
careful planning, understanding the process and
monitoring both surface and downhole measure- > Surface and downhole measurements for optimizing drilling in a Chevron Cabinda well.
ments in real time. Increases in surface (SWOB) and downhole (DWOB) weight on bit (track 2) correlate with the
Standard practice for increasing ROP was to onset of dangerous torsional vibrations (track 5) induced by stick-slip, first seen in the zone from
increase weight on bit (WOB). But increasing XX325 to XX330 ft. To avoid torsional vibrations, a stick-slip threshold weight was determined and
WOB can cause other problems, including tied to measured DWOB, which is more reliable than SWOB. This can be seen in the interval from
XX360 to XX369 ft: there are no torsional vibrations when DWOB is below the threshold, but
increased stick-slip and torsional vibration, SWOB is above the threshold and would have given a false alarm.
which in turn damage the drillstring and

18 Oilfield Review
 The well took 11 days and one bit run to drill Valhall Stuck-Pipe Risk
20
the 10,000 ft [3050 m] to the top of the Pinda, then
23 days with 6 bit runs to drill through the 3000-ft

Torque, klbf
[900-m] Pinda. The sporadic success of any par-
15
ticular bit and BHA combination in this field was
unexplainable. Sometimes one combination
would achieve excellent ROP and footage, and at
10
other times it would fail after the initial few feet.
The engineer combined data from surface and 3500

Standpipe pressure, psi

downhole measurements and rock strength anal-
3400
ysis and related these to previous bit and BHA
performance. This allowed estimation of optimal 3300
ranges for the while-drilling measurements and
3200
helped in subsequent bit and BHA selection.
Then specific drilling performance measurements 3100
were monitored in real time on the rig and kept
1
within the optimized range so as to achieve opti-
mal cost per foot.
The experience gained while drilling this well
Risk

0.5
was applied to subsequent wells. In all later
wells, the number of shocks measured with a
threshold shock sensor that detects shocks
0
greater than 100 G decreased from a range of X450 X475 X500 X525 X550 X575 X600 X625 X650
6 to 8 million in the Pinda formation to almost Depth, m
zero. The problems of eccentric drillpipe wear > Predicting the possibility of stuck pipe in the Valhall field. Torque (top) and standpipe pressure
disappeared completely and the learning curve (middle) measured while drilling are two elements, along with signal processing techniques,
for selecting the right bit and BHA sped up, that help identify well sections where the risk of stuck pipe is high (bottom). The shaded bars
resulting in improved drilling performance. indicate where the drilling team did experience drilling difficulties, mostly related to hole-
cleaning issues.
Tools for Success
The successes delivered by the Schlumberger well conditions, understanding the consequences Valhall wellbore stability case study discussed
PERFORM process stem from the combination of of a decision and being prepared for the future earlier, scientists at Schlumberger Cambridge
Schlumberger technical strengths in measure- with contingency plans. The Schlumberger Research have produced a stuck-pipe risk profile
ment and interpretation with the operator’s PERFORM initiative impacts this process most sig- that begins to foretell hole-cleaning problems
drilling expertise. High-quality while-drilling data nificantly by helping to provide an accurate view of (above). With further testing and experience,
and accompanying analyses are vital for success- the current conditions and a look ahead at poten- these advances will eventually change alarms
ful drilling, but they are most valuable when used tial hazards. The result is that better decisions can from signaling a surprising event when it occurs
in a consistent way to support decisions made be made by transferring the decision-making to advising drilling teams long before the prob-
during the drilling process. period from the stressful moments surrounding an lem becomes dangerous.
This process is a series of decisions and asso- incident to some earlier time when judgment is The oil industry, like all others, strives for
ciated actions taken during the planning and not impaired by anxiety and pressure. cost-effectiveness and productivity. Elimination
execution of a project that result in a completed Researchers are investigating ways to of waste and losses, whether in process or mate-
well. The degree of success or failure and effi- improve the decision-making process by making rials, is a key goal for all successful companies,
ciency of the well is determined by the quality of more data available faster and using knowledge regardless of prevailing economic conditions.
those decisions. Effective decision-making gained in other areas. For example, new tech- Increasing drilling efficiency by managing drilling
depends on having an accurate view of current niques are being devised for estimating the risk risk is a sure way for E&P companies to achieve
8. Jardine S, Malone D and Sheppard M: “Putting a of a drilling incident such as stuck pipe. Using that objective. —LS
Damper on Drilling’s Bad Vibrations,” Oilfield Review 6, standpipe pressure and torque data from the
no. 1 (January 1994): 15-20.

Summer 1999 19
 Predictive Simulations and Decisions

Reservoir Model
Simulation Model cts (Shared Earth Model)
ta
id con
Res Flu
erv
oir
vol
um
e
l
ode
og ic m
Syn G eol
Pro thet
per ic s
ty d eism
istr th Dep ogr
Sei ibu Dep ection th c am
sm tion onv s
r
ic p
roc cor ers
ion
ess
ing Gr
sts ma avity a
gne nd
tio n te t ysis
duc Env ic data nal
Pro p s iron te a
c r o
orr menta ibu
Out on
c
ect l Attr
h oriz B i ion
Seq ic ostr
uen S eism aps atig
ce m raph
stra Flu y
tigr
aph ana id
y lysi
s
ata
Nor ed
ma
liza Cor
Log tion
cor
rela sics
tion
s eophy
le g ll
-we odel
Spe eho gle
Pro cial Bor Sin sical m
pag cor hy
ate ea rop
m
fiel odel
nal
ysis pet
ing
d w thr
ells oug app
h i c al m
log
Geo

20 Oilfield Review
 Validating Reservoir Models to Improve Recovery

Valid and worthwhile reservoir simulation hinges on careful preparation of a

reservoir model and clear understanding of model uncertainty. Creative integration
and comparison of seemingly disparate data and interpretations elevate reservoir
models to new heights of reliability. Through this process, all data incorporated
into a model and ultimately fed into a simulator become more valuable, resulting
in increases on the order of 10% in predicted ultimate recovery.

Jack Bouska Mike Cooper Chip Corbett Alberto Malinverno Sarah Ryan
BP Amoco plc Andy O’Donovan Houston, Texas, USA Michael Prange Cambridge, England
Sunbury, England BP Amoco plc Ridgefield,
Aberdeen, Scotland Connecticut, USA

For help in preparation of this article, thanks to Ian Bryant, Making and testing predictions are part of our Working together, skilled interpreters use a
Schlumberger-Doll Research, Ridgefield, Connecticut, USA;
Henry Edmundson, Schlumberger Oilfield Services, Paris,
everyday existence and basic to most industries. simulator to predict reservoir behavior over time
France; Omer Gurpinar, Holditch-Reservoir Technologies, Safety equipment, medical treatments, weather and optimize field development strategies accord-
Denver, Colorado, USA; Steve McHugo, Geco-Prakla,
Gatwick, England; Claude Signer and Lars Sønneland,
forecasts and even interior designs are evaluated ingly. For instance, the effectiveness of infill
Geco-Prakla, Stavanger, Norway; and James Wang, by simulating situations and predicting the drilling locations and trajectories can be deter-
GeoQuest, Houston, Texas. We thank the Bureau of
Economic Geology, The University of Texas at Austin, for
results. Similarly, the oil and gas industry makes mined through simulations of multiple scenarios
permission to use the Stratton field 3D Seismic and Well predictions about hydrocarbon reservoirs to or assessment of the impact of the uncertainty of
Log Data Set, and BP Amoco and Shell UK E&P for
permission to use the Foinaven area data.
decide how to improve operations. specific parameters. Reservoir simulation is also
ARI (Azimuthal Resistivity Imager), FMI (Fullbore Reservoir optimization requires carefully useful in evaluating different completion tech-
Formation MicroImager), MDT (Modular Formation constructing a reservoir model and performing niques as well as deciding whether to maximize
Dynamics Tester), OFA (Optical Fluid Analyzer), RST
(Reservoir Saturation Tool) and TDT (Thermal Decay Time) simulations. Interpreting and integrating quality- production rate or ultimate recovery. In this arti-
are marks of Schlumberger. controlled data from a variety of sources and vin- cle, we consider how the integration of all avail-
tages, and at different scales, are prerequisites able data to validate and constrain reservoir
for preparing a comprehensive reservoir model. models leads to more realistic reservoir
Most computer simulators take the reservoir simulation (next page).
model and represent it as three-dimensional
blocks through which fluids flow (previous page).
The data, models and simulations provide a more
complete understanding of reservoir behavior.

Summer 1999 21
 In some cases, it is best to begin with the sim-
Traditional Approach Leading-Edge Approach
plest model that fits the data and the objectives of
Distributed disciplines Multidisciplinary teamwork the project and reproduces reservoir behavior. In
Use of data in isolation, obscuring relationships Integration of data and interpretations to confirm all cases, the starting point should be an evalua-
between data (for example, seismic and core data) reservoir models tion of what answers are required from reservoir
Inconsistent or poorly documented Archiving of interpretations and consistent methods simulation, the accuracy needed and the level of
interpretation techniques confidence or the acceptable range of quantita-
Overdependence on simple reservoir maps Seismic-guided reservoir property mapping tive predictions. The model complexity might be
increased as more data become available. The
Simulation dependent on computer availability Simulations run on personal computers
and capability or using massively parallel processing reward for increasing model complexity can be
evaluated after each simulation run to decide
Unlimited modification of simulation input values Reservoir models constrained by integrated data and
to achieve match with production history interpretations and prudent adjustment of inputs whether more complex simulation is justified.
Estimates of well flow rates and predictions
Limited use of simulation to guide data acquisition Modeling and simulation to determine optimal
timing for data gathering, such as 4D seismic surveys of reservoir performance from simulations affect
design of production facilities and should be
> Simulation approaches. In the past, single-discipline interpretation and lack of computing capability believed, even if they seem unlikely. For example,
limited the use of reservoir simulation. Now, a more sophisticated approach to simulation makes the a deepwater Gulf of Mexico field required expan-
most of multidisciplinary teams and nearly ubiquitous computers. sion and de-bottlenecking of facilities soon after
initial production because the initial reservoir
Reservoir simulation is a tool for reservoir These applications of simulation are made model was compromised by a pessimistic view of
management and risk reduction.1 Although the possible by new programs and computers that the interpreted reservoir continuity and flow
first simulations were performed during the 1950s, are faster and easier to use. (A full review of the rates. Better predictions allow operators to size
for a long time limited computer availability and advances in simulation software that have facilities correctly the first time rather than having
slow speed confined their use to only the most occurred in the last few years is beyond the to re-engineer them.
significant projects.2 scope of this article, but will be covered in a The quality of predevelopment reserve esti-
At present, reservoir simulation is performed future article in Oilfield Review.) The new simu- mates, field appraisals and development strate-
most commonly in high-risk, high-profile situa- lators run on less expensive computers and allow gies relates closely to reservoir architecture and
tions, but could improve virtually any project. The rapid studies to rank opportunities. Along with structural complexity; reserve estimates tend to
list of typical applications is varied and extensive these capabilities, however, arises the possibility be underestimated in large, less complex fields,
(next page): that simulations might be performed indiscrimi- whereas reserves in smaller, more complex fields
• New discoveries, to determine the number of nately or before a validated reservoir model has are commonly overstated. Poor reservoir models
wells and the type and specification of facilities been built, potentially prompting misleading or and resultant incorrect calculations of reserves,
needed for production. Particular attention is erroneous results and poor decision-making. whether too high or too low, have negative eco-
paid to the reservoir’s drive mechanism and the There is also the risk of performing simulations nomic consequences. In the North Sea, deficient
development of potential oil, gas and water pro- based on limited data. reservoir models have led to improper facilities
files. All assessments are subject to the risk of Developing a first-rate reservoir model from sizing and suboptimal well placement, even in
limited data, sometimes from only a single well. limited data at a variety of scales is difficult. In fields where simulation studies were carried out.3
• Deepwater exploration and other areas where its most basic form, model validation is achieved Better validation of models, particularly using 3D
initial test wells are expensive. Estimates draw through integrating different types of data. seismic data, might have averted over- or under-
on restricted data, such as seismic data and Researchers are investigating the best way to sizing production facilities or drilling unnecessary
results from a single well. integrate some new types of data, such as multi- wells in some cases. In other cases, reservoir
• Fields in which production surprises occur and component seismic data, into reservoir models. simulation has allowed identification of the key
development expenditures have already been A more sophisticated approach involves uncer- drivers of reservoir performance so that data-
incurred. New measurements or production tainty analysis (see “Validating Models Using gathering efforts can be targeted to reduce
strategies might be advisable. Diverse Data,“ page 24 ). uncertainty in those areas. Alternatively, facili-
• Secondary recovery implementation. Appro- ties can be designed to be flexible within a given
priate decisions are essential because of the reservoir uncertainty.
expense of enhanced production startup.
• Divestment and abandonment decisions.
Simulation can help determine whether a field
has reached the end of its economic life or
how remaining reserves might be exploited.

22 Oilfield Review
Increasing the Value of Data and trajectories. The reservoir model is strength- Seismic and well test data enable mapping of
Operating companies spend considerable time ened if a geological map of permeability values is permeability barriers, but are rarely used in tan-
and money acquiring data—from multimillion- created by applying a porosity-to-permeability dem. For example, horizon dip, azimuth, coherency
dollar seismic surveys and cores from costly transform to the porosity map according to per- or other seismic attributes might indicate fault
exploratory wells, to sophisticated well logs and meability values interpreted from well tests, well patterns.4 Such information is especially useful
production tests during and after drilling. Data logs or cores. Even more rigorous results are when contemplating the addition of directionally
acquisition presents potential risks—to both obtained when, in addition to inclusion of well drilled or multilateral wells. These types of inter-
project economics and the well itself—such as rates and produced or producible hydrocarbon pretations are just the beginning; all other data
logging or testing tools becoming stuck, a core volumes, all available production data are input types should be similarly scrutinized.
barrel malfunctioning or having to shut in or kill a into the model. These include pressure, gas/oil As mentioned earlier, the reliance of simula-
producing well. One would expect, then, that ratios, and fluid densities, saturations, viscosities tors on four simple subsurface maps has
data would be analyzed and incorporated into and compressibilities. impaired simulation effectiveness. Simulation
models as fully as possible or not collected in the In many instances, though, reservoir models becomes more realistic as additional data are
first place. Most reservoir simulations rely heav- fail to encompass the full diversity of reservoir incorporated into the reservoir model—reconcil-
ily on production data from wells and only four data because only a few basic geological and ing all available data tends to rule out some
types of geological or geophysical reservoir geophysical maps, constructed from a subset of interpretations. For example, permeability values
maps: structure of the top of the reservoir, reser- the data available, are used to describe varia- can be inferred from well logs and confirmed by
voir thickness, porosity and the ratio of net pay to tions in the data. Additional data and interpre- core and well test data, and possibly related to
gross pay. These maps are often constructed tations are needed to make reservoir models seismic attributes, rather than merely computed
from seismic and well log data alone. more robust. For example, core data can serve as from an empirical transform of a porosity map
Incorporating all available data, such as core calibrators for geological, petrophysical and and well test data. Reconciling conflicting data
analyses, seismic-guided reservoir property dis- engineering data and interpretations, but are requires acceptance of a hierarchy of data confi-
tributions and fluid analyses, is a cost-effective often used only as guides to permeability. Core dence. This hierarchy might be developed on the
way to strengthen and validate reservoir models analysis refines model values of porosity, perme- basis of probable measurement errors.
across disciplines. ability, capillary pressure and fluid saturation. (continued on page 26)
A reservoir model usually combines produc- Whole cores, while not necessarily represen-
1. For a general introduction to reservoir simulation:
tion rates and volumes with geological and geo- tative of the entire reservoir, offer tangible Adamson G, Crick M, Gane B, Gurpinar O, Hardiman J
physical maps of subsurface strata derived from evidence of grain size and composition, sorting, and Ponting D: “Simulation Throughout the Life of a
Reservoir,” Oilfield Review 8, no. 2 (Summer 1996): 16-27.
well logs and seismic data. Aquifers are often depositional environment and postdepositional 2. Watts JW: “Reservoir Simulation: Past, Present, and
included in the model and sealing rocks are typi- reservoir history, such as bioturbation, cementa- Future,” paper SPE 38441, presented at the SPE
cally treated as zero-permeability layers. The tion or diagenesis. Reservoir Simulation Symposium, Dallas, Texas, USA,
June 8-11, 1997.
subsurface maps take into account well locations 3. Dromgoole P and Speers R: “Geoscore: A Method for
Quantifying Uncertainty in Field Reserve Estimates,”
Petroleum Geoscience 3, no. 1 (February 1997): 1-12.
Situation Desired Results Pitfalls or Other Considerations 4. Key SC, Nielsen HH, Signer C, Sønneland L, Waagbø K
and Veire HH: “Fault and Fracture Classification Using
Artificial Neural Networks – Case Study from the Ekofisk
New discoveries Determine optimal number of Limited data, sometimes from
Field,” Expanded Abstracts, 67th SEG Annual
infill wells only a single well
International Meeting and Exposition, Dallas, Texas, USA,
November 2-7, 1997: 623-626.
Size and type of production Drive mechanism
facilities

Decide whether to maximize Terms of operating license

production rate or or lease
ultimate recovery

Deepwater exploration Prospect evaluation Limited data, no wells available

Scenario planning

Mature fields Answers to sudden Relatively inexpensive way

production problems to extract maximum value
from development costs

Implementation of secondary Determine appropriate

recovery recovery method

Divestment or abandonment Determine future Unanticipated future produc-

production volumes tion problems might reduce
property value

> Simulation uses. Reservoir simulation is useful during all phases of the life of a reservoir and in
both high- and low-risk projects.

Summer 1999 23
 Validating Models Using Diverse Data

A shared earth model is a model of the geome-

try and properties of the reservoir constrained
by a variety of measurements. To be predictive,
the model should approximate the key features
of the actual reservoir as closely as possible
(right). In a valid reservoir model, predictions
from the model agree with the measured data. A
Reservoir Model Measured Data
good fit between predictions and measurements
is not sufficient, though. Several models might
agree equally well with the data. The best model
is the one that agrees best with the data and Seismic data
with prior information on the model parame- Well log data
ters. The uncertainty of the model is defined as Dynamic data
the range of model parameters consistent with
the measurements.
Shared Earth Model Predicted Data
Consider a thin bed imaged by seismic data
(below right). The uncertainty of the shared > Model inputs. A shared earth model begins with seismic, well log and dynamic data from the
earth model in this case is described by the actual reservoir. In this example, the reservoir is represented by a model (top left). Measured
range of thickness and impedance values that seismic data (top right) are compared with data predicted by the model (bottom right),
which can be adjusted to improve the fit. The final three-dimensional shared earth model
satisfy the data. This range defines a probability (bottom left) incorporates all available data.
density function (PDF).

ρV
P

Uncertainty ellipse
Poor fit
ρV
P
h
Good fit

Best fit
h

> Quantifying uncertainty. The thin bed shown as a red layer (left) has thickness h
and acoustic impedance ρVP. The plot to the right displays the posterior probability
density function. Thickness and impedance values within the red uncertainty ellipse
satisfy the data, and within that ellipse are red circles denoting a good fit and the
best fit. The red circle outside the uncertainty ellipse does not satisfy the model.

24 Oilfield Review
 Researchers at Schlumberger-Doll Research,
Ridgefield, Connecticut, USA, are using a m2 m2 m2
Bayesian approach to quantify the uncertainty Uncertainty
ellipse
of reservoir models (right). A prior PDF
Best
represents initial information on model para- model
meters—the vector m. This prior PDF can
be combined with a likelihood PDF, which m1 m1 m1
quantifies information provided by additional
Prior PDF Likelihood Posterior PDF
data, to obtain a posterior PDF. When new data p (m) L (m d[1]) p (m d[1])
become available, the posterior PDF is used as
the initial or prior PDF, and the model is again m2 m2 m2
refined. As additional measurements are incor-
porated in the model, the uncertainty decreases
and a better reservoir description follows.
Effective model validation using a Bayesian
approach requires three modes of operation: m1 m1 m1
interactive, optimization and uncertainty. In the Prior PDF Likelihood Posterior PDF
interactive mode of the prototype application p (m d[1]) L (m d[2]) p (m d[1],d[2])
in development, the user modifies the reservoir
model and observes the consequences of inter- > Reducing uncertainty. In a Bayesian approach, a prior PDF quantifies the initial
pretation decisions on the data (below right). information on model parameters, expressed as vector m. The prior PDF (top left) is
In the example shown, predicted seismic data refined by the inclusion of new data (top center) to create the posterior PDF (top right).
The uncertainty of the model is shown in the red uncertainty ellipse. The blue circle
are compared with measured data. In the opti- represents the best model. The posterior PDF then becomes the prior PDF (bottom left)
mization mode, the user selects the model when more data become available (bottom center). The next posterior PDF
parameters to optimize. The software finds the (bottom right) has a smaller uncertainty ellipse and a slightly different optimal model.
best local fit of the model to the data. Finally,
in the uncertainty mode, the uncertainty ellipse
of selected reservoir properties is computed and
displayed. The uncertainty ellipse represents
the range of acceptable models.
The prototype application has been used to
test a Bayesian validation approach against
diverse data types, including seismic data, well
logs, while-drilling data and production infor-
mation. By validating a reservoir model against
all available data before beginning the history-
matching phase, the range of admissible models
can be reduced substantially. The result is a
more predictive reservoir model.

