Oilfield Review

Spring 2008

Intelligent Gas Storage

Zonal Isolation
Smart Materials
Seismic Inversion
08-OR-002-0
Zonal Isolation—Where Do We Go from Here?
One of the most important functions of a cement sheath is
to provide hydraulic zonal isolation between different loca-
tions in the wellbore. In the 1970s, the E&P industry faced
a significant challenge: understanding a phenomenon
known variously as gas migration, annular flow after
cementing or, simply, flow after cementing. In the simplest
terms, flow after cementing is a loss of well control experi-
enced in the first few hours following a cement job.
Wells were cemented using slurry with a density greater
than that of the mud that had been used for drilling the wells.
Yet these wells, which had been successfully controlled by
the drilling-mud density, could flow. This was puzzling.
Operators and service companies spent roughly two
decades performing cement research and trying field solu-
tions to solve the puzzle. We studied many aspects of slurry
design and performance, including fluid loss, free fluid,
permeability, static gel-strength development and shrink-
age in small- and large-scale laboratory tests. Instru-
mented field tests documented the loss of pressure in
actual wells during cement hydration.
Mud-removal techniques were improved through better
centralization, casing movement, fluid properties (density
and rheological hierarchies) and the power of computer
simulators to replace rules of thumb to engineer cement
placement for specific well applications. The result is that
today’s industry is well-equipped to meet the challenge of
providing zonal isolation in the short term through effec-
tive slurry design and efficient mud-removal technique.
Now the industry faces another important challenge:
improved understanding and maintenance of zonal isola-
tion throughout a well’s productive life and even beyond. It
has become apparent that the materials and techniques
that successfully achieve short-term zonal isolation are not
always sufficient to maintain that isolation in the longer
term. Sustained casing pressure (SCP) is pressure
between the well’s casing and tubing, or between strings of
casing, that rebuilds after being bled down. SCP and its
mirror image, casing vent flows, have been documented in
many fields worldwide, and cement-sheath damage is one
of the possible causes for these long-term problems.
As early as 1989, a field case was reported in which zonal
isolation was lost because of extensive fractures and fine
fissures in the cement matrix. This damage resulted from
repeated thermal cycling of the cement sheath in a geo-
thermal well. In the latter half of the 1990s, the first publi-
cations appeared dealing with damage of the cement
sheath caused by stresses imposed after its initial place-
ment and hydration. The stresses described result primar-
ily from changes in temperature and pressure during the
productive life of the well. Examples of such changes
include increasing temperature due to production of
hydrocarbons, temperature variations from cyclic-steam
enhanced-oil recovery, pressure changes due to required
casing-pressure tests, and pressure changes due to
change-out of well fluids from heavier drilling fluids to
lighter completion brines. These temperature and pres-
sure changes can cause loss of zonal isolation through the
formation of microannuli or stress fractures in the cement
sheath, or both. The small size of such defects makes them
difficult, if not impossible, to identify and repair using con-
ventional remedial techniques. Thus, prevention of the ini-
tial failure is important.
With respect to long-term isolation, the industry is now in
a position similar to that of the short-term flow-after-
cementing puzzle in the late 1970s. The problem has been
identified and the first steps in prediction and prevention
have been taken (see “Ensuring Zonal Isolation Beyond the
Life of the Well,” page 18). Much work remains to be done.
Test methods used to determine parameters—such as
Young’s modulus and Poisson’s ratio—need to be refined
and standardized so they can reliably be used in predictive
software. The most commonly measured cement mechani-
cal property is unconfined uniaxial compressive
strength—a result of a test adapted from the construction
industry and based on geometries common in construc-
tion. The measured value does not directly apply to the
performance of a cement sheath under confined downhole
conditions. The long-entrenched paradigm that higher
compressive strength is always better compressive
strength will gradually change.
The software models themselves (like the early versions of
cement-slurry placement software) have potential for growth
in predictive capability and refinement of the stresses
modeled. For example, what effect do compaction and other
formation changes caused by production have on isolation?
Inevitably, new techniques and materials will develop to
meet the market’s needs. Exciting times are ahead!
Craig Gardner
Cementing Team Leader and Cementing Consultant
Chevron
Houston, Texas, USA
Craig Gardner is a Consultant in cementing and Cement Team Leader at
Chevron in Houston. After receiving a BS degree in chemistry from the Univer-
sity of Houston, he worked for a major drilling fluids company prior to joining
Gulf Oil as a drilling supervisor in 1980. He is involved in Chevron’s worldwide
cementing operations through technical services, technology development and
training. Craig is a member of SPE, API and ISO and is a former chairman of
the API Subcommittee on Well Cements.
1
 Schlumberger

Oilfield Review
Executive Editor 4 Intelligent Well Technology in Underground
Mark A. Andersen Gas Storage
Advisory Editor
Lisa Stewart
Although the intelligent well technologies used in natural gas-
storage wells are similar to those used in producing wells, the
Editors information is often utilized in ways that are quite different.
Matt Varhaug
Rick von Flatern As underground gas-storage operations evolve, these new
Vladislav Glyanchenko approaches are proving to be an ideal platform for innovative
Tony Smithson applications. The results are storage facilities that look more
Contributing Editors like the hydrocarbon fields of the future than the winter supply
Rana Rottenberg depots of the past.
Judy Jones

Design/Production
Herring Design
Steve Freeman

Illustration
Tom McNeff
Mike Messinger
George Stewart 18 Ensuring Zonal Isolation Beyond the Life of the Well Gas
Liquid
Printing The cement sheath behind casing not only must support Solid
Wetmore Printing Company
Curtis Weeks
a well’s structure, but more importantly it must prevent
unwanted fluid flow. And it must do so for many years beyond
the well’s lifetime. A new, self-healing cement does just that,
and a new logging tool helps boost operator confidence in the
success of the primary cement job.

On the cover:

The WesternGeco Western Spirit is

equipped with automated source and
streamer steering to provide repeatable
time-lapse seismic studies and sophisti-
cated over/under and rich- or wide-azimuth
surveys. The inset shows an inversion of
seismic data used to characterize com- Useful links: Address editorial Address distribution inquiries to:
plex lithologies. correspondence to: Tony Smithson
Schlumberger Oilfield Review Oilfield Review
www.slb.com 5599 San Felipe 12149 Lakeview Manor Dr.
Houston, Texas 77056 USA Northport, Alabama 35475 USA
Oilfield Review Archive (1) 713-513-1194 (1) 832-886-5217
www.slb.com/oilfieldreview Fax: (1) 713-513-2057 Fax: (1) 281-285-0065
E-mail: editorOilfieldReview@slb.com E-mail: DistributionOR@slb.com
Oilfield Glossary
www.glossary.oilfield.slb.com

2
 Spring 2008
Volume 20
Number 1

32 Intelligence in Novel Materials Magnetorheological Liquid Advisory Panel

Abdulla I. Al-Kubaisy
Smart materials have properties that respond in a controlled Saudi Aramco
manner to changes in their environment. These materials can Ras Tanura, Saudi Arabia
perform sophisticated functions, sometimes several simultane- Dilip M. Kale
ously. This capability makes smart materials promising for use ONGC Energy Centre
New Delhi, India
in a variety of devices, from household appliances to complex
scientific instruments such as downhole intervention tools. Roland Hamp
Woodside Energy, Ltd.
Perth, Australia
George King
Rimrock Energy LLC
Denver, Colorado, USA
Eteng A. Salam
PERTAMINA
Jakarta, Indonesia
Jacques Braile Saliés
Petrobras
Houston, Texas, USA
42 Seismic Inversion: Reading Between the Lines Richard Woodhouse
Independent consultant
Seismic waves are primarily used to identify subsurface struc- Surrey, England
ture, but they also contain valuable information about the rock
and fluid properties of the formations they pass through. The
process of seismic inversion uses calibration information from
wells to extract formation properties from seismic reflection
amplitudes. This article examines inversion and presents
examples from 3D, time-lapse and multicomponent surveys.

64 Contributors Oilfield Review is pleased to welcome

68 New Books and Coming in Oilfield Review
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
www.oilfieldreview.com
Annual subscriptions, including postage,
are 200.00 US dollars, subject to
exchange-rate fluctuations.
Oilfield Review is published quarterly by
Schlumberger to communicate technical
advances in finding and producing hydro-
carbons to oilfield professionals. Oilfield
Review is distributed by Schlumberger to
its employees and clients. Oilfield Review
is printed in the USA.
Contributors listed with only geographic
location are employees of Schlumberger
or its affiliates.
© 2008 Schlumberger. All rights reserved.
No part of this publication may be repro-
duced, 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.
Jacques Braile Saliés to its Advisory
Panel. Jacques is Wells Operation
Manager of Petrobras America for the
Gulf of Mexico. His 27-year career at
Petrobras has been spent in various
engineering and management positions
in E&P, including coordination of the
Petrobras Technological Program on
Ultra-Deepwater Exploitation Systems—
PROCAP 3000. He served on the SPE
Board of Directors for Brazil and has
authored or coauthored papers on
drilling and subsea technology. Jacques
received a BS degree in mechanical
engineering from the Military Institute
of Engineering (IME), Rio de Janeiro, an
MS degree in petroleum engineering
from the Federal University of Ouro
Preto (UFOP), Brazil, and a PhD degree
in petroleum engineering from the
University of Tulsa.

3
 Intelligent Well Technology in
Underground Gas Storage

Intelligent well technologies are ideal for underground gas-storage facilities.

Formation properties have been determined; storage capacity and deliverability
can be modeled; and analytical tools can track historical production trends.
These technologies provide efficient, cost-effective storage and delivery systems,
helping secure the position of natural gas as a dependable energy resource.

Kenneth Brown When it comes to applying intelligent well
Pittsburgh, Pennsylvania, USA technologies to oil and gas production, a prime
objective is maximizing the value of a continually
Keith W. Chandler diminishing asset. For underground natural gas-
John M. Hopper storage facilities, the application of these smart
Lowell Thronson
technologies can be significantly different,
Falcon Gas Storage Company, Inc.
primarily because the gas reservoir can be
Houston, Texas, USA
replenished. Thus, it is the ability to repeatedly
James Hawkins inject natural gas into and withdraw it from
Midland, Texas underground storage at high rates that must be
optimized and intelligently managed.
Taoufik Manai Formation properties define the optimal level
Paris, France at which a well flows at high recovery rates. As the
stored natural gas is recovered from the reservoir,
Vladimír Onderka the pressure decreases and flow rates fall. Cushion
RWE Transgas Net gas, the gas that remains in place between
Brno, Czech Republic injection and withdrawal cycles, ensures that
there is sufficient pressure to maintain the desired
Joachim Wallbrecht
BEB Transport und Speicher Service GmbH minimum flow rates on withdrawal. The pressure
Hannover, Germany and volume provided by the cushion gas also
diminish the likelihood of water influx into the gas
Georg Zangl cap and can prevent gas/water contact movement.
Baden, Austria Because the most expensive component of an
underground gas-storage (UGS) facility can be the
For help in preparation of this article, thanks to Edmund cushion gas, minimizing its volume and
Knolle, Falcon Gas Storage Company, Houston; and
Monsurat Ottun, Houston. understanding the reservoir well enough to define
BorView, BlueField, DECIDE!, ECLIPSE, ELANPlus, FMI the efficient operating range can reduce the
(Fullbore Formation MicroImager), NODAL, Petrel and overall development cost of a storage project, as
PIPESIM are marks of Schlumberger.
Excel is a mark of Microsoft Corporation. well as greatly enhance project profitability.
In hydrocarbon production, intelligent well
technologies allow reservoir engineers to use
information such as decline curves, material-
balance relationships, inflow-performance-
relationship (IPR) curves and reservoir
simulations and models—all in real time or
almost real time.1 A sophisticated system may
automatically take corrective action or alert the
operator that intervention is warranted. The
ultimate goal of intelligent production wells is to
deliver more oil and gas with greater efficiency—
at a lower cost.
While UGS facilities also benefit from effi-
ciencies and lower cost provided by intelligent
well technologies, they are not operated to
maximize the recovery of hydrocarbon. In fact,
gas-storage operations in many parts of the world
are more analogous to a bank than a producing
reservoir. Just as currency flows into and out of a
bank, assets in the form of natural gas flow into
and out of the storage reservoir. When called
upon, sometimes months after injection into a
storage reservoir, but increasingly within a few
days or even hours, the gas is delivered to a buyer
who supplies it to industrial and residential
customers. Banks have automated the flow of
currency and capital between institutions and
users; similarly, storage facilities are automating
the flow of natural gas between producers
and consumers.

4 Oilfield Review
 > Image courtesy of Falcon Gas Storage Company, Inc.
UGS operations also differ from traditional
gas production because the wells must be able to
withstand high injection pressures, something
rarely experienced in producing wells, and the
withdrawal rates from UGS can be 5 to 10 times
greater. UGS wells have long life expectancy;
therefore maintaining well integrity and reservoir
integrity are crucial aspects of successful
operations. Due to the rapidly changing opera-
tional modes—injection to withdrawal—the
operator must be reactive and act quickly to avoid
well and reservoir mechanical damage. As a vital
1. Hydrocarbon depletion occurs in a predictable manner
based on formation properties and completion hardware.
The rate of decline of reserves can be plotted to define a
decline curve.
Material balance is an expression for conservation
of mass. The amount of mass leaving a control volume
Spring 2008
component in the secure delivery of natural gas,
these facilities must be managed appropriately.
Maintaining reliable natural gas supplies has
recently become a geopolitical priority in many
areas of the world. Governmental regulations,
such as those in the European Union, have had
increased influence on how the UGS industry
conducts business. Intelligent well technology is
being adopted as a natural by-product of these
developments because it helps facilitate
automatic storage and delivery of natural gas,
improves the operational efficiencies of these
is equal to the amount of mass entering the volume,
minus the amount of mass accumulated in the volume.
Through material balance, reservoir pressures measured
over time can be used to estimate the volume of hydro-
carbons remaining.
Inflow-performance relationship, IPR, is a tool used in
production engineering to assess gas-well performance
facilities, and assists in optimizing the
management of assets (gas) in the ground.
The degree of implementation varies
between different operations, and not all UGS
facilities operate using these relatively new
technologies. However, the improved operating
performance that has been demonstrated is
prompting operators to retrofit and upgrade
many older storage operations—sometimes
yielding unexpected benefits.
by plotting the well production rate against the flowing
bottomhole pressure (BHP). The data required to create
the IPR are obtained by measuring the production rates
under various drawdown pressures during a multirate
test. The reservoir fluid composition and behavior of the
fluid phases under flowing conditions determine the
shape of the curve.
5
 After a brief review of gas-storage basics, this An Underground Gas-Storage Primer
article examines the different levels of intelligent Traditionally, natural gas has been considered a
well technology being applied to underground seasonal fuel because of the higher demand for
gas-storage operations in North America and heating during winter months. Beginning in the
Europe. We present case studies showing how 1940s, the US natural gas industry recognized
real-time data are used to identify damage in that long-distance pipeline capacity was not
storage wells and how the implementation of new sufficient to supply natural gas to large
optimization and surveillance techniques has population centers during peak-demand periods.
improved performance. Also included is a To balance the gas-demand cycle, a gas-storage
description of a cutting-edge, automated opera- network was developed to inject gas into
tion that integrates three levels of intelligent underground storage facilities when demand was
well technology.2 low and to release gas during periods of high
demand. This buffering of demand is referred to
as peak-shaving.
Global Working-Gas Volume Distribution by Storage Types
Depleted oil and gas fields
81.6%
Aquifers
14.5%
Salt caverns 3.9% Abandoned mines
0.02%
> Underground gas storage by type. UGS facilities can take several forms,
but depleted hydrocarbon reservoirs and saline aquifers make up 96% of
the global supply. The choice of storage type can be driven by availability:
aquifers and salt caverns make up 34% of Western Europe’s storage capacity
compared with just 14% in the USA where there is greater access to
depleted fields. (Adapted from Wallbrecht, reference 6.)
Residential
Electric power
1.0 Commercial 28
24
0.8
20
Gas usage, billion m3
Underground gas storage, however, has been
available almost as long as long-distance
pipelines. In 1915, natural gas was first
successfully stored underground in Welland
County, Ontario, Canada. Several wells in a
partially depleted gas field were reconditioned,
and gas was injected into the reservoir during the
summer and withdrawn the following winter.
In 1916, Iroquois Gas Company placed the
Zoar field, south of Buffalo, New York, USA, into
operation as a storage site, and it is still in
operation today. In 1919, the Central Kentucky
Natural Gas Company injected gas into the
depleted Menifee gas field in Kentucky, USA. By
1930, nine storage sites in six different states
were in operation with a total capacity of about
18 Bcf [510 million m3]. Before 1950, essentially
all underground gas storage consisted of reused
partially or fully depleted gas reservoirs.
Today, the two primary types of underground
gas-storage locations are caverns and porous
reservoirs. Leached salt caverns and abandoned
mines account for a small portion of the total
storage capacity, while depleted oil and gas
reservoirs and saline aquifers are by far the most
common UGS medium (left). Salt-cavern storage,
better suited for high-rate delivery and injection,
is primarily used for peak-day delivery purposes.3
Typically, 20 to 30% of the gas must remain in
place to maintain the structural stability of the
cavern. Saline aquifers can provide high-rate
delivery, but cushion-gas requirements are
significant, ranging from 50 to 80% of the total
storage capacity. By far the most common type of
storage, depleted hydrocarbon reservoirs are
used for seasonal delivery or buffering high
demand. Typically, 30 to 50% of the storage
capacity must be maintained as cushion gas.4
In recent years, UGS withdrawal practices
have changed in the USA because of the
increased use of natural gas for electricity
generation. Drawdown during summer months is

higher than in the past because natural gas is

Gas usage, Tcf

0.6
16
being used to generate electricity for air
12
conditioning and cooling requirements (left). In
0.4
many ways, this has altered the scope of gas
8 storage in the USA. UGS facilities that are
0.2 located in proximity to gas-fired power plants are
4
used to moderate the supply for seasonal, as well
0 0 as hourly, variations. On a daily basis, gas in
2004 2005 2006 2007 2008
storage can be tapped during high-demand
Year
periods and stored during low-demand periods.
> The cyclicity of natural gas usage. As a source of residential heating in
Commercial pipelines may be incapable of
the USA (blue), natural gas from storage peaks during winter months. When
used for generating electricity to provide cooling (red), usage peaks during supplying sufficient quantities of gas during the
summer months. Commercial usage, driven by temperature (black), tracks peak periods—or putting away gas during
residential demand. Also note the rise in natural gas usage for electricity periods of low demand—but the UGS facility can
generation in successive years. [Adapted from http://www.eia.doe.gov/oil_gas/
make up the shortfall in either case.
natural_gas/info_glance/natural_gas.html (accessed February 29, 2008.)]

6 Oilfield Review
 UGS is not just a North American Working-Gas Volume Distribution by Region in 2004
phenomenon—storage facilities are currently Middle East 2%
operating in 33 countries—although the USA has
by far the greatest number. As an energy source,
Eastern Europe 42%
natural gas utilization in residential and North and South
commercial sectors of Western Europe has America 35%
exceeded 44%, highlighting the importance of
maintaining a secure, uninterrupted supply. In Western Europe 19%
France, storage facilities have at times supplied
more than half the residential gas needed to Asia 2%
meet temperature-driven demand.5 Total
Western Europe has recently experienced a Eastern Europe
350 and Middle East

Working-gas volume, 109 m3

rise in gas trading between holding companies 300 Americas
and market suppliers. The use of gas-storage 250
Western Europe
facilities is often driven by short-term buying and 200
selling, rather than traditional peak-shaving. 150
Profitability for both buyers and suppliers is 100
determined by the ability of the UGS facility to
50
store and deliver gas on demand in a cost- 0
effective manner. 1970 1975 1980 1985 1990 1995 2000 2005

In 1997, there were 580 UGS sites worldwide, Year

of which 448 were in depleted reservoirs. In 2006,
of the estimated 606 UGS sites, the number in
depleted reservoirs had grown to 495.6 In 1996,
there were 92 UGS operations in Europe,
excluding Russia. By 2006, the total number had
grown to 127—a 38% increase. The working-gas
volume in storage facilities in the same area grew
from 60.6 million m3 [2.14 Bcf] to 110.5 million m3
[3.9 Bcf], an 82% increase.
Although the USA has had a slight decrease in
the number of UGS sites between 1995 and 2004,
its total storage capacity has experienced a
marginal increase through improved field
utilization and retrofitting of existing facilities
(top right). Many of these older UGS operations
were developed before the introduction of the
reservoir modeling tools available today. Advances
in sensor technology and surface equipment are
being applied to these older facilities, making
them “smarter” and more versatile.7
How Smart Is an Intelligent Well?
The relative intelligence of gas-storage
operations can be grouped into three levels.
Level I, automated data flow, is reactive: receive
data, analyze data and respond. Level II,
surveillance and optimization, is reflective but
focuses on action: analyze data, compare and
validate models, manage models and determine
necessary courses of action. Level III can be
described as the digital oil field: integrate
processes, optimize, automate and operate
remotely, where it is applicable, in a proactive
manner (bottom right).
> A growing supply. The working-gas volume has grown steadily in the past 35 years with most of the
increase occurring outside North and South America (blue), especially in Eastern Europe and the
Middle East (black), which includes Russia. Current projections indicate that capacity is insufficient to
meet the long-term demand and increased growth is required. (Adapted from Wallbrecht, reference 6.)
Level I, Reactive
Level II, Active
Level III, Proactive
Reservoir Digital oil field
Surveillance and optimization
Wells
Dynamic data: SCADA system
Static data: well, reservoir
Gathering and technology
system
Facilities
> Levels of intelligence. Three levels can be identified in the
implementation of intelligent well technology. Each level brings added
complexity and builds upon the others. The most comprehensive is the
digital oil field with optimization and opportunities for automation.
2. For more on intelligent field applications in producing 5. Chabrelie MF, Dussaud M, Bourjas D and Hugout B:
wells: Dyer S, El-Khazindar Y, Reyes A, Huber M, Raw I “Underground Gas Storage: Technological Innovations
and Reed D: “Intelligent Completions—A Hands-Off for Increased Efficiency,” http://217.206.197.194:8190/
Management Style,” Oilfield Review 19, no. 4 wec-geis/publications/default/tech_papers/17th_
(Winter 2007/2008): 4–17. congress/2_2_09.asp (accessed December 12, 2007).
3. For more on underground gas storage: Bary A, 6. Wallbrecht J: “Underground Gas Storage,” International
Crotogino F, Prevedel B, Berger H, Brown, K, Frantz J, Gas Union, Report of Working Committee 2, Basic UGS
Sawyer W, Henzell M, Mohmeyer K-U, Ren NK, Stiles K Activities, presented at the 23rd World Gas Conference,
and Xiong H: “Storing Natural Gas Underground,” Oilfield Amsterdam, June 5–9, 2006, http://www.igu.org/html/
Review 14, no. 2 (Summer 2002): 2–17. wgc2006/WOC2database/index.htm (accessed
4. For more on current trends in UGS: https://www.ferc.gov/ March 27, 2008).
EventCalendar/Files/20041020081349-final-gs-report.pdf 7. For more on recent sensor and instrumentation develop-
(accessed January 23, 2008). ments in intelligent wells: Bouleau C, Gehin H, Gutierrez F,
Landgren K, Miller G, Peterson R, Sperandio U,
Traboulay I and Bravo da Silva L: “The Big Picture:
Integrated Asset Management,” Oilfield Review 19, no. 4
(Winter 2007/2008): 34–48.

Spring 2008 7
Data Sampled at High Frequency Data Cleansed and Reduced

Bottomhole pressure
Choke size

Wellhead pressure
Wellhead temperature

Time Time
> Cleaning and reducing noisy data. Data sampled at high frequency may exhibit considerable noise; sensor errors also may occur. Even though most of
these data are usable, the spikes and noise in the choke size (top left, blue), bottomhole pressure (top left, red), wellhead pressure (bottom left, green) and
wellhead temperature (bottom left, black) make them difficult to use with modeling programs. After cleansing and reduction, the data (right) are usable in
petrophysical software. (Adapted from Holland et al, reference 8.)

Level-I intelligence begins by developing
confidence in the data. Supervisory control and
data acquisition (SCADA) systems can be found
in most UGS operations. These computerized
networks remotely acquire well data such as flow
rate, pressure and flowing volume, and control
transmission of gas throughout a pipeline system.
With millions of data points thus acquired, it is
impossible to manually validate all the
information. Automating quality control of the
data flow is a necessity.
Software for traditional oil and gas produc-
tion is often used for UGS applications to identify
performance problems as well as monitor
individual wells, evaluate completions and
characterize the reservoir. Trend analysis and
type-curve matching are frequently used in these
programs. However, most petrophysical programs
are poorly equipped to handle the huge volume of
SCADA data coming from UGS operations. Also,
they cannot effectively deal with noisy data
resulting from sensor errors and spurious
responses (above).8 Since proper use of these
applications often depends on the ability to
identify the onset of linear trends over time or
clearly identify subtle features in various type-
curves, the data must be cleansed and reduced so
that proper identification of such features can be
accomplished (next page). Therefore, intelligent
data reduction is applied before importing the
data into these programs.
The data provide insight for evaluating the
relative health of individual wells, as well as that
of the producing field. The repeated cycling
ability of gas-storage wells—periods of injection
followed by periods of production—is a
fundamental difference between producing
reservoirs and storage reservoirs. Occasionally,
the storage wells remain static for varying
lengths of time and the collected data can be
treated as a conventional short-time buildup
test. Changes that occur from cycle to cycle can
be indicative of problems developing in
individual wells, in the reservoir or in the surface
equipment. By analyzing these data, the
presence of damage can be recognized, and
remediation plans implemented.
Early uses of electronic flow measurement
(EFM) data in UGS fields clearly demonstrated
their value in monitoring well performance,
conducting routine surveillance and identifying
operational problems.9 In 2002, as part of a Gas
Technology Institute (GTI) sponsored study,
Schlumberger engineers used EFM data to
develop a reasonably accurate, cost-effective
means of detecting wellbore damage in gas-
storage wells.10 The impetus for this work was the
fairly common practice of performing surface
backpressure tests in UGS wells to evaluate
damage on a very infrequent basis—testing
every 1 to 3 years was typical. The infrequent
nature of such testing made it impossible to
determine incremental damage or sudden
changes over reasonable time frames, such as
during an injection or withdrawal cycle.
Determining changes in damage in near real
time is important, since damage might be
occurring during injection or withdrawal, or
during the changeover from one to the other. This
work made it possible to estimate damage levels

8. Holland J, Oberwinkler C, Huber M and Zangl G: "Utilizing
the Value of Continuously Measured Data," paper SPE
90404, presented at the SPE Annual Technical Confer-
ence and Exhibition, Houston, September 26–29, 2004.
9. Brown KG and Meikle DE: “The Value of Wellhead
Electronic Flow Measurement in Gas Storage Fields,”
paper SPE 31000, presented at the SPE Eastern Regional
Conference and Exhibition, Morgantown, West Virginia,
USA, September 17–21, 1995.
10. Brown KG and Sawyer WK: “Novel Surveillance Helps
Operators Track Damage,” paper SPE 75713, presented
at the SPE Gas Technology Symposium, Calgary,
April 30–May 2, 2002.
11. Spivey JP, Brown KG, Sawyer WK and Gilmore RG:
“Identifying the Timing and Sources of Damage in Gas
Storage Wells Using Smart Storage Technology,” paper
SPE 97070, presented at the SPE Annual Technical
Conference and Exhibition, Dallas, October 9–12, 2005.
12. The pseudocritical temperature and the pseudocritical
pressure are the pressure and temperature conditions
of a multicomponent mixture at which liquid and vapor
cannot be distinguished (because the properties are
identical at this combination of pressure and temperature).
The z-factor, or ideal-gas deviation factor, is the departure
of a gas behavior from that of the ideal gas law.
The gas formation volume factor represents the fractional
change of volume per unit change in pressure. The
coefficient of isothermal compressibility is a measure of
the relative volume change of a fluid or solid in response
to a pressure (or mean stress) change. The gas formation
volume factor is used to convert a volume of gas at reser-
voir conditions to a volume of gas at standard (surface)
conditions, since the volume of any gas depends on its
pressure and temperature.