> Validation modes. Prototype software developed by Schlumberger includes an interactive mode in which the
user assesses the effects of interpretation decisions on reservoir models. In this case, the center of the upper
panel shows predicted seismic data as dotted lines and measured data as solid lines after the upper horizon,
shown in green to the left, has been moved. The lower panel shows a better fit between the predicted and
measured data (center) and the model uncertainty in the ellipse to the right.

Summer 1999 25
Limitations of Reservoir Models conduits might be miscalculated. Formation of the reservoir model. Horizontal upscaling in the
Generating and fine-tuning the model entail close thicknesses are usually defined by integrating absence of horizontal wellbores is typically simpler
collaboration by the reservoir team. As in other seismic and averaged well log data, although because there is generally less fine detail in seis-
phases of exploration and production, such as the resolution of seismic data is on the order of mic data, whereas vertical upscaling is compli-
geological and geophysical interpretation or tens to hundreds of feet, whereas well logs show cated by the greater amount of detail available at
drilling preparations, handing off results from variations at the scale of inches. Under- or over- the wellbores. Thickness and porosity, whose vari-
one team member to the next along a chain is estimating pay thicknesses directly impacts ations typically follow simple, linear averaging
less effective than working together from the simulation reliability. laws, are less prone to upscaling problems than
outset.5 Reservoir teams analyze data and per- Averaging techniques also affect simulation permeability. The ”average“ permeability of a two-
form simulations more rapidly as their experience results, especially when reservoir properties are layer system in which one layer has zero perme-
increases. Working as a team also ensures that highly variable. Also, problems may occur when ability is not one layer with the average
no one gets bogged down in endless tinkering averaging fine detail, such as interpretations permeability of the two layers. The reservoir model
with input parameters to try to obtain a match from well logs, to integrate with data of lower must be built around such impermeable layers.
with production history. resolution, such as seismic data. For example, a No matter how carefully a model is prepared
In addition to working interactively, the team reservoir that consists of several distinct layers and simulation performed, the dynamics of pro-
must employ consistent methods to ensure that with different properties might not behave like a duction might affect the reservoir in ways that
normalization and interpretation are performed single layer of the same overall thickness and reservoir simulation might not predict. History
properly. If data are not normalized and inter- average properties. The uncertainty of many matching, or comparing actual production vol-
preted consistently, relationships between data measurements increases dramatically with dis- umes and measured pressures with predictions
might be obscured, such as that between porosity tance from the wellbore. Even though there is a from simulations, is the most common method
and seismic attributes (see “Model Validation,” different level of uncertainty with each data type, for judging the quality of the reservoir model. The
next page).6 proper model validation forces comparisons of assumption is made that if the model yields a
Any uncertainty in the data limits confidence independent data and interpretations. simulation that matches past production, then
in reservoir models and reservoir simulations. Upscaling, or representing the data at a com- the model is more likely to be a useful tool for
Permeability barriers, pinchouts, faults and other mon scale, coarsens the fine-scale reservoir forecasting.7 Certainly, a model that does not
geological features are not always apparent from description in the shared earth model to the degree match past production history or reservoir
well, seismic and production data. Their exact that a computer can cope with it (below). This step response to past production is unlikely to cor-
locations might be off by tens or hundreds of usually reduces the number of cells, or subdivisions rectly predict future production.
meters and their effectiveness as flow barriers or

Simulation model

Upscaling

Geological modeling Well logs 3D and 4D seismic data Drilling data Classification system Reservoir simulations
Petrophysical modeling Seismic modeling

> Shared earth model. A numerical representation of the subsurface, housed in a database shared by multidisciplinary team members, allows constant
access to and updating of the reservoir model used for simulation. As databases and software improve, the simulation model and the shared earth model,
which now must be upscaled before being used as a simulation model, will be the same.

26 Oilfield Review
 VSP well tie Time horizon Attribute extraction
interpretation

Data loading Depth grids Reservoir property Model construction

distribution

Petrophysical Geologic Weighted average

interpretation correlation

> Model construction workflow. Once data from the Stratton field were loaded, the team worked together from the outset to correlate well logs,
vertical seismic profile (VSP) data and seismic data. Interpreted seismic horizons, depth conversion results and extracted attributes were compared with
normalized well log porosities and geologic log correlations. The consistent relationship between the weighted average porosity and seismic amplitude
prompted generation of a reservoir property distribution map, a seismic-guided map of porosity distribution in this case, to complete the reservoir model.

Obtaining a good match between the produc- Model Validation 5. Galas C: “The Future of Reservoir Simulation,”
Journal of Canadian Petroleum Technology 36, no. 1
tion history and predictions from simulations is A data set from the Stratton field of south Texas (January 1997): 5, 23.
inexpensive in some cases, but can become time (USA) demonstrates the value of cross-disciplinary 6. Corbett C: “Improved Reservoir Characterization Through
consuming when the model is continuously interpretation and model validation in calculating Cross-Discipline Multiwell Petrophysical Interpretation,”
presented at the SPWLA Houston Chapter Symposium,
refined and simulated. In certain situations, such in-situ gas reserves in the Frio formation (above).8 Houston, Texas, USA, May 18, 1999.
as waterfloods, tracers in the form of chlorides, The data include 3D seismic data, logs from nine 7. This assumption does not always hold. For example,
a reservoir model might match the production history
isotopes or brines are introduced into injected wells, correlations of geological markers and a even when there is bypassed oil. Additional seismic
water to reveal patterns in the reservoir. vertical seismic profile (VSP). Resistivity, neutron data might reveal undrained reservoir compartments
in this case.
Comparisons of these patterns with expected porosity, bulk density, and spontaneous potential,
8. Corbett C, Plato JS, Chalupsky GF and Finley RJ:
patterns can be used to reevaluate input values, gamma ray or both curves were available for the “Improved Reservoir Characterization Through Cross-
for example, porosity, permeability and transmis- nine wells. Preliminary examination of the well Discipline Multiwell Petrophysical Interpretation,”
Transactions of the SPWLA 37th Annual Logging
sibility—the ease with which fluid flows from logs and VSP data guided selection of horizons in Symposium, New Orleans, Louisiana, USA, June 16-19,
one model cell to another, to improve the history the Frio formation for seismic horizon tracking. 1996, paper WW.
9. Phase refers to the motion of, or means of comparison
match. Whenever a new well is drilled, it offers The VSP provided a good tie between the well of, periodic waves such as seismic waves. Waves that
an opportunity to check the quality of a reservoir and the seismic data along with good under- have the same shape, symmetry and frequency and that
reach maximum and minimum values simultaneously
simulation, principally by comparison of observed standing of the phase of the seismic data.9 are in phase. Waves that are not in phase are typically
pressure with the pressure predicted by the A thin, clean Frio sand that is easy to correlate described by the angular difference between them, such
as “180 degrees out of phase.” Zero-phase wavelets
model at the drilling location. and ties to a mappable seismic event was are symmetrical in shape about zero time whereas non-
The difficulty of simulating a reservoir selected for both well-by-well analysis and multi- zero-phase wavelets are asymmetrical. Non-zero-phase
wavelets are converted to zero-phase wavelets to
underscores the need to constrain the reservoir well petrophysical interpretation that ensured achieve the best resolution of the seismic data. Known
model with all available data. A reservoir model consistent analysis of all the logs. The interpreters (zero) phase well synthetics and vertical seismic profiles
(VSPs) can be compared with local surface seismic data
constrained and validated by geological, geo- observed that porous zones seemed to correspond to determine the relative phase of the surface seismic
physical and reservoir data before initiating wavelets. Such knowledge allows the surface seismic
data to be corrected to zero phase.
simulation extracts as much information as pos-
For more on combining vertical seismic profiles with
sible from the data and provides a better result. other geophysical data: Hope R, Ireson D, Leaney S,
Also, understanding the range and impact of Meyer J, Tittle W and Willis M: “Seismic Integration
to Reduce Risk,” Oilfield Review 10, no. 3
reservoir uncertainty allows a quantitative and (Autumn 1998): 2-15.
qualitative judgment of the accuracy or range of
model predictions.

Summer 1999 27
 to high seismic amplitude. To confirm this obser-
vation, crossplots of effective porosity and ampli-
tude were prepared. The crossplots of the
Well 18 well-by-well petrophysical analysis showed sig-
Well 7
0.10 nificant scatter, whereas the multiwell analysis
Well 13
Well 20 demonstrated a clear relationship between seis-
0.09 mic amplitude and effective porosity (left).
Well 19 Next, an equation that related effective
Single-well effective porosity

0.08 porosity to amplitude was used to generate a

Well 11 map of effective porosity. The mathematical rela-
0.07 tionship between the weighted average porosity
values at the wellheads and the seismic ampli-
Well 12
0.06 tudes at those locations guided the mapping.
Well 9
Combining carefully integrated core porosity, log-
0.05 derived porosity and seismic attributes—in this
case, amplitude—produced a single, validated
0.04 porosity map constrained by several independent
Well 10
sources of porosity information (next page).
Using each type of data in isolation in the
30 35 40 45 50 55 60 65
Stratton field example obscured relationships
Amplitude between data and probably would have resulted
in a set of incompatible subsurface maps that
were not physically realistic.
The difference between the single-well ana-
0.17 lytical approach and the consistent, normalized
Well 10 petrophysical analysis in the Stratton field
0.16 affects the economic evaluation of the reservoir.
Well 19
Well 9 The single-well approach precluded integrating
0.15 the well logs with the seismic data to generate a
Well 13
seismic-guided porosity map because the cross-
Multiwell effective porosity

Well 12
0.14
plot of effective porosity and amplitude indicated
no consistent relationship between the well logs
0.13
and seismic data. The in-situ gas volume calcu-
lated by single-well petrophysical analysis is
0.12
12% greater than that calculated from the vali-
Well 20
dated, seismic-guided porosity distribution. An
0.11
overstated gas volume might lead to unnecessary
Well 11 infill drilling.
0.10
In another case offshore Malaysia, 3D seis-
Well 7
Well 18 mic data, well logs, wellbore image logs and
0.09
30 35 40 45 50 55 60 65
core data enabled generation of time-depth
Amplitude
relationships and synthetic seismograms to tie
logs to seismic data.10 The relationship between
effective porosity, seismic amplitude and acous-
> Single well versus multiwell interpretation. Well-by-well petrophysical analysis tic impedance, expressed as a calibration func-
(top) obscures the relationship between porosity and seismic amplitude. In this tion, allowed prediction of effective porosity
example from the Stratton field, the plot of effective porosity versus seismic
amplitude shows considerable scatter around the line of best fit because the well throughout the 3D seismic data, similar to the
logs were not analyzed consistently. The relationship between seismic amplitude previous Stratton field example. Additional
and porosity is clear when the logs are normalized and consistent analytical data, such as pressure measurements from
methods are used (bottom). The observed relationship between seismic
wireline tools or well tests, make the reservoir
amplitude and effective porosity allowed interpreters to use the seismic data
to generate a map of effective porosity. model more robust and improve confidence in
the predictions from simulation.11

28 Oilfield Review
 7000

6000
100
5000 90
Well 9 Well 7
80
Well 20
4000
70
60
3000 Well 12 Well 11
50
0.040 Well 18
40 Well 10
0.060 2000
0.080 Well 13
0.100 30
0.120 20
0.140 1000
0.160 10
0.180 Well 19
0.200 0
0.220 20 40 60 80 100 120 140 160 180 200

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10,000 11,000

> Seismic-guided porosity distribution. In the Stratton field of south Texas, USA, a clear relationship between effective
porosity and seismic amplitude permitted seismic-guided mapping of effective porosity. This map could not have been
created without consistent, multiwell petrophysical analysis. Yellow and orange represent areas of high seismic
amplitude; blue represents low amplitude.

Model Manipulation Simulation experts use a three-stage related. Poor-quality seismic data, a common
Because simulation inputs are subject to revision approach to fine-tune a reservoir model, begin- problem in structurally complex areas, can
by the project team to improve the match between ning with the energy balance, then an adjust- hamper horizon tracking. Reprocessing can
the simulation and production history, it is impor- ment for multiple fluid phases, and finally the improve seismic data quality.
tant to restrict the input model as much as the well productivity. The energy balance stage The depth to the reservoir should also be well
data permit and avoid unnecessary adjustments of accounts for the reservoir pressure. The relative constrained if log and seismic data are inter-
input values. Simulation software typically allows permeabilities of different fluid phases are preted diligently. Comparing well logs or syn-
interpreters to change not only the geological and adjusted in the second stage. The final step uses thetic seismograms generated from logs with
geophysical maps used to build a reservoir model, recent productivity test data, such as bottomhole seismic data improves depth conversion.
but also variables such as pressure, temperature, flowing pressure, tubing surface pressure and Additional data, such as VSPs, also tend to
fluid composition and saturation, permeability, total fluid production rate, to further improve the improve depth conversion. In structurally com-
transmissibility, skin, productivity index and rock history match. plex areas, however, depth-based processing
compressibility. Seasoned interpreters have differ- Reservoir thickness values typically are con- from the outset is preferable to depth conversion.
ent opinions about what changes to simulation strained by seismic data and well logs, but are The enhanced integrity that validation brings to a
inputs are acceptable, but prudently adjusting wrong if the interpreter tracks seismic horizons depth-converted structure map—that is, a map
simulation input parameters often improves the incorrectly, if logs and seismic data are not tied displayed in units of depth rather than the seis-
history match. properly, or if well logs are off-depth or miscor- mic unit of time—is demonstrated by integrating
10. Corbett C, Solomon GJ, Sonrexa K, Ujang S and Ariffin T: dipmeter data with the structure map or by com-
“Application of Seismic-Guided Reservoir Property paring depth-converted seismic sections to dip
Mapping to the Dulang West Field, Offshore Peninsular
Malaysia,” paper SPE 30568, presented at the SPE
Annual Technical Conference and Exhibition, Dallas,
Texas, USA, October 22-25, 1995.
11. For another example of seismic-guided property mapping:
Hart BS: “Predicting Reservoir Properties from 3-D
Seismic Attributes With Little Well Control—Jurassic
Smackover Formation,” AAPG Explorer 20, no. 4
(April 1999): 50-51.

Summer 1999 29
 interpretations from wellbore images, such as
FMI Fullbore Formation MicroImager or ARI
Azimuthal Resistivity Imager logs (left). By inter-
preting seismic data, well logs and wellbore
images together rather than independently, the
interpreter ensures that the final structure map
has been rigorously checked.
Though typically not introduced in the large-
-800
scale model-construction phase, information
about reservoir fluids can offer important insight.
-1200
-800 Formation tester data, such as MDT Modular
-1600
Formation Dynamics Tester results, indicate the
-1000
location of a fluid contact. This information, com-
-1400 -1800 bined with well log and seismic data, yields a
more constrained starting model for simulation.
-1200 Structure contours in meters Other fluid information may be used to con-
Well location strain the fine-scale model in the vicinity of the
Strike and dip from dipmeter wellbore. Perforation locations are considered
known, but the effectiveness of the perforations
may be evaluated with production logs, and
changes in fluid saturations monitored with RST
Depth, ft 415 413 412 410 408 406 404
1443 1443 1443 1443 1443 1443 1443
Reservoir Saturation Tool or TDT Thermal Decay
Time data.
11,050 18,000 The ratio of net pay to gross pay can vary
11,150 21,000 widely across a reservoir, but like other simula-
11,250 24,000 tion input values, should not be altered during
27,000 the history-matching stage without good reason.
11,350
30,000 The net-to-gross ratio might be adjusted if sup-
11,450
33,000 ported by drilling results, such as well logs and
11,550 36,000 cores, or production logs.
11,650 39,000 Permeability values are obtained in several
42,000 ways, including core and log analysis and well
11,750
45,000 tests, so comparisons of the values from each of
11,850 these approaches can limit the range of input val-
48,000
11,950 51,000 ues, at least at well locations. Effective assimila-
12,050 54,000
tion of wellbore image logs, probe permeability
data and core data allows characterization of
12,150
horizontal permeability near the wellbore and
T67V f/s*g/cm3 prediction of vertical permeability.12 Saturation
values, established through well log analysis, are
> Confirming depth conversion. Dipmeter data reduce interpretive contouring options for structure maps verified by capillary pressure data from special
if the mapper honors the data (top). Dipmeter data from the depth of interest, plotted at each well, show core analysis, wireline formation tester results or
reasonable conformity with structure contours in the upper right and lower sections of the map, but RST measurements. All of these input parame-
refute the contouring of the upper left area in this fictitious example. Dip interpretation from an image log, ters, within reason, are considered adjustable by
tied to an actual depth-converted seismic section, confirms dip direction and magnitude at horizons of
interest (bottom). The color variation in the seismic section represents acoustic impedance. simulation experts.
12. Thomas S, Corbett P and Jensen J: “Permeability and
Permeability Anisotropy Characterisation in the Near
Wellbore: A Numerical Model Using the Probe
Permeameter and Micro-Resistivity Image Data,”
Transactions of the SPWLA 37th Annual Logging
Symposium, New Orleans, Louisiana, USA,
June 16-19, 1996, paper JJJ.
13. Crombie A, Halford F, Hashem M, McNeil R, Thomas EC,
Melbourne G and Mullins OC: “Innovations in
Wireline Fluid Sampling,” Oilfield Review 10, no. 3
(Autumn 1998): 26-41.
14. For more on skin: Hegeman P and Pelissier-Combescure J:
“Production Logging for Reservoir Testing,”
Oilfield Review 9, no. 2 (Summer 1997): 16-20.

30 Oilfield Review
 Some reservoir engineers minimize adjust-
ments to PVT samples, which indicate reservoir
fluid composition and behavior at the pressure,
volume and temperature conditions of the reser- . . . Validating interpretations and models across disciplines
voir. In cases of surface recombination of the
sample or sample collection long after initial pro- addresses complex problems that are difficult to solve
duction, however, the engineer might decide to
adjust PVT values. At the other extreme, produc- within the confines of a single discipline . . . Multidisciplinary
tion rates and volumes, and pressure data from
wells are considered inalterable by some validation of reservoir models increases the value of
experts, although exceptions are made at times,
such as when production measurement equip- data beyond the cost of data-gathering activities alone . . .
ment fails. Many experts choose to honor the
most accurate representation of production data.
Placing restrictions on the alteration of input
values makes a good history match from simula-
tion more elusive, but many input values may be Adjustment of the productivity index, another Multidisciplinary validation of reservoir mod-
adjusted during simulation. Transmissibility is input parameter, affects the quality of a history els increases the value of data beyond the cost of
computed by the simulator using the input poros- match. The productivity index, often expressed in data-gathering activities alone. In 1998, for
ity and permeability. A high computed transmis- units of B/D/psi or Mcf/D/psi, is a measure of example, Geco-Prakla acquired 3D multicompo-
sibility value can be overridden if well tests, how much a well is likely to produce. If the skin nent seismic data for Chevron in the Alba field in
formation tester data or seismic data provide evi- value is known, the productivity index—usually the North Sea. The objectives of the survey were
dence of separate sand bodies, stratigraphic computed from model inputs that include the to better image the sandstone reservoir, identify
changes, faults or other types of reservoir com- skin—can be computed more accurately. When intrareservoir shales that affect movement of
partmentalization. Differences in fluid chemistry differing stimulation or completion techniques injected water and map waterflood progress.
or pressure from one well to another also sug- among field wells are used, productivity index After integration of the new shear-wave data to
gest reservoir compartmentalization. In-situ fluid values often vary from well to well. For example, improve the reservoir model, two additional
samples obtained from the OFA Optical Fluid hydraulic fracturing of a single well in a field might wells were drilled in the field. The first well is
Analyzer component of the MDT tool are uncon- enhance permeability and, therefore, productivity producing up to 20,000 B/D [3200 m3/d]; the sec-
taminated and can be brought to surface without of that well alone. ond well is being completed and has resulted in
changing phase for chemical analysis.13 The options available to change a reservoir the discovery of Alba’s highest net sand. Both
Production logs, well tests and pressure tran- model to improve the match between one simu- wells have confirmed some of the features
sient analyses indicate skin, which is a dimen- lation run and a field’s production history might observed on the converted-wave data. Because
sionless measure of the formation damage appear endless. At some point, practical limits on the first well was drilled less than a year after
frequently caused by invasion of drilling fluids or data collection, computational power and time seismic acquisition started, Chevron felt the new
perforation residue.14 When the location, pene- for modifying input parameters curtail simulation data arrived in time to make a significant com-
tration and effectiveness of perforations are of iterations. Independent analyses that support the mercial impact on the field’s development.17
concern, production logs provide information that interpretations of other team members increase 15. For more on the shared earth model and integrated
may affect the model input for skin. If a field is confidence in reservoir simulation. A proper sim- interpretation: Beardsell M, Vernay P, Buscher H,
Denver L, Gras R and Tushingham K: “Streamlining
located in a geological trend of similar accumu- ulation workflow helps accomplish this goal. Interpretation Workflow,” Oilfield Review 10, no. 1
lations, skin values in the trend might be a useful Working as a team ensures that all data are used (Spring 1998): 22-39.
starting assumption if data within the field are to validate the reservoir model. 16. Major MJ: “3-D Gets Heavy (Oil) Duty Workout,” AAPG
Explorer 20, no. 6 (June 1999): 26-27.
initially scarce. Validating interpretations and models across O’Rourke ST and Ikwumonu A: “The Benefits of
disciplines addresses complex problems that are Enhanced Integration Capabilities in 3-D Reservoir
Modeling and Simulation,” paper SPE 36539, presented
difficult to solve within the confines of a single at the SPE Annual Technical Conference and Exhibition,
discipline. A multidisciplinary team sharing a Denver, Colorado, USA, October 6-9, 1996.
database and iteratively validating and updating Sibley MJ, Bent JV and Davis DW: “Reservoir Modeling
and Simulation of a Middle Eastern Carbonate
shared earth models, or geomodels, achieves this Reservoir,” SPE Reservoir Engineering 12, no. 2
goal.15 Operating companies report increases on (May 1997): 75-81.
17. For more on the Alba field survey and shear wave
the order of 10% in predicted ultimate recovery seismic data: MacLeod MK, Hanson RA, Bell CR and
through proper data integration, simulation and McHugo S: “The Alba Field Ocean Bottom Cable Seismic
Survey: Impact on Development,” paper SPE 56977,
reservoir development.16 Cycle time also prepared for presentation at the 1999 Offshore European
decreases in many cases, probably because of Conference, Aberdeen, Scotland, September 7-9, 1999.
ready access to data and interpretations for Caldwell J, Christie P, Engelmark F, McHugo S, Özdemir H,
Kristiansen P and MacLeod M: “Shear Waves Shine
everyone involved in the project. Brightly,” Oilfield Review 11, no. 1 (Spring 1999): 2-15.