8 Oilfield Review
 in UGS wells on a more frequent basis than was
previously possible.
In a more recent GTI study, Schlumberger
engineers developed a method to utilize EFM data
to continuously track and identify wellbore
damage over time.11 Developments in intelligent
well technology, including sensor improvements
and real-time streaming data, have been
combined with the experience gained from earlier
EFM studies to develop damage-identification
techniques for use in UGS facilities, such as one
operated by Columbia Gas Transmission
Corporation in the northeast region of the USA.
The reservoir consisted of a consolidated
sandstone, which formed a stratigraphic trap with
an average thickness of 10 ft [3 m], average
porosity of 10% and average permeability of
150 mD. Of the 20 wells in the field, five were
identified as key wells for the purpose of the study.
A SCADA system collected high-frequency
pressure and flow-rate EFM data from the
wellheads at 15-second intervals. The operator
collected monthly records, with as many as
115,000 data points per well per day, and supplied
them to Schlumberger engineers. A software
routine parsed the field-wide data into individual
files for each well, reduced the datasets to a more
manageable volume and automated the process of
making the information useable.
Because each well generated 3 million data
points over the course of the study, a routine was
developed just to handle the raw data. It per-
formed three primary functions: a gas-properties
correlation, a bottomhole-pressure calculation and
data processing. The gas-properties correlation
module calculates pseudocritical temperature and
pressure, z-factor, coefficient of isothermal
compressibility, gas formation volume factor and
gas density, viscosity and pseudopressure.12 It also
formats the data for export to Excel workbooks.
The high-frequency EFM data are aggregated
over 10-minute intervals during flow. The
software routine computes average flow rate and
average pressure, along with the standard
deviation of these quantities. It flags outlying
data points as invalid if there is a flow reversal or
if there is a mix of zero and nonzero EFM-
rate measurements.
During shut-in periods, a variable-width
window is applied to the data to give approxi-
mately the same number of points for each cycle.
The software routine fits the data to establish
weight factors. To qualify the data as a valid shut-
in period for buildup or falloff analysis, a series of
filters is applied based on the length of the prior
injection or production period, the length of the
shut-in period and the length of time necessary

High-Frequency EFM Radial Flow Plot Cleansed and 10-Minute Averaged Radial Flow Plot
123 123
Pseudopressure
Pseudopressure

122 122

121 121
10 100 10 100
Time function Time function

High-Frequency EFM Log-Log Plot Cleansed and 10-Minute Averaged Log-Log Plot

0.1 0.1
Buildup
Δpressure/Δrate

Δpressure/Δrate

Derivative
0.01 0.01

0.001 0.001

0.0001 0.0001
0.01 0.1 1 10 0.01 0.1 1 10
Equivalent time, h Equivalent time, h
> Cleaner, more coherent data. High-frequency electronic flow measurement (EFM) data (left) obtained from SCADA systems display considerable scatter,
making them difficult to use. It is almost impossible to identify the linear portions of the data as well as determine subtle features that are critical for type-
curve matching. The object of the radial flow plot (top left) is to determine the slope of the linear portion of the data. By eliminating outlying data points
and averaging the data over 10-minute intervals (top right), the scatter is reduced, and the slope of the line is more obvious. Wellbore-storage effects,
hidden in the raw data, can now be seen on the far right of the curve. Similar improvement in the log-log plot of the buildup and derivative curves (bottom)
helps to make sense of the data. The noise and scatter seen in the original data (bottom left) would make curve fitting and identification of the linear
portion difficult. With the aggregated and cleansed data (bottom right), the linear section can be determined and used for permeability estimation. The
subtle changes in the shape of the right half of the derivative plot, which is used for determining reservoir boundary conditions, are discernable. Prior to
cleansing and reduction, these data would be extremely difficult to interpret. (Adapted from Brown and Meikle, reference 9.)

Spring 2008 9

Apparent mechanical skin damage factor, S
Shift in both slope and
intercept: s and D changed
Flow rate
> Monitoring well performance. If mechanical skin damage factor, s, and non-Darcy flow coefficient, D, remain constant, apparent mechanical skin damage
factor, S, plotted versus flow rate should be a straight line (left, black). If s or D changes, the plot changes depending on the type of damage. From EFM data
(top right), a step change in s occurs between the two withdrawal cycles (arrows). Plotting flow rate (bottom right) versus injection S (buildup) and
withdrawal S (falloff) reveals changes occurring during injection but not during withdrawal. Had s and D remained constant, the data would have fallen
along the reference line (red). In this example, the change in slope indicates an increase in D. (Adapted from Spivey et al, reference 11.)
for sufficient data to be acquired for the analysis
to be valid. Additional controls were
implemented to avoid limitations when exporting
data to an Excel worksheet.
At this point in the data-handling process, a
user must intervene to select the plots and adjust
the scales as applicable. Apparent mechanical
skin damage factor, S, can be plotted as a
function of time or rate. If the mechanical skin
damage factor, s, and non-Darcy flow coefficient,
D, are constant with time, then the apparent
mechanical skin damage factor, S, will be a linear
function of rate and can be used to estimate s
and D.13 If s or D changes with time, the data will
not be a linear function of rate. Abrupt changes
will cause clusters of data points, and gradual
changes will result in the data drifting away from
the baseline model (above).
One study found that 60% of the reservoirs
evaluated had wells in which non-Darcy flow was
identified as a damage mechanism.14 If the
damage is assumed to be related to mechanical
10
0
Mechanical skin damage factor, s
–1
–2
–3
Shift in slope: –4
D changed
Shift in intercept: Withdrawal Injection Withdrawal
s changed –5
11/1/03 1/16/04 4/1/04 6/16/04 8/31/04 11/15/04 1/30/05 4/16/05
Date
Apparent mechanical skin damage factor, S
Falloff
3 Buildup
2 Best fit
1
0
–1
S = –2.67
–2
D = 1.124 D/MMcf
–3
–4
–5,000 –4,000 –3,000 –2,000 –1,000 0 1,000 2,000 3,000 4,000 5,000
Flow rate, Mcf/d
skin damage factor alone, erroneous conclusions warranted. Had the problem been detected
might lead to inappropriate or ineffective during an annual test conducted on an individual
intervention. Mechanical skin damage is often well, rather than by continuous monitoring of all
improved by pumping acid into the perforations the wells in the field, it is possible that the
or by hydraulically fracturing the formation. results could have led the operator to the wrong
These types of remediation would not effectively conclusion and needless expenditures.
treat non-Darcy flow effects, and performing Existing data sources and new data mining
them could be a waste of time and money. techniques were used to perform the analyses,
A change in non-Darcy flow effects during allowing the operator to determine the source of
injection was observed during the Columbia Gas perceived damage and make the proper decision
Transmission study period. The onset of damage for dealing with it. In this case study, the wells in
was isolated to a particular week in 2004. Several question were able to remain in operation and no
wells in the field exhibited this increased non- remediation was necessary.15
Darcy flow effect, which was identified by an
abrupt change of the slope on the apparent skin Practical Improvement Leads to Level II
damage versus rate plot, while other wells in the A systemic approach to the processes of
field did not experience this change. The analysis injection, storage and delivery for UGS facilities
suggests that the cause of this perceived damage is likely to provide the greatest benefit to
is related to an increase in turbulent flow during operators. Individual-well analysis, reservoir
injection, rather than mechanical damage to modeling, surface hardware and system ineffi-
individual wells. Thus, no remediation was ciencies need to be fully evaluated, but it is
not enough to focus on one or two aspects of
Oilfield Review
the operation, such as EFM data, reservoir
characterization or surface hardware. The
optimization process requires that all the
components be considered together to develop a
model that represents the total system.
Optimization, modeling and surveillance are key
components in developing a Level-II intelligent
UGS facility.
Most testing experts would agree that the
first step in optimization is to characterize
individual wells or completions. The condition
and quality of the reservoir in the vicinity of the
wellbore and the level of damage affect the flow
efficiency of the completion. Once these are
determined, a predictive model can be developed
to provide the expected performance under
various operating conditions. Multirate well
testing is the standard method for full
characterization of well damage and flow
performance, and, with regard to developing a
realistic model, it is indispensable.
Multirate pressure-transient testing quantifies
mechanical skin damage factor, s, and non-Darcy
flow effects, D, and establishes a baseline for
future comparison. Once the individual wells
have been tested, the next step is to characterize
the production properties of the system with a
multirate test across the entire field. Using
operational data, such as flow rates, tempera-
surface-facility effects must then be measured
and included. Nonetheless, the total-field
method will yield the most accurate results.
Finally, an inventory model for the storage
field can be developed that describes the total gas
inventory as a function of reservoir pressure. This
is accomplished by maintaining a constant flow
rate from the storage field for a period of time
sufficient to reach pseudosteady state, and using
the pressure-decline rate observed during
pseudosteady-state flow to calculate the reservoir
pore volume. In other words, this technique can
determine the size of the underground storage
“tank” available for storage and delivery. Once this
reservoir pore volume is quantified, the operator
can estimate the total gas contained in the
reservoir for a given average reservoir pressure.
With existing technology, the average
reservoir pressure can be estimated from a
buildup test any time there is a sufficiently long
shut-in period during normal operations. Since
500
Inflow at sandface
A D
400
Pressure, psi
the field-inventory model quantifies the
relationship between average reservoir pressure
and total inventory, the operator can perform
more frequent inventory audits than were
possible using data acquired only during routine
spring and fall shut-in periods. Calculation of
average reservoir pressure from flowing data,
perhaps available with currently developing
technology, may allow real-time inventory
auditing in UGS fields rather than waiting for
long- or short-term buildup data. This is
especially important in cases where frequent
cycling of gas is occurring and there is no
routine shut-in.
Once the pore volume has been established
by collecting both individual well and field-wide
data from extended tests, two independent
estimates of pore volume can be derived and
compared. This information is integrated with
the deliverability-testing data to create the total
system model.
B C
E

tures and pressures from wellheads, treatment
facilities and metering stations, an operator can
design an effective field-wide multirate test. If all
the wells in the field can flow simultaneously,
this type of test provides a field-wide delivera-
bility curve. The field-wide flow rate should be
high enough to identify the first bottleneck in
the system.
Bottlenecks can be characterized as system
inefficiencies that affect overall performance at
some specific operating condition. Wellbore size,
tubulars, wellhead equipment, gathering lines
and treatment facilities impact the system and
may act as bottlenecks (right). Once these
bottlenecks are identified, economics determine
whether it is worthwhile to remove the cause of
13. Mechanical skin damage factor, s, is a dimensionless
the bottleneck. For example, if a multirate test
indicates that larger tubing will eliminate a
bottleneck but the wellbore size limits the
tubulars that can be installed, there is little that
can be done to fix the problem. Adding wells or
replacing existing wells might be the only
solution, albeit an expensive one.
If a total-field flow test cannot be conducted,
a deliverability curve can be constructed from
tests performed on individual wells. However,
300
Outflow, tubing ID, in.
200 Case 1 (A) - 3.092
Case 2 (B) - 4.000
Case 3 (C) - 5.000
100 Case 4 (D) - 6.000
Case 5 (E) - 8.000
0
0 5,000 10,000 15,000 20,000 25,000 30,000
Gas rate, Mcf/d
> Identifying bottlenecks with multirate flow testing. NODAL production system
analysis indicates that the tubing size is a bottleneck in this test. Tubing inside
diameter (ID) is varied from the existing size of 3.092 in. (green) to 8.0 in. (blue).
The analysis confirms that increasing the tubing ID from 3.092 in. to 5 in.
(red) would provide a 54% increase in flow capacity, from 11 MMcf/d
[311,000 m3/d] to 17 MMcf/d [481,000 m3/d]. Above 5-in. ID, only a nominal flow
increase would be added. (Adapted from Brown and Sawyer, reference 14.)
turbulent flow in producing formations occurs near
number calculated to determine the production efficiency the wellbore, the effect of non-Darcy flow can be
of a well by comparing actual conditions with theoretical represented as a rate-dependent skin effect, D.
or ideal conditions. A positive skin value indicates that Apparent mechanical skin damage factor, S, is similar to
some damage or influences are impairing well produc- s but may be a result of mechanical damage or non-
tivity. A negative skin value indicates enhanced Darcy flow effects such as a rate-dependent restriction.
productivity, typically resulting from stimulation.
14. Brown KG and Sawyer WK: “Practical Methods to
The non-Darcy flow coefficient, D, is calculated from Improve Storage Operations—A Case Study,” paper
fluid flow that deviates from Darcy's law. Darcy’s law SPE 57460, presented at the SPE Eastern Regional
assumes laminar flow in the formation, and if the flow is Conference and Exhibition, Charleston, West Virginia,
turbulent rather than laminar, it is referred to as non- October 20–22, 1999.
Darcy flow. Typically observed near high-rate gas wells,
15. For more on recent sensor and data-handling develop-
turbulent flow occurs when the flow converging to the
ments: Bouleau et al, reference 7.
wellbore reaches velocities exceeding the Reynolds
number for laminar or Darcy flow. Since most of the

Spring 2008 11
 Various programs, such as the PIPESIM and from well logs and well tests. Characteristics of pressure data obtained during the well testing,
NODAL production system analysis software, can the gathering systems, processing equipment focusing on bottlenecks.
be used to construct a nodal analysis model and surface equipment are added to the model, Trend analysis compares actual results from
(below). Model creation begins with identifying which is then calibrated for system variables ongoing, routine operations with the model-based
the physical components of the system and such as pipe roughness and gathering-line results established during the characterization
integrating the reservoir properties obtained lengths. The model is adjusted to match the phase. More advanced predictive tools, such as
the DECIDE! data mining based production
optimization software, have been utilized to
Additional
automate the process of comparing model-
group of wells derived data with operational data.
Additional
group of wells Starting from Scratch—or Not
The potentially long lifetime of a UGS facility—
the Zoar field, for example, has been in operation
for more than 90 years—may demand well
construction practices that differ from
Compressor nonstorage wells. Storage wells must be able to
station tolerate high injection rates, high production
pressures and frequent cycling. Reuse of existing
downhole and surface equipment may be
Producing well
possible, but more common is a mixture of
Node
existing wellbores and newly drilled wells. The
original producing operation may, however,
dictate well placement and facility location. The
transition from production to storage should
focus on optimization and thorough
understanding of the reservoir.
Ideally, optimization begins during the
identification of prospective fields to be used for
gas storage. The first step in choosing a
candidate is to understand the reservoir.
Characteristics for UGS prospects are good
porosity and permeability and an effective
trapping mechanism. If the choice of reservoir is
> PIPESIM production system analysis model. The PIPESIM program can be used to create a visual
not an option, such as an existing gas-storage
model of the mechanical components of the facilities. Reservoir and production information can be
integrated into the model. Field-wide multirate tests are used to calibrate the model. In this example, facility in need of upgrading or improving, new
the wells are producing. Reversing the direction of flow provides an injection model. technologies can still be employed to enhance
the value of an UGS operation.
An example of optimizing an existing gas-
storage field with intelligent well technologies is
the Falcon Gas Storage Company, Hill-Lake field
operation in Texas (left). Formerly an oil
producer, this field was discovered in the 1920s
and by the 1950s had reached the end of its
productive life. It was converted to a gas-storage
facility in the 1960s. When Falcon took over the
operations in 2001, there were 21 wells in the
field, 10 of which were active in the gas-storage
operation. No development had taken place since
the 1950s, and 2.5 Bcf [71 million m3] of gas
could not be accounted for by previous operators,
attributed to “meter error.”
The original interpretation provided to
> Falcon Gas Storage Company Hill-Lake underground gas-storage Falcon was fairly simplistic. From limited well
operation. Located in Eastland County, Texas, the facility has the capacity to control and old electric logs obtained in the
deliver 515 MMcf/d [15 million m3/d] and inject 300 MMcf/d [8.5 million m3/d]. vicinity of the injection site, the structure was
(Photograph courtesy of Falcon Gas Storage Company.)

12 Oilfield Review
mapped as a fluvial delta (right). Between 2003 Sand Facies Thickness Maps
and 2006, Falcon added 16 wells to the field, Mitcham Est 1 HLGSU 1 HLGSU 1
with well placement aided by FMI Fullbore Mitcham 1A Main lobe
Formation MicroImager analysis. The images
HLGSU 3
from the FMI data suggested a new HLGSU 2
Mitcham Est 3 HLGSU 4
Storm 1
interpretation, that the reservoir was a braided
Mitcham Est 3
channel sand, not a delta.

40
HLGSU 7

30
20
In 2006, Schlumberger geologists used Petrel

10
HLGSU 11

0
seismic-to-simulation software to develop a basic HLGSU 5 HLGSU 6

50
40
30
20
10
0
HLGSU 12
geological model (bottom right). The existing HLGSU 8 AJ August 1A
HLGSU 9
wells were incorporated into the model, and an Wells productive
from Hill-Lake sand August 3
additional 17 wells, drilled by Falcon in 2006 and appeared connected 40
August 3
2007, were included in the analysis. 30
Williams 1 20
These new wells, drilled as step-outs from the Mitcham 3 Mitcham 1F
Williams 1
original injection site, followed trends indicated
10

40
30
20
by the interpretation of the FMI images. The

10
Fluvial Formation

0
additional wells led to some interesting Mitcham 2 Mitcham 2
discoveries, such as a previously unknown sand
lobe to the southwest of the main injection wells. Southern lobe
(connected)
As an isolated sand, it should have been depleted
by earlier production but, unexpectedly, the Braided Channel Sand
pressure was similar to that in the rest of the
Hill-Lake reservoir, proving that they were in
communication. Based on information derived
from the Petrel geological model, this structure
is believed to have contained the 2.5 Bcf of
missing gas.
Not only was unaccounted gas discovered, but > Changing interpretations. The Hill-Lake field was originally mapped as a fluvial delta (left). Two
as gas was injected into the field, the originally structural highs were identified on the isopach map (top left). New well locations were drilled
depleted reservoir was recharged. New wells according to the original structural map, which was based on wells drilled before 1960. An updated
interpretation, using Petrel seismic-to-simulation software (right), included 17 new wells and identified
penetrating down-dip sand sections encountered
a previously unknown southern lobe (top right). The original sand body’s geometry could be more
oil left behind during the initial production accurately visualized, and the structure was characterized as a braided channel sand (bottom right).
phase that could now be recovered because of Because the FMI interpretation provided flow direction, well placement was improved. The southern
the increased reservoir pressure. lobe contained 2.5 Bcf of stored natural gas that the original operators assumed had been lost
because of metering errors.
A by-product of the recovery of stored gas is
additional oil production in the form of natural
gas liquids (NGLs). When gas is injected and
recovered, it is enriched by hydrocarbon liquids
Petrel Model Construction Workflow
that were left behind after the original oil
production ceased. These liquids are stripped out Data inventory and
Petrel Petrel
loading (logs, markers,
of the recycled gas using a cryogenic gas surfaces ) Geological Reservoir
Modeling Modeling
processing plant, and then recovered and sold,
adding to the profitability of the UGS operation. ECLIPSE
Petrophysics and Fluid PVT software
The Petrel reservoir model identifies candidate
borehole geology Interwell
locations for future field development where the (ELANPlus, BorView correlation Software
sand properties are most conducive to the software) modules
production of liquids during gas withdrawal.
In 2007, because of the insight provided by the Rock type Framework modeling Property modeling Reservoir
identification (faults, surfaces, grids, (facies, net to gross, porosity, simulation model
preliminary model, Falcon initiated a detailed (facies properties) zonation layering) permeability, saturation) (history and forecasting)
Petrel geological and reservoir model, incorpor-
ating a total of 72 wells. The ELANPlus advanced
multimineral log analysis program and FMI Volumetrics
analysis were used to interpret 29 of the wells.
Core data from five wells provided calibration for > Modeling a reservoir. Geologists used Petrel software to model the Hill-Lake reservoir.
the model. With advanced petrophysical analysis, Once the geological model is fully developed, the reservoir properties can be used to develop
simulations for volumetrics and field performance.
the initial simplistic interpretation of a deltaic
deposition evolved into a more realistic model of
the reservoir.

Spring 2008 13
 As additional wells were drilled, greater
insight into the reservoir geometry was
gained, and the depositional environment was
recharacterized as an ancient riverbed with
anastomosed channel sands.16 Optimal well
placement depends on understanding the reser-
voir structure. Visualization of the subsurface
topography provided by the Petrel model was
crucial in identifying the correct depositional
environment and ruling out the two previous
interpretations (below). Thorough under-
standing of the reservoir has added new storage
capacity to the field, defined new exploration
areas to recover oil left behind after the initial
production ceased, and helped optimize future
field development. Production and capacity
improvements in the USA, such as those found
with the Hill-Lake operation, help explain the
UGS capacity increase, despite a decrease in the
number of sites.
Falcon’s use of new applications and tech-
nology is not limited to subsurface modeling and
optimization. Because this facility had been in
operation since the 1960s, production facilities
needed an upgrade. A SCADA system was
Computer-postulated lobes
(potential future storage capacity)
Mitcham Est 3
Main lobe
(original unit)
Cooper A-3 lobe
(separate lobe)
20
10
0
installed, providing instantaneous information
about temperature, pressure and flow rates. The
flow of gas can be managed from the wellhead, at
individual compressor sites or at the central field
control room. Although the SCADA system is not
used to remotely control the facility at present, it
has the capability to do so.
Falcon’s Hill-Lake facility is now a state-of-
the-art, multicycle UGS operation with the
capability of being used in a variety of ways,
including storage, high-rate delivery, peak-
shaving, “park and loan” and market trading.17
The maximum capacity is now 15 Bcf
[425 million m3], representing 11 Bcf
[311 million m3] of working gas and 4 Bcf
[113 million m3] of cushion gas. The field can
deliver 515 MMcf/d [15 million m3/d] and inject
300 MMcf/d [8.5 million m3/d]. Injecting in
summer and supplying in winter have been
replaced by a flexible operation capable of on-
demand delivery and storage as required by
customers, while recovering oil and NGLs that
were left behind during initial production.
Falcon’s success with Hill-Lake resulted in
the recent retrofitting of another gas-storage
0 program has two major components—a data hub
10
HLGSU 1 20
30
Northeast lobe
(limited connectivity)
0
20
10
30
0
10
20
30
40
50
August 3
field in north Texas, its Worsham-Steed facility.
Utilizing an abandoned oil and gas field originally
converted to gas storage by another operator, this
retrofit is a multicycle UGS operation employing
similar intelligent well technologies. This field
produces oil and NGLs along with providing
24 Bcf [680 million m3] of working-gas capacity.
Level III—Automated Reservoir Surveillance
Whether UGS facilities are buffering demand
cycles or acting as gas repositories, the ability to
automate the process is an attractive reason for
implementing intelligent well technologies. An
operator in Europe, working with Schlumberger
reservoir geologists and engineers, designed and
implemented an automated reservoir surveil-
lance operation using an integrated platform
built around DECIDE! software. An operator can
optimize and perform predictive modeling for
highly complex systems using the artificial
intelligence and software simulation of the
PC-based DECIDE! software (next page, top).18
This software provides a way to bring together
people, technology, processes and information in
a secure, global system—reducing cost, lowering
risk and enhancing operations. The DECIDE!
and an engineer’s desktop. Responsibility for
retrieving, storing and distributing data, as well
as automation of tasks, lies primarily with the
data hub. The engineer’s desktop uses state-of-
the-art data mining techniques to perform
analytical petroleum-engineering calculations,
giving the operator a powerful tool for managing
the asset.
RWE Transgas Net, an independent natural
gas operator in the Czech Republic, has installed
DECIDE! software to manage and optimize all of
its depleted reservoir and aquifer gas-storage
facilities. Implementation began in 2004, and
was completed in 2007 (next page, bottom).

Williams 1
16. Anastomosed channel sands are found in multichannel
rivers that have relatively low gradients, deep and
Anastomosed Channel Sand narrow channels, and stable banks characterized as
50
40 Mitcham 2 having been deposited in slow-moving rivers. Braided
30
20
channel sands are found in high-energy environments,
10 characterized by excessive deposition of sand bars or
0
gravel bars or both, so that water flows in many branch-
ing and reuniting channels.
Southern lobe
(connected) Computer-postulated lobes 17. Park and loan refers to storing gas for later use (parking)
(potential future storage capacity) and taking gas (loaning) to avoid purchasing gas at high
spot-market prices. Designed as a balancing service,
customers save money by using the service at times
> The final answer. A clearer picture developed as more wells were added to the field. With 72 wells, when they are out of balance on the pipeline. Customers
and FMI data from 29 of those wells, the final Petrel geological model was created. The structure was also can take advantage of short-term price swings.
characterized as an anastomosed channel sand (bottom left). Compared with braided channel sands, 18. For more details on applying ECLIPSE software
anastomosed channel sands are deposited from lower energy water flow. The more tortuous path of and DECIDE! software: Barber A, Shippen ME,
Barua S, Cruz Velázquez J, Garrido Hernández AM,
the riverbed required future wells to follow a more curved path than would have been indicated had
Klumpen HE, Moitra SK, Morales FL, Raphael S, Sauvé B,
the reservoir been either a delta or a braided channel sand. The FMI tool provided the direction to drill Sagli JR and Weber M: “Optimizing Production from
step-outs and develop the field. Knowledge of the reservoir also indicated the best areas to drill to Reservoir to Process Plant,” Oilfield Review 19, no. 4
recover oil left behind by previous production. New structure was also identified, locating untapped (Winter 2007/2008): 18–29.
reservoir potential and adding storage capacity for the facility.

14 Oilfield Review
 Model Integration and Control

Material balance Simulation Function of influence Proxy model Well simulator Network simulator

Design workflow

Operational workflow

Core of Expert System

• Monitoring and data QC Data hub
• Event detection and
performance monitoring
• Automated data transfer Database
and model execution components

> Level-III intelligence. Level III combines all the components of intelligent well operations together. By integrating processes such as trend analysis,
modeling and simulations, the UGS facility can be optimally managed with high levels of automation.