Summer 1999 31
Modeling for Data Acquisition t
2 Predicted Seismic Properties t
3
In addition to the general question of determin-
ing how best to produce a reservoir, simulation
can demonstrate the best time to acquire addi-
tional data. Time-lapse (4D) seismic surveys,
which are repeated 3D surveys, can be acquired
at optimal times predicted by careful model con-
struction and simulation.18 As oil and gas are pro-
duced from a reservoir, the traveltime, amplitude
and other attributes of reflected seismic waves
change; simulation can demonstrate when these
changes become visible.19 Because 4D seismic t t
2 Predicted Seismic Data 3
data acquisition, processing and interpretation
can cost millions of dollars, it is critical to deter-
mine from modeling studies whether reservoir
variations will be discernible in the new survey.
This type of prediction differs from routine simu-
lation techniques.
Traditionally, reservoir engineers have been
interested in matching simulated well production t Actual Seismic Data t
2 3
with actual production data. These data come
from one point in space—the wellhead—but are
nearly continuous in time. There are other types
of history matching, though. Seismic data repre-
sent a single point in time but offer almost com-
plete spatial coverage of the reservoir. Also, the
modeled and observed parameters are differ- > Forward modeling to optimize data acquisition. Predicted properties of seismic data at
ent—seismic amplitudes are of concern rather time t2 (top left) are used to predict the appearance of seismic data (middle left). These
predictions are revisited after acquisition of actual seismic data at time t2 (bottom left).
than fluid pressures, for example. Seismic properties at time t3 (top right) are predicted next from actual t2 seismic data.
To perform seismic history matching, first, the By considering fluid changes in the reservoir and their effects on seismic waves, and
seismic response to a saturated reservoir is then modeling the seismic data that would result from surveying at time t3 (middle right),
modeled. After some period of production, the additional seismic surveys for reservoir monitoring will be acquired at the optimal time
t3 (bottom right).
seismic response to the depleted reservoir is
calculated. The seismic response might be com-
plex and include a combination of changes in
amplitude, phase, attenuation and traveltime. Multicomponent seismic data and amplitude Interpretation of a seismic response change
The initial and depleted reservoir responses dif- variation with offset (AVO) analysis of compres- from an initial seismic survey to a repeated sur-
fer because the composition of the pore fluids sional-wave data both reduce the ambiguity of vey enables detection and spatial calibration of
and the reservoir pressure change during produc- distinguishing the effects of pressure changes additional faults, movement of oil-water contacts
tion, both of which affect the seismic velocity of from the effects of pore fluids. Without AVO and gas coming out of solution. Small faults in
the reservoir. The synthetic responses are com- processing, ordinary marine 3D seismic data are the reservoir section are often visible as linear
pared with recorded seismic data—the initial 3D fundamentally ambiguous because they respond features of decreased amplitude in a seismic
survey and the subsequent repeated survey. The to compressional waves only, but compressional amplitude or coherency plot. Sealing faults also
difference in seismic character of the reservoir waves respond to both pressure and saturation. appear as patches of undrained hydrocarbons
from the initial survey to the later survey is a Multicomponent seismic data separate compres- whose well-defined edges represent the fault.
function of the compressional (P) and shear (S) sional and shear components, allowing the inter- Movements of oil-water contacts are visible as
seismic velocities and is interpreted as a change preter to separate saturation effects and pressure changes in amplitude and possibly as ”flat
in fluid content and pressure. effects that influence porosity, because shear spots.“ When the reservoir pressure drops below
18. Gawith DE and Gutteridge PA: “Seismic Validation of waves do not respond to pore fluids.20
Reservoir Simulation Using a Shared Earth Model,”
Petroleum Geoscience 2, no. 2 (1996): 97-103.
19. Pedersen L, Ryan S, Sayers C, Sonneland L and Veire HH:
“Seismic Snapshots for Reservoir Monitoring,”
Oilfield Review 8, no. 4 (Winter 1996): 32-43.
20. In the case of the Magnus field, located on the UK conti-
nental shelf, the effect of pressure changes on time-
lapse seismic data is greater than the effect of changes
in pore fluids. For more information: Watts GFT, Jizba D,
Gawith DE and Gutteridge P: “Reservoir Monitoring of
the Magnus Field through 4D Time-Lapse Seismic
Analysis,” Petroleum Geoscience 2, no. 4
(November 1996): 361-372.
21. Pedersen et al, reference 19: 42.

32 Oilfield Review
bubblepoint and gas comes out of solution, ence of a gas cap predicted by seismic data. seismic data, such as maps of seismic attributes,
P-wave seismic data typically show a significant Statoil also reentered and sidetracked an aban- prestack gathers for AVO studies and so on, but
brightening of seismic amplitude. Multicomponent doned well and produced 6300 B/D [1000 m3/d]. each approach has the common goal of deter-
seismic data, which include shear-wave data, Currently, the scientists at Schlumberger mining the area of mismatch between predicted
show no brightening because the shear waves Cambridge Research, Cambridge, England, are and recorded seismic data and analyzing the rea-
do not respond to pore fluids. Such a response combining information on the distribution of fluids, sons for the differences.
confirms the presence of gas. pressure and other properties from the reservoir One major challenge in interpretation of an
Seismic history matching has benefited reser- simulator with rock properties from the geo- observed time-lapse seismic response is that the
voir management decisions in the Gullfaks field, model, or earth model, to generate a forward non-uniqueness in a particular seismic response
where interpretation of 4D seismic data indi- model of seismic response (previous page). In must be considered. For example, an observed
cated the existence of previously unseen sealing particular, the porosity, bulk modulus and shear change in amplitude might represent a change in
faults and the presence of bypassed oil. The modulus from the geomodel are combined with saturation of oil or free gas, a change in the
faults themselves were not seen on the new saturation and pressure information from the amount of gas dissolved in the oil, a change in
data, but were interpreted where fluid content simulator. The forward model provides informa- pressure, or, most likely, a combination of these.
changed abruptly in geophysical maps showing tion about the elastic moduli and density of the Clearly, it is important to know which of these
the time-lapse results. After fluid transmissibility reservoir, from which P-wave velocity, density factors are significant in the reservoir, and seismic
across faults was reevaluated, the potential for and acoustic impedance, or other properties, can modeling can help determine this.
bypassed pay supported drilling an additional be derived. In the Foinaven study area, West of
well.21 That well, drilled by Statoil, initially pro- Next, the synthetic seismic data are com- Shetlands, UK, normal-faulted, layered turbidite
duced 12,000 B/D [1900 m3/d] from a formerly pared with actual seismic data. Numerous com- reservoirs form separate reservoir compartments
undrained compartment and confirmed the pres- parisons can be made between vintages of (below). A preproduction baseline 3D survey was

Gas cap

DC2

Faroe
Islands

DC1

Clair

Schiehallion
0 1 km
Foinaven 0 1 mile Study area
Shetland Reservoir DC1 DC2
Islands T35
South North
T34
T32
T31
Orkney
Islands

0 20 40 60 80 100 km
Horizontal scale
SCOTLAND 0 20 40 60 mi 0 1 km 0
Vertical 50
0 1 mile scale, m
100

> Foinaven field. Located West of Shetlands (left), the Foinaven field produces from four main turbidite reservoirs. The reservoir map (top right) shows gas
caps in red and the strike of the normal faults as black lines. The platforms and well locations are shown in black. The Foinaven study area is indicated by
the blue box. The cross section (bottom right), which extends from south of platform DC1 to north of platform DC2, shows the layered reservoirs that have
been compartmentalized by normal faulting that must be drained by carefully constructed directional wells.

Summer 1999 33
 Synthetic Seismic 4D Response Surface Seismic 4D Response
Repeat Repeat T35 sands not yet
3D survey 3D survey on production
1998 1998

T32 sand after 10

Oil + Free gas (~5%)
months of production
Bright
Water

1000 2000 3000 4000 5000 6000 7000

Baseline Baseline
3D survey 3D survey
1995 1995

T32 sand before

Oil production
Dim
Water

1000 2000 3000 4000 5000 6000 7000

> Visible changes in repeated surveys. Cross-sectional synthetic seismic displays for the baseline survey and repeated survey (left)
show the development of a gas cap. The actual seismic sections confirm the predictions from seismic modeling (right).

acquired in 1995 and a repeat survey was shot in

1998 after 10 months of oil production. The
amplitude brightening in the synthetic seismic
data indicated that development of gas caps
would be visible as the reservoir pressure
Small dropped and gas came out of solution during pro-
gas caps
OWC duction (above). The repeat survey verified the
existence of the expanded gas caps as high-
amplitude events (left). Because seismic property
modeling predicted a good match between the
-3.40 baseline survey and the repeat survey—synthetic
-3.17
-2.94 seismic sections from simulator model matched
-2.71
-2.48
-2.25
real seismic data at appropriate times—the reser-
0 500 1000 1500 2000 m -2.02
voir team had more confidence in the reservoir
-1.79
-1.56
-1.33 simulator model, both in this area and, by extrap-
-1.10
-0.87
-0.65
olation, elsewhere in the field. As a result, addi-
-0.42
-0.19
tional 4D seismic data will be acquired across
0.04
0.27 Foinaven and neighboring Schiehallion field to val-
0.50
idate the reservoir models and ultimately support
the placement of a number of infill production and
Free gas injection wells.
Enlarged Correct timing of data acquisition maximizes
gas caps
the value of the data. Careful study before repeat
OWC 3D survey acquisition can ensure optimal acqui-
sition timing, which is field-dependent. In the
Free gas Foinaven case, timely acquisition of 4D seismic
data helped determine where the reservoir was
N
connected or segmented and where flow barriers
existed so the team could select optimal infill and
0 500 1000 1500 2000 m
water-injection well placement.
> Amplitude changes. Map views of the 1995 baseline survey and the 1998 repeat survey clearly
display the changes in seismic amplitude that result from gas cap development. The small gas
caps in the original survey and the synthetic data shown above it (top) enlarged significantly
after 10 months of production (bottom). The oil-water contact (OWC) remains consistent
between the two surveys.

34 Oilfield Review
 Reservoir Characterization

Reservoir modeling Interpretive data

Development and Production Planning

Reservoir Monitoring and Control

Reservoir performance Flow simulation Development scenario

History Reservoir
matching development

Production Production and reserves Uncertainty analysis

forecasting and risk management

Field Implementation

> Future reservoir management. Reservoir optimization is an iterative process that normally begins with reservoir characteriza-
tion of a new discovery, but can be implemented at any stage in an existing field. Reservoir management will rely increasingly on
monitoring and modeling reservoir performance to optimize oil and gas production. The key additional element will be ongoing
collection of data at the reservoir scale, including seismic data and wellbore measurements, so that the development plan can
be assessed and, where necessary, modified. Monitoring the reservoir closely will overcome the current problems of history
matching using only the loose constraints of production data.

Future Possibilities A prudent, efficient team that works together In the future, new software that validates the
Reservoir simulation has already helped oil and to develop a field reduces cycle time and shared earth model will incorporate measured
gas producers increase predicted ultimate recov- expense. Sharing data and interpretations allows uncertainties of data and interpretations. As the
ery, and further improvement is likely. In addition the team to maximize the value of its achieve- model is refined with the capture of new data,
to ongoing software and shared earth model ments for more realistic reservoir simulation and any change in uncertainty will be addressed
enhancements, reservoir monitoring with down- improved understanding of the reservoir. These in automatically. Forward modeling will further
hole sensors, 4D seismic surveys or other meth- turn advance reservoir management, reserve reduce uncertainty and risk, and maximize the
ods is becoming increasingly cost-effective, recovery and project economics. Currently, this value of additional data. If the shared earth
particularly when new data are acquired at opti- method relies heavily on the team’s motivation to model is consistently updated and new data and
mal times (above).22 work together. Software provides strong support, interpretations are incorporated, project team
22. Watts et al: reference 20.
but is not yet fully integrated to handle the com- members will have another tool to better cope
plete spectrum of oilfield data simultaneously. with increases in both the volume of data and the
productivity expectations of their companies.
—GMG

Summer 1999 35
 Real-Time Openhole Evaluation

Tom Barber Operators are gaining an accurate, at-the-wellsite first look at pay zones, thanks
Sugar Land, Texas, USA
to new technology that incorporates logging measurements environmentally
Laurent Jammes corrected in real time, fast forward modeling and inversion techniques. At
Jan Wouter Smits
Clamart, France the heart of this technology are good science and innovative engineering.

Werner Klopf
Milan, Italy

Anchala Ramasamy
BP Amoco Exploration
Aberdeen, Scotland
To maximize asset value, oil and gas operators the wellsite—opening up a wealth of opportuni-
Laurence Reynolds continually strive to characterize the location and ties for locating and tapping additional reserves.
Aberdeen, Scotland extent of recoverable reserves. Traditionally, open- Improved tool response in thin beds, better pad
hole logs generated with ”triple-combo“ tool application in poor holes, greater accuracy in
Alan Sibbit strings have provided key information: porosity high-weight muds and real-time corrected forma-
Houston, Texas from neutron-density measurements and satura- tion evaluation—all presented in clear, easy-to-
tion from resistivity measurements. The introduc- understand formats—aid decision-making. New
Robert Terry tion of an advanced logging system in 1995, tools, calibration methods and processing tech-
BP Amoco Exploration based on a platform of integrated sensors, has niques, combined with a comprehensive log qual-
Houston, Texas resulted in a quantum change in data acquisition ity-control (LQC) system, allow engineers to
For help in preparation of this article, thanks to David Allen,
capability, reliability and efficiency. The shorter, monitor tool measurements and environmental
Schlumberger-Doll Research, Ridgefield, Connecticut, lighter, operator-friendly tool string is capable of conditions—validating data acquisition and
USA; Gordon Ballard, Kevin Eyl and Alison Goligher, faster rig-up at the wellsite and better access to
Schlumberger Wireline & Testing, Montrouge, France;
ensuring that high-quality analysis can be per-
Vincent Belougne, Geco-Prakla, Gatwick, England; Kees deviated holes. Flexible mechanical design and formed over the entire logged zone. In many
Castelijns, Schlumberger Wireline & Testing, New Orleans, short tool length enable the operator to drill shal-
Louisiana, USA; Ollivier Faivre and Pascal Rothnemer,
wells, this saves valuable time by eliminating the
Schlumberger-Riboud Product Center, Clamart, France; lower wellbores while retaining the ability to log need for repeat passes for log verification. Faster
Pierre Roulle, Schlumberger Wireline & Testing, Pau, important pay zones at the bottom of the well.
France; Jay Russell, Schlumberger Wireline & Testing,
wellsite calibrations, real-time environmental
Livingston, Scotland; Bob Mitchell, Schlumberger Array-resistivity measurements, microresistivity corrections, quality control, depth matching and a
Wireline & Testing, Sugar Land, Texas, USA; Jim White, and three-detector density tools are improving
Schlumberger Wireline & Testing, Aberdeen, Scotland;
complete wellsite quick-look contribute to well-
and Technical editing Services (TeS), Chester, England. accuracy in difficult environments without sacri- site efficiency and put formation evaluation data
AIT (Array Induction Imager Tool), DLL (Dual Laterolog ficing logging speed. into the operator’s hands more quickly.
Resistivity), EPT (Electromagnetic Propagation Tool), FMI
(Fullbore Formation MicroImager), GeoFrame, HRLA (High- But there is more to the story than more flex- In this article, we look at three aspects of
Resolution Laterolog Array), InterACT, Platform Express, ible operations, reduced rig time and improved new platform logging technology and illustrate
SlimAccess, UBI (Ultrasonic Borehole Imager) and Xtreme
are marks of Schlumberger. HDLL (High-Definition Lateral accuracy of standard measurements. Using the the simultaneous improvement in operational
Log) is a mark of Baker Hughes, Inc. latest technology, new-generation tools provide efficiency that can be achieved while accurately
more complete reservoir characterization right at determining formation characteristics under
increasingly difficult environmental conditions.

36 Oilfield Review
 First, we discuss the foundations supporting Real-Time Corrections As the tool is pulled up the wellbore, chang-
real-time environmental corrections, including Every logging tool suffers ing wellbore conditions such as caves or fric-
speed and depth, as a vital first step toward the from environmental effects of tional drag will cause the tool speed to change
use of forward models and inversion techniques one sort or another. Real-time cor- erratically, and even to stick and slip, while the
for making real-time environmental corrections rections are essential to get accurate logging cable speed measured at surface remains con-
to basic logging measurements. Model-based information into the hands of the operator effi- stant. Thus each sensor’s motion across the for-
inversions—along with new measurements, ciently. As a first step, every tool raw sensor mation may not correlate with the motion of the
such as those from mud resistivity sensors, measurement requires a speed-derived depth cable at the surface. The high tool speed—up to
microresistivity measurements, array tools with correction. Next, forward models are used to pre- five times the normal logging speed—occurring
multiple-depth measurements and density dict each measurement response for a given set after a stuck zone often results in lost data.
backscatter detectors—contribute to a clearer of borehole and formation properties. Finally, by Speed-derived depth corrections are essential at
picture of the borehole and formation. comparing predicted sensor responses with log- two basic stages.
Second, the latest developments in openhole ging measurements—a process called inver- First, at the measurement stage, some logging
logging are highlighted, including the new HRLA sion—the environmentally corrected formation measurements depend on integrating raw data
High-Resolution Laterolog Array tool with multiple properties are determined. This sequence of from multiple sensors located at different loca-
depths of investigation, a tool capable of resolving steps is performed during data acquisition and tions along the tool string. Irregular or nonuniform
the effects of shoulder beds, invasion and dipping applied to openhole logging measurements made tool motion during the measurement cycle will
formations, thereby providing better resistivity with the Platform Express tool system first invalidate the assumption that all the data come
evaluation in complex environments with saline reviewed in the Oilfield Review three years ago.1 from the same volume of the formation. Log pro-
mud. Post-log processing techniques, to interpret It may seem surprising that real-time depth cessing will produce an incorrect or unstable
logs in extreme environments, are discussed along correction is such an important issue. After all, result, especially visible at layer boundaries.
with a new maximum-entropy-based inversion most log analysts can shift logs at the computing
technique developed to quantitatively interpret center simply by ”eye“ or with automatic mathe- 1. Goligher A, Scanlan B, Standen E and Wylie AS:
“A First Look at Platform Express Measurements,”
induction logs from highly deviated wells or with matical correlation algorithms. The depth of the Oilfield Review 8, no. 2 (Summer 1996): 4-15.
large shoulder-bed contrasts. logging tool is traditionally determined from the
Finally, we look at two important benefits of length of unwound cable with an approximate
real-time environmentally corrected logging adjustment for cable stretch. However, one of the
data—complete wellsite log interpretation and greatest uncertainties in wireline logging has
log quality control. We illustrate how LQC is been the assignment of petrophysical data to the
enhanced by real-time environmental information. correct depth of the subsurface.

Summer 1999 37
 Standoff At the second stage, depth corrections are
Gamma ray, no speed correction (density) equally important when integrating logging data
0 API 75 1 in. 0 from different tools to perform petrophysical inter-
Caliper Density, no speed correction pretations. For example, log analysts frequently
4 in. 14 Standoff
look for gas by comparing density and neutron
Gamma ray, speed corrected MD Density, speed corrected porosity logs. For the characteristic crossover to
A 1:200
0 API 75 ft 1.95 g/cm3 2.95 be meaningful, the spectral count rates of each
gamma ray detector in the density tool must
change in phase with the count rates in each neu-
tron detector in the neutron porosity tool as each
passes the gas-saturated bed in the formation.
In addition to speed-based depth corrections,
resolution matching is equally important. Since
X150 the neutron measurement samples a slightly
thicker region in the formation than does the den-
sity measurement, these measurements must be
matched volumetrically. This resolution-matching
process is important in all high-resolution inter-
X169 pretations, because it ensures that the sensors of
each tool used in a combined measurement see
exactly the same formation thickness.
A Finally, measurements with high vertical res-
X183 olution can have errors amplified by irregular
tool motion because the acquisition system
obtains data at sampling rates that vary as a
function of the required bed resolution. When
tool speed differs substantially from the recom-
X200
mended, then over- or under-sampling results.
Rapid acceleration following a stuck-tool
episode can result in lost data. Tool-speed-
based depth corrections are a prerequisite to
good high-resolution measurements.
Obtaining properly depth-matched high-reso-
lution measurements is critical for Ocean
Energy’s efforts to evaluate thinly bedded reser-
voirs in the Gulf of Mexico (next page). In one
well, high-resolution invaded-zone resistivity,
Rxo, measurements from the MicroCylindrically
Focused Log (MCFL) tool in combination with
density logs clearly show the many thin beds—
> Speed-corrected density logs. This example, from a 97º-deviated borehole in the North Sea, compares some less than 1-ft thick [0.3 m]—throughout the
the speed-corrected Platform Express density log (red) with the uncorrected log (black) in track 2. reservoir. Comparison with the FMI Fullbore
The green band on the left of track 1 is the accelerometer LQC flag indicating—by turning black— Formation MicroImager images confirms the
excessive stick and slip experienced in Zone A. The gamma ray in track 1 also shows shifts when presence of thin beds, and the real-time porosity
speed-corrected. The logs highlight the benefits of speed corrections in poor hole conditions. The two
high-porosity thin beds seen at X169 ft and X183 ft are incorrectly identified in the uncorrected log. derived from the high-resolution density log
Potential hydrocarbons in these zones could be missed if uncorrected depths were used to guide the enables accurate reserve calculations.
perforation interval. Crucial depth corrections are implemented in
the Platform Express system during acquisition
Some tool designs, such as neutron and den- Likewise, the AIT-H Array Induction Imager using a built in tool-axis accelerometer. This
sity tools, have multiple asymmetrically posi- tool filter-based processing algorithm assumes device measures instantaneous tool acceleration
tioned sources and detectors and unequal that the data from the eight asymmetrical arrays to determine tool velocity and the true depth at
source-detector spacings, and their measure- are regularly sampled every 3 in. Irregular tool which all the other tool measurements were
ments—based on comparing the count rates in motion can give rise to artifacts on the logs. recorded. Stability of the depth-correction algo-
each detector—can be affected by nonuniform Caliper measurements and auxiliary mud resistiv- rithm is maintained by the use of a Kalman filter-
tool motion. An example from the North Sea ity measurements must be correctly depth based optimization.2 This optimization minimizes
illustrates the effects of speed-corrected density aligned with the AIT array measurements to 2. Belougne V, Faivre O, Jammes L and Whittaker S: “Real-
measurements (above). derive AIT borehole corrections. Time Speed Correction of Logging Data,” Transactions of
the SPWLA 37th Annual Logging Symposium, New
Orleans, Louisiana, USA, June 16-19, 1996, paper F.

38 Oilfield Review
 Gamma ray
0 API 150
Rxo (1 in.) Density (1 in.)
g/cm3
Rxo (8 in.) Density
g/cm3
North Rxo Density (8 in.)
1.65 g/cm3 2.65
0 FMI image 360 Rxo AIT 90-in. resistivity Neutron porosity
Depth
0.5 8 0.2 ohm-m 20 60 p.u. 0
ft ohm-m

X760

X770

X780

X790

X800

X810

X820

X830

> Detecting thin beds in the Gulf of Mexico. The FMI Fullbore Formation MicroImager tool image shown in track 1 confirms the
presence of many beds less than 1-ft thick detected by the high-resolution Micro-Cylindrically Focused Log (MCFL) Rxo log shown
in track 2. The high-resolution (black) and very high-resolution (blue) density logs are shown in track 3. The high-resolution density
logs are quantitative in beds over 8-in. thick. The high-resolution, deep induction log shown in track 2 detects many of the thin
beds and can be used quantitatively in beds 1 ft or thicker. Track 4 compares all the density logs with the neutron porosity log.

the overall error in the depth correction by solv- Forward Modeling for understood, and with enough knowledge of a
ing a system of simultaneous equations linking Environmental Corrections resistivity tool design and its environment, the
the exact moment each sensor measurement In the language of log analysts, the phrase for- voltages and currents that make up the tool
was made with the true tool position—derived ward modeling refers to computing a logging responses are predictable.
from the instantaneous accelerometer measure- sensor response in the presence of the environ- Forward modeling is important because it
ment. Cable depth is used as a constraint to help ment surrounding the logging tool. Almost any allows prediction of tool response under any
stabilize the solution. All raw sensor measure- tool response can be linked to formation proper- given conditions. These predictions can then be
ments and optimization solutions are performed ties through the physics of the sensor measure- compared with observed measurements, in a pro-
in the time domain. This allows them to be easily ment and its interaction with the materials of the cess known as inversion—described later in this
converted to true depth or cable depth, formation and borehole environment. Compton article—to understand the real conditions under
whichever is required. Time-domain processing scattering and photoelectric absorption govern which the measurements were made. In this arti-
also helps to overcome limitations encountered the interaction of low-energy gamma rays used cle, unless otherwise specified, tool or sensor
in high-frequency depth sampling during high- to measure formation density. The physical prin- response means the raw measurement, such as
resolution logging operations. ciples embodied in Maxwell’s equations are well count rate in a nuclear detector, or voltage and
current measured on an electrode or induction
tool antenna coil.

Summer 1999 39
 Mud Mudcake Formation

Window count rates (W)

a W1 Long-spacing
detector
a
Count rate

W2 spectrum

W3
W4

b
b W1 Short-spacing
W2 detector
Count rate

W3 spectrum

W4 c

c W1 Backscatter
W2 detector
Count rate

spectrum
W3
> Density calibration database. The density for-
ward-model database was recorded in the Envi-
Energy
ronmental Effects Calibration Facility in Houston,
Texas, USA, which was built for the characteriza-
tion of nuclear logging tools. This database cov-
Density
ers the range of environments to which the tool
will be exposed. The facility manager, John Spal-
Formation lone, is shown lowering the Platform Express
density tool into one of the calibration blocks.
Mudcake Since the Platform Express tool first became
available, a continuous effort has been under
Mud way to enhance the density measurement capa-
bility in heavy mud environments.
Distance

> Three-detector density logging sonde. Multiple Compton scattering and photoelectric absorption A forward model is used in Platform Express
lead to a spectrum of gamma ray photons entering the detector windows from the borehole and forma- real-time density analysis to calculate each
tion environment (right). The forward model represents a homogeneous formation behind a thick layer
of mudcake (bottom). Each detector spectrum is partitioned into broad count-rate windows used to window count rate as a function of formation
estimate properties of the gamma ray scattering environment (left). and mudcake properties. The formation model
geometry consists of a homogeneous forma-
tion and mudcake corresponding to a one-
Density forward models—The three-detector degraded photons from multiple Compton scatter- dimensional (1D) radial step profile varying in
density tool in the Platform Express tool string ing and photoelectric absorption contribute to an density and photoelectric properties. Within this
uses a gamma ray source that emits 662-keV overall continuous gamma ray spectrum seen in framework, the different detector count-rate
photons from a source capsule located in the the formation by each of the three detectors. responses depend on only five environmental
logging pad (above).3 Although density measure- Typically, increasing density causes an
ments are sensitive to a relatively small volume of increase in the gamma ray flux near the source 3. Allioli F, Faivre O, Jammes L and Evans M: “A New
Approach to Computing Formation Density and Pe
the environment between the source and detector, because there are more scattering targets in a (Photoelectric Factor) Free of Mudcake Effects,”
the increasing source-to-detector spacing of each higher density material. This increases the Transactions of the SPWLA 38th Annual Logging
Symposium, Houston, Texas, USA, June 15-18, 1997,
detector enables each to see progressively deeper observed backscatter detector count rate. On the paper K.
into the mudcake and formation. other hand, increasing density tends to cause a 4. Ellis D: Well Logging for Earth Scientists. New York, New
York, USA: Elsevier Science Publishing Co., Inc, 1987.
Gamma rays from the source enter the mud- decrease in the observed count rates in the two
5. Anderson B, Druskin V, Habashy T, Lee P, Luling M,
cake and formation and typically scatter several detectors spaced farther from the source because Barber T, Grove G, Lovell J, Rosthal R, Tabanou J,
times before being detected by each detector in of the long attenuating path to these detectors. Kennedy D and Shen L: “New Dimensions in Modeling
Resistivity,” Oilfield Review 9, no. 1 (Spring 1997): 40-56.
the logging pad. Each Compton scattering Also, changes in formation lithology can be 6. In this article, we shall confine our discussion to the
encounter causes the incident gamma ray to lose detected by variations in the low-energy window dimensionality of the formation model—the number of
independent coordinates needed to describe the way in
energy and change direction, eventually bending count rates due to photoelectric absorption. which resistivity varies. See “The Vocabulary of
many gamma rays back towards the detector These count rates are also strongly affected by Resistivity Modeling,” Anderson et al, reference 5.
aperture in the tool. The comined effect of energy- barite in the mudcake, and its presence can make
the photoelectric effect measurement intractable.

40 Oilfield Review
parameters—formation density and photoelec- solution requires the use of numerical methods. A
tric factor; mudcake density and photoelectric The two methods most frequently employed are
factor; and mudcake thickness. the Finite Element Method (FEM) and the Finite
An ideal implementation of the gamma ray Difference Method (FDM). Development of for-
physics in this formation forward model would be ward modeling techniques using FEM and FDM
given by an exact solution to the Boltzmann for resistivity measurements was discussed two
equation for gamma ray transport. Unfortunately, years ago in the Oilfield Review.5
the Boltzmann equation has no simple analytic Both FDM and FEM divide the environmental 1D-radial
form for this environment.4 Instead, a proxy for domain into grid cells and solve for the potential
the exact sensor response physics parameterizes in each cell. The interactions between neighbor-
the detector window count rates as exponential ing cells are controlled by Laplace’s equation.
functions in terms of formation and mudcake Combining all the cell interactions together with
properties. Each window count-rate response the boundary conditions yields a large system of
function contains empirical coefficients, which linear equations, which is solved by the computer B
account for source strength, detector collimation, to find the electric potential at the vertices
average gamma ray track length for each partic- of each cell. Depending on the complexity
ular source-detector spacing, and energy-depen- and dimension of the formation model, either
dent Compton scattering and photoelectric two-dimensional (2D) or three-dimensional
absorption cross sections. These coefficients (3D) solutions of the electromagnetic field are
form the calibration for this nonlinear parametric needed (right).6
1D-layer
forward model, and are determined by a Multidimensional forward modeling tech-
weighted least-squares fit to a database of labo- niques have been successful in rapidly and accu-
ratory measurements. rately predicting the tool response of resistivity
The database measurements were obtained tools. The models are used to optimize tool
in a laboratory with a calibrated density tool designs, characterize their response and provide
in known formations and mudcake conditions the basis for inversion procedures designed to C

(previous page, right). Today, over 1130 calibra- find formation characteristics. Although modeling
tions in boreholes with barite mudcake and 420 speed has increased considerably in recent years,
calibrations with nonbarite mudcake have been it is frequently too slow to allow real-time inver-
made, and recent calibrations for the density for- sion of measurements while logging, especially
ward model have extended the operating range when a 3D model is needed. When the number of
of the density tool to mud weights of up to formation parameters is limited, it is often more
17 lbm/gal [2.04 g/cm3]. practical to create a database of the tool 2D
Resistivity forward models—In contrast to response to variations in parameters and inter-
nuclear measurements with their small volume of polate from this database to build the numerical
investigation, all deep-resistivity responses, forward model used during the inversion.
whether from induction or laterolog tools, are
Formation model dimensions. In HALS laterolog
>

influenced by the resistivity distribution in a large resistivity modeling, a 1D radial formation model D
volume surrounding the logging instrument. is concerned only with the influences of the
Correct interpretation requires corrections for bore- radial invasion profile (A). It assumes that the
reading is being taken in an infinitely thick
hole, invasion and other large-scale environmental formation bed. At other times, a 1D layered for-
or geometrical effects such as shoulder effects. mation is used to model the effects of thick and
The response of a laterolog, such as the High- thin shoulder beds with resistivity contrasts (B).
Resolution Azimuthal Laterolog Sonde (HALS) A 2D model assumes that the formation is com-
posed of homogeneous layers perpendicular to 3D (2D
measurement, is computed by solving Laplace’s the borehole (C). In this model, both the invasion plus dip)
equation for the electrostatic potential in the of each layer and the shoulder-bed effect of
borehole and formation environment. Laplace’s adjacent layers are taken into account, resulting
equation follows from Maxwell’s equations in in a more accurate Rt calculation in beds where
significant shoulder-bed effects exist. A 2D model
the low-frequency limit and dictates conserva- is defined by the values of Rxo, Rt, the invasion
tion of current everywhere in the environmental radii and thicknesses of the formation layers.
The 2D model can be taken one step further by E
domain. Knowing the voltage and currents every-
where along the tool allows prediction of the incorporating the formation dip relative to the
axis of the borehole. This results in a 3D-formation
apparent resistivity reading. Unfortunately, ana- model (D). This scenario could be due to struc-
lytical solutions exist for only a limited number of tural formation dip, deviated borehole or both.
problems with simple geometries. In practice, the More complicated 3D formation models can be
constructed in various ways. One way is to parti-
tion the layers into azimuthal sectors (E). This
model takes into account azimuthal anisotropy,
shoulder-bed effects and invasion as well as 3D
variations in layer thickness with distance from
the wellbore.

Summer 1999 41
 < AIT borehole correction model. Density inversion—Density inversion is
A tool can be located anywhere based on a maximum likelihood method that uses
in the borehole with any value the nonlinear parametric forward model dis-
standoff. The borehole
corrections are based cussed above to link the depth-corrected
on an iterative inversion observed count rates in each of the detector
processing. count-rate windows to the formation parameters.
At every depth, the inversion predicts the count
rates in each detector window and compares
them with those measured by the tool.
Minimizing a cost function containing three
terms optimizes the solution.9
Borehole
radius AIT tool The first term in the cost function is based on
Borehole
the best fit of all the observed window count
mud resistivity r rates. This is done by minimizing the ”recon-
Rm
struction error,“ which is a term proportional to
From Forward Models to Inversions the average squared difference between all the
Given a forward model with a system of equa- measured and modeled count rates in each
Formation tions governing tool responses, one simply enters detector energy window—each weighted by the
resistivity
Rf the formation and borehole parameters into the expected error based on counting statistics and
Standoff
model and, after computation, the desired tool model uncertainty.
The HALS tool uses two such FEM-derived measurements are predicted. The prediction is a The second term in the cost function mea-
databases for processing its measurements.7 The set of tool responses that would be observed if a sures the difference between the current model-
first database is used to correct apparent resis- physical experiment were performed in the given predicted environmental parameters and those at
tivities for borehole effects. It models the tool formation and borehole environment. When the the previous logging depth. This term, called the
response in an uninvaded formation as a function governing equations are linear, the process is smoothness condition, helps ensure compatibil-
of formation-to-mud-resistivity contrast, bore- reversible—or invertible in one step. Iteration is ity between sampling rate and measurement of
hole size and tool eccentricity. Because of tool not needed. Given the tool response, the model vertical resolution.
eccentering, a 3D model was needed to construct parameters can be estimated by multiplying the The final cost term helps control the stability
this database. The second database is used to vector of observed tool responses by the gener- of the solution when formation and mudcake
invert apparent resistivities for Rt. It describes alized inverse of the same matrix used in the parameter estimates are far from the database
the tool response as a function of invasion radius corresponding forward problem. range, which can happen when large standoffs
and contrast between formation resistivity, Rt, Unfortunately, nature does not always pose occur. This term vanishes when the solution is
and invaded zone resistivity Rxo. Even though the linear problems. Density and resistivity tool within the limits of the database.
response was originally determined using a responses are not linear with respect to forma- A powerful example of density inversion
3D-based computation, the forward model is tion properties. For these responses, a more robustness can be seen when comparing density
called 1D because it models resistivity variations versatile technique is to find a solution by itera- measurements derived during extreme condi-
only in the radial direction. tively solving the forward problem. Inversion is tions (down logging) with normal measurements
The AIT Array Induction Imager tool bases its the process of creating a model, mathematically taken while logging uphole (next page).
borehole corrections on a forward model that modeling the physical response to that model, Occasionally, for precautionary reasons, logging
includes a 2D plus eccentricity effects (above).8 and then varying the parameters in the model
7. Smits J, Benimeli D, Dubourg I, Faivre O, Hoyle D,
The borehole forward model is based on solutions until the modeled response matches the one Tourillon V and Trouiller J-C: “High Resolution From
to Maxwell’s equations in a cylindrical borehole seen by the logging tool. There are two ways to a New Laterolog with Azimuthal Imaging,” paper
SPE 30584, presented at the SPE Annual Technical
with resistivity Rm surrounded by a homogeneous perform this task, manually and automatically. A Conference and Exhibition, Dallas, Texas, USA,
formation of resistivity Rt. The tool can be located log analyst often uses manual inversion, adjust- October 22-25, 1995.
with a standoff anywhere in the borehole, but ing model parameters based on previous experi- 8. Anderson BI and Barber TD: Induction Logging.
Sugar Land, Texas, USA: Schlumberger Wireline &
is assumed to be parallel to the borehole axis. In ence or knowledge, with advanced post-logging Testing, 1997.
this model, the signal in any given AIT array is processing programs at a computing center. 9. The concept of a cost function comes from the mathe-
matics of operations research and making decisions
predicted as a function of four environmental Automatic inversion is performed by an algo- with multiple objectives. A cost function is an equation
parameters—Rm, Rt, borehole size and tool stand- rithm that computes the response to a model, then that contains a measure of the error in the decision-
making or optimization problem. The function can
off. This model is used to develop a database of follows certain rules to modify the model to con- contain many terms, one of which usually describes
tool responses that provides calibration for the verge to a solution. Common algorithms minimize the overall reconstruction (or fit) of the forward model
predictions to the observed sensor responses. The
fitted forward model used in real time to predict the difference between the observed response magnitude of the cost function decreases as the opti-
tool responses in any borehole environment. and the modeled synthetic response by adjusting mization improves.
the model parameters to reduce the differences at 10. Statistical (count rate) uncertainties and forward model
errors are propagated through the density inversion pro-
each step. Real-time environmental borehole cor- cessing by the measurement covariance calculation.
rections in the Platform Express system all depend
on automatic inversion techniques.

42 Oilfield Review
measurements are taken while going into the and ribs algorithm cannot account for these The use of a parametric forward-model-based
borehole. The caliper arm and pad are closed to extreme standoff conditions. The different radial inversion for density measurements has several
prevent getting stuck against the borehole wall sensitivity of each of the three density detectors advantages. It makes the most efficient use of all
while going down the borehole. This results in a helps the inversion algorithm provide a final envi- the sensor measurements while simultaneously
large standoff with excessive mud and mudcake ronmentally corrected density log that compen- obtaining formation and mudcake properties. For
between the density pad and formation. Under sates for the extra standoff. This corrected log example, all density, photoelectric and standoff
such large standoff conditions, two-detector den- agrees well with the density log obtained under information contained in the entire spectrum
sity measurements based on a graphical spine normal logging conditions. from each gamma ray detector is automatically
taken into account in the inversion algorithm,
Long-spacing density (up) providing a strong degree of redundancy in the
information available from the density tool. Both
Short-spacing density (up) statistical uncertainties and forward model
errors are considered in the inversion calculation,
Backscatter-spacing density (up) and provide realistic output uncertainties.10
These take into account count-rate statistics, cal-
Standoff
up Long-spacing density (down) ibration errors and model errors, and establish
1 in. 0 reliable confidence limits for LQC and subsequent
Caliper Depth Short-spacing density (down) Inversion density logging up petrophysical analysis.
4 in. 14 ft Resistivity inversion—Inversion is central to
Standoff
Gamma ray down Backscatter density (down) Inversion density logging down the borehole correction algorithm in the process-
0 API 150 1 in. 0 1.65 g/cm3 2.65 1.65 g/cm3 2.65 ing chain for AIT logs. It uses the AIT polynomial
forward model to correct the eight depth-
corrected raw array measurements for tool
effects in a nonstandard borehole environment.
The parameters for the inversion are the four
components of the database—borehole radius,
tool standoff, mud resistivity and formation resis-
tivity. The inversion is an optimization that finds
the set of borehole parameters that best repro-
X575
duces the four shortest arrays (6-, 9- 12- and
15-in. measurements). However, since these
measurements overlap considerably in their
investigation range, their information content is
not sufficient to solve for all borehole parameters
simultaneously.
In practice, the inversion process reliably
determines only two of the four parameters. The
other two parameters are always measured or
fixed. Since formation resistivity is always an
unknown, there is only one additional free
parameter for the inversion to determine.
Accordingly, there are three modes of borehole
correction—depending on which parameter is
sought. If an accurate hole diameter from the
density caliper and mud resistivity from the aux-
iliary mud resistivity, Rm, sensor are used, then
the borehole correction inversion determines tool
standoff. Mud resistivity needs to be measured
within 5% of its true value, which can be met
X600
with the AIT mud sensor. By solving for formation
> Using inversion to obtain the correct answer under extreme conditions. When a density tool is low- resistivity and standoff, the borehole correction
ered into the borehole, the caliper is kept closed to prevent getting stuck in the hole. Under these
conditions, the density pad is not pushed firmly against the borehole wall, resulting in a large standoff problem can be solved with no intervention from
(black dashed) derived from the inversion algorithm shown in the depth track. All three detectors (blue the engineer. Likewise, hole diameter can be
curves) see low density in track 2 because of extra mud and mudcake encountered in this configura- computed with an accurate standoff and mud-
tion. However, the inversion algorithm is still capable of accounting for the extra mud and standoff in resistivity measurement, or else the mud resistiv-
this environment. It produces a density curve (blue) in track 3 that agrees with the density reading
(red) obtained under normal conditions—logging up the borehole with the caliper open and pad ity can be computed from accurate hole size and
pushed against the formation. standoff information.