RWE Transgas Net, working with Schlumberger
engineers, began the process of implementing
the DECIDE! program by first developing an
integration platform. A SCADA system was
installed to provide continuous high-frequency
measurements (on the order of seconds), which
are grouped into 15-minute increments and
streamed in real time from individual wells,
gathering systems and facilities. At this Level-I
step in the process, the software system checks
that a connection to a datastream has been
established and generates a notification if there
is a failure. When a valid connection is
confirmed, the high-frequency data are
imported, filtered, quality checked and
aggregated over longer time intervals to reduce
the size of the dataset. The software filters
sensor errors and transmission errors prior to
data aggregation and generates statistical
reports to allow the engineer to evaluate the
reliability of the information. Artificial intelli-

Short-term
SCADA system control module Engineer’s desktop

Legacy data

Manually
captured data
Surveillance and reporting modules

Data buffer Data hub Database

NN proxy Workflow design and automation

Network modeling Simulation

Material balance proxy

Proxy models

Seconds Minutes to hours One hour to one day One month to one year

> DECIDE! program workflow. SCADA data are streamed into the data buffer where they are quality checked, cleansed and reduced using a neural
network (NN) proxy model. Data are fed to various software modules for automated surveillance, report generation and preparation. Proxy models process
the information and use trend analysis and simulation-matching to look for optimization opportunities and to detect developing system problems. Reports
are available in almost real time. History-matching is available to determine the ongoing health of the operation. The majority of these processes are
carried on in the background with little operator intervention required.

Spring 2008 15
 Measured BHP deviates
from calculated BHP

85
Calculated BHP
80
Measured BHP

Pressure, MPa
NN Architecture for Well Model 75

70
Reservoir pressure
65
Aquifer influx

Gas injection
55
Gas production
50

Jan 1999 April July Oct Jan 2000 April July Oct Jan 2001
Time
> Neural networks (NN) and problem identification. Multiple parameters are required to calculate the bottomhole pressure
(BHP). Neural networks, functioning as proxy models, are trained with data from operations and can reduce the total iterations
required to produce results in a fraction of the time required by full-scale modeling programs. Measured BHP data (red) during
injection and withdrawal are compared with NN-derived BHP data (blue). The two sets of data compare very favorably until
September 2000 when the calculated BHP during injection increased and remained higher for the next four months because of
reservoir damage. The higher backpressure indicated well intervention would be required to maintain deliverability.

gence has been developed to automate these
routine tasks as well as increase the speed of
delivery. With the newly acquired data, key
performance indicators are available to evaluate
the ongoing operation.
At Level II, the cleansed data are fed to
software modules to validate proper system
performance. The DECIDE! program can integrate
external applications that allow data exchange,
including ECLIPSE reservoir simulation software,
PIPESIM production system analysis software and
various modules available in the DECIDE!
software. The process automatically conducts
history-matching for trend analysis, provides
individual well status and determines production
and capacity constraints. Current and future
delivery requests for injection and withdrawal—
inputs into a dispatcher’s module—are passed
to a DECIDE! software module. This module
then provides all the necessary calculations
and predictions to ensure that the reservoir
has sufficient capacity to meet the dis-
patcher’s requests.19
The advanced programs used for the reservoir-
surveillance models require computer-intensive
calculations. Running optimization iterations
cannot provide satisfactory results in the
required time frame using the large volume of
data, even after the data have been aggregated
and cleansed. Proxy models, although not as
accurate, are substituted for full-scale
simulations and can provide results in seconds
or minutes.20
Proxy models, in the form of trained neural
networks (NNs) optimized to require a reduced
number of input parameters, use artificial
intelligence to mimic large-scale simulators. The
NN learns to behave like the simulator and, once
trained, it can perform a set of calculations in a
fraction of a second for a given set of input
parameters. The NN drastically reduces the
computation time necessary, allowing real-time
history-matching with the optimized original
outputs of full-scale numerical models.
An example of the use of a NN is a production
forecast and optimization simulation. If small
changes of the input parameters are involved,
such as tubing-head pressure, a forecast can be
calculated immediately, rather than waiting for a
time-intensive full simulation to be performed.
Multiple iterations can also be run quickly to
determine the best course of action. Additionally,
NNs are used to evaluate uncertainties in the
input data provided in the Level-I data-
acquisition phase. This speeds up the
quality-control and data-cleansing functions
prior to inputting data into the proxy models.
An automated surveillance system compares
calculated results with measured results. If
information from proxy models indicates that a
well or surface component has failed to perform
as expected, an event alarm is triggered and
reported to the operator by way of the engineer’s
desktop. For the parameters that are set to
trigger alarms, a deviation range can be
established and adjusted as required. Once an
alarm has been triggered, the reservoir engineer
can react in a timely manner to investigate the
source of the problem (above).
The level beyond monitoring and surveillance
is Level-III intelligence—an example of the
digital oil field. Although oil companies and
service providers have found it difficult to
provide a single definition for this term, the
digital oil field essentially provides a high level of
automation, simulation modeling, decision-
making tools (the faster, the better) and an
integrated approach that does not lose sight of
the small details (or at least has a system to
monitor them). Schlumberger refers to this level
of field operation as the BlueField intelligent
asset integration service.21
At the BlueField level, data are acquired and
prepared for processing, and integrated-
simulation models are run from various
performance modules. System checks are carried
out at the highest level, and reports concerning
the health of the complete operation, using
deterministic models, are generated and
delivered to the DECIDE! engineer’s desktop
(next page). The service provides the ability to
oversee scheduled automated tasks or those
triggered by event alarms.

19. Onderka V, Dressler M, Severýn O, Giovannoli M and
Zangl G: “Expert System of UGS—An Efficient Tool for
Online Performance Management and Optimization,”
presented at the 23rd World Gas Conference,
Amsterdam, June 5–9, 2006.
20. Zangl G, Giovannoli M and Stundner M: “Application of
Artificial Intelligence in Gas Storage Management,”
paper SPE 100133, presented at the SPE Europec/EAGE
Annual Conference and Exhibition, Vienna, Austria,
June 12–15, 2006.
21. Bouleau et al, reference 7.
22. Foh SE: “The Use of Inert Gas as Cushion Gas in
Underground Storage: Practical and Economic Issues,”
presented at the Gas Supply Planning and Management:
1991 and Beyond Conference, Lake Buena Vista, Florida,
USA, February 25–27, 1991.
23. “Supply on Demand,” http://www.falcongasstorage.com/
fw/main/MoBay_Storage_Hub_LLC-28.html (accessed
January 15, 2008).
24. National Petroleum Council (ed): Hard Truths: Facing the
Hard Truths about Energy. Washington, DC: National
Petroleum Council (2007): 36.

16 Oilfield Review
 In the system implemented by RWE Transgas
Net, automated tasks have the following
structure: first run scheduled tasks or trigger
automation tasks, then run predictive data-
mining proxy models and apply rules. Triggering
events are either discrepancies from expected
trends or violations of predefined constraints.
The actions triggered by the alarm include
system notification, retrieval and execution of
surveillance software, exchange of data with
third-party software, initiation of subordinate
tasks and generation of e-mail or text messages
to alert the operator of an error condition.
Along with the alarms, the software
automatically provides key performance
indicators to the operator at the engineer’s
desktop. It formats the data for visualization and
provides forecasts based on current performance
of the field. Reservoir performance modules
identify bottlenecks, like facility constraints, and
report on optimization opportunities along with
recommended courses of action. With dramati-
cally reduced analysis cycle time, the engineer
can react almost instantaneously. Automated
data flow and transparently updated models
allow the engineer to focus on system
optimization and problem elimination. Proactive,
intelligent reservoir management—a BlueField
application—becomes a reality.
Into the Future
Underground gas storage in depleted reservoirs
has proved to be well-suited for many of the
intelligent well and intelligent field technologies
that are being developed for traditional oil and
gas production. The UGS industry has achieved
great success in the adoption of these
techniques. The lessons learned by UGS
operators are being applied with greater
confidence by the oil and gas production side of
the business because these new technologies
have demonstrated their ability to provide
reduced costs and increased efficiencies.
Because UGS fields have long life expectancies,
they afford a long-term outlook for payback.
Compared with conventional hydrocarbon fields,
gas-storage fields do not experience the same
decline in their asset’s value as the reservoir
depletes because gas-storage fields can be
repeatedly recharged. Retrofitting older facilities
with modern intelligent-field equipment makes
financial sense, increasing the value of the
existing asset.
Maximizing the asset, above and below the
ground, leads to innovative approaches like those
discussed in this article, but there are still more
technologies and techniques to apply. For
Spring 2008
SCADA system
01010101010101 Data Data System Data
01010101010101 acquisition cleansing health check storage
01010101010101
System acquires data every 15 minutes
Model validation loop
Run simulation
Data Write ECLIPSE Simulation
schedule file results
storage
Optimization and prediction loop–DECIDE! desktop
Run optimization Run forecast
Optimization Generate
results production
curves
Current time plus Current day plus
previous 24 hours previous 100 days To dispatcher
Dispatcher request
> Data in, decision out. SCADA data arrive in 15-minute intervals and are cleansed and quality
checked. Data from simulation software are compared with model predictions. The engineer at the
DECIDE! desktop receives results, forecasts and production information. This information is received
automatically or can be generated upon request.
instance, cushion gas can be the most expensive UGS to the first offshore UGS facility in North
component in a UGS facility and realistically will America.23 The reservoir has been characterized
be returned to the operator only when the field is and modeled using Petrel software. The surface
decommissioned. As an example, UGS in a equipment has been designed, and operations
depleted field with 20 Bcf [566 million m3] of are expected to commence during 2009.
total capacity would require 30 to 50% of the gas This high-deliverability, multicycle facility is
to remain in place as cushion gas. Borrowing designed to have a working volume of 50 Bcf
from the bank analogy, that equates, at the high [1.4 billion m3], with injection and withdrawal
end, to 10 Bcf [283 million m3] left in an interest- capabilities of 1 Bcf [28 million m3] per day.
free checking account. At 2008 price levels, that By 2030, global demand for natural gas is
comes to more than US $80,000,000. projected to range from 356 to 581 Bcf [10 billion
Even if the reservoir is operated in the most to 16.4 billion m3] per day, up from 243 Bcf/d
efficient pressure and flow range, some cushion [6.9 billion m3/d] in 2000.24 The Middle East has
gas must be left in the ground to enable high-rate by far the largest natural gas reserves with an
delivery. Reservoir engineers have tested the estimated 2,566 Tcf [72.7 trillion m3]—or 41% of
feasibility of injecting inert gas into the reservoir the world’s total. Russia, second in proven
to function as cushion gas. This approach is reserves, has extensive pipelines into Europe and
especially practical considering current natural has proposed pipeline construction to China and
gas prices.22 The technique does, however, require other countries. As demand for natural gas
detailed understanding of the reservoir storage grows, new methods will be developed to
properties and flow characteristics, the conse- transport, store and deliver it. Because the
quences of mixing different gases and the supply is generally far removed from most users,
long-term effects of the inert gas on the reservoir. UGS facilities are a major component in
This is yet another example of UGS operators providing a stable, secure source of natural gas
applying novel reservoir management techniques. for industrial and residential consumption.
As operators develop UGS fields and attempt As the character of UGS evolves from peak-
innovative approaches, the emphasis on reser- shaving to flexible applications, intelligent field
voir characterization, process optimization and technologies are assisting operators in the quest
automated operation bring greater flexibility and for greater efficiency, lower costs and innovative
opportunities to UGS projects. As an example, methods of doing business. As a result, the digital
Falcon Gas Storage Company is applying much of oil field has become a reality in the underground
its experience with intelligent field technology in gas-storage industry. —TS
17
 Ensuring Zonal Isolation
Beyond the Life of the Well

When zonal isolation fails, production or injection efficiency is severely degraded.

In some cases, the well is lost entirely. No less significantly, such failures present
environmental and safety implications since hydrocarbons or previously injected
fluids may flow to the surface or into nearby aquifers. Therefore, it is not sufficient
to obtain good zonal isolation; the resulting seal must last many years beyond the life
of the well.

Mario Bellabarba Placed between casing and wellbore, a cement
Hélène Bulte-Loyer sheath is expected to provide zonal isolation
Benoit Froelich throughout the life of a well. But its ability to do so
Sylvaine Le Roy-Delage depends on the proper placement of the cement,
Robert van Kuijk the mechanical behavior of the cement and the
Smaine Zeroug stress conditions in the wellbore. Even if the
Clamart, France
slurry was properly placed, changes in downhole
conditions can induce sufficient stresses to
Dominique Guillot
Cambridge, Massachusetts, USA destroy the integrity of the cement sheath. Over
time, stresses are imposed on the cement by
Nevio Moroni pressure integrity tests, increased mud weight,
Eni S.p.A casing perforation, stimulation, gas production or
Milan, Italy a large increase in wellbore temperature.1 Any of
these events can damage the sheath.
Slavo Pastor Often, damage to the cement sheath resulting
Tyumen, Russia from these forces manifests as microannuli so
small as to be nearly impossible to pinpoint and
Augusto Zanchi
even harder to repair. Even the smallest micro-
Stogit
annulus can be large enough to provide a
Crema, Italy
pathway for fluid migration. Remedial work
For help in preparation of this article, thanks to Matthew for such cement failures has been estimated to
Andruchow, Clamart, France; and Martin Isaacs and Ali Mazen, cost more than $50 million annually in the
Sugar Land, Texas, USA.
CemCRETE, CemSTONE, CemSTRESS, FlexSTONE, FUTUR,
United States.2
Isolation Scanner, LiteCRETE, PS Platform, SCMT (Slim Despite changes in downhole conditions over
Cement Mapping Tool), SlimXtreme, SlurryDesigner and USI
(UltraSonic Imager) are marks of Schlumberger.
time—both predictable and unexpected—
Fann is a trademark of Fann Instrument Company. obtaining long-term zonal isolation is not only
possible, in today’s fiscally and environmentally
sensitive oil industry, it is mandatory. To do so
requires the right technology, processes and
evaluation because drilling a well disturbs long-
settled and precariously balanced stresses.
Drillers must compensate for this disturbance, to
the degree that it is possible, by using drilling
fluids to exert hydrostatic pressure on the
formation. However, this pressure may be
insufficient to maintain equilibrium with the far-
field stresses, and the formation surrounding the
removed volume will deform.3
Draining fluids from a formation during pro-
duction may also change formation pore pressure
and related stresses. Within the rock, the result-
ing increased loading leads to varying degrees of
deformation or failure that can cause cement to
break or debond at the formation interface.
Production-induced stresses can also result in
reservoir compaction, which may lead to tubular
shearing and buckling of completion components.4
An obvious key to long-term zonal isolation is
obtaining a good seal in the first place. To
determine whether that objective has been
achieved, standard sonic and ultrasonic logging
tools have been developed and improved over time
in an effort to quantify the cement-to-casing bond.
Recent versions of ultrasonic tools can now detect
the presence of channels within the cement
sheath through which hydrocarbons can flow.
In this article, we will highlight the most
recent development of these ultrasonic tools that

18 Oilfield Review
can also indicate casing eccentricity, evaluate
the material in the casing annulus and
distinguish between new lightweight cements
and drilling fluids of similar acoustic impedance.
Case histories will also demonstrate the new
ultrasonic logging tool’s ability to offer improved
characterizations of cement-to-casing bonds and
Oil and Gas
annular fill.
This article also examines industry efforts to
achieve long-term zonal isolation using specially
formulated cements as annular sealing material.
Of primary interest is a new, long-life, self-
healing cement. Developed in response to what
has been called the weak link in zonal
isolation—the inability to correct defects after
the cement has set—the new sealant swells in
the presence of hydrocarbons to close cracks and
Surface Sediments
microannuli that may form in cement sheaths as Surface Casing Fresh Water
a result of changing downhole conditions. We
also present laboratory tests and case histories Shale Intermediate Casing
that demonstrate the success of this develop-
ment effort. Salty Water and Sandstone

Preparing the Ground

Shale
Obtaining a good cement sheath demands
adherence to well-established operating practices
of hole preparation, casing centralization and Hole Cement Sheath
casing rotation and reciprocation.5 Successful
zonal isolation first requires removing
contaminants—principally drilling mud—from Production Casing Limestone
the wellbore. Since formation pressure must be
contained during this hole-cleaning operation, the
Production Tubing
drilling fluids being removed must be displaced
with a fluid of higher density—a spacer—pumped
Annulus
behind the mud and ahead of the cement.

1. Le Roy-Delage S, Baumgarte C, Thiercelin M and Sandstone

Vidick B: “New Cement Systems for Durable Zonal
Isolation,” paper IADC/SPE 59132, presented at the
IADC/SPE Drilling Conference, New Orleans,
February 23–25, 2000.
2. Cavanagh P, Johnson CR, Le Roy-Delage S, DeBruijn G, Dolomite
Cooper I, Guillot D, Bulte H and Dargaud B: “Self-Healing
Cement—Novel Technology to Achieve Leak-Free Wells,”
paper IADC/SPE 105781, presented at the IADC/SPE
Drilling Conference, Amsterdam, February 20–22, 2007.
3. Gray KE, Podnos E and Becker E: “Finite Element Studies
Shale
of Near-Wellbore Region During Cementing Operations:
Part I,” paper SPE 106998, presented at the SPE
Production and Operations Symposium, Oklahoma City,
Oklahoma, USA, March 31–April 3, 2007. Limestone
4. For more on formation stresses: Cook J, Frederiksen RA,
Hasbo K, Green S, Judzis A, Martin JW, Suarez-Rivera R,
Herwanger J, Hooyman P, Lee D, Noeth S, Sayers C,
Packer
Koutsabeloulis N, Marsden R, Stage MG and Tan CP:
“Rocks Matter: Ground Truth in Geomechanics,”
Oilfield Review 19, no. 3 (Autumn 2007): 36–55.
For more on reservoir compaction: Doornhof D,
Kristiansen TG, Nagel NB, Pattillo PD and Sayers C: Productive Formation
“Compaction and Subsidence,” Oilfield Review 18, no. 3 Productive Formation
(Autumn 2006): 50–68.
5. Casing rotation and reciprocation refer to any movement
of the casing to help remove drilling fluids from the
annulus while cement is being pumped downhole. Shale

Spring 2008 19

Centralizer
Casing
Good cementing
where casing
is centered
B
A
C measurements can be compared with traditional
> Failed isolation. Problems that occur while
running casing and cementing can create
conditions that may lead to loss of zonal
isolation. Among the most common of these is
casing eccentricity from poor centralizer
positioning. Cement, like all fluids, seeks the path
of least resistance and so flows to the more open
side of the casing, creating a narrow space
between the casing and the formation that can
become fluid-migration paths (A). Inadequate
slurry density can also allow formation gas (red)
to enter the wellbore (B) and create weak points
or gaps within the cement that fail when stresses
are imposed on the cement sheath by changing
downhole temperatures and pressures. The
geometry of washed-out areas (C) often results
in inefficient flow rates during wellbore cleaning
operations that leave drilling fluids behind. These
contaminants also lead to weak spots in the
cement sheath and, if large or numerous, can
create channels through which formation fluids
may flow.
liquid state, they are liable to form channels of
communication between zones along the
borehole or casing (left).6
Efficient borehole cleaning is not the only
requirement for good zonal isolation. A poorly
drilled hole, for example, may have washed-out
areas that are difficult to clean and that may
contain pockets of gelled drilling fluids. These
gelled fluids can be pulled into and contaminate
the passing cement slurry. Poor casing centrali-
zation can contribute to a poorly placed cement
since it can be difficult to remove fluids from the
side that is closest to the borehole wall in
eccentrically positioned casing. Since the 1940s,
research and development efforts have gone into
developing recommended standards for
centralizer placement along the casing string to
be cemented. Those practices are now being
tested by new cement evaluation tools that
provide casing eccentricity measurements. These
calculated standoff values that rely on
unlikely assumptions such as a perfectly
in-gauge wellbore.7
To avoid leaving behind a heavy filtercake
that is impossible to remove, the properties of
the drilling fluid must be altered to match those
more suited to hole cleaning. For best results,
the mud density, yield stress, plastic viscosity and
gel strength should all be reduced.
Mud rheology may be reduced by adding
water or dispersants to the system and
circulating the fluid until its properties reach the
desired range. This requires circulation of at
Tangential
stress
least one borehole volume and, when possible,
should be performed before removing the
drillpipe to prevent mud from gelling while it is
static during pipe-pulling operations.
Mechanical steps are also recommended to
help remove contaminating fluids prior to
cementing. Moving the casing frees mud trapped
in narrow sections of the annulus. Attaching
scratchers, scrapers and wipers to the casing
also helps remove gelled and dehydrated mud as
the casing is rotated and reciprocated.
The optimum wellbore for cementing
purposes, then, is one with controlled subsurface
pressures and minimum doglegs, is in-gauge,
stabilized and free of drill cuttings, and has a
thin dynamic filtercake across permeable zones.8
Sound Technologies
Following industry best practices does not
guarantee that the resulting cement sheath will
be up to the tasks—casing support, corrosion
protection and, most critically, zonal isolation—
for which it was designed. Determining
contamination, continuity and bonding quality of
the cement behind the pipe is therefore
tantamount to protecting the asset and the
environment by recognizing the need for
remedial operations before the well is brought
on production.
Finding the top of the cement behind pipe
where expected is a reasonable indication that
the volumes displaced match those calculated
and that the annulus is filled with the correct
amount of cement. Since cement hydration is an
Microannuli
Tensile
failure

The spacer is designed to keep the drilling

fluids and cement apart while the cement is
being pumped through the casing and into the
annulus, and is generally formulated with a
viscosity close to or greater than that of the
drilling fluid. Besides maintaining well control,
the spacer also serves as a chemical wash to
clean leftover drilling mud from the casing-
casing and casing-wellbore annuli. If the spacer
leaves drilling fluids behind, or if it allows them
to mix with the cement, then good bonding
between cement and formation or casing is
unlikely. Since these contaminants remain in a
Radial
stress Compressional
failure
> Cracks and microannuli. Over time, as downhole stress conditions change, primarily in response to
temperature and pressure changes, even a successful cementing operation can fail. Large increases
in wellbore pressure or temperature and tectonic stresses can crack the sheath and even reduce it to
rubble. The interplay of tangential and radial stress changes may be caused by displacement of
casing as a result of cement bulk shrinkage or temperature or pressure decreases (left). These stress
changes can cause the cement to fail in tension or compression, or to debond from the casing or
formation, creating microannuli (right).

20 Oilfield Review
exothermic reaction, this can be done using a
temperature survey. This method, however,
reveals little else about the results of the
cementing operation.
Hydraulic testing—a common test of zonal
isolation—applies internal pressure along the
entire casing string. But pressure can expand the
casing, causing the cement sheath to experience
tensile failure. This may lead to radial cracks and
local debonding of the cement and casing in
areas where the cracks are near the casing wall
(previous page, bottom).
Because of the limitations of the other
techniques, acoustic logging has become the
industry’s tool of choice for detecting cement
behind casing and assessing the quality of the
bonds between casing, cement and formation.
Acoustic logs help indicate nonintrusively the
depth interval at which cement has been placed
around the casing, measure acoustic impedance
of the cement bonded to the casing, and quantify
the percentage of pipe circumference bonded to
the cement.
These characteristics inform the operator of
faults in the cement sheath that may require
remedial measures—commonly a squeeze
operation in which cement is pushed through
perforations into the annulus to fill gaps along
interfaces at the casing, formation or within the
annular material itself.
Cement bond logs (CBLs) and variable density
logs (VDLs) are acquired using a sonic logging tool
(right). Standard CBL tools, which comprise those
DSLT
that measure signal amplitude or attenuation, have
a common theory of measurement and
interpretation. The principle behind them is to
measure the amplitude of a sonic signal produced
by a transmitter emitting a 10- to 20-kHz acoustic SSLT SlimXtreme
tool SCMT
wave after it has traveled through a section of the tool
casing as an extensional mode.9
Measurements are displayed on the CBL log
in millivolts (mV) or decibel (dB) attenuation, or HSLT
both. Increased attenuation indicates better
quality bonding of the cement to the outer casing
> Traditional sonic cement bond log tools. The slim array sonic tool (SSLT) is
6. For more on hole cleaning: Abbas R, Cunningham E,
Munk T, Bjelland B, Chukwueke V, Ferri A, Garrison G, a digital sonic tool that provides conventional openhole sonic measurements,
Hollies D, Labat C and Moussa O: “Solutions for standard cement bond log (CBL) amplitude and a variable density log (VDL).
Long-Term Zonal Isolation,” Oilfield Review 14, no. 3 The SlimXtreme slimhole well logging platform provides the same
(Autumn 2002): 16–29. measurements as the SSLT for evaluation of the cement bond quality in high-
7. Guillot DJ, Froelich B, Caceres E and Verbakel R: pressure and high-temperature environments. The digital sonic logging tool
“Are Current Casing Centralization Calculations Really (DSLT) uses the sonic logging sonde to measure the cement bond amplitude
Conservative?” paper IADC/SPE 112725, presented at the
IADC/SPE Drilling Conference, Orlando, Florida, USA, and provide a VDL display for evaluation of the cement bond quality. The
March 4–6, 2008. hostile environment sonic logging tool (HSLT) provides the same
8. Nelson EB and Guillot D: Well Cementing, 2nd ed. measurements of the cement bond amplitude and the same variable density
Sugar Land, Texas: Schlumberger, 2006. display as the SSLT in standard wellbore sizes. The SCMT Slim Cement
9. Morris C, Sabbagh L, Wydrinski R, Hupp J, van Kuijk R Mapping Tool is a through-tubing cement evaluation tool combinable with
and Froelich B: “Application of Enhanced Ultrasonic the PS Platform new-generation production services tool. It is sized so that it
Measurements for Cement and Casing Evaluation,” may be used to evaluate the cement behind casing in workover operations
paper SPE/IADC 105648, presented at the SPE/IADC without having to first pull the tubing.
Drilling Conference, Amsterdam, February 20–22, 2007.

Spring 2008 21
 dependent on casing thickness and with an
amplitude decay dependent on the acoustic
impedances of the media on either side of the
Casing
casing. The cement acoustic impedance is then
classified as gas, liquid or cement based on the
Bonded cement sheath
thresholds set for acoustic impedance
boundaries between these materials.