Summer 1999 43
 Step profile invasion profiles. Typically, radial inversion is correction can be derived from the tool itself or
another 1D four-parameter optimization that uses from an external mud-resistivity measurement.
Rxo
a monotonic-conductivity invasion profile model Following borehole correction, an inversion
to produce logs of Rxo, Rt, and the limits of a tran- based on a 1D three-parameter step-profile inva-
sition zone (left). sion model is used to determine the formation
Rt Real-time wellsite processing for laterologs parameters—Rt and the invaded-zone radius—
ri also involves borehole corrections followed by a that best describe the borehole-corrected deep
1D inversion for Rt. For example, HALS borehole and shallow measurements. Since only two
corrections adjust for the presence of the bore- formation parameters can be determined from
Ramp profile
hole, taking into account the borehole size and the two measurements (shallow and deep resis-
the ratio of apparent resistivity to mud resistivity tivity), the value of Rxo must be supplied to the
Ra /Rm. It also includes an eccentricity correction inversion. It is obtained from the MCFL tool,
Rxo that allows for the eccentered position of the tool which gives a resolution-matched microresis-
Invasion
in the borehole. The mud resistivity needed in the tivity measurement.
Rt
midpoint
Formation resistivity profile

ri
HRLA array resistivity 1

HRLA array resistivity 2

Slope profile
HRLA array resistivity 3
Rxo
r1 Gamma ray HRLA array resistivity 4 DLL shallow resistivity
0 API 150
Caliper Depth HRLA array resistivity 5 DLL deep resistivity
Rt 6 in. 11 m 0.1 ohm-m 10 0.1 ohm-m 10
r2

X150
Annulus profile

Rann
Rxo
r1

Rt
r2

Distance from wellbore

> Resistivity invasion models. The simple piston-

invasion, or step profile, and the ramp profile
require three parameters; the slope profile
is a four-parameter model; the annulus profile
is a five-parameter model. A step profile is used X250
for real-time HALS and HRLA radial inverse inva-
sion modeling, and the slope profile is used for
real-time AIT radial invasion inversion.

After all the raw array measurements have

been corrected for nonstandard borehole effects,
they are processed by the usual AIT log-forming
techniques. This involves generating the stan-
dard depth (10-, 20-, 30-, 60- and 90-in.) and
simultaneously resolution-matched (1-, 2- and 4-
ft) logs by convolving the corrected raw array
measurements with the Born-approximation-
based weighting functions.11 Subsequently, real-
time interpretations are based on inverting these > Eliminating the Groningen effect. The DLL Dual Laterolog Resistivity deep-resistivity (red)
processed measurements radially to obtain an and shallow-resistivity (blue dashed) curves in track 3 show a large separation (yellow
shaded) caused by the Groningen effect. The HRLA curves in track 2 are not affected by
estimate of Rt.12 This radial processing also gives the Groningen effect because all the currents return to the tool string itself, rather than the
a quantitative estimate of invasion geometry as surface. The fact that all five HRLA curves are reading the same resistivity over this interval
well as an accurate estimate of Rt in complex also establishes that the formation is uninvaded.

44 Oilfield Review
 A similar Rt inversion technique is used for the Reference effects—Traditionally, laterolog
new HRLA High-Resolution Laterolog Array tool tools operate in a deep mode with currents
discussed below.13 For this tool, there is enough returning to a reference electrode at the surface.
information in the five array measurements to This requires isolating the logging cable from the
allow an accurate estimate of Rt, and determine tool by use of a long insulating bridle. A system-
the invasion profile—independent of an external atic shift in the resistivity measurement, called
Rxo under most conditions. However, if an addi- the Groningen effect, arises when high-resistivity
tional Rxo measurement is used, it helps constrain formation layers above the tool force returning
the inversion processing and improve the derived currents—following the path of least resis-
Rt. In the HRLA inversion, weights are assigned to tance—into the borehole. This leads to a poten-

Squeeze
each measurement based on the magnitude of its tial drop along the cable, and subsequently the
borehole correction. The measurement with the voltage reference of the tool can no longer be
smallest borehole correction is given the highest considered to be infinitely far away. As a result,
weight in the inversion algorithm. formation resistivity computed using this refer-
ence reads artificially high (previous page, right).
New High-Resolution Long tool strings and drillpipe have a similar
Resistivity Technology effect—artificially increasing the measured for-
Although the concept of an array laterolog tool mation resistivity.
has existed since the 1950s, Shell was first to Shoulder-bed effects—Shoulder beds with
recognize that an array-resistivity tool could large resistivity contrasts have a strong influence
improve thin-bed saturation evaluation, and pro- on most laterolog measurements. The measure-
posed a point-electrode array tool called the ment and focusing currents from the laterolog
Multi-Electrode Resistivity Tool (MERT).14 Their tool tend to flow along zones of least resistance
tool design is similar to a multiple-spaced (left). A distortion in the focusing current distribu-
”normal“ measurement to which lateral and tion allows the measurement current to flow dif-
second-difference voltage measurements are fusely across intervals with significant resistivity
added.15 Recently, Baker Atlas commercialized variations. This defocusing introduces a coupling
their HDLL High-Definition Lateral Log service, between the vertical and radial response charac-
which is an implementation of the MERT teristics of the resistivity measurement.
architecture.16 Antisqueeze 11. Generalized (finite conductivity) 1D tool geometrical
In response to this need, Schlumberger response functions are derived using a forward model
solution to similar to the single scattering Born
developed a new HRLA High-Resolution Laterolog approximation formalism traditionally used in quantum
Array tool that can be used with the Platform mechanics. See Gianzero S and Anderson B: “A New
Look at Skin Effect,” The Log Analyst 23, no. 1 (January-
Express system. This tool achieves multiple February, 1982): 20-34.
depths of investigation through a segmented Barber TA and Rosthal R: “Using Multiarray Induction
array of six simultaneous, symmetrical and Tool to Achieve High-Resolution Logs with Minimum
Environmental Effects,” paper SPE 22725, presented at
actively focused laterolog measurements. This the 66th SPE Annual Technical Conference and
design gives a coherent set of high-resolution Exhibition, Dallas, Texas, USA, October 6-9, 1991.
12. Howard AQ: “A New Invasion Model for Resistivity Log
resistivity measurements that can be inverted Interpretation,” The Log Analyst 33, no. 2 (March-April,
to correct the deepest measurements for the 1992): 96-110.
environmental influences of invasion and shoulder 13. Griffiths R, Smits J, Faivre O, Dubourg I, Legendre E and
Doduy J: “Better Saturation from a New Array
beds. By having all the laterolog currents return Laterolog,” Transactions of the SPWLA 40th Annual
to the tool body, the HRLA tool minimizes the > Squeeze and antisqueeze. Squeeze (top) occurs Logging Symposium, Oslo, Norway, May 31-June 3, 1999,
paper DDD.
two most unwanted laterolog parasitic distor- when adjacent outer beds have significantly
14. Vallinga PM, Harris JR and Yuratich MA:”A Multi-
higher resistivity than the middle bed. Laterolog
tions—reference effects and shoulder-bed focusing currents tend to migrate toward the bed
Electrode Tool, Allowing More Flexibility in Resistivity
Logging,” Transactions of the SPWLA 14th European
effects. In addition, the surface current return of interest and squeeze into this bed of compara- Formation Evaluation Symposium, London, England,
and insulated bridle are no longer needed— tively low resistivity. This results in a deeper mea- December 9-11, 1991, paper E.
reducing cost and risk. surement than in a homogeneous formation. The Vallinga PM and Yuratich MA: “Accurate Assessment
squeeze can result in resistivity curve separation of Hydrocarbon Saturation in Complex Reservoirs
that imitates an invasion profile even in the From Multi-Electrode Resistivity Measurements,”
Transactions of the SPWLA/CWLS 14th Formation
absence of invasion. When both invasion and Evaluation Symposium, Calgary, Alberta, Canada, June
squeeze are present, the resistivity curve separa- 13-16, 1993, paper E.
tion will indicate deeper invasion than is actually 15. The use of computed focusing makes it possible, in prin-
present. The reverse situation, antisqueeze (bot- ciple, to obtain a Laterolog-7-style measurement from
tom), results when adjacent outer beds have lower the MERT tool.
resistivity than the middle bed. In this case, mea- 16. Itskovich GB, Mezzatesta AG, Strack KM and Tabarovsky
surement currents tend to flow into beds of lower LA: “High-Definition Lateral Log-Resistivity Device: Basic
resistivity rather than flow through higher resistiv- Physics and Resolution,” Transactions of the SPWLA
39th Annual Logging Symposium, Keystone, Colorado,
ity beds. The resulting defocusing causes the deep USA, May 26-29, 1999, paper V.
measurements to have a reduced depth of investi-
gation and thus read a lower formation resistivity if
the invaded-zone resistivity is lower than Rt. This
effect can lead to underestimation of reserves.

Summer 1999 45
 The HRLA tool addresses these problems
with multiple modes of tightly focused array
150
measurements.17 Multifrequency operation of the
100 segmented electrodes enables the simultaneous
resistivity measurement modes to be distin-
50 guished (left). Software focusing by linear super-
position of each mode is used to provide active
0 focusing. The shallowest mode is the most sen-
sitive to the borehole and is used to estimate
-50
mud resistivity, Rm. The deepest mode has a
response comparable to the deep-resistivity
-100
measurement of the HALS tool. The spacing of
-150 the other arrays is such that they have response
characteristics that optimize the information con-
Mode-0 Mode-1 Mode-2 tent of their measurements with respect to the
invasion profile.
The addition of shallower curves improves
150 the radial sensitivity to resistivity change, which
results in greater log curve separation in the
100
presence of invasion (below). This is especially
50
helpful in thin beds, where deeper measure-
ments tend to lose both depth of investigation
0 and vertical resolution because of antisqueeze
effects. In addition, with reference effects gone
-50 and shoulder-bed effects reduced, the separation
between the deep and shallow measurements
-100 caused by these effects is also reduced. The
improved invasion discrimination in thinly bed-
-150
ded formations leads to better vertical resolution
and more accurate inversion processing for
-100 -50 0 50 100 -100 -50 0 50 100 -100 -50 0 50 100
Mode-3 Mode-4 Mode-5 formation resistivity.
> Current distributions for HRLA focusing modes. By increasing the number of central 17. Smits et al, reference 7.
electrodes that are kept at the same potential, the tool current return in the formation is
moved farther away, and the depth of investigation is increased. Six modes with increas-
ing depth of investigation are used. In Mode-0, current flows directly from the central
electrode to the nearest array electrodes. This mode is sensitive to the mud column envi-
ronment and is used to estimate mud resistivity and borehole diameters. The deepest
mode, Mode-5, sends current out from all but the outermost electrodes. The spacing of
the array has been designed to optimize the information content of the measurement data
with respect to the invasion profile.

10
Radial response of resistivity tools. The bore-
>

hole-corrected HALS deep-resistivity radial

response compares well with the Mode-5
response from the HRLA array measurement,
while the HALS shallow-resistivity response is
intermediate between the Mode-2 and Mode-3
HALS deep HRLA responses. The additional HRLA resistivity
Apparent resistivity, ohm-m

measurements provide improved definition of

HRLA mode 5 radial resistivity changes, which helps evaluate
invasion profiles.
HRLA mode 4

HRLA mode 3

HRLA mode 1
HALS shallow

HRLA mode 2

1.0
0 5 10 15 20 25 30 35 40 45 50
Invasion radius, in.

46 Oilfield Review
 Rxo > Rt

Rxo < Rt
Getting More Pay from
HRLA array resistivity 1
Resistivity Logs
HRLA array resistivity 2
Rxo > Rt

HRLA array resistivity 3

Rxo < Rt

HRLA array resistivity 4 Rt from 2D inversion

An inversion for the HRLA tool based 2D invasion
EPT
on a 2D-formation model with a piston-invasion Invasion radius Rxo from 2D inversion Rxo from 2D inversion
profile is currently being developed for the Crossover Washout
GeoFrame log interpretation system. The inver- Propagation time Differential Bit radius HRLA array resistivity 5 Rt from 1D inversion
Caliper
sion processing proceeds by detecting bed 27 ns/m 7 0 in. 3
Attenuation Hole radius Rt from 2D inversion MCFL microresistivity
boundaries and segmenting the log into discrete
beds with average thickness of 1 to 2 ft [0.3 to 100 dB/m 700 0 in. 20 1 ohm-m 30 1 ohm-m 30

0.6 m]. The inversion will iteratively refine for- XX00

mation parameters until the forward model
accurately reproduces the input logs. Because
the formation model includes the borehole
shape and mud resistivity, the input logs do not XX10
need to be borehole-corrected.
More accurate representations of the
formation environment lead to more accurate
estimates of Rt, especially in thinly bedded XX20
formations. When the reservoir and water layer
thickness are on the order of 2 to 5 ft [0.6 to
1.5 m], it is not unusual to see differences
of 50 to 100% between the 1D- and 2D-inversion
XX30
resistivity estimates (right). Usually the 1D-
inversion estimates are too pessimistic. An obvi-
ous extension of these models includes dipping
layers with 3D models.
XX40

XX50

Improved Rt estimate with 2D inversion. The 2D

>

XX60
inverted formation resistivity Rt (wide red) and
invasion resistivity Rxo (green) are shown in track 3
along with the raw HRLA curves. The shading
between Rxo and Rt indicates where the invasion is
normal (Rxo < Rt) or reversed (Rxo > Rt). In track 4,
XX70
the 2D inverted resistivities Rt (red) and Rxo
(green) are compared with the 1D inverted forma-
tion resistivity Rt (magenta) and the Rxo (black)
from the MCFL tool. The 2D inversion shows a sig-
nificant increase in Rt obtained in thin beds—such
as those between XX30 and XX70 ft—over the 1D- XX80
inversion results. A good match between the 2D
inversion-derived Rxo and the one independently
obtained from the MCFL measurement—adds
confidence to the inversion results. The EPT Elec-
tromagnetic Propagation Tool dielectric attenuation XX90
and propagation time curves, confirming the pres-
ence of thin beds in track 1, were used to constrain
the inversion for the uninvaded formation model in
the shales.
X100

Summer 1999 47
 Optimal array focusing is enhanced by the Tackling Difficult Environments shoulder beds, spiral boreholes and dipping for-
symmetric tool design, ensuring that all the sig- Model-based inversion processing is a delicate mation beds with invasion—can continue to
nals are measured at exactly the same time and balancing act. Two conflicting factors need to be cause errors when computing Rt. For these, post-
at the same logging tool position. This helps considered. On one hand, the accuracy of the logging processing techniques available at the
avoid horns and oscillations produced by irregu- result depends on how much additional informa- computing center can help.
lar tool motion, and ensures that the measure- tion can be built into the model. On the other, the Shoulder beds—The presence of shoulder
ments are exactly depth aligned. The coherent speed with which the result is delivered beds can lead to overestimating Rt in the
nature of the focused, depth-aligned, resolution- decreases with the complexity of the model. Fast ”squeeze“ case and underestimating it in the
matched measurements from the HRLA tool 1D-inversion models are needed for real-time ”antisqueeze“ case. Consequently, in both cases,
produces a more intuitive LQC, as the curves sep- environmental corrections to help the operator water saturation, Sw, estimation will be affected.
arate following the invasion resistivity profile. process, interpret and evaluate logs quickly at For example, a log analyst will estimate the
(below). At the wellsite, operational safety and the wellsite. However, sometimes the wellsite water-filled resistivity, Ro, from a water-saturated
efficiency are improved by the elimination of the answer isn’t enough. For example, various para- bed (squeeze case), and estimate Rt in the pay
bridle and surface current system. sitic effects on resistivity measurements— zone (antisqueeze) using Archie’s equation. Errors
in both resistivities contribute to overestimating
HRLA array resistivity 1 water saturation in the pay zone, which can lead
to overlooked hydrocarbons.
HRLA array resistivity 2
How to tackle this problem? The 1D radial
inversions used in real time are for simpler cases
and do not address the fact that nearby adjacent
HRLA array resistivity 3 high-contrast beds may influence the resistivity
reading. Improved methods involve the combina-
tion of 1D shoulder-bed corrections followed by a
Bit size HRLA array resistivity 4 HALS deep resistivity 1D radial inversion, but they do not address the
0 in. 60
fact that the radial and vertical responses are
Invasion
diameter HRLA array resistivity 5 HALS shallow resistivity 1D Rt from HALS
coupled, leading to significant errors in resistivity
HALS determination. It has been long recognized that
0 in. 60 resistivity estimation can be improved by the use
HALS invasion MCFL microresistivity MCFL microresistivity 1D Rt from HRLA of inversion techniques that take into account true
1 ohm-m 100 2D or 3D formation structure (see ”Getting More
Invasion
diameter Pay from Resistivity Logs,“ previous page).
HRLA Depth 1D Rt from HRLA 1D Rt from HALS Application of 2D formation models to the inver-
1D Rt Increase
0 60 ft 1 ohm-m 100 1 ohm-m 100 sion technique can double the calculated reserves,
in.
particularly in thinly bedded formations.18
Spiral boreholes—Some drilling practices
produce a borehole with a 3D shape that has a
XX00 spiral groove on top of the bit-sized hole. Such
A boreholes produce a quasi-periodic character in
the logs, and are variously referred to as
”corkscrew“ or ”threaded hole.“ These have
been associated with downhole drill motors and
high-angle wells. Displaced stabilizers can pro-
duce the same effect in vertical wells. The effect
XX20 is seen as a periodic oscillation on the caliper
log. The effect on other logging tools depends
on the physics of the measurement. The impact
> Reducing shoulder-bed effects on laterolog measurements. In this example, HALS resistivities, of spiral boreholes on induction measurements
shown in track 3, are out of sequence because of shoulder-bed effects. Therefore, the HALS real-time has been extensively studied by BP Amoco
processing forces invasion diameter to bit size—thus computing zero invasion. As a result, the HALS- Exploration.19 Recently, filter techniques have
derived formation resistivity, Rt, defaults to the deep-reading resistivity value because the data are
inconsistent with the 1D formation forward model. Over the same interval, the HRLA tool is much less been developed to reduce the effects of spiral
sensitive to shoulder-bed effects, and the additional data from the array measurements, shown in borehole rugosity on logging tool measurements
track 2, provide a realistic estimation of the invaded zone. The invasion corrections are applied to all (see ”Dealing with Spiral Boreholes,“ next page).
HRLA array resistivity measurements for an accurate Rt evaluation. Results show that the maximum
For density tools, the effect of the spiral-grooved
resistivity from HRLA real-time inversion is 45% higher than that estimated by the HALS measurement,
leading to a 16% increase in reserves estimated in Zone A. hole is to produce a cyclic mudcake effect.
18. Smits et al, reference 7.
19. Webster M: “AIT Performance in Prudhoe Bay,” BP
Exploration Memorandum, October 1997.

48 Oilfield Review
 Dealing with Spiral Boreholes

For the AIT tool, saline mud combined with signal-processing techniques effectively cancel soidal harmonics are detected in AIT logs. Once
spiral or corkscrew boreholes can produce a the unwanted periodic signal. the harmonic components are determined, the
strong signal and completely smear an array There are three main steps: automatically task is to filter them out. For each array data
induction tool log (below). The origin of the detecting the primary signal and its harmonics, segment, a notch filter is designed with the
induction tool response distortion is likely due estimation of their frequency, and subsequent appropriate transfer and phase characteristics
to changes in standoff.1 Practical experience removal of all unwanted components. For AIT to remove—without phase distortion—all har-
shows that the effect is worst in 6-in. holes measurements, the method involves frequency monics from the spiral borehole (bottom). The
where the standoff distance is limited. As the spectrum estimation and peak identification filter is applied to the raw signal just after bore-
tool moves along the borehole, the standoff ribs over short logging segments, in which each array hole correction. Similar methods have been
on the tool tend to fall into the grooves, allowing data segment is replaced by a parametric proposed for handling the effects of corkscrew
the induction sonde to approach the borehole autoregressive model. The advantage of this rugosity on nuclear logs.2
wall—periodically reducing tool standoff. Using approach is automatic detection and frequency 1. Barber et al, reference 21, main text.
the fact that the spiral borehole introduces estimation of the dominant periodic components 2. Betts P, Blount C, Broman B, Clark B, Hibbard L,
Louis A and Oostoek P: “Acquiring and Interpreting
a periodic effect or distortion to the log, several in each segment. Typically, up to three sinu- Logs in Horizontal Wells,” Oilfield Review 2, no. 3 (July
1990): 34-51.

Before filtering After filtering

Resistivity, ohm-m Resistivity, ohm-m
CBM 1.0 10.0 100.0 1000.0 CBM 1.0 10.0 100.0 1000.0
Induction logs in a spiral borehole. On

>
LQC LQC
the left is a Platform Express induction
Valid flag Valid flag
1-ft 1-ft log from a Canadian well with a severe
50 2-ft 50 2-ft spiral borehole (left). Filtering the data
4-ft 4-ft from the spiral borehole well has
OR OR removed unwanted harmonic borehole
Mag mud flag Mag mud flag
Non Non
noise, and now the environmental flags
Mag Mag indicate that the 2-ft resolution logs are
valid (right). Chart-based C, bad hole B
and magnetic mud M log quality-control
100 100 (LQC) flags are shown to the left of
each log. The color of the chart-based
and bad hole flag shows the recom-
mended resistivity bed resolution. The
color of the magnetic mud flag indicates
either magnetic mud (red) or nonmag-
150 150 netic mud (yellow).

Hole size Array resistivity Hole size Array resistivity

4 in. 14 10 in. 60 in. 4 in. 14 10 in. 60 in.
Rm 20 in. 90 in. Rm 20 in. 90 in.
0 ohm-m 10 30 in. 0 ohm-m 10 30 in.

50 0 Frequency spectrum and notch filter.

>
Attenuation response, dB

-5 A power spectrum of the logging data

40 -10 shows several peaks (arrows) caused
-15 by cyclic standoff variations in tool
30 -20 standoff. A frequency response shows
Spectral power, dB

-25 the attenuation (top right) and phase-

20 -30 shift (bottom right) characteristics of
-35 a three-frequency notch filter designed
10
to remove periodic noise found in log-
100
ging data.
0
Phase shift, degrees

50
-10
0

-20 -50
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Frequency
-100
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Normalized frequency, Nyquist = 1

Summer 1999 49
 Processing flag statistical analysis
PEF flags up 0.7%
Density flags up 0.0%
Window 1 W2 W3 W4
BS average reconstruction error 1.93% 0.50% 1.29%
SS average reconstruction error -0.53% 0.03% 0.41% 0.70%
LS average reconstruction error 3.32% 2.29% -0.29% -0.58%

Cost function
Detector reconstruction Black areas show that the corresponding error flag is set
errors 0 200
Offset error
Density Tau loop error
2 g/cm3 3 Stabilization loop or crystal resolution error
Short spacing, Window 1 Long spacing, Window 1 Backscatter detector Short-spacing Long-spacing
Backscatter Low-energy window CR Low-energy window CR Low-energy window CR
Window 1 Window 2 Window 2 0 cps 10K 0 cps 5K 0 cps 1K
High voltage Total count rate Total count rate
Window 2 Window 3 Window 3 1600 V 1700 0 cps 500K 0 cps 50K
Total count rate Crystal resolution Crystal resolution
0 cps 1M 5 % 25 5 % 25
Window 3 Window 4 Window 4 Crystal resolution Form factor Form factor
-10 % 10 -10 % 10 -20 % 20 5 % 25 -0.5 % 0.5 -0.5 % 0.5
Form factor Caliper High voltage High voltage
-0.5 % 0.5 6 in. 16 1600 V 1700 1600 V 1700
XX300 XX300

XX400 XX400 A

> Density LQC logs. In a well drilled with nonbarite mud in the North Sea, the density inversion processing reconstruction errors (left) for all detectors are seen
to be nearly zero—closely tracking the center of each track, as expected. The density curve (yellow) from this interval is superimposed to highlight the large
change in densities computed across this interval. The global cost function log is low throughout the entire interval indicating good reconstruction and high
confidence in the inversion results. Statistical analysis of the reconstruction errors shows that every energy window is below its maximum bias level. The
hardware LQC logs (right) from this well show stable tool operation. As expected, the backscatter total count rates (black) in track 1 anticorrelate in Zones A
and B with those of the from the short-spacing detector (black) in track 2 and long-spacing detector (black) in track 3. No detector hardware error flags were
displayed in the green columns shown at the left side of each track.

Handling dip—The large volume of investi- is to blur the log response and to introduce has been excessively time-consuming for long
gation of induction tools complicates the inter- horns at the bed boundaries. log sections or when many thin beds are
pretation of their logs. Modern induction Traditional dip-correction algorithms for encountered. Recently, a new algorithm based
devices such as the AIT tool are designed for induction logs are limited in practice to angles on a maximum-entropy inversion of raw, bore-
use in vertical wells, and are carefully focused less than 50º because of an increasing nonlinear hole-corrected array data through a fast 1D
to limit their response to a relatively thin forma- response to dip and adjacent bed-boundary con- forward model has dramatically improved the
tion layer perpendicular to the borehole. trasts. As a result, interpreting resistivity from speed and ability to interpret multiarray
However, in wells at high relative dip, the induction logs at apparent dip angles over 50º induction logs in invaded formations at high rel-
response cuts across several beds, and the has been limited to iterative inversion using 1D ative dip angles (see ”Interpreting Induction
measurement is no longer focused in an iso- forward models. The presence of invasion has Logs in High Dip Angle Formations with
lated single layer. The effect of high relative dip added additional complexity to the geometry, Invasion,“ page 54).
requiring processing based on 3D inversion
codes. Even with fast computers, the processing

50 Oilfield Review
 Processing flag statistical analysis Black areas show that the corresponding error flag is set
PEF flags up 4.9% Offset error
Density flags up 1.4% Tau loop error
Window 1 W2 W3 W4 Stabilization loop or crystal resolution error
BS average reconstruction error -0.53% 1.44% 2.91% Backscatter Short-spacing Long-spacing
SS average reconstruction error -0.72% 1.44% 0.87% 0.48% Total count rate Total count rate Total count rate
LS average reconstruction error -12.59% -1.04% -2.13% -2.48% 1650 V 1750 0 cps 500K 0 cps 50K
Detector reconstruction Cost function Total count rate Crystal resolution Crystal resolution
errors 0 200 0 cps 1M 5 % 25 5 % 25
Short spacing, Window 1 Long spacing, Window 1 Crystal resolution Form factor Form factor
Backscatter Standoff
Window 1 Window 2 Window 2 5 % 25 -0.5 % 0.5 -0.5 % 0.5

Window 2 Window 3 Window 3 Form factor Density High voltage High voltage
Window 3 Window 4 Window 4 -0.5 % 0.5 standoff 1600 V 1700 1620 V 1720
-10 10 -10 10 -20 2.5 in. 0
% % % 20

XX800 XX800

XX900 Backscatter XX900 Short-spacing Long-spacing

> Density LQC logs. In another 52º deviated well drilled with heavy barite-weighted mud for a different operator in the North Sea, a section of the density pro-
cessing LQC logs tells a different story. The long-spacing and backscatter detector reconstruction errors (left) shown in tracks 1 and 3 are large because of
the presence of heavy barite mud (14 lbm/gal) [1.67 g/cm3] and a highly fractured borehole environment through a known coal bed. In other parts of the
well—with better borehole conditions—reconstruction errors were lower. The density algorithm automatically detects the presence of barite and changes
the weighting of the window count rates from each of the detectors to obtain the most accurate answer in this difficult environment. The hardware LQC logs
(right) from a higher section in this well show an unusual phenomenon. The total count rates (black) in the lower coal bed in Zone A, from the backscatter
detector (track 1) anticorrelate with those from the short-spacing detector (track 2) and long-spacing detector (track 3) as before, and as expected. However,
in Zone B, they all decrease uncharacteristically. This is apparently due to increased attenuation in the backscatter gamma ray flux due to thick barite mud-
cake in this zone. The density standoff curve (red) in the depth track confirms the increased thickness of mudcake over this zone.

Active Log Quality Control These literally form a log quality-control hierar- indicates a problem in calibration, excessive pad
Many questions arise when logs appear strange. chy from the top level of environmental process- wear or tool standoff. High intermittent values in
Raw data may be fine, but the computed forma- ing down to the bottom level of tool-specific the reconstruction errors indicate an abnormal
tion parameter, such as density, may look abnor- sensor performance and calibration. Following noise level on the measurement (hardware prob-
mal. This leads to questions: Is the tool in an are some examples showing how LQC actively lems), or instabilities in the inversion process
unusual formation? Is the software correct? Is the works to provide better logging answers both at that may be associated with bad borehole condi-
calibration correct? Is the tool working properly? the wellsite and afterwards.20 tions. A log of the cost function used in the inver-
Good log quality control (LQC) resolves these Density LQC—Window count-rate reconstruc- sion modeling helps to evaluate confidence in the
issues, and with the addition of real-time envi- tion errors from the density inversion algorithm estimations. Hardware LQC logs along with
ronmental corrections provides insight during are a measure of the ”health“ of the inversion detector count-rate logs are used to help confirm
acquisition into both the effects of the logging process (previous page). They pinpoint significant tool response, calibrations and stabilization in
environment on every tool measurement and the differences between modeled window count unusual environments such as in thick mudcake
way these measurements are being processed. rates and those measured for each detector. A with heavy muds (above).
Numerous LQC analyses, logs and flags are large systematic bias in the observed reconstruc-
20. For more on LQC and tool specifics see: Platform
available in the Platform Express system to tion error log for more than one energy window Express User’s Guide. Houston, Texas, USA:
ensure quality measurements and processing. Schlumberger Wireline & Testing, 1999.

Summer 1999 51
 Standoff
Low UBI amplitude High density
0.5 in. 0

LQC flags Long-spacing density

Standoff
PEF computation 2 g/cm3 3
Density computation
Short-spacing density Depth Low UBI transit time High
Density detector
2 g/cm3 3 ft
Accelerometer
Backscatter density Gamma ray Density
ADRP 2 g/cm3 3 0 API 150 2 g/cm3 3

A typical example of significantly high recon- XX850

struction errors occurs when the logging engi-
neer selects the wrong mud-type-processing
option—Barite or No-Barite. These two modes
correspond to different tool-response models as
well as specific inversion schemes, and the
choice has to be made according to the impor-
tance of photoelectric absorption in the mud, A
which is linked to the barite content. If the engi-
neer does not have exact knowledge of the mud
composition, the wrong option might be selected,
XX900
resulting in poor estimation of density and Pe.
However, the raw count rates are always
recorded, and since processing is applied on the
raw counts, it is always possible to recompute
correct results regardless of the acquisition mode.
Recently, a new ”switchless“ mud-density
algorithm was designed to eliminate problems
with incorrect mud-algorithm selection. Currently
being field-tested, the algorithm uses a unique
tool-response model, valid for all mud properties,
and a generalized inversion scheme based on a
XX950
barite indicator. The barite indicator provides an
estimation of the amount of barite in the mud at
each depth, and is based on low- and high-
energy window count-rate ratios in the back-
scatter detector.
Another example of how LQC provides active
control of the log acquisition process is the
automatic offset recalibration for high-resolution
density logging measurements. The robust low-
resolution window count-rate reconstruction
error helps identify slowly varying count-rate off- X1000
set errors on the backscatter and short-spacing
> Density profile caused by pipe groove. In this 97º-deviated wellbore, pipe groove filled with
detector measurements—used primarily for the mudcake produces a mudcake density profile, shown in the depth track. The backscatter (dotted),
high-resolution density logs. Any offset count- short-spacing (short dashed) and long-spacing (long dash) detector density curves show an
rate error detected in these measurements is cor- increasing density profile as they each see deeper into the formation. LQC flags, shown on the
rected before being used in the inversion left of track 1, verify that the tool readings in this environment are normal. The UBI Ultrasonic
Borehole Imager tool images—amplitude (track 1) and transit time (track 2)—confirm that the
algorithm to estimate the high-resolution forma- pipe groove contains mudcake. In other sections of the wellbore that have been scraped clean
tion parameters. Whenever LQC detects extreme of mudcake, each detector reads the same density. The characteristic parallel lines caused by
rugosity or bad borehole, the more robust low- the drillpipe scraping the borehole wall shown in Zone A confirm this. The standard resolution
resolution density logs are recommended. formation density (solid blue) derived from inverse modeling is shown in track 2.

52 Oilfield Review
 Limit of 4-ft logs Possible large errors on all logs
1000 1000
Possible large errors on shallow logs – 2-ft limit Possible large errors on 2-ft logs

Limit of 1-ft logs Possible large errors on 1-ft logs

100 100
Rt, ohm-m

Rt, ohm-m
AIT-family tools Recommended range
recommended operating range
using computed standoff method
10 10

Possible large
1 errors 1
on all logs

0.01 0.1 1 10 100 1,000 10,000 0.01 0.1 1 10 100 1,000 10,000
(Rt/Rm)(hole diameter/8)2 (1.5/standoff) (Rt/Rm)(Bit size/8)2

> Operating range for AIT induction measurements. Borehole resistivity, Rm, hole diameter and standoff limit the range of acceptable
AIT induction measurements (left). The chart shows that the effect of the borehole environment is most significant on the high-resolu-
tion logs. A web-based job planner can be used to determine what resolution is usable for a range of expected formation resistivity and
borehole parameters: Rt, Rm and bit size (right). The scatter of data represents the range of uncertainty on the input parameters.

An example from the North Sea shows how with the environment (above).21 The output tion with the InterACT communications system.
LQC helps build confidence in the tool measure- of the logic has four possible states—1-ft valid, The InterACT system provides the capability for
ments when the unexpected happens in unusual 2-ft valid, 4-ft valid or “Out of Range.” The last real-time transmission of log data and wellsite
borehole conditions (previous page). Each of the state is flagged when environmental parameters graphics to distant locations. This allows direct
three detectors in the Platform Express density are completely outside the range of the least and immediate communication and interaction
tool sees progressively farther into the formation. restrictive induction tool measurement. This between the offshore wellsite and consultants in
Like the invasion profile produced by an array- means laterolog tools would provide better resis- Aberdeen and London during log acquisition for
resistivity tool, array density measurements pro- tivity measurements. better and more timely decision-making.
duce density profiles. These profiles depend on However, the chart-based induction tool LQC For example, in one well, a wireline logging
the mud weight used. In wells drilled with light algorithm is based on smooth boreholes, and tool was unable to reach target depth due to
nonbarite mud, the density profile tends to does not always detect when the environment is wellbore deviation. The situation was confirmed
increase from low to higher density, as the detec- unfavorable for induction measurements. In while logging. An immediate decision was made
tors look deeper into the formation. Typically, in wells where the borehole is very rough or when to pull out of hole and go straight to a drillpipe-
wells drilled with high-density barite mud, the standoff is inadequate, spurious spikes and other conveyed logging option. In other cases, irregu-
normal density profile will be from high to low as anomalies might render high-resolution logs larities in borehole dimensions shown on the
each detector looks farther into the formation. unusable. Research has shown that high-fre- caliper were witnessed during logging and a
When first seen in boreholes, unexpected quency induction array signals come from near quick decision was taken to pull out of the hole
apparent density profiles were thought to be due the borehole, confirmed by the extremely sharp and rig down—eliminating the possibility of a
to hardware problems. However, LQC quickly ver- spikes near the tool axis seen on the shortest stuck tool. In all cases, real-time LQC provided
ified that the tool was functioning correctly. array Born-response tool sensitivity function. In confidence in the tool measurements during log-
Further examination revealed the answer: In these cases, a rugosity-detection algorithm com- ging, enabling operators to make appropriate
deviated wells, where pipe grooves frequently bines high-pass-filtered, short-array data with decisions based on environmental constraints,
occur, the borehole is scraped clean of mudcake, mud-resistivity information to make certain that and not on limitations in the tool performance.
giving an unpredicted density response. Now rugosity detection is dominant. The default well- Remote witnessing has also decreased costs
that the phenomenon is well understood, the site presentation is the most appropriate com- and improved safety by reducing personnel and
inversion algorithms are designed to accommo- posite-resolution log, which varies smoothly transportation requirements at the wellsite.
date this effect. between the 1-ft to 4-ft resolution log—based Logs and evaluations are immediately available
Environmental LQC for resistivity—The oper- on the combination of the chart and hole-rugosity to the experts who need them, and real-time
ational limits of the AIT induction measurement logic. In all cases, the three basic-resolution and LQC ensures that logging measurements are
have been incorporated into a ”fuzzy-logic“ composite-resolution logs are always recorded. valid and can be trusted. If problems occur,
algorithm that uses real-time inputs of caliper Remote witnessing—Recently, BP Amoco expert opinions are available to help with con-
and mud-resistivity measurements to determine Exploration initiated a program of remote wit- tingency plans and decisions.
the best resolution logging output consistent nessing on their wells in the Andrew field in the 21. Barber T, Sijercic Z, Darling H and Wu X: “Interpreting
North Sea by combining the capabilities of Multiarray Induction Logs in Difficult Environments,”
Transactions of the SPWLA 40th Annual Logging
Platform Express real-time LQC and interpreta- Symposium, Oslo, Norway, May 31-June 3, 1999,
paper YY.

Summer 1999 53
Interpreting Induction Logs in High Dip Angle Formations with Invasion

A new algorithm based on maximum-entropy measurement—weighted by the expected error term in the cost function—the results can have
inversion of borehole-corrected multiarray or noise in each measurement. For example, many high-frequency “wiggles” (below).
induction data through a fast 1D forward model if the model predicts array voltages that agree Including the entropy term in the cost function
has been developed and tailored for highly devi- well with those observed on each AIT receiver helps smooth out all the extraneous high-
ated wells.1 This algorithm provides the same coil, then its contribution to the cost function frequency information content—meaningless
interpretation for invasion that has previously is low. The second term is one proportional information below the resolution capability
been available only for vertical wells. The key to the total entropy in the resistivity log. This of the tool.
to maximum-entropy inversion is a fast forward term adds stability to the solution. Finally, The implementation of maximum-entropy
model. For this model, an analytical solution there is an empirical smoothing term included inversion processing is more robust when pairs
is used to compute the response of the AIT tool in the cost function. The smoothing term of arrays are inverted at the same time to obtain
in a layered formation with dip. The response also helps add stability to the solution, but formation resistivity. As a result, rational sets
of each array is computed by finding exact is used sparingly because it tends to decrease of array pairs are inverted together, with the
solutions to Maxwell’s equations for the beds vertical resolution. surprising property that the depth of investiga-
at a given dip angle. In implementation, The concept of entropy as applied to log data tion of each pair is the same as that of the
it is desirable to have the layer thickness less is not intuitive. In physics, entropy is a measure deeper-reading array. This means that radial
than the resolution of the sensors—typically of the degree of disorder in a system, and the response functions for the arrays can be used
layers 6- to 12-in. [15- to 30-cm] thick. second law of thermodynamics states that the to define the radial response of the inverted
The inversion is formulated on finding the total entropy of a system can never decrease logs. Furthermore, by weighting the results
unknown formation conductivity that minimizes during a change. Applied to log data, entropy of inverting the array pairs, the inverted logs
a cost function. Like the other inversions used is a measure of the departure of log values from can be focused radially to give logs with stan-
in the Platform Express system, the first term locally averaged values. For example, by setting dard AIT depths of investigation.
in this cost function is a measure of how well up a simple least-squares inversion of the thinly The combination of maximum-entropy
the forward model predicts each array raw layered model—done by including only the first processing, resolution-matched inverted forma-
tion resistivities and radial focusing is called
Maximum Entropy Resistivity Log INversion
Resistivity, ohm-m Depth Resistivity, ohm-m (MERLIN). This processing works on AIT
1.0 10.0 100.0 ft 1.0 10.0 100.0 data at all dip angles from 0º to approximately
0
80º (next page, top). In principle, it will work
Low 10 High up to 90º, but in practice, the parameterization
entropy entropy requires that the wellbore cut all beds
20
of interest. MERLIN processing replaces the
30 Born-response function filter-based standard
40 AIT processing. The resulting logs can
be inverted for invasion parameters as if they
50
were in a vertical well, but at any dip angle.
60 In addition, although maximum-entropy inver-
sion was developed to remove the effects
70
of high dip, the exact solution forward model
80 at the heart of the method works at any dip
90 angle, including zero.

100
Maximum-entropy solution. A simple least-squares
>

110 inversion of the thinly layered model results in a log

with a high noise component (left). In the high-
120 entropy solution (right), the dotted curve indicates
the increased entropy found in the maximum-
130 entropy solution.
140

150

54 Oilfield Review
 Wellsite-processed MERLIN-processed MERLIN-processed
resistivity resistivity (75 deg) resistivity (65 deg)
AIT 10 in. AIT 10 in. AIT 10 in.

Depth Hole deviation AIT 20 in. AIT 20 in. AIT 20 in.

m 70 degrees 120
Bit size AIT 30 in. AIT 30 in. AIT 30 in.
4 14 Correcting for dip. In a section of a deviated well

>
Caliper Azimuth AIT 60 in. AIT 60 in. AIT 60 in.
drilled with oil-base mud, real-time processed AIT
C2 -40 degrees 360
induction wellsite logs show non-normal invasion
C1 Corrected gamma ray AIT 90 in. AIT 90 in. AIT 90 in.
profiles, resistivity horns and overshoots in four zones,
4 in. 14 0 API 150 0.2 ohm-m 2K 0.2 ohm-m 2K 0.2 ohm-m 2K
shown in track 2. The first post-log MERLIN processing
X830 results in track 3 assumed that the wellbore was at 75º
relative to the beds, but better results in track 4 were
obtained assuming a 65º deviation. The final result
shows coherent resistivity separation in the top zone
and a dramatic increase in Rt, indicating more pay
in each zone.

X840 < Resolving shoulder-bed anomalies. The gamma ray

curve (green) in track 1 clearly shows shale beds
above and below the reservoir section. The wellsite
AIT 1-ft induction logs in track 2 show anomalous
behavior with large resistivity oscillations and a non-
coherent separation over a large zone near the bottom
and top of the reservoir. The 2-ft and 4-ft logs (tracks
4 and 5) display similar behavior. The anomalies are
X850
caused by high shoulder-bed contrast coupled with low-
resistivity shoulders which violates the Born-approxi-
mation used in the wellsite processing. After MERLIN
processing, the anomalies are eliminated from the
logs in track 3, leaving the middle thick section of
the reservoir unaffected.

Wellsite (1-ft) MERLIN-processed Wellsite (2-ft) Wellsite (4-ft)

resistivity resistivity resistivity resistivity
AIT 10 in. AIT 10 in. AIT 10 in. AIT 10 in.