Transmitter Sonic pulse path Strengths and Weaknesses

These sonic and ultrasonic logging tools have
Early Signal Arrival had shortcomings. The traditional sonic CBL-VDL

Amplitude, mV
tool does not provide radial or azimuthal
Detection level information to differentiate among channels,
3-ft receiver
gives CBL E1 contaminated cement, microannuli and tool
Transmitter
firing eccentricity; this makes confident data interpre-
Time
Transit time tation difficult.11
5-ft receiver
gives VDL
Ultrasonic imaging tools that are based on the
Partially Cemented Pipe pulse-echo technique are limited when logging in
Amplitude, mV

highly attenuative muds because of low signal-to-

Casing Formation Mud noise ratios. Their radial probing power is limited
arrival arrival arrival to the cement region adjacent to the casing.12
E1 Because of the high acoustic impedance
Transmitter
firing Time
contrast between steel and the surrounding
material—mud inside the casing and cement
outside—the signal dies away so rapidly that
echoes arising from acoustic contrasts outside of
the casing are typically undetectable unless they
> Sonic logging tools. Cement bond logs (CBLs) and variable density logs are very close to the casing and strongly
(VDLs) are acquired using a sonic logging tool with a monopole transducer reflective surfaces.
and two monopole receivers placed at 3 and 5 feet [0.9 and 1.5 m] from the Additionally, the pulse-echo technique has
transmitter (left). The monopole sonic transmitter sends an omnidirectional difficulty differentiating between a drilling fluid
pulse at relatively low frequency (10 to 20 kHz) that induces a longitudinal and a lightweight or mud-contaminated cement of
vibration of the casing. The recorded amplitude of the first positive peak
(E1) of the sonic waveform received at 3 ft and the full waveform received similar acoustic impedance. Even under favorable
at 5 ft represent the average values over the circumference of the casing conditions, the acoustic impedance contrast
(top right). In well-cemented pipe, the sonic signal in the casing is attenuated, between drilling fluid and cement typically must
and the CBL E1 amplitude is small. In free pipe, the casing arrivals are strong.
be larger than 0.5 Mrayl for the pulse-echo
The transit time is the time it takes the wave to travel from transmitter to
receiver. It is used for quality control of the tool centralization and to set technique to distinguish between them.
parameters for material detection. In partially cemented pipe (bottom right), To overcome tool limitations, and depending on
casing, formation and mud arrivals may be present and can occur in the well conditions, an utrasonic and standard CBL-
presence of a microannulus at the casing/cement interface. The VDL (bottom
VDL tool may be run together. But even then,
inset) provides visualization of arrivals that propagate in the casing as
extensional waves and in the formation as refracted waves. experience from various wells around the world
has shown that an unambiguous conclusion about
the quality of the cement bond may be elusive. This
wall. In simple cases, the interpreted log ultrasonic imaging tools, such as the is particularly true in the case of lightweight and
response can provide good information about Schlumberger USI UltraSonic Imager, use a contaminated cements.
cement quality (above). rotating transducer that emits a broadband This issue has become increasingly urgent
About 25 years ago, engineers developed ultrasonic wave perpendicular to the casing wall with the proliferation of lightweight cements in
cased-hole ultrasonic imaging tools that used a with a frequency that can be adjusted between deepwater wells and in sealing across formations
high-frequency pulse-echo technique (next page, 250 and 700 kHz (next page, right). The effect is to with low pore pressure. To deal with this
top left).10 More recent versions of these excite a casing resonance mode at a frequency problem, Schlumberger has developed a
measurement technique that is the basis of the
10. Sheives TC, Tello LN, Maki VE Jr, Standley TE and presented at the SPE Latin American and Caribbean Isolation Scanner cement evaluation service. The
Blankinship TJ: “A Comparison of New Ultrasonic Petroleum Engineering Conference and Exhibition,
Cement and Casing Evaluation Logs with Standard Rio de Janeiro, August 30–September 3, 1997. tool combines the classic pulse-echo technique
Cement Bond Logs,” paper SPE 15436, presented at the 12. Van Kuijk R, Zeroug S, Froelich B, Allouche M, Bose S, with an ultrasonic imaging technique that
SPE Annual Technical Conference and Exhibition, Miller D, le Calvez J-L, Schoepf V and Pagnin A:
New Orleans, October 5–8, 1986. “A Novel Ultrasonic Cased-Hole Imager for Enhanced provides more effective imaging of the annular
11. Coelho de Souza Padilha ST and Gomes da Silva Araujo R: Cement Evaluation,” paper IPTC 10546, presented at the fill including reflection echoes at the
“New Approach on Cement Evaluation for Oil and Gas International Petroleum Technology Conference, Doha,
Reservoirs Using Ultrasonic Images,” paper SPE 38981, Qatar, November 21–23, 2005. cement/formation interface.

22 Oilfield Review
 Tranducer Mud Casing Cement Formation

Echo amplitude Transit time Resonance frequency Resonance decay

Internal casing Internal radius Casing Cement acoustic

condition thickness impedance

Formation Acoustic
Casing resonance Casing beam
Cement

Time

ion
tat
Ro

Resonance
decay Transducer Metal plate

> Cased-hole ultrasonic tool basics. An ultrasonic tool’s transducer sends a

slightly divergent beam—an acoustic wave generated by a transducer when
electrical power is applied to it—toward the casing to excite the casing into
its thickness resonance mode. The USI UltraSonic Imager tool scans the
casing at 7½ revolutions per second to render an azimuthal resolution of 5 or
10 degrees. This yields 36 or 72 separate waveforms at each depth. These are
processed to yield the casing thickness, internal radius and inner wall
smoothness—from the initial echo—as well as an azimuthal image of the
cement acoustic impedance—from the signal resonance decay (top). The
acoustic impedance of the cement (essentially the quality of the cement
sheath) can be derived from the resonance decay (bottom). A good casing-
cement bond results in immediate resonance decay, while free pipe rings
(generates echoes) for an extended period.

A
C
D

B > USI tool. The Schlumberger USI tool improved

on earlier versions of the ultrasonic imaging tool
by using a single rotating transducer mounted on
the bottom of the tool (A).

Sounds of Success
The Isolation Scanner tool includes a rotating
subassembly supporting four transducers (left).
A A normally aligned transducer for generating
and detecting the pulse echo is positioned on
one side of the tool. The other three transducers
are on the opposite side of the tool and are
> Isolation Scanner subassembly. The Isolation Scanner sub combines the
aligned obliquely. One of these transducers
traditional pulse-echo technique using an acoustic transmitter and receiver
normal to the casing (A), while adding flexural-wave imaging with one transmits a high-frequency pulsed beam of about
transmitter (B) and two receivers (C) aligned obliquely. This configuration 250 kHz to excite a flexural mode in the casing.
excites the casing flexural mode (D). The subassembly, mounted on the same
platform as the USI tool and with updated signal generation and acquisition
software, is the basis for Isolation Scanner tool.

Spring 2008 23
 SLG Flexural
Annulus Measured Attenuation, Impedance, Channel Channel
CBL Sonic VDL map
R depth, m L dB/cm Mrayl map width, %

–5,000 0 5,000 0 0.5 1 0 1 2 3 4

S G
0 50 100 0 180 0 180 0 180 0 180 0 50100
X,440

USI X,450
Formation
tool

X,460

X,470

T Casing
X,480

Casing arrival TIE

X,490

X,500

X,510

80 90 100 110 120 130 140 150 160 170

Time, µs
Isolation Scanner tool flexural-wave imaging
USI tool pulse-echo imaging
> Geometrical interpretation of USI measurements
and Isolation Scanner flexural-wave imaging.
Shown here is a geometrical ray interpretation of
the propagation of the signal for the pulse-echo
(top, red) and from the transmitter (T) to a
receiver (R) for the flexural-wave pitch-catch
techniques (top, blue). A typical waveform from
the latter technique comprises an early echo,
called casing arrival, and third-interface echoes
(TIEs) (bottom, blue). The attenuation of the
casing arrival amplitude is used to complement
the pulse-echo measurement (bottom, red) in
distinguishing unambiguously between fluid and
solid behind the casing. The properties of the TIE
provide an enhanced characterization of the
cased-hole environment, indicating the acoustic
properties of the material filling the annulus, the
position of the casing within the hole and the
geometrical shape of the hole.
X,520
X,530
X,540
> Isolation Scanner and CBL-VDL measurements. In 2003, the Isolation
Scanner prototype was tested in an In-Salah Gas vertical well. The 95⁄8-in.
casing was cemented in a 12¼-in. hole using the low-density (low-
impedance) LiteCRETE slurry system. The CBL (Track 1) and VDL (Track 2)
show a nearly free-pipe response with strong casing arrivals in the VDL. The
pulse-echo impedance map (Track 5) shows fluid with patches of solid.
Obtaining an adequate interpretation from both measurements was made
difficult by the low-impedance LiteCRETE cement. The flexural-wave
attenuation map (Track 4), on the other hand, provides a correct diagnosis of
the solid behind casing. It also reveals the existence of a fluid-filled channel
between X,465 and X,485 m. The solid-liquid-gas (SLG) map (Track 3) supports
and simplifies this information. The azimuthal and axial extent of the channel
is reported automatically in Tracks 6 and 7.

As it propogates, this mode radiates acoustic
energy into the annulus; this energy reflects at
interfaces that present an acoustic contrast,
such as the cement/formation interface, and
propagates back through the casing predomi-
nantly as a flexural wave to reradiate energy into
the casing fluid. The two receiving transducers
are placed to allow optimal acquisition of these
signals (above left).
This new technique is termed pitch-catch.
Processing of the resulting signals provides
information about the nature and acoustic
velocity of the material filling the annulus, the
position of the casing in the hole and the
geometrical shape of the hole.
The first aim of processing Isolation Scanner
logs is to obtain a robust interpretation of the
material immediately behind the casing. The
inputs to this processing sequence are the

24 Oilfield Review
 150 Casing
1.8 Gas
Liquid
1.6 Solid 100

1.4
50

Flexural-wave attenuation, dB/cm

1.2

Time, µs
0
1.0

0.8 –50
Channel
0.6
–100
0.4 Formation reflection
within channel
–150
0.2 –150 –100 –50 0 50 100 150
Time, µs
0
> Waveform polar plot across the fluid-filled
–0.2
–0.2 0 2 4 6 8 10 channel at a depth of X,477 m on the Isolation
Scanner log (previous page, top right). The
Pulse-echo impedance, Mrayl curvature of the annulus-formation echo reveals
that the casing is slightly eccentered in the
> Solid-liquid-gas mapping of the measurement plane for a Class G cement. borehole and that the channel is located on the
Once the expected impedance values are defined for the cement, liquid and narrow side (direction of blue arrows). The
gas through a laboratory-measured database and the material selection is absence of a third-interface echo across the
converted into acoustic properties, the next step is to predict the cement azimuth may be due to a low acoustic
measurements from the expected acoustic material properties. Then, multiple contrast between the cement and formation.
realizations of the measurement noise are added to generate three clouds of
points (solid, liquid and gas) in the bidimensional measurements plane.

cement impedance, as delivered by the pulse-
echo measurement, and the flexural-wave
attenuation computed from the amplitude of the
casing arrivals on the obliquely aligned receivers.
These two inputs are independent measure-
ments linked through an invertible relation to
the properties of the fluids inside both the
casing and annular fill. The inputs are first
combined to eliminate the effect of the inside
fluid, thus obviating the need for specific
hardware for fluid-property measurements
required by the USI tool.
The output of the Isolation Scanner service is
a solid-liquid-gas (SLG) map displaying the most
likely material state behind the casing. The state
is obtained for each azimuth by locating the two
measurements, corrected for the effects due to
the inside fluid, on a crossplot of attentuation
and acoustic impedance, giving the area
encompassed by each state (above left). The
measurement plane can be mapped out in
different regions with three colors corresponding
to the different states (previous page, top right).
The white-colored areas in the SLG map
correspond to locations with nonsolvable
inconsistencies between measurements, such as
might appear at the casing collars.
In addition to evaluating the material behind
the casing, a second objective of processing is to
extract relevant information from the annulus-
formation reflection echo or echoes and further
characterize the annulus between the casing
and formation.
First, the software detects the echoes on the
waveform envelope following the casing arrival
and then measures their time of arrival and
amplitude. From the time differences between
the reflection echoes and the casing arrival—
provided enough echo azimuthal presence is
available in the data—it is straightforward to
determine how well the casing is centered within
the borehole. This is presented as a percentage
in which 100% represents perfect centering, and
0% is casing in contact with the formation wall.
Additionally, if the borehole diameter is known,
the time-difference processing can be further
converted into a material-wave velocity and
displayed as an annulus-velocity map or cement
azimuthal thickness.
A polar plot of the flexural waveforms from
the variable density log provides a picture of the
geometry of the casing within the borehole
(above right).
New Cements
Even the most sophisticated logs present only a
snapshot of the cement condition and its ability
to provide zonal isolation. Over the long life span
of a well, changing downhole conditions remain
the enemy of cement sheaths and may cause
even well-placed sheaths to fail over time.
Throughout the industry’s long history of
using cement in well construction, addressing
these failures first focused on placement of the
slurry and later on its chemical makeup. During
the 1980s, engineers and scientists began to
consider ways to deliver specific set-cement
properties with the aim of increasing the

Spring 2008 25
 likelihood of attaining good zonal isolation. From
Old wells New wells this strategy came the idea of reducing cement
density through the injection of nitrogen into the
slurry while pumping, and of introducing ceramic
Pressure and Formation and Pressure and microspheres into the cement blend. The latter
temperature history casing properties temperature prediction design was the precursor to the Schlumberger
CemCRETE concrete-based oilwell cementing
technology, including the LiteCRETE slurry
Pressure uncertainties
system, and CemSTONE advanced cement
technology. These innovations allowed engineers
to increase or decrease slurry density without
Cement properties Cement properties
significantly affecting the permeability of the
Laboratory tests set cement.
Laboratory tests
The new cement systems were accompanied by
CemSTRESS Database of properties
development and deployment of software to
software
analyze and improve fluid displacement behind
casing and simulate stresses on the cement over
the life of the well (left). Beginning in 2000,
Predicts Young’s
modulus of new continual improvements to cementing software
cement system provided engineers with a tool to tailor slurries
for input to based on gas-migration risks and wellbore stresses.
CemSTRESS
software In 2002, Schlumberger introduced FlexSTONE
advanced flexible cement technology to handle
SlurryDesigner changing stresses imposed on cement sheaths
software
over time. Expected stress changes from drilling,
Robustness criteria
production and abandonment activities are
predicted by numerical modeling. The system’s
mechanical properties are customized using
FlexSTONE trimodal particle-size distribution
technology. The resulting mechanical flexibility
Appropriate? Appropriate? allows these cement systems to resist failure
through a variety of changes that may occur
Yes No Yes No during the drilling, production and abandonment
cycles of a well.13
While such methods increase the resistance of
the cement matrix to physical stresses, they are
ineffectual once the cement sheath fails. Even if
the sheath is intact during the well’s lifetime, the
increased emphasis on environmental
responsibility dictates that hydrocarbon-
Continue to operate Mitigate bearing formations remain sealed for many years
Blend, test and
well within safe consequences of
pump cement after the asset has been plugged and abandoned.
stress limits failed cement sheath
This extended period of service significantly
increases the chances that even the most
appropriate and resilient cement sheath may fail.
> Designing cement systems. Cementing experts can use the CemSTRESS cement sheath stress
In response to these concerns, Schlumberger
analysis software to analyze the radial and tangential stresses imposed on each casing string during
events such as treating and pressure testing. In addition to indicating cement sheath performance in engineers have taken another step in the
compression, tension, or both, the software has the ability to establish parameters, including set- evolution of zonal isolation systems with the
cement flexibility, support and standoff. It can also be used to identify both inner and outer microannuli introduction of self-healing cement (SHC). As
and show their size and development over time. CemSTRESS software uses a three-stage the name implies, when cracks or microannuli
methodology to aid in selecting and designing a cement system that can extend well life. In the first
stage of the method, a cement expert determines whether the well requires a conventional cement occur at the interface between the cement
system or a specialized system. This provides direction for the next two stages. The second stage of sheath and the casing or formation, self-healing
the methodology analyzes scenarios to design a cement system whose Young’s modulus is below the components within the set-cement matrix swell
stress level that the software predicted would induce failure. In the third stage, Schlumberger
cementing engineers use proprietary software, such as SlurryDesigner cement blend and slurry
design software, to optimize the cement slurry design.

26 Oilfield Review
to close the gaps without any outside inter-
Formulation and Properties Design 1 Design 2 Design 3
vention. This FUTUR active set-cement
BHCT, °C [°F] 60 [140] 25 [77] 60 [140]
technology reacts specifically to the presence of
Density, kg/m3 [lbm/galUS] 1,870 [15.8] 1,700 [14.2] 1,400 [11.7]
hydrocarbons. When the integrity of the cement
Antifoam, L/t [gal/sk]* 2.66 [0.03] 0.2% BWOB 4.2 [0.05]
sheath is compromised and zonal isolation is
Dispersant, L/t [gal/sk] 6.22 [0.07] 4.2 [0.05]
breached, the cement reacts to the presence of
Retarder, L/t [gal/sk] 2.66 [0.03] 0.5% BWOB
hydrocarbons by swelling. This effectively closes
Gas-migration control additive, %BWOC*** 0.77
the gap and shuts off formation fluid movement.
Fluid-loss control agent, %BWOB** 0.7
Except for its self-healing abilities, FUTUR
Gelling control agent, %BWOB 0.5
cement is similar to traditional cement.
Successful placement requires adherence to the
same best practices as any oilfield cementing
Mixing Rheology
operation, and the cement itself requires no
special mixing or pumping equipment. FUTUR Ty, Pa [lbf/ft2] 1.9 [4.0] 11.4 [23.9] 5.9 [12.3]

slurries are compatible with all standard PV, MPa [thousand psi] 237 [34.3] 202.7 [29.4] 55 [8.0]

additives and spacers. Standard mixing and

slurry tests of rheological properties, free fluid,
API Rheology
sedimentation, fluid-loss control, thickening
time and development of compressive strength Ty, Pa [lbf/ft2] 5.7 [11.8] 9.0 [11.9] 7.8 [16.3]
all apply. PV, MPa [thousand psi] 151 [21.9] 148.7 [21.6] 60 [8.7]
Once placed in the well, FUTUR cement 10-s gel, Pa [lbf/100 ft2] 5.1 [10.7] 18.2 [38] 7.4 [15.4]
behaves in the same way as classic cements when 10-min gel, Pa [lbf/100 ft2] 13.7 [28.5] 10.8 [22.6] 10.2 [21.4]
not in the presence of hydrocarbon, and its set- 1-min stirring gel, Pa [lbf/100 ft2] 11.4 [23.7]
cement properties are equivalent to those of API free fluid at 60°C [140°F], mL Traces 58
traditional cements (right). API free fluid at 25°C [77°F], mL 0.5
API fluid loss at 25°C [77°F], mL 30
Laboratory Work API sedimentation test, lbm/galUS 0.2 – 0.2 – 0.15
FUTUR cement technology is designed to provide
well integrity for the very long term. Therefore,
laboratory testing to replicate downhole sheath Thickening Time

failure was critical in proving that the cement At BHCT, h:min 6:16 8:05 6:33
would indeed heal itself and that it would Time 30 to 100 Bc****, h:min 0:54 1:09 4:13
continue do so for years after placement. The
cement also had to be checked for any problems
its self-healing characteristic might create. Compressive Strength Development
To test swelling properties, the cement was Time to reach 50 psi at BHST, h:min 9:23 6:00
placed in an annular expansion mold. These tests Time to reach 500 psi at BHST, h:min 35:44 11:16
simulate normal setting of the cement matrix in 24-h compressive strength, MPa [psi] 2.5 [363] 4 [637]
the well, followed by an invasion of hydrocarbon
such as would be expected when cracks or
creation of a microannulus causes a loss of zonal Mechanical Properties of SHC Matrix After 7 Days Curing in Water at Atmospheric Pressure
isolation. The FUTUR cement was cured in water Temperature, °C [°F] 60 [140] 25 [75] 25 [75]
for seven days prior to immersion in oil, and Compressive strength, MPa [thousand psi] 20 ± 5 [2.9 ± 0.7] 10 ± 0.8 [1.5 ± 0.1] 4.5 ± 0.5 [0.65 ± 0.07]
identical temperatures and pressures were Young’s modulus, MPa [thousand psi] 6,500 ± 500 [940 ± 73] 2,800 ± 400 [400 ± 60] 1,300 ± 300 [190 ± 44]
used in water and oil. Results showed that the
linear swelling increased with temperature at
*gal/sk = gallon of additive per sack of cement **%BWOB = by weight of blend
constant pressure. ***%BWOC = by weight of cement ****Bc = Bearden’s unit of consistency
To evaluate the FUTUR system’s self-healing
> Slurry designs. Laboratory tests determined the properties of three FUTUR slurry designs. Designs 1
properties, engineers at the Schlumberger
and 3 were tested at 60°C, and Design 2 was tested at 25°C. All designs used Class G cement and
Riboud Product Center in Clamart, France, were prepared with fresh water. The slurry rheologies were measured with a Fann 35 viscometer
developed a flow loop to simulate downhole after mixing at ambient conditions and after 20 minutes of conditioning at the bottomhole circulating
conditions and installed an SHC cell designed to temperature (BHCT). The plastic viscosity (PV) and the yield value (Ty) were calculated using the
Bingham plastic model. Thickening times of these systems were controllable, and no free water was
observed. For all three designs, a compressive strength of 3.44 MPa [500 psi] was achieved in less
13. For more on new cements: Abbas et al, reference 6.
than 48 hours, as measured using an ultrasonic cement analyzer. The compressive strengths of the
designs ranged from 4.5 to 20 MPa.

Spring 2008 27
 100

Microannulus width, µm
80

60
Cement
40

Outer ring 20

0
Inner ring 0 1 2 3
Time, days

100
Expandable and
retractable
center assembly 80

Normalized flow rate, %

60

40

> Healing check. An SHC cell with two concentric cylinders simulates an 20
annular volume. The outer cylinder, or ring, is a thin, steel sleeve (green). The
inner cylinder (purple) is made of a deformable elastic material into which a 0
radially expandable core assembly (gray) is inserted, allowing expansion of 0 0.5 1.0 1.5
the inner sleeve in a controlled manner. Top and bottom plugs sealing the Time, days
annular volume are equipped with fittings that allow fluid to enter and to Neat system
escape the cell. With the inner cylinder expanded by a core assembly, the Self-healing cement
cement is injected into the annular space. Once the cement has set, the
inner-cylinder expansion is released. The inner cylinder shrinks back to its
original shape, generating a microannulus of a controlled size. Radial cracks > Test cell results. The SHC cell was installed in
within the cement are created by expanding the inner core assembly after a flow loop to investigate FUTUR self-healing
the cement has set. efficiency. Oil was injected through samples in
the SHC cell at pressures up to 0.4 MPa [58 psi],
corresponding to a differential pressure of
5.3 MPa/m [234 psi/ft] across the sample. In one
test, a microannulus of 100 microns was created
between the casing and the cement inside the
SHC cell. The neat cement system (green)
evaluate self-healing capability in an annular capabilities of the system, which maintained allowed oil to flow through the sample, whereas
configuration (above). Oil was injected through integrity and continued to block the flow of oil at the SHC system (blue) responded to oil invasion
samples to test both cracks and microannuli. In high pressure. The tests were repeated with with an efficient closure of the microannulus in
less than 6 hours.
one test, a microannulus of 100 microns was differential pressures up to 5.3 MPa/m, and the
created between the casing and cement inside cement’s self-healing property was confirmed in
the SHC cell. While a conventional cement system every test.
tested in this apparatus allowed oil to flow A specialized testing system was devised to
through the sample, the FUTUR system reacted to study self-healing properties in dynamic Swelling tests performed after prolonged
oil invasion with an efficient closure of the conditions with dry gas under realistic reservoir immersion of FUTUR cement in water confirmed
microannulus in less than six hours (above right). conditions. Test results highlighted the efficiency that the self-healing properties were maintained.
The self-healing system has also been tested of the FUTUR system when exposed to natural In this test, cement placed in an expansion cell
in a cyclic failure scenario. Successive micro- gas under dynamic conditions. In less than one was cured in water for several months, then
annuli were created and repaired with the SHC hour, the self-healing cement caused a immersed in oil at 60°C. Results showed that the
at a differential pressure of 1.4 MPa/m significant decrease of flow rate from reactivity of the self-healing matrix remained
[62 psi/ft]. Using the same flow loop, the flow of 425 mL/min [26 in.3/min] to 0.52 mL/min effective even after resting dormant for a year.
oil through successively generated cracks was [0.03 in.3/min] (next page, top right). Testing of the matrix integrity, after one year of
repeatedly shut off by the SHC. The same test Finally, researchers investigated the exposure in oil, also showed no indication that
using a conventional neat cement system did not durability of the self-healing cement. The the integrity of the matrix was deteriorating. The
show any decrease in flow (next page, top left). durability—or aging—test consisted of two mechanical properties remained within the same
FUTUR cement also can handle oil flows at parts. The first part, using a swelling test, was to range even after immersion in oil for a year.
higher pressure. A pressure increase to 3 MPa/m check that the self-healing property is
[133 psi/ft] did not diminish the self-healing maintained over time. The second was to
evaluate whether matrix integrity is maintained
when the cement is immersed in oil for a long
period of time.

28 Oilfield Review
 Flow rate of nitrogen through cement
60
Microannulus width, µm Flow rate of natural gas through cement
50
1,000
40
30
20 100

Flow rate, mL/min

10
0
0 10 20 30 40 50 60 70 80 90 100 10
Time, min
120
1
100
Crack width, mm

80 Self-healing cement Neat cement

60
> Testing SHC with gas flow. SHC seals in the same way for hydrocarbon gas
40 as it does for oil. A test cell containing cement was cured in such a way that
20
a microannulus of an arbitrary size was present. With SHC in the holder, the
flow rate through the cement drops significantly in less than one hour from
0 the time the fluid was switched from inert nitrogen to natural gas (left).
0 10 20 30 40 50 60 70 80 Traditional cement tested in the same fashion experienced almost no loss
Time, min of flow rate in that time (right). The specialized testing system is based on a
Neat system
Hassler sleeve-type core holder to prevent gas passing around the outside
Self-healing cement
of the core.