AIT 20 in. AIT 20 in. AIT 20 in. AIT 20 in.

For example, formations with high shoulder
conductivity and with shoulder-bed contrasts Bit size AIT 30 in. AIT 30 in. AIT 30 in. AIT 30 in.
can be modeled correctly with the exact-
solution-based MERLIN processing. In a vertical Caliper AIT 60 in. AIT 60 in. AIT 60 in. AIT 60 in.
well drilled with oil-base mud, anomalies are 12 in. 17
Depth Gamma ray AIT 90 in. AIT 90 in. AIT 90 in. AIT 90 in.
seen on the bottom and top of the reservoir m 0 API 150 0.2 ohm-m 2K 0.2 ohm-m 2K 0.2 ohm-m 2K 0.2 ohm-m 2K
section in the real-time AIT logs (right). The
high contrast between the reservoir resistivity X240
(100 ohm-m) and that of the conductive clay
shoulder beds (less than 1 ohm-m) causes the
standard logs to overshoot in the resistive reser- X245
voir. This happens because standard AIT pro-
cessing—based on Born-based response
functions—includes contributions from many X250
depth intervals around the tool. Although these
functions are accurate models of the induction
response in low-to-moderate-contrast envi-
X255
ronments, when the shoulder-bed contrast
approaches 500:1, the Born-based approxima-
tion for the tool response is poor. The MERLIN
X260
logs are well-behaved in these difficult induction-
logging environments.

1. Barber TD, Broussard T, Minerbo G, Sijercic Z and X265

Murgatroyd D: “Interpretation of Multiarray Induction
Logs in Invaded Formations at High Relative Dip Angles,”
The Log Analyst 40, no. 3 (May-June, 1999): 202-217.

Summer 1999 55
 Real-Time Interpretations
1500 Wellsite interpretation during log acquisition is
another benefit of real-time environmental cor-
rections. Platform Express interpretations feature
integrated petrophysical computations and
graphic presentations that help operators make
timely decisions about the reserves in their field
during the logging run. These include a complete
formation volume and lithology evaluation, fluid
saturation analysis, invaded-zone gas saturation
and a special horizontal well presentation.
For example, formation porosity is derived
from the traditional crossplot of tool-measured
bulk density, ρB, and the thermal neutron poros-
ity. Comparing neutron response with that of the
density measurement, a lithology-corrected for-
mation porosity is determined. With these
results, porosity-corrected formation grain den-
sity, ρmaa, and volumetric matrix photoelectric
factor, Umaa, are computed, to provide inputs
needed for a standard ρmaa-Umaa crossplot-based
1600
mineral interpretation. By its nature, this cross-
plot method is independent of porosity. Finally,
clay volume, Vcl, derived from gamma ray or
spontaneous potential (SP) measurements,
provides a third dimension to the standard
mineralogy crossplot (below).

Gamma ray Computed Rt Radial Density

Depth Sw image
0 API 150 ft 70 ohm-m 120 1.95 g/cm3 2.95
10.00
Caliper Standoff AIT 90-in. resistivity 20.00
Neutron porosity
(resistivity)
6 in. 16 0.2 ohm-m 2K 30.00
45 p.u. -15
40.00
SP Standoff Rxo 50.00
(density) 60.00
70.00
0 MV 200 0.2 ohm-m 2K 80.00
90.00
Tension Pe
0 lbf 5K 0 10
90 0 90 Zone of interest

> Formation lithology analysis. The ρmaa-Umaa

crossplot, derived from neutron and density tool
measurements, forms the basis for a standard
mineralogy interpretation (right). Porosity is
effectively removed from these inputs, resulting
in a crossplot that is a function only of the miner-
alogy in the formation. End-points for pure anhy-
drite, sandstone (quartz), dolomite and limestone VCI
(calcite) are shown. The lithology column color Quartz increases
change depends on clay volume, Vcl , and is
derived from the SP or gamma ray measurement.
Colors assigned to the plot are derived from a ρmaa
color cube with the corners of yellow, green,
cyan and white on one face, which represents
Dolomite Calcite
zero Vcl, and corners of red, black, blue and
magenta on the opposite face, which represents Anhydrite
100% Vcl. The Platform Express lithology column
shows significant lithology changes in the lower Umaa
section of the BP Amoco Catoosa test well (above).

56 Oilfield Review
 Horizontal well presentation. In this well,

>
Well trajectory Depth, m Resistivity Porosity the wellbore penetrated salt and anhydrite
layers, and entered the sand reservoir. Below
the 75-m sand reservoir, the lithology log in
track 3 indicates that the well appears to
have entered another anhydrite layer and
then turned back into a second sand interval.
However, the well trajectory plot (track 1)
X100 clearly shows that the wellbore turned
upwards prior to entering the anhydrite
Anhydrite layer—suggesting that the well simply reen-
tered the caprock anhydrite and then turned
down—back into the first sand reservoir.
Without the well trajectory plot, analysis
would have been delayed until other logs and
deviation information could be correlated to
the log data to explain the lithology changes.
Sand reservoir X150

X200

Anhydrite

X250 The Road Ahead

More than two-thirds of the ”triple-combo“
logging operations performed by Schlumberger
in 1998 were done with the new-generation
platform logging technology. Real-time depth
and environmental corrections provide the key to
getting accurate formation information into the
X300 hands of the operator when and where it does
the most good—at the wellsite. Real-time,
speed-based corrections are being implemented
on every tool platform that contains a built-in
accelerometer, such as the Xtreme quad-combo
tool string designed for high-pressure and high-
The clay volume and the mineralogy crossplot Finally, a special wellsite Platform Express pre- temperature environments and the SlimAccess
drive the real-time lithology column profile. The sentation has been developed for highly deviated quad-combo tool string designed for the slim and
color change depends on Vcl and the position of and horizontal wellbores (above). A real-time true complex-geometry borehole environment. New
each point on the ρmaa-Umaa crossplot. This lithol- vertical depth (TVD) computation—based on the technology, such as the HRLA tool, addresses the
ogy image provides a valuable method to observe deviation derived from the built-in tool-axis traditional problems encountered in conductive
lithology changes in the formation. Although not accelerometer—is used to plot a wellbore shape boreholes and promises to bring more accurate
intended to provide a volumetric interpretation of versus cable depth using the lithology image as an results in a wider range of environments.
the mineralogy of the formation, it can be used area fill in the wellbore trajectory curve. The avail- Improved 2D and 3D forward models are
effectively for well-to-well correlation. ability of a real-time well deviation and lithology helping to clarify increasingly more difficult
Next, porosity and resistivity measurements profile at the well helps to explain unexpected log- logging environments. Breakthroughs in com-
are coupled with Archie’s equation to determine ging results often encountered when crossing puting speed will eventually lead to their
water saturation. An optional invaded-zone gas or beds and fault zones in highly deviated wells. increased application. Advanced processing, such
steam saturation—computed from the density- as MERLIN and HRLA 2D processing are available
neutron crossover separation—is used to com- in computing centers and may be available
pute a gas-corrected neutron-density porosity. someday for real-time operation. These develop-
ments promise an exciting time for geologists,
reservoir engineers and petrophysicists who use
openhole-logging measurements. —RH

Summer 1999 57
 Contributors

Walt Aldred, Drilling Performance Product Cham- John Cook leads the Geomechanics group in the Vidhya Gholkar, a Senior Research Scientist with
pion, is based in Houston, Texas, USA. Previously, Well Construction department at Schlumberger Schlumberger Oilfield Research, is based at
he was seconded from Anadrill to Schlumberger Cambridge Research in England, where he works Schlumberger Cambridge Research in England. He
Cambridge Research in England, where he worked on on wellbore stability control, sand management, joined the laboratory in 1992 and currently works
the Real-Time Wellbore Stability project and projects perforating strategies and improvements to the in the Real-Time Drilling Decisions group. He has
related to drilling learning. This involved leading the drilling process. John is a graduate of the University contributed to various projects, including MSM*
Schlumberger PERFORM* team, a group dedicated of Cambridge with a BA degree in materials science (muds solids monitor), sonic labeling, quantitative
to reducing drilling costs and improving drilling and a PhD degree in physics. risk analysis, drillstring failure, multiphase flow
efficiency. In 1980, he joined the company, working in Mike Cooper, a consulting geophysicist with a strong water-cut determination, real-time wellbore stability
West Africa and the North Sea. He later spent five background in seismic processing, does 4D seismic and drilling problem avoidance. Two years ago, Vidhya
years in Sugar Land, Texas, developing drilling technical consulting, seismic inversion studies, organized and cochaired the Schlumberger Signal
interpretation products and then was in Nigeria amplitude variation with offset (AVO) analysis and Processing Conference in Austin, Texas. He has
for four years working on drilling optimization and seismic processing quality control. Before this (1996 BS and PhD degrees in electrical and electronic
engineering. Holder of a patent in drilling motor to 1999), he was a development geophysicist on the engineering from the University of Birmingham,
optimization, Walt earned a BS degree (Hons) in Foinhaven field for BP Amoco. He also served as a England, and a Certificate in Management from the
geology and chemistry from the University of geophysical analyst for the West of Shetlands area and Open University in Milton Keynes, England.
Durham, England. spent four years as seismic processing geophysicist Shuja Goraya is working as a Schlumberger
Tom Barber has worked on induction modeling, array (central North Sea), responsible for seismic data PERFORM engineer in Cabinda, Angola, on several
design and environmental corrections since he joined quality in the central North Sea asset group (ETAP). projects involving drilling efficiency improvement,
Schlumberger Well Services in Houston, Texas, in 1978. He began his career in London, England, with Shell and on the design and execution of a shallow gas
There he developed log processing algorithms for the UK Exploration & Production Ltd. as an assistant geo- extended-reach drilling project. Shuja joined Anadrill
AIT* Array Induction Imager family of tools, and physicist (1981 to 1982). Following this he moved to in 1994 as a drilling services engineer and worked as
the first commercial signal-processing algorithm for Seismograph Services Ltd. in Bromley, Kent, as senior measurements-while-drilling and logging-while-
resistivity tools, Phasor* processing. His most recent seismologist. His next assignment was as seismic drilling, directional drilling and drilling engineer in
work has involved interpreting AIT logs at high well processing geophysicist in Glasgow, Scotland (1988 to various parts of West Africa. He obtained a BS degree
deviation and other difficult environments. Author 1990). Mike earned a BS degree in physics and (Hons) in electronics engineering from University of
of numerous papers and holder of eight patents, he geophysics (Hons) from University of Bath, England. Engineering and Technology, Lahore, Pakistan.
was awarded the SPWLA Distinguished Technical Chip Corbett, Principal Engineer for Schlumberger Laurent Jammes, who is in the Interpretation
Achievement Award in 1993 for significant contribu- Holditch-Reservoir Technologies (H-RT) in Houston, Engineering department in Clamart, France, is leader
tions in electromagnetic logging. He previously worked Texas, joined Schlumberger in 1981 as a wireline of the Invasion-2.0 Interpretation project. He began
on magnetic susceptibility logging measurements at field engineer in Sacramento, California, USA. During with Schlumberger in 1988 as a development engineer
Schlumberger-Doll Research, Ridgefield, Connecticut, the next eight years he had various engineering and physicist, working on density measurements.
USA. Before joining Schlumberger he worked at the assignments, primarily in California. Since 1990 he From 1992 to 1998, he was project leader for Platform
National Aeronautics and Space Administration has been committed to software marketing for Express* software and data integration, responsible
(NASA), Marshall Flight Center, Huntsville, Alabama, GeoQuest and integrated field studies for H-RT, and for sensor physics studies, real-time measurement
USA, and at Brookhaven National Laboratory, Upton, has authored or co-authored several papers on processing and interpretation products for the
New York, USA. He has a BA degree in physics from reservoir characterization. Chip has a BS degree Platform Express tool. Laurent is a graduate of Ecole
Vanderbilt University, Nashville, Tennessee, USA, and in mechanical engineering from the University of Centrale de Paris and earned a doctorate degree in
did graduate work on low-temperature magnetism at California at Berkeley and an MS degree in petroleum nuclear physics at Commissariat à l’Energie Atomique
the University of Georgia, Athens, USA. engineering from the University of Houston. in Saclay, France.
Jack Bouska, Geophysical Associate at BP Amoco in William (Liam) Cousins, Drilling Superintendent for Werner Klopf, senior petrophysicist in Milan, Italy
Sunbury on Thames, England, has been involved in 3D the Mungo Wells team, works for BP Amoco in Dyce, for Schlumberger, works on interpretation develop-
acquisition design, and 4D and 4C seismic imaging at Aberdeen, Scotland. After joining BP in 1980, he held ment, particularly for nuclear magnetic resonance
BP Amoco since January 1999. In 1981 after receiving a various drilling engineering and drilling supervisory measurements. After joining the company in 1978, he
BS degree in geophysics from the University of Alberta positions in the North Sea and in Russia. He has been had various field assignments in South America. From
in Edmonton, Canada, and studying electronics engi- drilling superintendent since late 1997. Liam holds 1984 to 1994, he was a log analyst with several postings
neering at Southern Alberta Institute of Technology, a BS degree in geology from National University of in Milan and Paris. Werner has been in his current
Canada, he joined Seiscom Delta United. From 1983 to Ireland in Cork. position since 1994. His degree in petroleum engineer-
1985, he was with Western Geophysical in Calgary, ing was earned at University of Leoben, Austria.
Alberta, Canada, and spent the next three years with John Fuller is the Schlumberger Holditch-Reservoir
Dome Petroleum Ltd. in Calgary. He began with Amoco Technologies coordinator for geomechanics in Europe, Alberto Malinverno is a research scientist in the
in 1988, working first in Calgary and then in London, Africa and the CIS. He joined Schlumberger as a Reservoir Optimization department at Schlumberger-
England. In 1995 Jack won the best theme paper award wireline field engineer in 1980 in Abu Dhabi, United Doll Research, Ridgefield, Connecticut. He is working
at the Canadian Society of Exploration Geophysics Arab Emirates, and for the next 10 years had various on inverse techniques that combine different data
national convention. field assignments in the Middle East. In 1990 he sources to construct a model of the reservoir that is
moved to Paris, France, to work on applications and consistent with the available information and to
Ian Bradford is a senior research scientist in the Well products during the launch of the DSI* Dipole Shear quantify the uncertainty associated with the model
Construction department at Schlumberger Cambridge Sonic Imager tool. After transferring to the London description. Before joining the company in 1992,
Research in England. He joined Schlumberger in 1991 Computing Centre later that year, John worked in the Alberto spent three years doing postdoctoral studies
and, most recently, worked on the Real-Time Wellbore UK and Norway on several geomechanics projects and as an associate research scientist at the Lamont-
Stability project. Prior to that, his activities were supporting drilling and production operations. He has Doherty Earth Observatory of Columbia University,
focused on well planning, sanding and bit mechanics. served as Technical Vice-President of the London Palisades, New York. His undergraduate degree in
Ian holds BS and PhD degrees in applied mathemat- chapter of the SPWLA and is a member of the 1999 geological sciences is from Università degli Studi di
ics from the University of Nottingham, England. SPE Forum steering committee for sanding. John Milano, Italy; his MS and PhD degrees, also in geologi-
received a BS degree in physics from the University of cal sciences, are from Columbia University, New York.
Portsmouth, England.

58 Oilfield Review
Reginald Minton, Well Integrity Technology Challenge Anchala Ramasamy is a petrophysicist who works Jan Wouter Smits, who is based in Clamart, France,
Leader for BP Amoco and Project Manager for the on the completions team for high-value wells at BP has been project manager charged with developing
company’s “No Drilling Surprises” R&D objective, is Amoco in Aberdeen, Scotland. There she provides the new HRLA* High-Resolution Laterolog Array tool
presently based in Aberdeen, Scotland. He first joined petrophysical support for the company’s Intelligent and inversion answer products. He started with
BP in 1976 and worked in drilling fluids before moving Wells project, with particular emphasis on electrical Schlumberger in 1991 in Clamart, working as an
to Anchor Drilling Fluids in 1983 as technical director arrays, optical logging methods, and installation of electronics design engineer on the ARI* Azimuthal
and then UK operations manager. He returned to BP permanent sensors in wells and their integration with Resistivity Imager tool and subsequently on resistivity
in 1986 and has held a series of drilling engineering, other disciplines to maximize data value. She works on measurements for the Platform Express tool. Jan has
R&D project management and exploration drilling a team specializing in analysis of cased-hole nuclear an MS degree in electrical engineering from Delft
operations management posts prior to assuming his and production logs and their integration with other University of Technology in The Netherlands.
current position in February 1999. He has a BS degree data sources to optimize production and reduce Alan Sibbit is an Interpretation Advisor (Petrophysics)
from Hatfield Polytechnic, Hertfordshire, England, risks and uncertainties. She began her career with and manager of the Schlumberger Center for Advanced
and a PhD degree from the University of Aberdeen Schlumberger in 1990 as a field engineer working Formation Evaluation in Houston, Texas. Previously he
in Scotland. Reginald has been an SPE Distinguished with Shell in Aberdeen. From 1993 to 1994, she was a worked in various interpretation assignments in the
Lecturer and received the Stavanger SPE Engineer of senior field engineer for open- and cased-hole services field and in research. He joined Schlumberger in 1975.
the Year award in 1997. in Kuwait. In 1994 she joined GeoQuest in Aberdeen Alan holds BA and MA degrees in mathematics from
Andrew (Andy) O’Donovan is a reservoir engineer as geoscientist and petrophysicist for multiclient data St. John’s College, University of Cambridge, England.
and geophysicist with the BP Amoco West of processing and analysis for GeoFrame* environments.
Two years later, she became operational petrophysicist Robert Terry is a petrophysical associate working in
Shetlands Subsurface Team in Aberdeen, Scotland. international operations and doing nuclear magnetic
Prior to this (1996 to 1999), he worked on field involved in data acquisition for the BP Andrew field.
She moved to BP Amoco in 1998. Anchala holds a resonance processing and job planning for BP Amoco
development and production with the BP Exploration in Houston, Texas. After receiving a BS degree in
Foinaven Subsurface Team. His 11 years with BP BS degree (Hons) in aeronautical engineering from
City University, London, England. physics with a minor in geophysics at the Georgia
Exploration have been spent in frontier exploration, Institute of Technology in 1975, he joined Schlum-
prospect evaluation, appraisal, development and Laurence Reynolds, Petrophysicist for the Schlum- berger as a field engineer in West Texas. He spent three
production in the UK, Vietnam and China. Andy holds berger Formation Evaluation group in Aberdeen, years working in various open- and cased-hole field-
a BS degree (Hons) in physics with geophysics from Scotland, provides support for openhole wireline engineering assignments before taking a series of
the University of Bath in England and Masters degrees operations, sales and marketing, as well as interpre- management and log interpretation positions. In 1988
in petroleum engineering from Heriot-Watt University, tation training, in an integrated group involving LWD he joined Amoco to perform international log analysis.
Edinburgh, Scotland, and in geophysics from Imperial (Anadrill), Wireline and GeoQuest. He joined the He is a graduate of the Amoco Petrophysics training
College in London, England. company in 1986 and spent seven years as a field program, where in subsequent years he taught log
Dick Plumb is Principal Consultant, Geomechanics engineer in various European locations for land and analysis. An active member of the SPWLA and author
at Schlumberger Holditch-Reservoir Technologies offshore cased-hole and openhole wireline opera- of several papers, Bob is chairman of the Log
in Houston, Texas. Previously, he was responsible for tions. From 1993 to 1997, he was engineer in charge Characterization Consortium.
case studies in the interpretation and geomechanics for AGIP in Aberdeen, Scotland. Laurence has a
BS degree in engineering physics from Queen’s Dean Tucker, a senior well engineer at the
department at Schlumberger Cambridge Research, Schlumberger Integrated Project Management (IPM)
England. He also worked at Schlumberger-Doll University, Kingston, Ontario, Canada.
group in Aberdeen, Scotland, has spent the past four
Research, Ridgefield, Connecticut, where he Sarah Ryan has been the program manager of the years in research and operational assignments. He is
developed log interpretation techniques for fracture seismic reservoir characterization and monitoring currently Project Coordinator for the Upstream
characterization, in-situ stress measurements and group at Schlumberger Cambridge Research (SCR) Technology Group, BP Amoco. This is a joint venture of
hydraulic fracture containment. Dick received a in England since 1998. From 1996 to 1998, she was a Schlumberger and BP Amoco to develop a process for
BA degree in physics and geology from Wesleyan research scientist at SCR involved in seismic reservoir improving drilling performance using 3D visualization
University, Middletown, Connecticut; an MA degree characterization and time-lapse seismic studies. technology. He is also lead engineer in the review
in geology from Dartmouth College, Hanover, New She began with Schlumberger in 1990 as a wireline- of Mungo field drilling performance. In 1985, after
Hampshire, USA; and a PhD degree in geophysics logging engineer, and spent three years working in earning a BS degree in engineering from Memorial
from Columbia University in New York. Indonesia and Australia. Sarah has also worked for University of Newfoundland in St. John’s, Canada, he
Michael Prange is a senior research scientist at BHP petroleum and Woodside Offshore petroleum. joined Chevron Canada Resources, Calgary, Alberta,
Schlumberger Doll Research (SDR), Ridgefield, Her degrees include a BS degree in geology from the as a drilling engineer and spent the next five years
Connecticut, where he is working on the Validation University of Melbourne, Victoria, Australia; and a BS in field and office operations. From 1991 to 1995,
Gauntlet, a suite of tools to validate shared earth (Hons) degree and a PhD degree in petroleum geology he served as a petroleum engineer in technical,
models against all available data. A validated model and geophysics, both from the University of Adelaide, administrative and managerial assignments. Dean
will include a description of uncertainty on all South Australia, Australia. joined Sedco Forex in Paris, France, as a senior
model components. Michael earned a PhD degree drilling engineer in the engineering division to provide
in geophysics from Massachusetts Institute of An asterisk (*) is used to denote a mark of Schlumberger. drilling engineering support for drilling operations
Technology (MIT) in Cambridge, Massachusetts, worldwide. He transferred to Aberdeen two years
USA in 1989, and held postdoctoral positions as a later. He assumed his current position in 1997.
Fulbright Scholar at Elf Aquitaine in Pau, France, in
1990 and at the Earth Resources Laboratory at MIT.
He joined SDR in 1991.

Summer 1999 59
NEW BOOKS

• Sequence Stratigraphy of the • Structure, Stratigraphy and

Westphalian in the Northern Part Petroleum Geology of the SE Nam
of the Southern North Sea Con Son Basin, Offshore Vietnam
• Modelling of Sandbody Connectivity • Predicting Reservoir Quality During
in the Schooner Field Exploration: Lithic Grains, Porosity
• Application of High-Frequency and Permeability in Tertiary Clastics
Cycle Analysis in High-Resolution of the South China Sea Basin
Sequence Stratigraphy • Miocene Carbonate Buildups Offshore
• Index Socialist Republic of Vietnam
• Exploration History of the Offshore
The book is well illustrated and SE Sumatra Production Sharing
nicely produced. Contract, Java Sea, Indonesia
...it will be of great value to • N. Sumatra/Java Oil Families
geologists working in the southern • Fault Seal Analysis of SE
Petroleum Geology of North Sea. It also provides some Asian Basins with Examples Whole Earth Geophysics:
the Southern North Sea: useful lessons from this basin that from West Java An Introductory Textbook for
have broad applicability to other areas.
Future Potential • Permeability Heterogeneity Within Geologists and Geophysicists
K. Ziegler, P. Turner and Sweet M: AAPG Bulletin 82, no. 7 (July 1998): the Jerudong Formation: An Outcrop Robert J. Lillie
S. R. Daines (eds) 1440-1441. Analogue for Subsurface Miocene Prentice Hall, Inc.
Geological Society Publishing House Reservoirs in Brunei Upper Saddle River, New Jersey
Unit 7, Brassmill Enterprise Centre, 07458 USA
• The Tectonic Evolution and
Brassmill Lane 1999. 361 pages. $73.00
Associated Sedimentation History
Bath, Somerset BA1 3JN UK ISBN 0-13-490517-2
of Sarawak Basin, Eastern Malaysia:
1997. 210 pages. $99.00
A Guide for Future Hydrocarbon
($60.00 for AAPG members) The book focuses on how each geo-
Exploration
ISBN 1-897799-82-9 physical technique provides information
• Platform-Top and Ramp Deposits on the Earth’s internal structure and
The book reviews the topics that are
of the Tonasa Carbonate Platform, tectonic development. Each chapter
important in the exploitation of gas in
Sulawesi, Indonesia contains definitions of terms used and
the southern North Sea hydrocarbon
province. It describes the advanced • A Review of the Petroleum Potential has many diagrams, a bibliography and
techniques and technologies needed of Papua New Guinea with a Focus on a set of questions.
for stratigraphic correlation and for the Eastern Papuan Basin and the
Contents:
improving production efficiency. Pale Sandstone as a Potential
Reservoir Fairway • Introduction
Contents:
• An Overview of the Hydrocarbon • Plate Tectonics
• Introduction
Potential of the Spratly Islands • Seismic Waves
• History of Exploration in the Petroleum Geology of Archipelago and Its Implications
Southern North Sea Southeast Asia • Seismic Refraction Interpretation
for Regional Development
A.J. Fraser, S.J. Mathews and • Seismic Reflection: Acquisition,
• Recent Advances in Understanding • The Tectonostratigraphic Evolution
R.W. Murphy (eds) Processing, and Waveform Analysis
the Southern North Sea Basin: of SE Asia
Geological Society Publishing House
A Summary • Structural and Tectonic
Unit 7, Brassmill Enterprise Centre, • Cenozoic Deformation in Sumatra:
• Permian (Upper Rotliegend) Interpretation of Seismic
Brassmill Lane Oblique Subduction and the
Synsedimentary Tectonics, Basin Reflection Profiles
Bath, Somerset BA1 3JN UK Development of Sumatran
Development and Palaeogeography 1997. 436 pages. $125.00 ($65.00 for Fault System • Earthquake Seismology
of the Southern North Sea Geological Society members) • Gravity and Isostasy
• Tertiary Response to Oblique
• Climatic Cyclicity and ISBN 1-897799-91-8 Subduction and Indentation in • Magnetic Interpretation
Accommodation Space in Arid to This book comprises papers presented Sumatra, Indonesia: New Ideas for
Semiarid Depositional Systems: An • Heat Flow
at a two-day 1995 conference of the Hydrocarbon Exploration
Example from the Rotliegend Group Petroleum Group of the Geological • Appendices, Index
• A Tectonic Model for the Onshore
of the UK Southern North Sea Society of London. The papers cover Kutai Basin, East Kalimantan While not a replacement
• Compartmentalization of most of the active areas of petroleum
• New Observations on the Sedimentary for more traditional introductory
Rotliegendes Gas Reservoirs by exploration in Southeast Asia, with
and Tectonic Evolution of the Tertiary geophysics texts...a valuable addition
Sealing Faults, Jupiter Fields Area, the majority pertaining to Indonesia
Kutai Basin, East Kalimantan to the library of any geo-professionals
Southern North Sea and Vietnam.
charged with educating newcomers
Contents: • Gravity Anomalies and Deep
• Diagenetic Controls on Reservoir to the field or those moving from
Structural Controls at the Sabah-
Quality in Permian Rotliegendes • Energy Trends in SE Asia small-scale prospecting to regional
Palawan Margin, South China Sea
Sandstones, Jupiter Fields Area, geophysical interpretation.
• Cenozoic Plate Tectonic • Index
Southern North Sea
Reconstructions of SE Asia Avakian B: The Leading Edge 17 (November 1998):
• Probing the Lower Limits of the ...because of that diversity
• Characterizing Petroleum Charge 1610-1611.
Fairway: Further Pre-Permian [of topics], there should be something
Systems in the Tertiary of SE Asia
Potential in the Southern North Sea of interest for most explorationists
• Exploring the Lake Basins of East and attracted to Southeast Asia.
• The Structure of the Westphalian in
Southeast Asia
the Northern Part of the Southern Hatley AG: APG Bulletin 82, no. 9
North Sea • Exploration in the Gulf of Thailand (September 1998): 1763-1764.
in Deltaic Reservoirs, Related to the
Bongkot Field

60 Oilfield Review
 • Fracture Distribution in Faulted • Turbidite Flux, Architecture
Basement Blocks: Gulf of Suez, Egypt and Chemostratigraphy of the
Herodotus Basin, Levantine Sea,
• Polygonal Faulting in the Tertiary of
SE Mediterranean
the Central North Sea: Implication for
Reservoir Geology • Sediment Delivery to the Gulf of
Alaska: Source Mechanisms Along a
• Fault and Fracture Characteristics
Glaciated Transform Margin
of a Major Fault Zone in the
Northern North Sea: Analysis of 3D • Large Scale Debrites and Submarine
Seismic and Oriented Cores in the Landslides on the Barra Fan, W of
Brage Field (Block 31/4) Britain
• Structural Geology of the Gullfaks • Morphology and Sedimentation on
Field, Northern North Sea the Hebrides Slope and Barra Fan,
NW UK Continental Margin
• Index
• Debris Flows on the Sula Sgeir Fan,
It comes packaged in a slim 266
Structural Geology in NW of Scotland Unlocking the Stratigraphic
pages with clear illustrations...Every
Reservoir Characterization paper has an extensive bibliography. • Shallow Geotechnical Profiles, Record: Advances in
M.P. Coward, T.S. Daltaban and Acoustic Character and Depositional Modern Stratigraphy
I learned so much and heartily
H. Johnson (eds) History in Glacially Influenced Peter Doyle and
recommend it to all geoscientists and
Geological Society Publishing House Sediments from the Hebrides and Mathew R. Bennett (eds)
reservoir and petroleum engineers.
Unit 7, Brassmill Enterprise Centre, West Shetland Slopes John Wiley & Sons
Brassmill Lane Hossack J: Marine and Petroleum Geology 15, no. 8
• Mechanical Properties of 1605 Third Avenue
Bath, Somerset BA1 3JN UK (December 1998): 842-843.
Terrigenous Muds from Levee New York, New York 10158 USA
1998. 266 pages. $115.00 ($57.00 for Systems on the Amazon Fan 1998. 532 pages. $64.95
Geological Society members) ISBN 0-471-97463-3
• The Var Submarine Sedimentary
ISBN 1-897799-94-2
System: Understanding Holocene Sed-
iment Delivery Processes and Their This multiauthored book describes
Intended for geologists, geophysicists how new geophysical and laboratory
Importance to the Geological Record
and engineers, the book aims to techniques are being used to quantify
capture the wide range of research on • Cenozoic Changes in the properties and relationships among
reservoir characterization and also to Sedimentary Regime on the rocks and to interpret stratigraphic
promote synergy between geoscientists Northeastern Faeroes Margin successions.
and engineers. • The Southeast Greenland Glaciated
Margin: 3D Stratal Architecture of Contents:
Contents: Shelf and Deep Sea • Establishing the Sequence:
• The Role of Structural Geology in • Recent Geological Processes in Lithostratigraphy:
Reservoir Characterization the Central Bransfield Basin Principles and Practice: Remote
• Reservoir Characterization and (Western Antarctic Peninsula) Sensing and Lithostratigraphy;
Modelling: A Framework for Field • Seismic Stratigraphy of Palaeogene Field Interpretation of Complex
Development Geological Processes on Depositional Sequences: Northeast Tectonic Areas; Evolutionary
Continental Margins: Rockall Trough Concepts in Biostratigraphy; Event
• Fault Seal Prediction: The Gouge Stratigraphy: Recognition and
Ratio Method Sedimentation, Mass-Wasting • Sediment-Drift Development on the Interpretation of Sedimentary
• Experimental Fault Sealing: and Stability Continental Margin off NW Britain Event Horizons; Cyclostratigraphy;
Shear Band Permeability M.S. Stoker, D. Evans and • Cyclic Sedimentation on the Strontium Isotope Stratigraphy;
Dependency on Cataclastic A. Cramp (eds) Faeroe Drift 53-10 Ka BP Related to Borehole Data and Geophysical
Fault Gouge Characteristics Geological Society Publishing House Climatic Variations Log Stratigraphy; Principles of
Unit 7, Brassmill Enterprise Centre, Seismic Stratigraphy; Sequence
• Identification and Spatial Brassmill Lane • Late Quaternary Stratigraphy and
Distribution of Fractures in Porous, Palaeoceanographic Change in the Stratigraphy; Stratigraphical
Bath, Somerset BA1 3JN UK Applications of Radiogenic Isotope
Siliciclastic Sediments 1998. 355 pages. $120.00 Northern Rockall Trough, North
Atlantic Ocean Geochemistry; Chronostratigraphy
• Flow Through Fault Systems in ($72.00 for AAPG members) (Global Standard Stratigraphy):
High-Porosity Sandstones ISBN 1-987799-97-7 • Upper Slope Sand Deposits: A Personal Perspective
• Physical Character and Fluid-Flow The Example of Campos Basin, Latest
This collection of papers focuses on Pleistocene/Holocene Record of the • Interpreting the Record:
Properties of Sandstone-Derived continental margin processes that Interpreting the Record: Facies
Fault Zones Interaction Between Alongslope and
are becoming crucial to the safe Downslope Currents Analysis; Interpreting Sea Level;
• Assessment of the Effects of development of deep-water oil fields. Interpreting Palaeoenvironments
Sub-Seismic Faults on Bulk Perme- • Hemipelagites: Processes, Facies from Fossils; Interpreting
abilities of Reservoir Sequences Contents: and Model Paleoclimates; Interpreting
• Reservoir Characterization: • Geological Processes on • Late Glacial to Recent Accumulation Orogenic Belts: Principles
How Can Anisotropy Help? Continental Margins: Fluxes of Sediments at the Shelf Edge and Examples
Sedimentation, Mass-Wasting and Slope of NW Europe, 48-50 • Index
• Index
• Numerical Simulation of Fluid Flow
and Stability: An Introduction Degrees N
in Complex Faulted Regions Highly recommended for
• Curvature Analysis of Gridded • Large Submarine Slides at the The papers in this volume ... all earth science libraries.
Geological Surfaces NE Faeroe Continental Margin have considerable actual or potential Beck JH: Choice 36, no. 2 (October 1998): 346.
commercial value.
• Strain Partitioning During
Flexural-Slip Folding AAPG Bulletin 82, no. 10 (October 1998): 1880.

Summer 1999 61
Coming in Oilfield Review

Managing Production. Focused • Mechanisms

efforts are required to maximize
the economic potential of any • The Bronze Age
asset. By outsourcing oilfield • Alloys and Compounds
management, operators are now • Atoms in Motion
able to realize the most benefit
from their available infrastructure, • The Iron Age
resources and services. Utilizing • Iron and Steel
new technology, production • Degradation of Materials
analysts and operating personnel
from service companies develop • The Age of Electronic Materials
techniques to increase oil and • Electrical Properties of Materials
gas field longevity, productivity, • Magnetic Properties of Materials
profitability and reserve recovery.
• Optical Properties of Materials
Seismic-Data Acquisition. The • Thermal Properties of Materials
newest addition to the fleet Understanding Materials • No Ceramics Age? Petroleum Geology of the
of Schlumberger vessels—the Science: History, Properties, North Sea: Basic Concepts
Geco Eagle—was designed Applications • From Natural Fibers to
Man-Made Plastics
and Recent Advances
not only according to seismic- Rolf E. Hummel K.W. Glennie (ed)
survey acquisition requirements, Springer-Verlag • Gold Blackwell Science
but also with the needs of 175 Fifth Avenue • Economic and Environmental Commerce Place
workers in mind. In this article, New York, New York 10010 USA Considerations 350 Main Street
we explain how application of the 1998. 407 pages. $59.95 Malden, Massachusetts 02148 USA
principles of ergonomics results ISBN 0-387-98303-1
• What Does the Future Hold?
1998. 636 pages. $85.00
in the design of equipment and • Appendices, Index ISBN 0-632-03845-4
work-area environments that Intended for engineering, physics
help increase operational safety, and materials science students, He has succeeded admirably...in
Longer than previous versions, the 4th
efficiency and productivity. this textbook links elementary, but rendering intrinsically complicated
edition reflects expanded knowledge
quantitative explanations of phenom- topics, such as polymerization,
about North Sea hydrocarbons. It
Remote Well-Process Control. ena to historical eras, the history and palatable and digestible.
includes a historical review of North
Real-time reservoir management, growth of understanding and control, This mix of proper science and Sea exploration and a summary of
maximizing reserve recovery and and issues of environment, substitu- respectable history is something the geologic controls on hydrocarbon
increasing asset value require tion of materials, and recycling and new among the plethora of materials distribution in this area.
more than simply selecting and reserves of raw materials. science books....As a first-level
deploying completion equipment introduction to materials science, Contents:
in a wellbore. We discuss new Contents: I can recommend it unreservedly. • Historical Review of
technology—measurements • The First Materials (Stone Age Cahn R: MRS Bulletin 23, no. 12 North Sea Exploration
from sophisticated monitoring and Copper-Stone Age)
equipment and downhole (December 1998): 48. • Origin, Development and
sensors, software for modeling • Fundamental Mechanical Evolution of Structural Styles
and interpretation, and special Properties of Materials • Devonian
devices that control flow from
• Carboniferous
reservoirs remotely—to help
fine-tune well performance and • Lower Permian—Rotliegend
optimize field production. • Upper Permian—Zechstein
• Triassic
Scale Removal and Control. The
buildup of scale can eventually • Jurassic
shut down a producing well. • Cretaceous
Traditional scale treatments
• Cenozoic
have not always been successful
and were highly specialized for • Source Rocks and Hydrocarbons of
specific types of scale. This the North Sea
article discusses the causes of • North Sea Plays: Geologic Controls
scale formation and how new on Hydrocarbon Distribution
removal and control techniques
• References, Index
were developed. Armed with
these innovations, scale can be ...the overall presentation is
removed from wells quickly, and crisper and neater than before.
at low cost. The book will be essential to those
with an interest in the North Sea, but
it will also be valued by a far wider
stratigraphic, petroleum-geological
and structural readership.
Journal of Petroleum Geology 21, no. 4
(October 1998): 483.

62 Oilfield Review
 Call or Copy and FAX this form.
Multimedia Oilfield Review Archive

Multimedia, graphics and

D
C
Platform: ❏ Macintosh ❏ Windows ❏ UNIX

E
IV
Computer Model

O A 98
H
reprints from
C W 9
ER C
K E -1
V R
Quantity Ordered @$25.00 each = Total $

C I 2
A V 9
B RE : 19
Corporate Express Document and Print Management

Oilfield
D LD LE 3635 Willow Bend, Suite 300

MORA
SI IE B

Houston, Texas 77054

IN F LA

☎1(713)660-7226 • Fax 1(713)660-0362

E OIL AI

Review
E
SE IA AV

E-mail: rmartin1@idt.net
ED OW

Multimedia
IM N

Oilfield Review MORA or Oilfield Review Services Order Form

Archive
Name
LT
U

Company
M

MORA Address
Your electronic journey into the pages of
Oilfield Review begins here. City State Postal Code
This archive provides quick and easy access
Country
Did you know that to Oilfield Review articles published from 1992
to 1998. Select a button to explore articles by
topic, issue or search by keyword.
Telephone

Oilfield Review can Fax E-mail

Method of Payment: ❏ MasterCard ❏ Visa ❏ Check
be found on-line?
About Topics Issues Help Search

Account Number-Please include all digits

Upper sands
Producing Injection Lower sands

Point your browser to

(2 wells)
well wells (3 wells)
M
L N
Production
water Steam K O

www.connect.slb.com to find: J
1.0 1.5
2.0
2.5
P
VAT number (EEC countries only)
0.5

• technical presentations, forums and Steamflooded

upper sands
I A

H B
Please print name as it appears on credit card
publications, including Oilfield Review Waterflooded
lower sands G C

Expiration Date
F D
E

• detailed service information,

applications and case studies A D H 100
90
Order Date Signature

% upper sand production, steamflood

E
80
F L I Before steamflood
B

• log interpretation charts and K

J
70

60
After steamflood

Statement Address (if different from above)

G

nonexclusive seismic maps C 50

40

30
0 ft 300

• an oilfield dictionary, mnemonics Oil from lower sands only

20

10
Oil from upper sands only

database and other oilfield links Commingled oils

0
A B C D E F G H
Well
I J K L

• location and telephone directory

• DataLink* secure data transfer Viewgraph and Slide sets of Oilfield Review Illustrations
• InterACT* real-time data transmission Available from Corporate Express/Ross-Martin Article title No. of sets Subtotal
from wellsites Multimedia (call for prices)
• SuperVision* project monitoring. MORA—the Multimedia Oilfield Review Archive—provides 1.
Connect Schlumberger is the Internet quick and easy access to issues published from 1992 to 1998. 2.
communications resource for Schlumberger MORA preserves the look of the printed magazine in a cross-
clients. Updated daily, this password-protected platform format. Full-color articles can be printed or explored 3.
interface is growing rapidly, providing timely on-screen using key-word searches and cross-document links. 4.
and interactive access to the information, data Specify: ❏ Viewgraphs: $16 each ❏ Slides: $14 each Total $
and tools needed by the E&P industry. Available from Oilfield Review Services:
You can also visit the public Color 35-mm slides and viewgraphs Oilfield Review Back Issues and Articles
Schlumberger Web site at The ideal way to prepare high-quality sales presentations and
www.slb.com. Back-issue price:$15. Call for article reprint price.
lectures—Oilfield Review Services offers color 35-mm slides
* Mark of Schlumberger and viewgraphs of illustrations from Oilfield Review and The Issue date Price Quantity Subtotal
Technical Review. The cost of a slide or viewgraph set ranges 1.
from $120 to $450. Please call for more information. 2.
Article reprints and back issues 3.
Reprints of Oilfield Review articles can be supplied. Minimum 4.
order is 500. Back issues of the Oilfield Review and The
Call for added cost of delivery. Total $
Technical Review can also be ordered, subject to availability.
Specify: ❏ Courier ❏ Air mail
Exhibition posters $
Special display materials such as posters or large transparencies, ORDER TOTAL
incorporating illustrations from Oilfield Review, can be ordered.

Oilfield Review Services also creates tailor-made multimedia

Oilfield Review
Oilfield Review Services
presentations for exhibitions, conferences and seminars, and can
Barbour Square, High Street, Tattenhall,
provide staff and equipment to run the presentations.
Chester CH3 9RF England
☎ (44) 1829-770569 • Fax: (44) 1829-771354
Browse Oilfield Review Services online at www.t-e-s.co.uk E-mail: orservices@t-e-s.co.uk

SMP-6068
SCHLUMBERGER OILFIELD REVIEW
Oilfield Review
Summer 1999

SUMMER 1999

Drilling Risk Management

VOLUME 11 NUMBER 2

Reservoir Model Validation

Real-Time, Wellsite Log Corrections