> Repeated healing. The self-healing systems

were also tested in a cyclic failure scenario.
Successive microannuli were created and
repaired with the SHC system (top). At a
differential pressure of 1.4 MPa/m, the flow of oil
through successively generated cracks was
repeatedly shut off by the SHC system, while the
same test using a conventional system did not
show any decrease in flow. At increased
pressure, the FUTUR slurry design reacted to
stop the invasion of oil through a 100-micron
crack in less than 20 minutes (bottom).
Surface Solutions
The self-healing nature of FUTUR cement over
time, as demonstrated in the laboratory, makes it
particularly well-suited for long-term zonal
isolation. That same ability also means SHC is a
good solution for immediate or chronic gas-
migration problems.
For instance, because of highly varied geology
and several shallow, gas-bearing coal seams,
wells in the foothills of the Rocky Mountains of
Alberta, Canada, present a particular set of
cementing challenges. The coal seams may emit
gas that eventually migrates through the casing
annulus and manifests as surface casing vent
flows (SCVF). Depending on the extent of the
leak, operators may be required to shut in, repair
or even abandon their afflicted wells.
Remediation of SCVF on these wells costs from
US $250,000 to $1 million per well—a figure that
does not include the loss in production or
potential loss of the well.14
To address the problem, the operator of a
deep gas field in the west-central Alberta Grande
Cache area turned to FUTUR cement technology
for zonal isolation in two new wells. The self-
healing system was chosen to complement
cementing practices implemented to reduce the
risk of SCVF, which occurs in approximately 10%
of Grande Cache area wells.
Both Well 1 and Well 2 required cement to the
surface. The operator had a particular concern
about Well 1; a similar well located about 500 m
[1,640 ft] away had experienced SCVF. Losses
encountered while drilling Well 2 required the
use of a stage tool to ensure cement placement to
the surface across known nuisance gas zones.15
The SHC slurries were mixed and pumped using
standard oilfield equipment, easily achieving
continuous mixing rates of up to 0.95 m3/min
[6 bbl/min].
Immediately after cementing of Well 1, some
gas pressure was observed in the intermediate
casing annulus. However, gas pressure was not
evident after the well’s completion, suggesting
that the SHC had activated to contain a leak.
Well 2 displayed no leaks in the 12 months
following cementing operations. While that may
seem a short observation period, SCVF in this
area typically occurs within days or weeks of the
cementing operation.
During completion and production, the two
wells were subjected to various downhole stresses,
including downhole pressure of 64 MPa [9,282 psi]
applied to test the completion and stresses
related to cyclical temperature changes caused by
a heater string in the top 600 m [1,968 ft] of the
well. Throughout and after these events, no
surface casing vent flows were detected.
Elsewhere in central Alberta, another field
was also plagued by SCVF from zones above
the target formation. To complement other
cementing technologies already in use, the
operator selected FUTUR self-healing cement for
use in two wells. In the first, cement density was
constrained to a maximum 1,380 kg/m3
[11.5 lbm/galUS] because fluid losses had been
observed during drilling. With a nearly opposite
problem on the second well—nuisance gas
detected by mud loggers during drilling—drilling
fluid density was increased to 1,470 kg/m3
14. Roth J, Reeves C, Johnson CR, De Bruijn G, Bellabarba M,
Le Roy-Delage S and Bulte-Loyer H: “Innovative
Hydraulic Isolation Material Preserves Well Integrity,”
paper IADC/SPE 112715, presented at the IADC/SPE
Drilling Conference, Orlando, Florida, March 4–6, 2008.
15. Cementing stage tools allow slurry to be placed at
specific depths along the casing through sliding sleeves.
They are used when the hydrostatic pressure of the full
column of cement threatens to overcome the wellbore
fracture gradient beneath the stage tool.

Spring 2008 29
 Acoustic
L
Impedance Flexural Casing and Casing and
SLG Min Amplitude Max
Cement Attenuation Casing Third-Interface Third-Interface
Map
Map Map Centering Long Axis Short Axis CBL
Measured S G Variable Density
Depth, m 0 180 01234 0 0.5 1 0 % 100 5 in. 2.5 5 in. 2.5 0 mV 50 200 1,200
X,650

X,675

X,700

X,725

> Analysis of cement evaluation logs in the Cortemaggiore 155dir underground gas-storage (UGS) well. The brown color of the
SLG map (Track 1) of the ultrasonic flexural attenuation logging tool indicates solid (cement), resulting from a measured
acoustic impedance (Track 2) of about 5 Mrayl (close to the expected cement value) and high flexural attenuation. The CBL
(Track 7) is in agreement, showing 100% casing to cement bonding (average CBL value of 5 mV) and strong formation arrivals
on the VDL (Track 8), which is an indication of excellent cement-to-formation bonding. The optimal cement bonding is also
related to the fact that the liner is fairly well-centralized, as shown by casing centering (Track 4) and third interface short- and
long-axis outputs (Tracks 5 and 6). The casing centering curve is above 80% for most of the interval shown except near a depth
of X,720 m, where the casing nearly touches the formation. This is also indicated in the vanishingly small cement thickness
along the short axis. The horizontal features visible on the SLG map, cement map and flexural attenuation map (Track 3) are the
casing joints about every 14 m [45 ft] and two casing centralizers per joint. Together the logs show optimal cement bonding to
the casing and to the formation, providing assurance of effective hydraulic isolation across the permeable injection zones of
the UGS well.

30 Oilfield Review
[12.25 lbm/galUS]. This meant cement density
for that well had to be increased to 1,550 kg/m3
[12.9 lbm/galUS] to meet the requirements of the
mud-removal plan.16
Zonal isolation was achieved with the new
cementing system in both wells despite difficult
conditions—a low-pore-pressure zone in Well A,
and a narrow pore-pressure-fracture window
across a gas-influx zone in Well B. Cement
returns to the surface in Well A also
demonstrated that SHC can be applied in a
single-stage cement job in wells prone to lost
circulation. During and after subsequent drilling,
stimulation and completion operations, there
was no indication of annular gas flow.
Gas Storage
Sustaining self-healing characteristics over time
holds special attraction for engineers charged
with sealing underground gas-storage (UGS)
wells. Because these wells are used to both inject
and produce, the wellbores are repeatedly
subjected to considerable temperature and
pressure changes—often in short cycles—that
can induce stress-load changes on the casing
and cement.
Additionally, in contrast to producing wells
that have a life expectancy of perhaps 20 years,
UGS well plans are likely to include a production-
injection life span of 80 years or more. As a
consequence, zonal isolation failure in
underground gas storage wells is a significant
and ongoing operator concern (see “Intelligent
Well Technology in Underground Gas Storage,”
page 4). In many UGS facilities, poor hydraulic
isolation is caused by drilling fluid channeling as
a result of eccentric casing or through the
development of a dry microannulus.17
A combination of these factors had
historically resulted in poor zonal isolation in
UGS wells in an Eni S.p.A.-operated, depleted gas
field in northern Italy. The challenges facing Eni
and Schlumberger engineers included obtaining
seals across gas-injection zones and gas-tight
cement sheaths across deviated (49°), washed-
out sections.
For a new well in this field, Eni subsidiary
Stogit chose to cement the production casing
with FUTUR self-healing cement. The plan also
included proper placement of centralizers, a gas-
migration analysis software package and
16. Cavanagh et al, reference 2.
17. Moroni N, Panciera N, Zanchi A, Johnson CR,
LeRoy-Delage S, Bulte-Loyer H, Cantini S, Belleggia E
and Illuminati R: “Overcoming the Weak Link in Cemented
Hydraulic Isolation,” paper SPE 110523, presented at the
SPE Annual Technical Conference and Exhibition,
Anaheim, California, USA, November 11–14, 2007.
Spring 2008
Isolation Scanner logging tools to evaluate the
cement bond.
The solution to persistent problems of
sustained casing pressure in UGS wells drilled in
this area involved a multipronged approach:
• more centralizers to improve standoff
• a liner across the zone of interest to facilitate
casing rotation during cementing
• use of a software advisor tool for gas-migration
prediction and prevention during cement
hydration
• software to tailor the cement system to associ-
ated risk
• use of FUTUR SHC
• a full suite of postcementing logging tools that
included all ultrasonic and flexural-
wave measurements.
In 2007, the SHC cementing operation was
performed on the Cortemaggiore 155dir UGS
well. The team used a computer-aided design and
evaluation software program that optimized mud
removal and standoff, and fine-tuned the spacer
and slurry characteristics. A separate software
program was used to evaluate the risk of the
severity of gas migration based on the pressure-
decay limit—a measurement of how far the
hydrostatic pressure of the slurry will fall during
hydration before it is below the pore pressure
and might allow gas migration into the annulus.
Finally, a mechanical-stress modeling
software tool that simulated pressure and
temperature variations during the life of the well
was used to evaluate cement sheath integrity
over time. The software modeled the three
mechanisms of cement sheath failure—traction,
compressional failure and both internal and
external microannulus development.
Once the cement was placed, CBL-VDL and
Isolation Scanner tool analyses all indicated
optimal cement bonding at the casing and
formation interfaces (previous page). The
ultimate success of the SHC system is being
monitored over time as gas is cyclically injected
into and produced from the well.
In another recent FUTUR cement appli-
cation, the technology was applied in
environmentally sensitive areas of the Canadian
Rocky Mountains. Following the suspension of
drilling because of surface casing leaks and
resulting environmental concerns, one US
operator reevaluated its well construction
procedures in the area and then resumed
operations. Despite the revised strategy, three of
seven wells drilled had obvious gas leaks while
the other four were suspect; possible leaks were
masked by heavy cement poured around the
casing at the surface. The company then added
FUTUR cement to its drilling and completion
program. Of the 13 area wells drilled and
completed since then, only two wells had
detectable leaks; one was due to an operational
failure unrelated to the SHC and a second was
barely detectable.
Time Will Tell
Long-term performance of its self-healing
characteristics is key to the success of FUTUR
cement. Laboratory work shows that this SHC
will continue to close the pathways through
which gas migrates without intervention
throughout the life of the well and beyond. In
time, operators pressured by environmental
regulators—internal and external—will insist
that the cement sheaths in their wells prevent
hydrocarbons from escaping formations long
after the well has been plugged and abandoned.
The ability of FUTUR cement to react to and
repair the channels through which formation
fluids travel to the surface makes it an ideal
answer to such demands.
Operators, especially those in gas-migration
prone areas, will also come to expect an improved
view behind casing to eliminate other costly zonal
isolation tests in the face of conflicting or
ambiguous CBL-VDL logs. In drilling environ-
ments with proximate pore pressures and
fracture gradients, lightweight cements that pose
a significant challenge to traditional sonic logging
tools are required. The Isolation Scanner tool
offers a clear solution to these and other current
zonal isolation challenges. —RvF
31
 Intelligence in Novel Materials

Defined as materials whose properties can be varied controllably in response to

changes in their environment, smart materials can convert one type of energy to
another. This opens the way to use them for performing the complex functions of
sensors and actuators—sometimes several functions simultaneously—in a device
essentially consisting of a single piece of a single substance.

Rashmi Bhavsar Throughout history, people have shaped tools from The two main types of energy-conversion
Nitin Y. Vaidya the materials at hand. With better understanding devices are sensors and actuators, and these are
Rosharon, Texas, USA of material properties came the ability to fabri- the principal applications of smart materials. A
cate materials with designed characteristics. sensor converts an action to a signal, whereas an
Partha Ganguly Currently, a materials category that is experi- actuator converts a signal to an action.
Alan Humphreys
encing extensive research and some application is Conventional sensors and actuators are typically
Agathe Robisson
“smart materials.” constructed of multiple materials and have
Huilin Tu
Nathan Wicks Some smart materials are widely known. movable parts. Some smart materials can perform
Cambridge, Massachusetts, USA Piezoelectric lighters and igniters in gas stoves, the functions of several materials and parts
grills and other gas appliances produce a spark, simultaneously, thus simplifying the device design
Gareth H. McKinley or electric discharge, without using an electric and having fewer parts to break or wear down.
Massachusetts Institute of Technology circuit—just by striking a piezoelectric crystal From the standpoint of practical applications,
Cambridge, Massachusetts with a spring-loaded hammer. This property of of greatest interest are materials that convert
piezoelectric materials to “feel” pressure and mechanical energy to thermal, electric, magnetic
Frederic Pauchet respond by generating electric potential is used or chemical energy, and vice versa. In addition to
Clamart, France in a wide range of smart applications. Other piezoelectric materials converting mechanical
smart materials respond to different external energy into electricity, other smart materials that
For help in preparation of this article, thanks to Hiroshi Hori,
Sagamihara, Kanagawa, Japan. stimuli, such as temperature, electromagnetic are utilized in commercial applications include
CQG (Crystal Quartz Gauge), FUTUR, Isolation Scanner, fields and moisture. shape-memory alloys that respond mechanically
Q-Marine, Q-Seabed, Sonic Scanner and sonicVISION
are marks of Schlumberger. What all smart materials have in common is to applied heat; magnetorheological and
CryoFit is a mark of Aerofit Products Inc. the ability to convert one type of energy to magnetostrictive materials, whose properties are
Motion Master is a mark of LORD Corporation. another. Piezoelectric materials can convert controlled by the application of magnetic fields;
Simon Nitinol Filter is a mark of C.R. Bard, Inc. or an affiliate. mechanical energy to electric energy, and vice and materials that swell when chemically
Smart Magnetix is a mark of Biedermann Motech GmbH.
versa. Other smart materials convert between activated. This article will look at some of these
1. Schwartz MM (ed): Encyclopedia of Smart Materials.
New York City: John Wiley & Sons, 2002. other types of energy. A key to practical appli- materials, their current applications and their
2. Otsuka K and Wayman CM (eds): Shape Memory cations is the fact that this conversion can occur potential for use in future oilfield applications.
Materials. Cambridge, England: Cambridge University
Press, 1998.
in a controlled manner. Materials that manifest
3. Kauffman GB and Mayo I: “The Metal with a Memory,” this property of responding in a controllable
Invention & Technology Magazine 9, no. 2 (Fall 1993): fashion to changes in the environment are
18–23, http://www.americanheritage.com/articles/
magazine/it/1993/2/1993_2_18.shtml (accessed commonly termed smart materials.1
December 4, 2007).

32 Oilfield Review
Thermally Activated Materials: Total Recall
Some materials can be deformed but then regain
their original shape when heat is applied. These
are called shape-memory materials. Alloys with
properties later found to be directly related to
the shape-memory phenomenon have been
known since the 1930s.2 However, technological
application of this phenomenon did not occur
until almost three decades later.3 In early
1958, William J. Buehler, a metallurgist at the
Naval Ordnance Laboratory (NOL), White Oak,
Maryland, USA, began to test various alloys to be
used for the nose cone of a submarine rocket. He
determined that a nickel-titanium alloy had the
highest impact resistance and other beneficial
properties, such as elasticity, malleability and
fatigue resistance. Buehler named this alloy
Spring 2008
Nitinol, combining the chemical symbols of
nickel, Ni, and titanium, Ti, with the laboratory’s
acronym, NOL.
The first hint of the unusual properties of
Nitinol was seen in 1959 when Buehler
discovered the alloy’s exceptional temperature-
dependent acoustic-damping characteristics,
which suggested temperature-dependent changes
in the alloy’s atomic structure. But the final step
toward the discovery of shape memory was made
in 1960 at a meeting of NOL management. They
were presented with a Nitinol specimen that
would demonstrate the alloy’s favorable fatigue-
resistance properties. The specimen was a long
Nitinol strip folded repeatedly to form a zigzag
profile. The directors bent and unbent the
specimen and were satisfied with its mechanical
characteristics. One of the managers decided to
check the alloy’s thermal properties using a
cigarette lighter. Amazingly, when the compressed
strip got hot, it stretched out longitudinally.
It took a few more years to understand the
mechanism of shape memory. One important
discovery was that Nitinol can exist as two
different temperature-dependent phases; shape
memory is possible because of phase transitions
between these phases. To fix the original shape,
or to “train” a specimen to “remember” this
shape, the Nitinol specimen must be annealed at
approximately 500°C [932°F] for an hour while it
is held in a fixed position. Heating gives rise to a
high-temperature, hard, inelastic phase called
austenite. Subsequent cooling, or quenching, of
33
 Austenite Twinned Martensite

Cooling

Untwinned Martensite
Heating Deformation

A
> Mechanism of shape-memory effect. On cooling, the high-temperature
austenite phase with a face-centered cubic lattice transforms into the
low-temperature martensite phase. Because of stresses experienced
during cooling, the martensite produced from austenite undergoes crystal
twinning: the formation of adjacent layers related by mirror symmetry.
B
Deformation removes twinning. Untwinned martensite has a tetragonal
crystal lattice. Heating the deformed untwinned martensite converts it
back to the austenite phase.

the specimen produces a low-temperature, elastic,
more deformable phase called martensite. If the
trained specimen is deformed and heated again,
the thermal motion causes the atoms to form the
austenite lattice, thus restoring the original
shape of the specimen (above). The annealing
and quenching temperatures, as well as other
properties, depend strongly on the alloy compo-
sition and additives used.
The above procedure describes the so-called
one-way shape-memory effect, in which the
material remembers a single shape. By appro-
priate training, some shape-memory materials can
remember two different shapes, one at lower
temperature and the other at higher temperature,
thus exhibiting a two-way shape-memory effect.
To date, shape memory has been observed in
dozens of two- and three-component metal
alloys, of which, along with Nitinol, copper-zinc-
aluminum [CuZnAl] and copper-aluminum-nickel
[CuAlNi] alloys are most widely used. Another
promising group of materials is shape-memory
polymers, which became commercially available
in the 1990s.4
The first commercial application of shape-
memory materials was CryoFit shrink-to-fit pipe
couplings developed in 1969 to join hydraulic
lines in F-14 fighter aircraft.5 The tubular
couplings are easily installed by positioning the
machined and liquid-nitrogen-cooled coupling on
the pipe ends to be joined and allowing it to
warm to ambient temperature. As the coupling
warms, it shrinks and crimps down on the pipes
to form a tight joint (right).6 Following this, the
use of shape-memory couplings was extended to
oil and gas pipelines, water pipes and other types
of pipes and tubes. A wide range of various shape-
memory fasteners, such as rings and clamps, was
also developed.7
Another important area of application of
shape-memory materials is medicine. The most
readily observable medical shape-memory device
is dental braces. Nitinol-based braces were first
used in patients in 1975 and patented in 1977.8
Traditional dental braces include a stainless-
steel wire, which is insufficiently springy and
requires frequent readjustments. In contrast, a
Nitinol wire not only is springier, but also
provides a constant load on the teeth, thus
requiring fewer or no readjustments. A Nitinol
wire is initially molded to obtain a correct shape;
then an orthodontist attaches it to the patient’s
teeth, bending it as necessary. Body heat
activates the Nitinol wire, restoring it to the
originally molded shape.
A similar procedure is used in shape-memory
orthopedic staples and plates, which accelerate
the healing of bone fractures. However, perhaps
the most important, truly vital medical
applications of shape memory are in cardio-
C
D
> Photograph of CryoFit shrink-to-fit coupling
(top) and the principle of its use (bottom). The
coupling is machined at ambient temperature
until its inner diameter is somewhat smaller than
the outer diameter of the pipes to be joined (A).
Then the coupling is cooled in liquid nitrogen and
mechanically expanded so that its inner diameter
is slightly larger than the outer diameter of the
pipes (B). The expanded coupling easily slips
over the pipe ends (C). The coupling is properly
positioned and allowed to warm to ambient
temperature. During warming, it shrinks back to
its original smaller size to form a tight joint (D).
(Photograph courtesy of Intrinsic Devices, Inc.,
reference 5. Drawings courtesy of ATI Wah
Chang, reference 6.)
vascular surgery.9 An example is the Simon
Nitinol Filter device, a Nitinol wire sieve that is
inserted into a blood vessel to trap clots traveling
in the bloodstream.10 The trapped clots gradually
dissolve and an embolism, or obstruction of the
blood vessel, is thus prevented. The Simon

34 Oilfield Review
Nitinol Filter sieve is inserted using a catheter Simon Nitinol Filter Being Deployed
while being in a cooled, deformed martensitic
state, and then it expands to full size when
warmed by body heat (right).

Electrically Activated Materials:

Smart as Paint
A large range of applications has been created
using piezoelectric smart materials. The piezo-
electric effect, also called the direct piezoelectric
effect, is the ability of certain materials—
minerals, ceramics and some polymers—to
produce an electric charge in response to an
applied mechanical stress. The converse effect Front View Side View
can also be seen, in which piezoelectric materials
are deformed in an applied electric field.
The direct piezoelectric effect was discovered
by the brothers Pierre and Jacques Curie in 1880.
They noticed that compressing a quartz crystal
plate cut at a certain orientation gave rise to
electric charges on plate faces opposite to the
compression direction: a positive charge on one
> Simon Nitinol Filter. The schematic (top) shows the deployment of the
face and a negative charge on the other.
device on a catheter. Also included are the front and side views in the
Expanding the crystal plate also generated deployed state (bottom). (Copyright Brazilian Journal of Medical and
electric charges, but if the charge on a face while Biological Research; used with permission, reference 9.)
compressing was negative, then the charge on
this face while expanding was positive, and vice
versa. The direct piezoelectric effect occurs if an
elastic deformation of a solid is accompanied by
an asymmetric distortion of the distribution of
positive and negative charges, dipoles or groups
of parallel dipoles (Weiss domains) in the
structure of the solid so that a total dipole
moment is induced; that is, the solid is polarized.
The converse piezoelectric effect takes place if
an applied electric field causes such a distortion
of the distribution of charges, dipoles or Weiss
domains that this leads to geometric distortions,
manifested as mechanical strains (right). No strain Compression Expansion

4. Lendlein A and Kelch S: “Shape-Memory Polymers,”

Angewandte Chemie International Edition 41, issue 2
(June 12, 2002): 2034–2057. Piezoelectricity Converse Piezoelectricity
5. “Use of Shape Memory Alloys in High Reliability
Fastening Applications,” http://www.intrinsicdevices.com/
history.html (accessed December 24, 2007).
6. Tuominen S and Wojcik C: “Unique Alloys for Aerospace
and Beyond,” Outlook 16, no. 2 (2nd Quarter 1995),
http://www.wahchang.com/pages/outlook/html/bkissues/
16_02.htm (accessed December 24, 2007).
7. Stöckel D: “The Shape Memory Effect: Phenomenon,
> Direct and converse piezoelectric effects. In the direct piezoelectric
Alloys, Applications,” Report (2000), NDC, Nitinol Devices effect, compressing and expanding a piezoelectric material sample
& Components, Inc., Fremont, California, USA, generate opposite electric charges on respective faces of the sample (top).
www.nitinol-europe.com/pdfs/smemory.pdf (accessed In the converse piezoelectric effect, application of voltage to a
December 24, 2007).
piezoelectric material sample causes deformation Δh (bottom right). This
8. Andreasen GF: “Method and System for Orthodontic
contrasts with the direct piezoelectric effect, in which deformation Δh
Moving of Teeth,” US Patent No. 4,037,324 (July 26, 1977).
produces voltage (bottom left).
9. Machado LG and Savi MA: “Medical Applications of
Shape Memory Alloys,” Brazilian Journal of Medical
and Biological Research 36, no. 6 (June 2003): 683–691,
www.scielo.br/pdf/bjmbr/v36n6/4720.pdf (accessed
December 19, 2007).
10. Duerig TW, Pelton AR and Stöckel D: “Superelastic
Nitinol for Medical Devices,” Medical Plastics and
Biomaterials 4, no. 2 (March 1997): 30–43.

Spring 2008 35
 Before Polarization
Ep
> Polarization effects. Dipoles in sintered ceramics are parallel only within
each domain, whereas the domains are polarized randomly (left). After
polarization in strong electric field Ep at elevated temperature, the domains
are approximately aligned and the substance becomes piezoelectric (right).
To date, piezoelectricity has been detected in
many types of materials. The Curie brothers
discovered piezoelectricity in naturally occurring
minerals, such as quartz, tourmaline, topaz and
Rochelle salt (sodium potassium tartrate
tetrahydrate, or KNaC4H4O6·4H2O). Of these, only
quartz is now used commercially. All other
practically important piezoelectric single
crystals—such as ammonium dihydrophosphate
[NH4H2PO4], gallium orthophosphate [GaPO4],
and lanthanum gallium complex oxides—are
grown artificially.
Although single-crystal piezoelectric materials
continue to be developed, the most extensively
used class of piezoelectric materials is now
polycrystalline piezoelectric ceramics, which
have much wider ranges of useful characteristics
and work under broader operating conditions.
Currently, the largest group of piezoelectric
ceramics is materials consisting of crystallites
with the perovskite structure.11 These are
complex metal oxides with the general formula
ABO3, where A and B are cations of different
sizes. The cation A includes elements such as Na,
K, Rb, Ca, Sr, Ba and Pb, and B includes Ti, Sn, Zr,
Nb, Ta and W. Sometimes A and B each may
represent two or more of these cations, provided
that the total stoichiometry is satisfied (for
example as in lead zirconate titanate,
PbZrxTi1−xO3). The main examples of perovskite-
like piezoelectric ceramics are barium titanate
[BaTiO3] (the first piezoelectric ceramic
discovered), lead titanate [PbTiO3], lead
zirconate titanate (the most widely used
piezoelectric ceramic to date), lead lanthanum
zirconate titanate [Pb1−xLax(ZryTi1−y)1−x/4O3]
and lead magnesium niobate [PbMg1/3Nb2/3O3].12
36
After Polarization
After the sintering stage in manufacturing,
dipoles in such ceramics are parallel only within
each domain, whereas the domains are polarized
randomly.13 An elastic deformation of a set of
randomly polarized dipoles cannot lead to an
asymmetric distortion of the charge distribution
and, therefore, cannot result in piezoelectricity.
Therefore, the last stage of manufacturing of
piezoelectric ceramics is always the application
of a strong electric field at elevated temperature,
after which the domains are polarized approxi-
mately identically and the substance becomes
piezoelectric (above).
Some polymers can be piezoelectric or can be
made so. Piezoelectricity was discovered or
developed in a number of natural polymers,
including keratin, collagen, some polypeptides and
oriented films of DNA, and synthetic polymers,
such as some nylons and polyurea. However,
currently the only commercially available
piezoelectric polymers are polyvinylidene
difluoride (PVDF) and its copolymers with
trifluoroethylene and tetrafluoroethylene.14 PVDF
is a semicrystalline synthetic polymer with the
chemical formula (CH2–CF2)n. PVDF is produced
in thin films, which are stretched along the film
plane and polarized perpendicular to this plane to
produce piezoelectric properties (right).
Because piezoelectric materials can convert
mechanical energy to electric energy and vice
versa, their applications are dominated by
various electromechanical sensors and actuators.
The piezoelectric effect is used in sensors for
various physical quantities (such as force,
pressure, acceleration, side impact and yaw rate),
and in microphones, hydrophones, ultrasonic
sensors, seismic sensors, acoustic pickups and
many other devices.
An interesting example of a continuously
distributed piezoelectric sensor is piezoelectric,
or smart, paints.15 Such paint can be prepared
using lead zirconate titanate ceramic powder as
a pigment with epoxy resin as a binder. The
mixture is coated on a surface and cured and
polarized at room temperature. The resulting
paint film acts as a vibration and acoustic-
emission sensor for the entire surface. These
smart paints can be used to cover large surface
areas of individual structural elements and even
entire constructions, such as bridges, to monitor
their integrity. Recent controlled weathering
trials on river-crossing bridges in the UK and
Finland have shown that the piezoelectric-paint
sensors can survive harsh outdoor conditions and
remain functional for at least six years.16
Melt-Cast
Crystalline
region
Amorphous
region
Mechanically Oriented
Stretch
direction
Electrically Polarized
Electric field direction
> Polyvinylidene difluoride (PVDF) treatment to
impart piezoelectric properties. In a melt-cast
polymer film, crystallites (tens to hundreds of
nanometers in size) are randomly distributed
among amorphous regions (top). Stretching the
polymer film (middle) significantly aligns polymer
chains in the amorphous regions in the sheet
plane and facilitates uniform rotation of the
crystallites by an electric field. Polarization
through the film thickness (such as by using
deposited metal electrodes) makes the film
piezoelectric (bottom). (Figure courtesy of NASA;
used with permission, reference 14.)
Oilfield Review
 Other important examples of piezoelectric Magnetorheological Liquid
actuators are loudspeakers, piezoelectric motors
and high-precision microactuators. High- Carrier fluid
precision microactuators use the fact that small
changes in voltage applied to piezoelectric Ferrous particles
materials cause small changes in their shape.
This allows fine control of positions and displace-
ments of parts and elements, which is critical in
the operation of a variety of devices from inkjet
printheads to guidance systems.

Magnetic field direction

The most significant class of piezoelectric
devices is piezoelectric ultrasonic generators,
which, unlike magnetostrictive or other types of
ultrasonic generators, provide the most efficient
generation of ultrasound with controlled power
and frequency. Ultrasound in such generators is
produced using the converse piezoelectric effect.
Here, a cyclic application of voltage to a
piezoelectric material causes it to expand and
contract, thus emitting a pressure wave.
The creation of piezoelectric ultrasonic
generators has opened the way for an extremely
wide array of applications. The first practical
application of piezoelectricity was a piezo-
electric quartz ultrasonic generator in an active
sonar designed to detect submarines during
World War I in 1915.17 Since then, this appli-
cation has grown to an extensive collection of
methods for detecting inhomogeneities in
various media. Flaw detection using ultrasonic
technology tests a broad range of materials and
constructions, including various pipes and pipe-
lines. For the general public, the most common
application is medical ultrasonography, a
technique for visualizing internal tissues and
organs of the body, especially obstetric ultra-
sonography for visualizing an embryo or fetus in
utero, which has become a standard procedure of
prenatal care in many countries.
Piezoelectric devices are also found in
numerous oilfield applications. A quartz
piezoelectric element is an important part of the
Schlumberger CQG Crystal Quartz Gauge, which
is used as a pressure sensor in a wide variety of
tools. Piezoelectric ceramic devices also play a
key role in Schlumberger seismic, sonic and
ultrasonic logging instrumentation: as pingers
Chains of particles
aligned with the field
> Applying a magnetic field to magnetorheological (MR) fluids. Without a
magnetic field, ferrous particles are randomly distributed in a nonmagnetic
oil to form an MR fluid (top). Once a magnetic field is applied, the particles
align with the magnetic field to form chains, dramatically increasing the
viscosity in the direction perpendicular to the field direction (bottom).
and hydrophones in the Q-Marine single-sensor A typical MR fluid is a suspension of micron-
marine seismic system and Q-Seabed multi- sized (usually 3 to 8 microns) magnetically
component seabed seismic system, and as susceptible particles (generally 20 to 40% by
receivers and monopole transmitters in the Sonic volume of pure iron particles) in a carrier fluid,
Scanner acoustic scanning platform, Isolation such as mineral oil, synthetic oil, water or
Scanner cement evaluation service and glycol.18 Various surfactants, including oleic and
sonicVISION sonic-while-drilling tool. Although citric acids, tetramethylammonium hydroxide
current applications are limited to sensors, and soy lecithin, are also added to MR fluids to
future oilfield applications might utilize the prevent particles from settling. MR materials
piezoelectric effect for energy harvesting and systems may be manufactured as gels, foams,
microactuators. powders, greases, and even solid elastomers.
Without a magnetic field, particles in an MR
Magnetically Activated Materials: Fast fluid are randomly distributed. Once a magnetic
Strength of Minute Particles field is applied, the particles align with the
Another category of smart materials is magnetic field to form chains, which resist
magnetorheological (MR) fluids. These fluids flow or shear deformation in the direction
have rheological properties that may be varied by perpendicular to the magnetic field direction
applying a magnetic field. The change is propor- and dramatically increase the viscosity (or more
tional to the magnetic field intensity, can be accurately, yield strength) in this direction
controlled very accurately by varying this (above). As soon as the magnetic field is
intensity, and is immediately reversible after removed, the chains of particles disintegrate
removing the field. (through random Brownian forces) and the
initial viscosity is restored.

11. Perovskite (named after Lev A. Perovski, a Russian
mineralogist) is a natural calcium titanate [CaTiO3] with
a pseudocubic lattice. This class of solids includes
many technologically important ceramics, such as
semiconductors and magnetic, ferroelectric and
piezoelectric materials.
12. Kholkin A, Jadidian B and Safari A: “Ceramics,
Piezoelectric and Electrostrictive,” in Schwartz MM
(ed): Encyclopedia of Smart Materials. New York City:
John Wiley & Sons (2002): 139–148.
13. Sintering is a method for forming objects from a
granular material by heating the material close to its
melting point until its particles adhere to one other.
14. Harrison JS and Ounaies Z: “Piezoelectric Polymers,”
ICASE Report No. 2001-43, NASA/CR-2001-211422,
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/
20020044745_2002075689.pdf (accessed December 24,
2007). (Color added to original figure.)
15. Egusa S and Iwasawa N: “Piezoelectric Paints as One
Approach to Smart Structural Materials with Health-
Monitoring Capabilities,” Smart Materials and
Structures 7, no. 4 (August 1998): 438–445.
Egusa S and Iwasawa N: “Piezoelectric Paints:
Preparation and Application as Built-In Vibration
Sensors of Structural Materials,” Journal of Materials
Science 28, no. 6 (March 1993): 1667–1672.
16. Hale JM and Lahtinen R: “Piezoelectric Paint: Effect of
Harsh Weathering on Aging,” Plastics, Rubber and
Composites 36, no. 9 (November 2007): 419–422.
17. Sonar, an acronym for sound navigation and ranging,
is a technique that uses underwater sound waves to
detect and locate submerged objects. Active sonars
produce a pulse of sound and then listen for reflections
of the pulse. Passive sonar equipment only listens for
underwater sounds without transmitting.
18. Henrie AJM and Carlson JD: “Magnetorheological Fluids,”
in Schwartz MM (ed): Encyclopedia of Smart Materials.
New York City: John Wiley & Sons (2002): 597–600.

Spring 2008 37
 Prosthetic Leg Magnetorheological Damping System

Wires to
electromagnet

Piston

Bearing
and seal

MR fluid
MR damping system

Control devices Coil Annular flow

channel

Diaphragm

Accumulator

> A schematic of the Motion Master MR fluid damper (LORD Corporation) in the artificial knee of the
Smart Magnetix prosthetic leg (Biedermann Motech) (left) and a schematic of the MR fluid damper
(right). (Used with permission from LORD Corporation, reference 22.)

MR fluids were discovered in the 1940s and
early 1950s at the National Bureau of Standards,
Gaithersburg, Maryland.19 A number of devices
were developed based on dry magnetic powders,
such as a magnetic-powder brake. However,
these early MR fluids and devices had limited life
and stability, and it was not until the early 1990s
that progress in materials science and control
electronics renewed interest in these materials.
MR fluids attract interest because of their
unique ability to undergo a rapid, abrupt
increase in viscosity, corresponding to an almost
instantaneous transition to a semisolid state, in
response to application of a magnetic field. The
restoration to the initial viscosity after removing
the magnetic field is equally rapid with response
times as short as 6.5 ms.20 Therefore, MR fluids
have mostly been used in various damping
systems. MR fluids were first commercialized in
1995 in fluid rotary brakes for aerobic exercise
equipment. Other commercially available
products using MR fluids are dampers for real-
time vibration-control systems in heavy-duty
trucks, adjustable shock absorbers for oval-
and dirt-track automobile racing, and linear
dampers for real-time gait control in advanced
prosthetic devices.21
An example of this last application is the
Motion Master MR fluid damper in the Smart
Magnetix prosthetic leg (above).22 The MR fluid
damper in the prosthetic responds 20 times
faster than prior state-of-the-art mechanical or
hydraulic designs. The total response time,
40 ms, is similar to the response time for signals
in the human knee.23 This improvement helps the
new prosthetic more closely mimic natural loco-
motion and makes it more convenient for the user.
Another class of magnetically activated
materials is magnetostrictive substances.
Magnetostriction is the property of ferromag-
netic materials to change their shape in
response to application of a magnetic field.24
Magnetostriction was discovered in 1842 by
James P. Joule, who noticed that the length of a
sample of iron changed after a magnetic field
was applied. Along with this effect, which is also
referred to as the Joule effect, there is a
reciprocal effect, called the Villari effect, in
which applying a stress to a material causes a
change in its magnetization.
This behavior resembles both the direct and
converse piezoelectric effects. In fact, the
macroscopic mechanisms of piezoelectricity and
magnetostriction resemble each other, with the
difference that piezoelectric effects are
determined by the action of an electric field on
charges, electric dipoles or domains of electric
dipoles, whereas the magnetostrictive effects are
controlled by the action of a magnetic field on
magnetic domains—regions of uniform magneti-
zation. A magnetic field applied to a
ferromagnetic specimen shifts magnetic
domains, causing macroscopically detectable
changes in the shape and size of the specimen.
And conversely, an applied stress causes a
mechanical shift of magnetic domains, thereby
altering the magnetization of the specimen.

38 Oilfield Review
 The direct magnetostrictive (Joule) effect is
used in magnetostrictive actuators, while the
Villari effect is used in magnetostrictive sensors.
Applications of magnetostriction include tele-
phone receivers, hydrophones, magnetostrictive
ultrasonic generators for sonars, linear and
rotational motors, and various sensors for
deformation, motion, position and force.

Chemically Activated Materials:

How Swell They Are
Chemical activation of materials is an almost
Antiextrusion caps
endless topic. Here, we address only the
chemical activation of polymers under exposure
to liquids. This phenomenon is general enough to
occur in everyday life and also specific enough to
underlie smart applications, including some in
the oil field. Most people have observed both
intentional and unintentional swelling of Swellable packer
polymers in ordinary life. For example, spilling
coffee or tea onto a book lets the natural polymer
cellulose contained in the book paper swell, and
preparing gelatin desserts makes use of the
property of the polymer gelatin to swell in water.
However, such swelling behavior can also be Swelling
deleterious: industrial companies may incur
great losses if even a small gasket made of a
polymer unsuitable for the existing operating
conditions swells and degrades, causing leakage
or other dangerous consequences. > Photograph (top), schematic (middle) and illustration of swelling (bottom) of a swellable packer.
On the other hand, people have long found
ways to use polymer swelling in a controllable
way, as in food processing, medicine (absorbent
materials), chemical spill kits and construction
(various fillers). An example of the modern
application of controllable polymer swelling in each other. For water control, an unswollen Swellable packers are advantageous com-
medicine is targeted drug-delivery systems.25 The water-sensitive polymer (elastomer or composite) pared with conventional ones as they are
simplest form of such a system is a capsule with packer is installed in the well. If water generally less expensive, contain no moving
a drug-containing core and a swellable coating. encroaches into the wellbore, the packer swells parts, and require no mechanical or hydraulic
The properties of the coating are designed so and seals the wellbore at that location, isolating actuation mechanism. All the functions of these
that the coating gradually swells and the drug is the interval so that water influx decreases and oil elements are performed by a single piece of
released at given rates at given places as the production increases.26 polymeric smart material.
capsule is transported through the gastro-
intestinal tract. More intricate designs include 19. Rabinow J: “Magnetic Fluid Torque and Force 22. http://www.lord.com/Home/MagnetoRheologicalMRFluid/
Transmitting Device,” US Patent No. 2,575,360 Applications/OtherMRApplicationSolutions/Medical/
multilayer and multidrug capsules, sometimes (November 20, 1951). tabid/3791/Default.aspx (accessed January 5, 2008).
provided with special drug-delivery ports. Rabinow J: “The Magnetic Fluid Clutch,” Transactions 23. Bullough WA: “Fluid Machines,” in Schwartz MM (ed):
of the American Institute of Electrical Engineers 67 Encyclopedia of Smart Materials. New York City: John
Swellable polymers are starting to be (1948): 1308–1315. Wiley & Sons (2002): 448–456.
employed in oilfield applications. They are used 20. Weiss KD, Duclos TG, Carlson JD, Chrzan MJ and 24. Not only can a ferromagnetic material be magnetized in
in swellable packers for zonal isolation and Margida AJ: “High Strength Magneto- and Electro- an external magnetic field, but it remains magnetized after
Rheological Fluids,” Society of Automotive Engineers removing the field. Examples of ferromagnetic materials
efficient borehole water control (above right). Technical Paper Series, no. 932451, Warendale, are iron, nickel, cobalt, some rare-earth elements and
For zonal isolation, a series of unswollen oil- Pennsylvania, USA (1993): 1–6. some alloys and compounds of these elements.
21. Carlson JD and Sproston JL: “Controllable Fluids in 25. Wise DL (ed): Handbook of Pharmaceutical Controlled
sensitive packers is run into the well. When they 2000—Status of ER and MR Fluid Technology,” paper Release Technology. New York City: Marcel Dekker, 2002.
are exposed to oil, they swell and seal off the presented at the Actuator 2000—7th International 26. http://www.tamintl.com/pdf/FreeCapAd1JPT.pdf
Conference on New Actuators, Bremen, Germany, (accessed January 11, 2008).
formation face, creating intervals isolated from June 19–21, 2000.

Spring 2008 39

Catalyst
Microcapsule
Crack
Healing agent
Polymerized healing agent
> Self-healing material, in which 200-micron microcapsules containing a
polymerizable healing agent and polymerization catalyst particles are
embedded. Damage causes crack propagation (top); the crack ruptures
microcapsules, releasing a healing agent (middle); the healing agent
contacts the catalyst, polymerizes and heals the damaged area (bottom).
(Adapted with permission from Macmillan Publishers Ltd., reference 29.)
An early success story of swellable polymers
occurred during World War II, when swellable
rubber materials were used in self-sealing fuel
tanks in aircraft.27 A self-sealing tank was made
of two layers of rubber; the outer layer was
composed of cured rubber and the inner one was
of oil-swellable uncured rubber. The inner layer
was lined with a fuel-impervious material to
prevent the uncured rubber from contacting the
fuel while the tank was intact. If a bullet or other
projectile punctured the tank, the fuel spilled
and contacted the uncured rubber, which
27. Gustin E: “Fighter Armour,” http://www.geocities.com/
CapeCanaveral/Hangar/8217/fgun/fgun-ar.html
(accessed February 28, 2008).
28. Shah AD and Baghdachi J: “Development and
Characterization of Self-Healing Coating Systems,”
http://www.emich.edu/public/coatings_research/
AmitPresentation.pdf (accessed January 14, 2008).
29. White SR, Sottos NR, Geubelle PH, Moore JS,
Kessler MR, Sriram SR, Brown EN and Viswanathan S:
“Autonomic Healing of Polymer Composites,” Nature 409
(February 15, 2001): 794–797.
40
swelled and thus sealed the puncture. Such self-
sealing fuel tanks are still being used.
These self-sealing rubber materials may be
considered predecessors of the modern concept
of self-healing materials. In the latter, a healing
agent does not form an adjacent layer but is
enclosed in microcontainers, such as micro-
capsules or hollow microfibers, and uniformly
distributed throughout the material to be
protected.28 In self-healing polymer materials,
the healing agent is typically the corresponding
uncured polymer. If a self-healing polymer
30. “Self-Healing Composite Materials,” http://www.aer.bris.
ac.uk/research/fibres/sr.html (accessed January 14, 2008).
31. Moroni N, Panciera N, Zanchi A, Johnson CR,
LeRoy-Delage S, Bulte-Loyer H, Cantini S, Belleggia E
and Illuminati R: “Overcoming the Weak Link in Cemented
Hydraulic Isolation,” paper SPE 110523, presented at the
SPE Annual Technical Conference and Exhibition,
Anaheim, California, November 11–14, 2007.
material is damaged, these microcontainers
rupture and release the healing agent, which
infiltrates into the damage site, polymerizes—if
necessary, a polymerization catalyst is also added
to the bulk of the material—and thus heals the
damaged area (left).29 This procedure mimics the
self-healing functions of biological tissues, whose
response to damage is often secretion of healing
fluids. To follow nature further, some proposals
suggest piercing a material with a vascular
network that can carry a circulating healing
agent throughout the material.30
Self-healing materials are also beginning to
make their mark in oilfield applications. For
example, Schlumberger recently announced the
availability of its FUTUR active set-cement
technology that automatically seals microleaks
in a cement sheath (see “Ensuring Zonal
Isolation Beyond the Life of the Well,” page 18).
The FUTUR cement system, pumped and placed
in the same way as any ordinary cement, contains
components that remain dormant until exposed
to hydrocarbons, such as those seeping though
microcracks in the cement sheath. The contact
with hydrocarbons activates the FUTUR cement
sheath, which self-repairs within hours without
intervention. This prevents many undesirable
events after the cement has set, such as annular
migration of fluids behind the casing between
zones, sustained casing pressure at surface,
surface casing leaks and crossflows.31
Toward Novel Smart Materials
These examples of materials and processes are
only a small sampling of the world of smart
materials and their applications. Smart
materials abound and can be encountered in a
variety of devices from simple piezoelectric
lighters and igniters to complex ultrasonic
instrumentation.
Even ordinary materials can be made smart
or responsive. Self-healing cement is an example
of an abundant, everyday material that has been
engineered to take on smart properties for
oilfield application. Promising candidates for
smart-material adaptation may be all around us,
waiting to be discovered.
Investigation into smart materials is one of
the new research directions at Schlumberger-Doll
Research Center in Cambridge, Massachusetts.
This includes defining and executing a road-
map for actuation technology in various
oilfield applications.
Oilfield Review
 A significant part of implementing the
roadmap involves defining and developing
common technological building blocks that can
be integrated in various ways to provide
actuation applications. This will be accomplished
by studying actuation systems—actuators,
sensors, system dynamics and control, and novel
mechanisms—and applying smart materials to
invent new actuating systems (right).
While materials-science researchers are
excited about the enormous potential of smart
materials, these new materials are unlikely to
supplant the standard materials we use every day.
The vast majority of materials are structural—
selected not only for their properties, but because
they are cheap and abundant. Smart materials,
like other functional materials, including

> A researcher (above) studying the thermo-

mechanical properties of a specimen of a
material at Schlumberger-Doll Research (SDR)
Center, Cambridge, Massachusetts (below left).

tungsten filaments in light bulbs, platinum-

rhodium wire in thermocouples and diamond tips
in drill bits, typically have small-volume
applications. These require unique properties for
which there are few or no substitutes, and thus
cost is less of an issue. For sophisticated oilfield
tools, smart materials may allow implementation
of new technologies, miniaturization of parts and
enhanced reliability in the increasingly harsh
downhole environment. —VG

Spring 2008 41
 Seismic Inversion:
Reading Between the Lines

Frazer Barclay The reflections of seismic waves from subsurface layers illuminate potential
Perth, Western Australia, Australia
hydrocarbon accumulations. As waves reflect, their amplitudes change to reveal
Anders Bruun important information about the underlying materials. Seismic amplitude inversion
Klaus Bolding Rasmussen
Copenhagen, Denmark uses reflection amplitudes, calibrated with well data, to extract details that can be

Jose Camara Alfaro correlated with porosity, lithology, fluid saturation and geomechanical parameters.
Pemex
Tampico, Tamaulipas, Mexico

Anthony Cooke
Aberdeen, Scotland The undisputed leader among tools for characterization, can help fill gaps in our
identifying potential exploration targets is the knowledge of formation properties between wells.
Dennis Cooke 3D seismic survey. These surveys probe great This article describes the science and art of
Darren Salter volumes of the subsurface, helping oil and gas seismic inversion, and how oil and gas companies
Santos companies map geological structures and select are using it to reduce risk in their exploration,
Perth, Western Australia drilling locations. development and production operations. After an
The original use of seismic data, and still the introduction to the uses of inversion, we present
Robert Godfrey
main use today, has been to identify the geometry its various types, from simple to more complex.
Dominic Lowden
Steve McHugo of reflectors and ascertain their depths. This is Examples from Mexico, Egypt, Australia and the
Hüseyin Özdemir possible because seismic waves reflect at North Sea demonstrate applications of inversion
Stephen Pickering interfaces between materials of different to fine-tune drilling locations, characterize hard-
Gatwick, England acoustic properties. to-image reservoirs, map water saturation,
However, seismic reflection data contain improve reservoir simulations and enhance
Francisco Gonzalez Pineda information beyond reflector location: every knowledge of geomechanical properties.
Pemex reflection changes the amplitude of the returned
Reynosa, Tamaulipas, Mexico wave. The controlling property in this change at Inversion Basics
the interface is the contrast in impedance, which Many measurements in the E&P industry rely to
Jorg Herwanger
is the product of density and velocity. Seismic some extent on inversion for their interpretation.
Stefano Volterrani
reflection amplitude information can be used to The reason is simple. For several measurement-
Houston, Texas, USA
back out, or invert for, the relative impedances of interpretation problems, no equation that directly
Andrea Murineddu the materials on both sides of the interface. By relates the multiple measurements—which
Andreas Rasmussen correlating these seismically derived properties include noise, losses and other inaccuracies—can
Stavanger, Norway with values measured in the borehole, inter- be solved with a unique answer. We then resort to
preters may be able to extend well information inversion, which is a mathematical way of
Ron Roberts throughout the entire seismic volume. This estimating an answer, checking it against
Apache Corporation process, called seismic inversion for reservoir observations and modifying it until the answer
Calgary, Alberta, Canada is acceptable.

For help in preparation of this article, thanks to Trine Alsos,
StatoilHydro, Harstad, Norway; Ted Bakamjian, SEG, Tulsa;
Richard Bottomley, Mexico City; Jonathan Bown, Henrik Juhl
Hansen and Kim Gunn Maver, Copenhagen; Tim Bunting,
Kuala Lumpur; Karen Sullivan Glaser, Houston; Jalal
Khazanehdari, Abu Dhabi, UAE; Hasbi Lubis, Gatwick,
England; Farid Mohamed, Aberdeen; Richard Patenall, Perth,
Western Australia; Pramesh Tyagi, Cairo; and Anke Simone
Wendt, Stavanger.
ECLIPSE, ISIS and Q-Marine are marks of Schlumberger.
1. Quirein J, Kimminau S, La Vigne J, Singer J and Wendel F:
“A Coherent Framework for Developing and Applying
Multiple Formation Evaluation Models,” Transactions of
the SPWLA 27th Annual Logging Symposium, Houston,
June 9–13, 1986, paper DD.
Jammes L, Faivre O, Legendre E, Rothnemer P,
Trouiller J-C, Galli M-T, Gonfalini M and Gossenberg P:
“Improved Saturation Determination in Thin-Bed
Environments Using 2D Parametric Inversion,” paper
SPE 62907, presented at the SPE Annual Technical
Conference and Exhibition, Dallas, October 1–4, 2000.
Faivre O, Barber T, Jammes L and Vuhoang D: “Using
Array Induction and Array Laterolog Data to Characterize
Resistivity Anisotropy in Vertical Wells,” Transactions of
the SPWLA 43rd Annual Logging Symposium, Oiso,
Japan, June 4–7, 2002, paper M.
Marsala AF, Al-Ruwaili S, Ma SM, Modiu SL, Ali Z,
Donadille J-M and Wilt M: “Crosswell Electromagnetic
Tomography in Haradh Field: Modeling to Measurements,”
paper SPE 110528, presented at the SPE Annual Technical
Conference and Exhibition, Anaheim, California, USA,
November 11–14, 2007.

42 Oilfield Review
 Inversion, as the name implies, can be
considered as the inverse of forward modeling,
sometimes simply called modeling. For the
purpose of this article, forward modeling begins
with a model of earth properties, then mathe-
matically simulates a physical experiment or
process—for example, electromagnetic, acoustic,
nuclear, chemical or optical—on the earth
model, and finally outputs a modeled response. If
the model and the assumptions are accurate, the
modeled response looks like real data. Inversion
does the reverse: it starts with actual measured
data, applies an operation that steps backward
through the physical experiment, and delivers an
earth model. If the inversion is done properly, the
earth model looks like the real earth.
Inversion is used by many E&P disciplines
and can be applied at a wide range of scales and
varying levels of complexity:
• calculating borehole-fluid invasion profiles from
induction logging measurements
• assessing cement-bond quality from ultrasonic
logs (see “Ensuring Zonal Isolation Beyond the
Life of the Well,” page 18).
• extracting layer lithologies and fluid satura-
tions from multiple log measurements
• interpreting gas, oil and water volumes from
production logs
• inferring reservoir permeability and extent
from pressure-transient data (see “Intelligent
Well Technology in Underground Gas Storage,”
page 4).
• mapping fluid fronts from crosswell electro-
magnetic measurements
• integrating electromagnetic and seismic
measurements for improved delineation of
subsalt sediments.1
E&P seismic specialists use different types of
inversion—velocity inversion and amplitude
inversion—to solve distinct types of problems.
The first type of inversion, velocity inversion,
sometimes known as traveltime inversion, is used
for depth imaging. Using seismic traces at widely
spaced locations, it generates a velocity-depth
earth model that fits recorded arrival times of
seismic waves. The result is a relatively coarse
velocity-depth model extending over several
kilometers in depth and perhaps hundreds of
kilometers in length and width. This solution is
applied in data-processing steps such as
migration and stacking, eventually producing
the type of seismic image that is familiar to
most readers. Seismic interpreters use these
images to determine the shape and depth of
subsurface reflectors.

Spring 2008 43
 Forward Modeling
Earth model Input Synthetic the required interval. If no borehole velocity data
of acoustic wavelet seismic are available, velocities from traveltime
impedance trace
inversion may serve as a substitute. Missing
800 density data may be estimated from empirical
Time, ms relationships. For nonvertical-incidence data,
the model must include both S-wave and
Reservoir
850
P-wave velocities.
For conventional inversion of vertical-
incidence data, the density-velocity model is then
converted to a reflectivity model. Reflectivity, the
900 ratio of the amplitude of the reflected wave to
Inversion
that of the incident wave, is the parameter that
governs reflection-driven changes in normal-
Recorded Estimated Earth model
seismic wavelet of acoustic incidence seismic amplitudes. It relates to the
trace impedance densities and velocities on each side of an
800 interface through the acoustic impedance con-
trast; reflectivity is the ratio of the difference in
acoustic impedances to their sum.3 The resulting
Time, ms

depth-based reflectivity model is converted to a

Reservoir
850
time-based model through the velocities.
Combining the time-based model with a
seismic pulse creates a synthetic trace. Mathe-
900 matically, this process is known as convolution.4
The seismic pulse, or wavelet, represents the
packet of energy arriving from a seismic source.
> Modeling and inversion. Forward modeling (top) takes a model of
A model wavelet is selected to match the
formation properties—in this case acoustic impedance developed from
well logs—combines this with a seismic wavelet, or pulse, and outputs a amplitude, phase and frequency characteristics
synthetic seismic trace. Conversely, inversion (bottom) begins with a of the processed seismic data. Convolution of the
recorded seismic data trace and removes the effect of an estimated wavelet with the reflectivity model yields a
wavelet, creating values of acoustic impedance at every time sample. synthetic seismic trace that represents the
response of the earth as modeled to the input
The second type of inversion, amplitude Models that include shear (S-wave) velocities seismic pulse. Additional steps are required if
inversion, is the focus of this article. This (Vs) can solve for S-wave, or elastic, impedance. noise, attenuation and multiple reflections are to
approach uses the arrival time and the amplitude The simple model is combined with a seismic be included in the modeled trace.
of reflected seismic waves at every reflection pulse to create a modeled seismic trace called The inverse operation starts with an actual
point to solve for the relative impedances of a synthetic (above). Inversion takes an actual seismic trace. Because the amplitude and shape
formations bounded by the imaged reflectors. seismic trace, removes the seismic pulse, and of each swing in the seismic trace affect the
This inversion, called seismic inversion for delivers an earth model for that trace location. outcome, it is vital that the processing steps up to
reservoir characterization, reads between the To arrive at the best-fit model, most inversion this point conserve signal phase and amplitude.
lines, or between reflecting interfaces, to routines iterate between forward modeling and Different types of inversion start with
produce detailed models of rock properties. For inversion, seeking to minimize the difference different types of traces. The main distinction is
simplicity, the following discussion describes between the synthetic trace and the data. between inversion performed before stacking
only model-based inversion. Some other In practice, each of these steps may be quite and inversion performed after it—prestack and
alternatives are space-adaptive and discrete involved and can depend on the type of seismic poststack. Most seismic surveys provide images
spike inversions.2 data being inverted. For vertical-incidence data, using data that have been stacked. Stacking is a
In principle, the first step in model-based creating the initial model requires bulk density signal-enhancement technique that averages
seismic inversion—forward modeling—begins measurements from density logs and compres- many seismic traces. The traces represent
with a model of layers with estimated formation sional velocities from sonic logs, both spanning recordings from a collection of different source-
depths, thicknesses, densities and velocities the interval to be inverted. Unfortunately, the receiver offsets with a common reflecting
derived from well logs. The simplest model, necessary logs often are acquired only in the midpoint (next page, top left). Each trace is
which involves only compressional (P-wave) reservoir. In the absence of sonic logs, borehole assumed to contain the same signal but different
velocities (Vp) and density ( ρ ), can be used to seismic surveys—vertical seismic profiles random noise. Stacking produces a single trace
invert for P-wave, or acoustic, impedance. (VSPs)—can provide average velocities across with minimal random noise and with signal
amplitude equal to the average of the signal in
2. “Space-Adaptive Inversion,” http://www.slb.com/content/ polarity as the incident wave. Negative reflectivity the stacked traces. The resulting stacked trace is
services/seismic/reservoir/inversion/space_adaptive.asp means the reflected wave has the opposite polarity
(accessed April 22, 2008). relative to the incident wave. taken to be the response of a normal-incidence
3. Reflectivity may be positive or negative. Positive 4. Yilmaz O and Doherty SM: Seismic Data Processing. reflection at the common midpoint (CMP).
reflectivity means the reflected wave has the same Tulsa: Society of Exploration Geophysicists, 1987.

44 Oilfield Review
 Relative Acoustic Impedance Absolute Acoustic Impedance
3,200

AI
LFM
H6 H5 H4 H3 H2 H1
S1 S2 S3 S4 S5 S6 3,400

R1 R2 R3 R4 R5 R6
3,600
R2 R3 R4 R5 R6 R7
R3 R4 R5 R6 R7 R8
R4 R5 R6 R7 R8 R9
R5 R6 R7 R8 R9 R10 3,800
R6 R7 R8 R9 R10 R11

Common midpoint (CMP) gather 4,000

Time, ms
H6 H5 H4 H3 H2 H1 S1 S2 S3 S4 S5 S6

4,200

CMP
Stacked
Trace Trace trace 4,400
1 2 3 4 5 6 1 2 3 4 5 6 1
Two-way traveltime

Before NMO correction After NMO correction

4,600

Offset Offset 4,800

> Stacking basics. Stacking enhances signal and reduces noise by
adding several traces together. The seismic vessel acquires traces at
many offsets from every source (top). S numbers represent sources, R
5,000
numbers represent reflection points, and H numbers represent
hydrophones. Stacking first gathers traces from all available source- > Relative and absolute acoustic impedance. Inversion of seismic
receiver offsets that reflect at a common midpoint (CMP) (middle). amplitudes yields relative acoustic impedance (AI) (left). However, the
Because arrivals from longer offsets have traveled farther, a time true absolute acoustic impedance (blue) contains a low-frequency
correction, called normal moveout (NMO) correction, is applied to each model (LFM) (red) that must be obtained from borehole data or modeled
gather to flatten the arrivals (bottom left). The flattened traces are by other means (right).
averaged (bottom right) to produce one stacked trace that represents the
normal-incidence (zero-offset) trace.

Stacking is a reasonable processing step if
certain assumptions hold: the velocity of the
medium overlying the reflector may vary only
gradually, and the average of the amplitudes in
the stacked traces must be equivalent to the
amplitude that would be recorded in a normal-
incidence trace. In many cases, these assumptions
are valid, and inversion may be performed on the
stacked data—in other words, poststack. In
contrast, when amplitude varies strongly with
offset, these assumptions do not hold, and
inversion is applied to unstacked traces—
prestack. Before discussing prestack situations in
detail, we continue with the simpler case of
poststack inversion.
A stacked trace is compared with the
synthetic trace computed from the reflectivity
model and wavelet. The differences between the
two traces are used to modify the reflectivity
model so that the next iteration of the synthetic
trace more closely resembles the stacked trace.
This process continues, repeating the generation
of a synthetic trace, comparison with the
stacked trace, and modification of the model
until the fit between the synthetic and stacked
traces is optimized.
There are many ways to construct synthetic
traces, and various methods may be used to
determine the best fit. A common approach for
determining fit is least-squares inversion, which
minimizes the sum of the squares of the
differences at every time sample. This inversion
technique operates on a trace-by-trace basis,
while others attempt to globally optimize the
inversion of the seismic volume. We discuss
global optimization later.
In the simplest case, inversion produces a
model of relative reflectivity at every time
sample, which can be inverted to yield relative
acoustic impedance. To obtain formation
properties such as velocity and density, a
conversion to absolute acoustic impedance is
necessary. However, such a conversion requires
frequencies down to near 0 Hz, lower than
contained in conventional seismic data. An
absolute acoustic impedance model can be
constructed by combining the relative acoustic
impedance model obtained from the seismic
frequency range with a low-frequency model
derived from borehole data (above right).

Spring 2008 45
 Relating seismically derived acoustic imped-
2.00
ances to formation properties makes use of
correlations between logging measurements. For
2.05
example, crossplotting acoustic impedance and
porosity measured in nearby wells establishes a
transform that allows seismically measured
2.10
acoustic impedance to be converted to porosity
values throughout the seismic volume. An
2.15
example from a carbonate reservoir in Mexico
demonstrates the power of this technique.
2.20
Two-way time, s

Inversion for Porosity in Mexico

2.25 Following the 2003 discovery of the Lobina
field offshore Mexico, Pemex contracted with
2.30 WesternGeco to obtain a seismic survey with
better resolution than one acquired in 1996.
2.35 Seismic data with increased frequency content
would significantly enhance the ability of
interpreters to map key reservoir layers. The
2.40
company’s objective was to identify high-porosity
zones within two carbonate layers: the primary
2.45
Jurassic San Andres (Jsa) limestone and the
Acoustic impedance
inferior shallower Cretaceous Tamaulipas (Kti)
2.50
2,700 2,800 carbonate target.
Common depth point (CDP) A Q-Marine high-resolution 3D seismic survey
> Absolute acoustic impedance from poststack inversion. Inversion of achieved a maximum frequency of 60 Hz,
seismic amplitudes generated the color-coded panel, with low acoustic doubling that of the 1996 survey.5 Inversion of
impedance in pink and red, and high acoustic impedance in blue and green. the new data generated porosity maps that
The acoustic impedance calculated from density and sonic logs, displayed
at the well location in the middle of the panel, shows a good correlation helped rank previously defined drilling locations,
with the seismically derived values. determine new potential locations and optimize
development drilling.
Crossplot Analysis Trace-by-trace inversion of the stacked
30 seismic data allowed geophysicists to obtain
relative acoustic impedance at every trace
throughout the seismic volume. Key horizons that
had been interpreted as strong acoustic events
were converted from time to depth by correlation
0.70 0.93 with formations seen in well logs. This
Water saturation
20 combination of interpreted horizons and values of
acoustic impedance at these points enabled the
Effective porosity, %

creation of a low-frequency model to convert the

relative acoustic impedance to an absolute
measurement (above left).6
Crossplotting porosity with acoustic imped-
10 ance from logs and core data in the survey area
revealed a strong correlation between the two
properties—an increase in porosity causes a
decrease in velocity, a decrease in density, and
therefore a corresponding decrease in acoustic
impedance (left). Porosity-acoustic impedance
0 functions were created for the Jsa and Kti
formations separately. Applying these correlations
8 10 12 14 16
Edited acoustic impedance (Zp), km/s .g/cm3
to the seismically derived acoustic impedance
volume, geophysicists created fieldwide maps of
> Acoustic impedance and porosity. The strong correlation between
porosity. The seismic porosity results were
porosity and acoustic impedance from logs and core data in the Jsa
formation indicates a robust transform for application to seismic inversion
results. As in other carbonate rocks, an increase in acoustic impedance is
related to a decrease in porosity. A separate porosity-acoustic impedance
function was created for the Kti formation.

46 Oilfield Review
checked using “blind wells,” that is, wells that
2,200
were not used in the inversion. The seismically
derived porosity closely matched the blind-well 2,250
porosity logs, adding confidence to the seismically
calculated results. 2,300
The porosity maps had a significant impact on

Two-way time, ms
defining infill drilling locations. In the nearby 2,350

Arenque field, covered by the same survey,

2,400
Pemex upgraded four previously identified
prospects. Increased priority was given to the 2,450
two locations corresponding to higher porosity in
the seismic volume. In one area, the inversion 2,500
calculations allowed identification of undrilled,
discrete porosity features (right). With these 2,550
Porosity
results, well placement could be designed to
2,600
maximize contact with high-porosity zones in the 1,075 1,100 1,125 1,150 1,175 1,200 1,225 1,250 1,275 1,300
Jsa formation. Crossline number
In another area where seismic porosity > Identifying undrilled high-porosity targets. Inversion revealed a high-
results were used to guide drilling, a well porosity interval (purple and red), helping Pemex delineate zones that could
produced oil from the Jsa formation at be reached with new wells. The black line is a possible well trajectory. An
existing well is shown in gold.
2,000 bbl/d [318 m3/d]. The seismically derived
results show excellent correlation with the
porosity measured in the well (below right).

Inversion When Offset Matters

In many cases, the stacking process does not
adequately preserve amplitude. For example,
when traces exhibit amplitude variation with N
Kti
offset (AVO), the trace that results from stacking
does not have the same amplitudes as the
vertical-incidence, or zero-offset, trace. Under
these conditions, inversion should be performed
on data that have not been stacked. Moreover,
the parameters that cause the amplitude to Jp

change can be modeled and used to further the

Porosity
inversion process.
Jsa
Data preparation for inversion of AVO traces
requires steps similar to those for preparation for
stacking. Traces reflecting at a common midpoint
are gathered and sorted by offset, which is
related to incidence angle. Then, a velocity Bas
model is applied to each gather to flatten events

5. Salter R, Shelander D, Beller M, Flack B, Gillespie D,

Moldoveanu N, Gonzalez Pineda F and Camara Alfaro J:
“Using High-Resolution Seismic for Carbonate Reservoir
Description,” World Oil 227, no. 3 (March 2006): 57–66.
6. Salter R, Shelander D, Beller M, Flack B, Gillespie D,
Moldoveanu N, Pineda F and Camara J: “The Impact of
High-Resolution Seismic Data on Carbonate Reservoir
Description, Offshore Mexico,” Expanded Abstracts,
75th SEG Annual International Meeting and Exposition,
Houston (November 6–11, 2005): 1347–1350.
> Results of drilling into a zone predicted to have high porosity. A well
penetrated both the Cretaceous (Kti) and Jurassic (Jsa) reservoirs,
encountering porosities that matched values predicted for the two
carbonate zones. The green circle marks the top of the Kti formation, and
the light blue circle marks the top of the Jsa formation. The porosity log,
projected on the well path, has the same color-coding as the seismically
predicted porosities.

Spring 2008 47
 Single-Layer Geometry: Direct Relationship Between and Offset To facilitate inversion, an AVO dataset may be
divided into subsets according to angle. For
Offset 4 example, the near-offset, mid-offset and far-offset
Offset 3 traces may form three separate datasets. For
Offset 2 each CMP gather, the near-offset traces are
Offset 1 stacked and then collected with the near-offset
traces from all the other CMPs, forming a near-
S4 S3 S2 S1 R1 R2 R3 R4
offset dataset. Similarly, the mid-offset and
θ1 far-offset traces from each CMP gather can be
grouped. Each offset group can be inverted
θ2
separately. Although some of the AVO information
is lost in these partial stacks, sometimes called
offset or angle stacks, in many cases enough
remains to obtain reasonable inversion results.
Inversion of traces with AVO is more
Common midpoint (CMP) Gas sand complicated than poststack inversion because
the reflectivity formula is more elaborate, with
Synthetic Traces: CMP Gather
dependence not only on density and compres-
sional velocity, but also on shear velocity and
Offset 4 angle of incidence. The general expressions for
Offset 3 the angular dependence of the reflection of
Offset 2
compressional and shear waves as functions of
densities, velocities and incident angle are
Offset 1
known as the Zoeppritz equations.7 Because the
full Zoeppritz formulation is cumbersome,
approximations are often used to generate
synthetics and facilitate fast inversion.8
Amplitude increases with offset
Each approximation method attempts to fit a
simplified formula to the curve of reflection
Multilayer Geometry: Complex Relationship Between and Offset
amplitude versus angle of incidence (next page,
S1 R1 top right). The simplified approaches differ in
the number of terms used in the
Shale 1 approximation—usually two or three—and in
the parameters solved for. Some two-parameter
inversions calculate P-wave impedance (Zp,
Shale 2 equal to ρ Vp) and S-wave impedance (Zs, equal to
ρ Vs). A three-parameter inversion might
determine Zp, Zs and density ( ρ), but a three-
CMP Gas sand
parameter inversion for Zp, Vp /Vs and ρ would
contain the same information. Some
> Amplitude variation with offset (AVO). In steps similar to preparation for
stacking, traces reflecting at a common midpoint are gathered and sorted approximations are expressed in terms of
by offset (top), then the arrivals are flattened using a normal moveout Poisson’s ratio (ν), shear modulus ( μ), bulk
velocity model while preserving the amplitude information (middle). Clearly, modulus (λ) and ρ, which again are related to Vp
in this case, averaging the four traces would produce a trace that does not and Vs.
resemble the zero-offset trace; in other words, stacking would not preserve
amplitudes. The offset versus angle (θ) relationship is determined by ray The number of parameters that can be solved
tracing (bottom). for depends on the range of offsets—or equiva-
lently, angles—available and on data quality. If a
large range of offsets or angles is available and
the signal-to-noise ratio at high offset is good,
to a common arrival time over all offsets (above). Most AVO inversion algorithms are based on three parameters may be resolved. If offsets are
For a given reflection, amplitudes are tracked the relationship between reflection amplitude limited, then inversion may deliver only two
and plotted against offset. The flattened gather and angle of incidence. Therefore, additional parameters reliably. Density is the most difficult
and the amplitude variation with offset comprise steps before inversion include converting the parameter to invert for; the process requires long
the data that will be compared with synthetics offset values to angles. The traces are initially offsets and high-quality data.
during the inversion process. labeled by source-receiver offset. The relation-
ship between angle and offset is calculated by
tracing a ray from source to receiver in an
accurate velocity model.

48 Oilfield Review
 A case study presents three-parameter Poisson’s
inversion on AVO data acquired offshore Egypt. Vp Vs ρ Ro Vp / Vs Ratio
Layer 1 2,800 1,700 2.3 1.647059 0.208081
1.0
Inversion in the Nile Delta Layer 2 3,000 2,000 2.2 0.01227 1.5 0.1
Apache Egypt Companies, with partners RWE
Dea and BP Egypt, recorded a 3D seismic survey
in a western Mediterranean deep marine Zoeppritz Shuey 2-Term
concession in the Nile Delta.9 The seismic data 1.2 Aki and Richards Gidlow 2-Term

Two-way time, s
exhibited strong amplitudes over a gas-charged Pan-Gardner Connolly
Gidlow 3-Term Anisotropic
complex of channel and levee sands. However,

Reflection coefficient
amplitude alone was not a reliable indicator of
0
gas saturation: distinct accumulations—one
1.4
with high gas saturation and the other with low
gas saturation—both displayed high amplitude.
Extracting density information from the seismic
data was a key to identifying commercial
gas sands. 1.6
5 20 35 50 0 20 40 60 80
The main objective of prestack inversion was Average angle, deg Average angle, deg
to improve the existing reservoir model in
preparation for optimizing the appraisal and
development plan. The survey featured long > Amplitude variation with offset data and reflection coefficient versus angle of incidence. Several
offsets, up to 6,000 m [19,690 ft], enabling AVO reflections in the CMP gather (left) exhibit amplitude variation with offset. These data come from the
inversion for three elastic parameters: P-wave North Sea example described on page 51. The nearly vertical black lines delimit angle ranges
computed by ray tracing. The reflection of interest is at 1.26 s (yellow). At zero offset (normal
impedance, S-wave impedance and density. incidence), the reflection has slightly positive amplitude—a swing to the right—then turns negative,
Correlation with log data would help Apache with a swing to the left. Several methods can be used to model the reflection coefficient versus angle
estimate rock and fluid properties throughout (right). The properties of the two-layer model are shown (top). R0 stands for reflection coefficient at
zero offset. The exact solution by the Zoeppritz equations is shown by the black curve. The other
the 1,500-km2 [580-mi2] study area.
curves are approximations taken from the work described in reference 8.
Rock-property correlations using log data from
the five wells in the concession discriminated
rock-fluid classes on the basis of Vp / Vs and
P-wave impedance (below right). The separation
between sands with high and low water satura-
tions suggested that fluid-content differences
would be apparent in the inversion results. 0.1 0.9
3.0 Sw
7. Zoeppritz K: “Über Erdbebenwellen, VIIB: Über
Reflexion und Durchgang seismicher Wellen durch Water-filled
Unstetigkeitsflächen,” Nachrichten der Königlichen sands
Gesellschaft der Wissenschaften zu Göttingen,
Mathematisch-physikalische Klasse (1919): 57–84.
8. Aki K and Richards PG: Quantitative Seismology: Theory 2.5
Vp / Vs

and Methods. San Francisco: W.H. Freeman and

Company, 1980.
Connolly P: “Elastic Impedance,” The Leading Edge 18,
no. 4 (April 1999): 438–452.
Pan ND and Gardner GF: “The Basic Equations of Plane
Elastic Wave Reflection and Scattering Applied to AVO 2.0
Gas sands
Analysis,” Report S-87-7, Seismic Acoustic Laboratory,
University of Houston, 1987.
Rüger A: “P-Wave Reflection Coefficients for
Transversely Isotropic Models with Vertical and
Horizontal Axis of Symmetry,” Geophysics 62, no. 3 12,500 15,000 17,500 20,000 22,500
(May–June 1997): 713–722.
P-wave impedance, ft/s . g/cm3
Shuey RT: “A Simplification of the Zoeppritz Equations,”
Geophysics 50, no. 4 (April 1985): 609–614. > Correlating acoustic properties with water saturation (Sw). Log
Smith GC and Gidlow PM: “Weighted Stacking for Rock measurements of P-wave impedance, water saturation and Vp / Vs are
Property Estimation and Detection of Gas,” Geophysical
Prospecting 35, no. 9 (November 1987): 993–1014.
crossplotted to show relationships that can be applied to seismic inversion
results. Clean gas sands are plotted in red, laminated sands in green, and
9. Roberts R, Bedingfield J, Phelps D, Lau A, Godfrey B,
Volterrani S, Engelmark F and Hughes K: “Hybrid water-filled sands in blue. (Adapted from Roberts et al, reference 9.)
Inversion Techniques Used to Derive Key Elastic
Parameters: A Case Study from the Nile Delta,”
The Leading Edge 24, no. 1 (January 2005): 86–92.

Spring 2008 49
 The inversion workflow combined full-
Poisson’s Density
Vp Ratio Synthetic Gather Observed Gather waveform prestack inversion with three-term
g/cm3
1.5 2.5 AVO inversion. The prestack inversion, per-
formed at sparsely sampled locations, provided
2.0 background Vp / Vs trends, which, with the well
data, were used to build low-frequency models to
merge with the results of the AVO inversion.
Agreement between synthetic predictions and
actual results was generally good (left).
Two-way time, s

The three-parameter AVO inversion results

were converted into relative impedances and
merged with the low-frequency background
2.5 models to generate 3D volumes of P-wave
impedance, S-wave impedance and density.
With transforms derived from rock-physics
analysis, these elastic attributes were then
converted to volumes of net-to-gross sand and
bulk water saturation.
The density volume was found to be a reliable
indicator of fluid saturation. For example, the
> Comparison between observed and synthetic AVO gathers. The observed
Abu Sir 2X well, drilled at a location of high
AVO gather (right) was inverted for Vp , Poisson’s ratio and density. The
results (left three panels) are plotted with associated uncertainty range
seismic amplitudes, encountered one zone with
(yellow). A synthetic gather generated from the Vp , Poisson’s ratio and high gas saturation and two deeper zones with
density models appears in the fourth panel. The close match between the low gas saturation (below left). A seismically
observed and synthetic AVO gathers indicates that the property models are derived density profile through the well predicts
good representations of actual earth properties. (Adapted from Roberts et
al, reference 9.) low gas saturation in the deeper layers. The
density results from seismic inversion delineate
a single high-saturation interval and also show its
limited lateral extent.
Abu Sir 2X The inversion results can be examined from a
variety of perspectives. For instance, tracking
one of the layers that was uneconomic in the Abu
Gamma Sir 2X well throughout the seismic volume
Ray Resistivity
reveals a region where that layer might contain
high gas saturation (next page, bottom).
Although this accumulation is downdip from the
reservoir encountered in other wells in the area,
the density and water-saturation maps support
the interpretation that the downdip area has
Zone 1 high gas saturation and is not water-filled. As a
consequence of this study, a new well, Abu Sir 3X,
Gas/water contact is planned for this area.

Zone 2 10. For more on drilling injectite targets: Chou L, Li Q,

Low-saturation gas Darquin A, Denichou J-M, Griffiths R, Hart N, McInally A,
Templeton G, Omeragic D, Tribe I, Watson K and
Zone 3 Wiig M: “Steering Toward Enhanced Production,”
Oilfield Review 17, no. 3 (Autumn 2005): 54–63.
11. Pickering S and McHugo S: “Reservoirs Come in All
Shapes and Sizes, and Some Are More Difficult Than
Others,” GEO ExPro no. 1 (June 2004): 34–36.
McHugo S, Cooke A and Pickering S: “Description of a
1.95 2.48
Density, g/cm3 Highly Complex Reservoir Using Single Sensor Seismic
Acquisition,” paper SPE 83965, presented at SPE
> Inversion for density. Inversion of AVO data over gas fields in the Nile Delta predicts low density Offshore Europe, Aberdeen, September 2–5, 2003.
(red) in the upper part of the reservoir (Zone 1) and higher densities (green and yellow) deeper in the
reservoir (Zones 2 and 3). The density measured at the well location is inserted in the center of the
panel and plotted on the same color scale as the seismic-inversion density. The well logs (inset right)
show where sands were logged (yellow shaded gamma ray) and where high resistivity (red curve)
indicates hydrocarbon. The seismic amplitude section, not shown, exhibited high amplitudes in all
zones of the reservoir, and so was unable to distinguish the low gas saturation in Zones 2 and 3 from
the high gas saturation of Zone 1. (Adapted from Roberts et al, reference 9.)

50 Oilfield Review
Inversion to Enhance Visibility Sand Injectite
In some cases, the acoustic impedance contrast
between two lithologies may be so small that the
interface between them generates almost no
normal-incidence reflection. For example, an oil-
filled sandstone with high density and low
P-wave velocity might have nearly the same
acoustic impedance as a shale with lower density
and higher P-wave velocity. Without an acoustic
impedance contrast, such oil reservoirs are
> Sand-injection features, or injectites. The remobilization of unconsolidated
extremely difficult to detect using traditional
sand (gold) into overlying shale layers (gray) can result in injectites. These
surface seismic acquisition and processing. sandstone features have irregular shapes and are difficult to image seismically.
An example of a low-contrast reservoir is the
Alba field in the North Sea. Alba and reservoirs
like it are interpreted to be injectites, formed by
the injection, or remobilization, of unconsoli-
dated sand into overlying shale layers during are not discovered by seismic imaging, but are seismic characterization of injected reservoir
periods of differential stress (above right). These encountered inadvertently while drilling to sands in the Balder interval that were
complex reservoirs are characterized by deeper targets.10 particularly difficult to image.11 Modeling studies
irregular morphology and chaotically distributed In one area of the central North Sea, an using rock properties from well data established
high-porosity sands. Often, such accumulations operating company wanted to improve the that prestack inversion of seismic data could

Conventional Amplitude P -Wave Impedance

Abu Sir 2X Abu Sir 2X

Abu Sir 1X Abu Sir 1X

Planned well Planned well

Density Water Saturation

Abu Sir 2X
Abu Sir 2X

Abu Sir 1X Abu Sir 1X

Planned well Planned well

0 Water saturation 1

> Tracking inversion results through the reservoir. Parameters extracted from the seismic data and its
inversion are displayed for Zone 2—one of the zones that had uneconomic gas saturations in the Abu
Sir 2X well. The amplitudes of the original seismic data (top left) show anomalies near the Abu Sir 2X
well, but the density plot (bottom left) does not. Low amplitudes are plotted in blue and green, and
high amplitudes are plotted in red and purple. Low densities are plotted in red and high densities are
plotted in blue and green. Amplitude, density and P-wave impedance (top right) all exhibit exceptional
values in the southeast corner, where a well is planned. Low P-wave impedances are plotted in red
and purple, and high impedances are plotted in blue and green. Conversion of the inversion results to
water saturation (bottom right) indicates that the planned well should encounter low water saturation.
(Adapted from Roberts et al, reference 9.)

Spring 2008 51
 Log Data potentially distinguish the clean sands from the
surrounding shales, but the existing surface
seismic data were of insufficient resolution for
this purpose.
A new survey was designed to map the distri-
0.05 0.95
Volume of sand bution and thickness of reservoir pay, delineate the
3.0 geometry of individual sand wings and assess
reservoir connectivity. Acquisition with the
Q-Marine system would allow accurate cable
positioning, fine spatial sampling and calibration
Vp / Vs

of sources and receivers. Together, these capabili-

2.5 ties facilitate precision imaging, improved noise
attenuation, increased bandwidth and preser-
vation of amplitude and phase information—all
important for successful inversion.
Log data from three wells intersecting the
2.0 reservoir were analyzed for correlations
between P- and S-wave velocities, ρ, μ, λ,
lithology and fluid saturation. For example,
crossplotting Vp / Vs with the product μρ, and
color-coding by lithology, showed that high sand
2 4 6 8 10 12 14
Shear modulus.density (μρ ), GPa.g/cm3 content correlated with low Vp / Vs and high μρ
values (left). These relationships were then
> Correlating acoustic properties with lithology. A crossplot of Vp / Vs with
applied to Vp / Vs calculated from seismic
the product of shear modulus () and density (ρ) shows a trend related to
sand volume: high sand content correlates with low Vp / Vs and high ρ inversion to map high sand content throughout
values. Applying this relation to Vp / Vs and ρ values obtained from the seismic volume.
inversion yields lithology maps of the subsurface. The prestack seismic data were divided
into seven angle stacks, each containing
reflections in a 7° range of incidence angles out
to 49° (next page, bottom). Three-parameter AVO
Seismic Values
inversion generated estimates of P- and S-wave
reflectivities and density contrast. These volumes
were inverted for P- and S-wave impedances and
density, from which volumes of μρ, Vp / Vs and λ /μ
0 100
Sand probability, % were generated.
3.0 Crossplots of seismically derived Vp / Vs and
μρ through the interval containing the injected
sands were color-coded by sand probability
(left). Applying the color-coding to the rock-
constant volumes obtained from seismic
Vp / Vs

2.5
inversion yielded interpretable 3D cubes of sand
probability. A close-up of a section through the
sand-probability volume highlights a steeply
dipping sand-injection feature (next page, top).
The sand-probability volume can be
illuminated by rendering the surrounding
2.0
shales—lithologies with low probability of

2 4 6 8 10 12 14
Shear modulus.density (μρ ), GPa.g/cm3

> Sand probability. The correlation between inversion outputs Vp / Vs and

ρ with sand probability shows a direct relationship: increasing ρ and
decreasing Vp / Vs point to higher probability of sand. This relationship was
applied to the seismic inversion results to obtain maps of sand probability.

52 Oilfield Review
 Seismic Amplitude Sand Probability

0 100
Sand probability, %

80 m
0 750 0 750
Distance, m Distance, m
> Comparing seismic reflection amplitudes with sand probability. A steeply dipping feature seen in the
center of the seismic reflection image (left) has a high probability of being sand (right). This structure,
which is 80 m [260 ft] high, has the shape and aspect expected of a sand injectite.

Input Data AVO Analysis Acoustic Properties Rock Constants

P-wave reflectivity P-wave impedance μρ

Seven angle gathers

(0 to 49°)

S-wave reflectivity S-wave impedance Vp /Vs

Conditioned well data

Synthetics Density Sonic

Density contrast Density λ/μ

> AVO inversion workflow. The input data consisted of prestack AVO gathers in 7° offset ranges, along with sonic and density well logs (left). The first step,
three-parameter AVO inversion, produced estimates of P-wave and S-wave reflectivities and density contrast. These volumes were inverted for P-wave
and S-wave impedances and density. The final step extracted rock properties, in the form of ρ, Vp / Vs and λ/.

Spring 2008 53
Sand-Intrusion Visualization
> Seeing the sand. Volumes with high sand probability are colored yellow, gold and red, and the portions
with low probability of sand have been made transparent. The top surface of the underlying sand
formation from which the injectite was ejected is blue. (Adapted from Pickering and McHugo,
reference 11.)
sand—transparent using 3D visualization
technology (above). This characterization of the
extent and quality of the injected sand
bodies can help optimize development of these
complex features.
Simultaneous Inversion
The examples presented so far have shown the
results of techniques that invert traces
separately and then combine the results in a
display of reflectivity. Geophysicists at the
Danish company Ødegaard, now part of
Schlumberger, have developed a simultaneous
inversion technique that examines all traces at
once to invert for a globally optimized model of
rock properties.12
Global optimization is a term describing
several methods designed to find the best overall
solution of a problem that has multiple local
solutions. An inversion problem may be cast as
finding the absolute minimum of a multi-
dimensional, nonlinear function (right). This can
be likened to placing a ball on a hilly surface and
letting it roll to the lowest level. Depending on
which hill the ball starts on and which direction
it rolls, it may get stuck in a nearby low spot—a
local minimum—or land in the lowest area in the
space—the global minimum.
Analogously, some inversion techniques
depend heavily on the starting model—which hill
they start on—and so may find a local minimum
54
rather than the absolute minimum. Global optimi-
zation attempts to find the absolute minimum by
adopting new ways of searching for solution
candidates. Various strategies may be utilized to
reach a solution. The approach taken by the ISIS
suite of reservoir characterization technology
developed by Ødegaard is simulated annealing.
Simulated annealing is based on a physical
analogy. In metallurgy, annealing is the process
of controlled heating and subsequent cooling of a
metal. Heating increases the internal energy of
the metal atoms, causing them to abandon their
places in the crystal structure. Gradual cooling
allows the atoms to reach lower energy states.
Under properly controlled heating and cooling,
the system becomes more ordered; crystal size
increases, and the resulting material has
minimal defects.
Instead of minimizing the thermodynamic
energy of a system, inversion by simulated
annealing aims to minimize an objective function,
also called a cost function. The algorithm replaces
the starting solution with another attempt by
selecting a random solution not far from the first.
If the new solution reduces the cost function, it is
kept, and the process is repeated. If the new
solution is not much better than the previous one,
another random solution is tested. However,
simulated annealing improves over some other
methods by allowing a “worse” solution if it helps
investigate more of the solution space.
The ISIS simultaneous inversion cost
function is made up of four penalty terms that
are collectively minimized to deliver the best
solution. The first term contains a penalty for
differences between the seismic data and the
synthetic. The second term includes the low-
frequency acoustic impedance trend in the
inversion through a penalty for deviation of the
estimated acoustic impedance model from the
low-frequency model. The third term attenuates
horizontally uncorrelated noise by introducing a
penalty for horizontal variations in the estimated
acoustic impedance model. The fourth term
introduces a sparsely parameterized background
model of layer boundaries. These terms can be
modified to include requirements of more
complex data types, such as time-lapse surveys
and shear waves.
Compared with trace-by-trace reflectivity
methods, simultaneous inversion has several
benefits. Honoring the full bandwidth of the
seismic signal—low and high frequencies
together—enhances resolution and accuracy.
The ISIS inversion algorithm can be used on
many types of seismic data (next page, top right).
The remainder of the article focuses on three
distinct applications: a 3D AVO study from
Australia, a time-lapse example from the North
Sea and a multicomponent case using seabottom
sensors also from the North Sea.
Revealing a Reservoir in Australia
Many seismic surveys are acquired and
processed purely for reflector-imaging purposes,
without inversion in mind. However, inversion
Z
X
Y
> Finding the minimum. Many inversion schemes
attempt to minimize a multidimensional,
nonlinear cost function with multiple minima. In
this case, the minima are shown as low points in
this 3D surface. Depending on the inversion
algorithm and starting point, the process might
end in a local minimum—the point that is the
lowest in a neighborhood—instead of the global
minimum—the lowest point of all.
Oilfield Review
can deliver even better results when survey
Data Type Physical Properties
design, acquisition and processing are tailored to
the requirements of the inversion scheme. Full-stack data P-wave impedance
Operating offshore Western Australia, Santos
Ltd. and partners wanted to enhance recovery Partial-stack AVO data P-wave impedance, Vp /Vs (or S-wave impedance)
and density, from which Poisson’s ratio, λ and μ
from their reservoir, and accurate mapping can be estimated.
would help achieve this.13 However, even after
reprocessing, the seismic data acquired in 1998 Intercept and gradient AVO data Acoustic impedance from the intercept data;
shear impedance from shear seismic dat a
were not of sufficient quality to allow interpre- calculated from the intercept and gradient data
tation of the top and base of the primary reservoir.14
Multicomponent full-stack data P-wave impedance and S-wave impedance
Rock-physics analysis of well logs revealed (P-to-P and P-to-S conversions)
that the contrast in P-wave impedance between
the reservoir and the overlying shale was subtle. Multicomponent partial-stack P-wave impedance, Vp /Vs (or S-wave impedance)
AVO data and density
This explained some of the difficulty in
identifying the reservoir in vertical-incidence Borehole seismic (VSP) data Acoustic impedance from the PP data; shear
reflection data. However, a large contrast in impedance from the PS data

Poisson’s ratio should be observable if AVO data
in the appropriate offset range were acquired.
In addition to the low-reflectivity problem,
the 1998 survey data were noisy. The continental
shelf offshore northwest Australia has a layer of
high acoustic impedance contrast near the
seafloor. This layer traps seismic energy,
generating reverberations called multiples that
contaminate the seismic record.
WesternGeco survey evaluation and design
(SED) specialists investigated ways to eliminate
noise and improve overall recording in a new
survey. Removing the noise from multiples
required an accurate image of the seafloor, which
could be acquired if extremely short offsets were
recorded. The 3.125-m [10.25-ft] spacing of
Time-lapse full-stack data (which
may include multicomponent
full-stack data)
Time-lapse partial-stack AVO data
(which may include multicomponent
partial-stack AVO data)
> Applications of ISIS simultaneous inversion.
1998 Survey with 2006 Reprocessing
1,700
1,800
Simultaneous time-lapse inversion for baseline
P-wave impedance and the changes in P-wave
impedance for each time interval: for
multicomponent data, inversion will also output
baseline S-wave impedance and changes in
S-wave impedance over the time interval.
Simultaneous time-lapse inversion for baseline
properties and the changes: for example, fo r
partial-stack data, inversion can determine
baseline P-wave impedance, Vp /Vs (or S-wave
impedance) and density and the changes in
these properties over the time interval.
2006 Q-Marine Survey

Q-Marine hydrophones would adequately sample 1,900

both the desired signal and the noise, facilitating
2,000
effective removal of the latter. Modeling showed
that a streamer length exceeding 5,000 m 2,100

[16,400 ft] would be needed to capture AVO 2,200

effects at the reservoir level. This length would
2,300
provide data over an incidence-angle range of
10 to 50°. 2,400
Time, ms

Comparison of a reflection-amplitude image 2,500

from the 2006 Q-Marine survey with one from the
2,600
reprocessed 1998 dataset shows improved
structural imaging and reduced noise (right). 2,700
Testing during processing identified the steps 2,800
that would optimize inversion.
2,900

12. Rasmussen KB, Brunn A and Pedersen JM: “Simultaneous 3,000

Seismic Inversion,” presented at the 66th EAGE
Conference and Exhibition, Paris, June 7–10, 2004. 3,100
13. Partners were Kuwait Foreign Petroleum Exploration
Company (KUFPEC), Nippon Oil Exploration and 3,200
Woodside Energy.
14. Barclay F, Patenall R and Bunting T: “Revealing the 3,300
5,700 4,200 5,625 3,135
Reservoir: Integrating Seismic Survey Design,
Acquisition, Processing and Inversion to Optimize Crossline number Crossline number
Reservoir Characterization,” presented at the 19th ASEG > Seismic images in an Australian field. Multiples generated by a high acoustic impedance layer near
International Geophysical Conference and Exhibition,
Perth, Western Australia, November 18–22, 2007. the seafloor make it difficult to image the low-impedance-contrast reservoir, which is within the
shaded interval, at approximately 2,100 to 2,200 ms. The Q-Marine image (right) exhibits less noise and
better resolution of structural features than the 1998 dataset (left). (Adapted from Barclay et al,
reference 14.)

Spring 2008 55
 The inversion for P-wave impedance gave
1.7 high-quality results that correlated strongly with
values measured in four of the field’s wells (left).
1.8
The subtle rise in acoustic impedance at the top
of the reservoir, although significantly smaller
1.9
than those in the overlying layers, is accurately
2.0 detected by the simultaneous inversion.
The P-wave impedance contrast at the top of
2.1 the reservoir is small, but the contrast in Poisson’s
Two-way time, s

ratio is significant, and so is a potentially more

2.2
useful indicator of reservoir quality. Poisson’s ratio
is more accurately estimated by including large
2.3
incidence angles in the inversion. A comparison of
2.4 Poisson’s ratio obtained by incorporating different
ranges of incidence angles showed greater
2.5 resolution and less noise when wider angles were
included (below).
2.6

2.7 Time-Lapse Inversion

Simultaneous inversion can incorporate data
5.7 8.2 from various vintages to highlight time-lapse
Acoustic impedance, km/s .g/cm3 changes in rock and fluid properties. This
> Simultaneous inversion for P-wave impedance. Impedance sections from inversion show excellent approach has recently been tested on the Norne
correlation with values in four wells. In each panel, the impedances measured in the well are color- field, where operator StatoilHydro is trying to
coded at the same scale as the inversion results and inserted in the middle of the panel. The top of
increase oil recovery from 40% to more than 50%.
the reservoir is marked with a nearly horizontal black line. The white curves are unscaled water
saturation logs, with water saturation decreasing to the left. To the right of each panel is a display of The Norne field has had multiple time-lapse,
the log acoustic impedance (red) and the seismically estimated acoustic impedance at the well or 4D, seismic surveys.15 The high-quality
location (blue). (Adapted from Barclay et al, reference 14.) sandstone reservoirs, with porosities of 25 to 32%

Cro 0.34 0.43

s Poisson’s ratio Cro
sli ssl
ne ine
ine

ine
Inl

Inl

450 460 470 450 460 470

> Inversion for Poisson’s ratio. In this field, Poisson’s ratio provides a better measure than acoustic impedance for assessing reservoir quality. Low
Poisson’s ratio (green) is generally indicative of higher quality sand. Reflection amplitudes are more affected by Poisson’s ratio at larger angles of
incidence. When a larger range of angles (5 to 42°) is included in the inversion (right), the estimation of Poisson’s ratio shows less noise, and the regions of
similar Poisson’s ratio appear more continuous than when inversion uses a smaller range (5 to 35°) of angles (left). White circles are well locations.
(Adapted from Barclay et al, reference 14.)

56 Oilfield Review
and permeabilities ranging from 200 to 2,000 mD, Well A Acoustic Impedance
1.9
are conducive to successful time-lapse moni-
toring; changes in fluid saturation and pressure 9

Acoustic impedance, 2003

B
give rise to noticeable differences in seismic A

Two-way time, s
amplitudes and elastic impedances.
The first 3D surface seismic survey over the
field was acquired in 1992. This large exploration
survey was acquired before production and water
and gas injection, but was not considered a
5
baseline for time-lapse monitoring. In 2001, 2.8
Norne’s first Q-Marine survey was acquired, with Well B
repeatable acquisition, forming the baseline for 1.9
the 2003, 2004 and 2006 monitor surveys—all
acquired with Q-Marine technology.

Two-way time, s
From the start, time-lapse monitoring
delivered crucial information for optimizing field
development. Differences in the AVO inversions
of the 2001 and 2003 surveys revealed changes in –5
Acoustic impedance 5
acoustic impedance that could be interpreted as change, 2006/2001, %
increases in water saturation.16 In one area, the
2.8
trajectory of a planned well was modified to avoid Acoustic impedance Log curve
a zone inferred to have high water saturation.17 Inversion result
Recently, the evaluation of changes in effec- 5 MPa.s/m 12 Low-frequency model
tive stress has become important for optimizing
reservoir depletion and injection strategies. To
understand the continuing effects of production Well A Poisson’s Ratio
on the field, StatoilHydro and Schlumberger 1.9
undertook a simultaneous inversion project that 0.36
incorporated all available seismic data, log data B
Poisson’s ratio, 2003 A
Two-way time, s

from seven wells and production data from the

ECLIPSE reservoir model.18
The ISIS simultaneous inversion estimated
baseline values and changes in acoustic
impedance and Poisson’s ratio from the time-
lapse seismic data (right). To compensate for the 0.18
2.8
lack of low-frequency information in the seismic
Well C
bandwidth—needed to determine absolute 1.9
elastic properties—background models were
constructed. For the baseline survey, the
Two-way time, s

background model was derived by propagating

borehole values of elastic properties throughout
the zone of interest, constrained by key
interpreted horizons and the seismic velocities
–10
in each interval. Poisson’s ratio change, 10
2006/2001, %

15. Osdal B, Husby O, Aronsen HA, Chen N and Alsos T: 2.8

“Mapping the Fluid Front and Pressure Buildup Using 4D
Data on Norne Field,” The Leading Edge 25, no. 9
(September 2006): 1134–1141.
16. Khazanehdari J, Curtis A and Goto R: “Quantitative Time-
Lapse Seismic Analysis Through Prestack Inversion and
Rock Physics,” Expanded Abstracts, 75th SEG Annual
International Meeting and Exposition, Houston,
November 6–11, 2005: 2476–2479.
17. Aronsen HA, Osdal B, Dahl T, Eiken O, Goto R,
Khazanehdari J, Pickering S and Smith P: “Time Will Tell:
New Insights from Time-Lapse Seismic Data,” Oilfield
Review 16, no. 2 (Summer 2004): 6–15.
18. Murineddu A, Bertrand-Biran V, Hope T, Westeng K and
Osdal B: “Reservoir Monitoring Using Time-Lapse
Seismic over the Norne Field: An Ongoing Story,”
presented at the Norsk Petroleumsforening Biennial
Geophysical Seminar, Kristiansand, Norway,
March 10–12, 2008.
Log curve
Inversion result
0.20 Poisson’s ratio 0.45 Low-frequency model
> Time-lapse inversion. Results for acoustic impedance (top) and Poisson’s ratio (bottom) use a low-
frequency model based on simulation results. In the 3D volume (top right), the back and side panels
show acoustic impedance values from the 2003 survey. The horizontal surface is a time-slice of the
ratio of acoustic impedance in 2006 to that in 2001. The increase (red) has been interpreted as
replacement of oil by water. Absolute acoustic impedance comparisons at two wells (top left) show
good correlation between well measurements and the 2003 acoustic impedance values. The red
arrows in each log track point to the top of the horizon of interest. The log tracks display well data
(red), seismically derived values (blue) and the low-frequency model (green). Results for Poisson’s
ratio (bottom) are plotted similarly. Well C is outside the 3D volume.

Spring 2008 57
 Background Model with Background Model with PZ CMP Gather
Time-Lapse Information No Time-Lapse Information
2.2
Acoustic impedance change Acoustic impedance change

2.3
Two-way time, s

2.4 Permeable fault? Impermeable fault?

2.5

PZ time
2.6
1,545 1,645 1,745 1,845 1,945 2,045 2,145 1,545 1,645 1,745 1,845 1,945 2,045 2,145
Crossline number Crossline number

2.2
Poisson’s ratio change Poisson’s ratio change

2.3
Two-way time, s

2.4

2.5
PS CMP Gather

2.6
1,545 1,645 1,745 1,845 1,945 2,045 2,145 1,545 1,645 1,745 1,845 1,945 2,045 2,145
Crossline number Crossline number

0.94 Amplitude 1.06

> Effect of background models on inversion. Time-lapse inversion for acoustic impedance (top) and
Poisson’s ratio (bottom) shows different results using different background models. These panels
focus on a region where the reservoir simulation model contains a transmissible fault that allows gas
PS time

migration. The acoustic impedance section calculated with a background model that incorporated
time-lapse effects (top left) indicates a decrease in acoustic impedance (red) across the fault. In the
acoustic impedance section calculated without a time-lapse background model (top right), the
decrease in acoustic impedance is constrained to the area above the fault, suggesting the fault is not
transmissible. The inversion for Poisson’s ratio also suggests a transmissible fault, but only when a
background model is used that honors simulator data (bottom left). The black curve on each panel is
the top of the formation indicated by red arrows in the previous figure. The amplitudes are the ratio of
the 2004 values to the 2001 values.

For the time-lapse low-frequency models,
estimates of elastic properties were obtained
from the ECLIPSE reservoir simulator in three
steps: reservoir properties were converted from
depth to time using the velocity model, then
converted to elastic-property changes using rock-
physics models. Finally, the spatial and temporal
distributions of the property changes were
constrained by seismic-velocity changes
observed in time-lapse traveltime differences.
This unique combination of time-converted
reservoir properties with seismic-derived
traveltime changes delivered accurate changes
in elastic properties consistent with the reservoir
simulation. Significant differences were found
between inversion results that did and did not
use updated background models (above).
The StatoilHydro reservoir management team
plans to use these results to track the movement
of the waterflood front, evaluate the progress of
water and gas injection, estimate the pressure
> Multicomponent seismic data. Common
distribution and update the reservoir model.
midpoint (CMP) gathers of PZ (top) and PS (bottom)
reflection data show traces at increasing offset
Multicomponent Inversion from left to right. Color bands delineate angle
The previous examples dealt with inversion of ranges. Several reflections exhibit AVO effects,
P-wave data. Towed-streamer seismic surveys are which may differ in their expressions on PZ and
PS gathers. For example, in the PZ gather, the
designed to generate and record only P-waves; reflection at the dotted red line is slightly positive
S-waves do not propagate in fluids. Compressional at zero offset, and decreases to nearly zero
waves generated by the source may convert to amplitude with increasing offset. Arrivals from
the same reflector in the PS gather are strongly
shear waves at the seafloor or below and then
positive at zero offset and decrease gradually
travel as such through the solid formations of the with increasing offset.
subsurface, but they must convert again to
P-waves to travel through the water and be
recorded by the receivers. Information about
S-wave velocity and shear modulus, μ, may be

58 Oilfield Review
gleaned through AVO analysis, but S-waves PZ Inversion
Acoustic impedance Density
themselves are not recorded.
However, it is possible to acquire S-wave data
if the receivers are coupled to the seafloor.
Ocean-bottom cables (OBCs) are designed for
this purpose. Typically, these cables contain four
multicomponent sensors—three geophones and
a hydrophone—spaced at intervals determined
by the survey requirements.19 The geophones
detect the multiple components of S-wave
motion, and the hydrophone—like towed-
streamer hydrophones—detects P-wave signals,
designated as PP arrivals. The P-wave is also
detected by the geophones, mainly on the
vertical component, giving rise to PZ signals.
The sources used in these surveys are the
same as those in towed-streamer surveys,
generating P-waves that convert to S-waves at
the seafloor or deeper. The resulting signals are Acoustic impedance Log curve Density Log curve
called PS data. Although multicomponent Inversion result Inversion result
surveys are more complex to acquire and process
5 km/s .g/cm3 10 Low-frequency 2,000 kg/m3 3,000 Low-frequency
model model
than single-component surveys, they provide
data that single-component surveys cannot.
Schlumberger has inverted multicomponent
seismic data from a gas and condensate field in PZ and PS Inversion
Acoustic impedance Density
the North Sea. The main objective of the
inversion study was to generate elastic
properties—P-wave impedance, Vp / Vs and
density—from the seismic datasets as input for
calculating large-scale geomechanical properties.
The geomechanical properties would be used for
building a 3D mechanical earth model (MEM).20
Processing of PZ and PS data is far more
complex than conventional single-component
dataset processing. The two data types came
from the same survey, but showed many
differences. For example, amplitudes, velocities
and AVO behavior were markedly different
between the two datasets (previous page, right).
To assess the value of the PS data, simul-
taneous inversion of the PZ data was compared
with simultaneous inversion of the combined PZ
and PS datasets (right).21 The acoustic > Simultaneous inversion of multicomponent data. Acoustic impedance (left) and density (right) from
impedance and density derived from the PZ and inversion using only PZ data (top) lack the resolution and continuity of the results of inversion using
PZ and PS data (bottom). In particular, compared with PZ inversion, the densities predicted by
PS reflection amplitudes were much better inversion of PZ and PS data showed much better correlation with log values. In the panels showing
resolved and matched the well values better than inversion results, the nearly horizontal black lines are interpreted horizons. (Adapted from Rasmussen
those calculated from PZ arrivals alone. et al, reference 21.)

19. Surveys that acquire such multicomponent data are also
called 4C surveys. For more on 4C surveys: Barkved O,
Bartman B, Compani B, Gaiser J, Van Dok R, Johns T,
Kristiansen P, Probert T and Thompson M: “The Many
Facets of Multicomponent Seismic Data,” Oilfield
Review 16, no. 2 (Summer 2004): 42–56.
20. Mohamed FR, Rasmussen A, Wendt AS, Murineddu A
and Nickel M: “High Resolution 3D Mechanical Earth
Model Using Seismic Neural Netmodeling: Integrating
Geological, Petrophysical and Geophysical Data,” paper
A043, prepared for presentation at the 70th EAGE
Conference and Exhibition, Rome, June 9–12, 2008.
21. Rasmussen A, Mohamed FR, Murineddu A and
Wendt AS: “Event Matching and Simultaneous
Inversion—A Critical Input to 3D Mechanical Earth
Modeling,” paper P348, prepared for presentation
at the 70th EAGE Conference and Exhibition, Rome,
June 9–12, 2008.

Spring 2008 59
 Original New Workflow SGS Kriging
The acoustic impedances from seismic inver-
sion improved the accuracy of the mechanical
AI AI Error AI Error AI Error
MPa. s/m MPa. s/m % MPa. s/m % MPa. s/m % earth model. In a blind-well test, acoustic imped-
3 12 3 12 0 100 3 12 0 100 3 12 0 100
Well log and ances predicted by inversion were compared
seismically derived with those measured in a well that had not been
properties
used for inversion calibration (left). In the 10
layers of the MEM, the well log acoustic
impedances showed an extremely close match
with impedances from the seismic inversion.
Well X Correlations with models built using conven-
tional methods of generating geomechanical
Depth

properties—methods that do not incorporate

seismically calculated properties—did not
match as well and exhibited large errors in
Acoustic impedance several layers.
4.5 MPa.s/m 8.5
Looking Forward with Inversion
Seismic inversion is a powerful tool for extracting
reservoir rock and fluid information from seismic
data. Although most seismic surveys are designed
for imaging alone, companies are increasingly
> Acoustic properties in a 3D mechanical earth model (MEM). Seismically derived acoustic applying inversion to get more out of their
impedances helped populate a 3D MEM with mechanical properties. The inset (right) shows one of 10 investments in seismic data. Some companies
layers in the model. Acoustic impedance (AI) values extracted along a wellbore (Well X) that had not now perform inversion on every seismic dataset
been used in building the model (Track 1) are compared with values predicted by three methods: and won’t drill without it.
seismic inversion (Track 2), sequential Gaussian simulation (SGS) (Track 3) and kriging (Track 4). SGS Inversion for reservoir characterization is a
and kriging do not use seismic data as input. Error bars (red) in each track display percentage error.
The match with seismically derived acoustic impedance is significantly better than with results from multistep process that requires, in addition to
the other two methods. (Adapted from Mohamed et al, reference 20.) the inversion algorithm itself, careful data
preparation, seismic data processing, log editing
and calibration, rock-property correlation and
Wedge Model
visualization. Workflows are being developed to
CDP
1 5 11 21 31 41 51 61 71 81 combine these steps for optimal results.
The addition of new measurements from other
1,600
disciplines, such as from deep electromagnetic
sensing, promises to bring enhancements to
seismic inversion results. Work on magneto-
1,800
Two-way time, ms

tellurics and controlled-source electromagnetics

2,000
Acoustic impedance
2,200 4,750 m/s . g/cm3 7,750
> Wedge model of acoustic impedance. Layers thicken from left to right.
Synthetic wells are shown as black curves at selected CDP numbers. The
curves represent water saturations. This model was used to generate
synthetic seismic sections. The acoustic impedance at CDP 5 was the basis
of the background model used to invert the synthetic sections.
22. Camara Alfaro J, Corcoran C, Davies K, Gonzalez
Pineda F, Hampson G, Hill D, Howard M, Kapoor J,
Moldoveanu N and Kragh E: “Reducing Exploration
Risk,” Oilfield Review 19, no. 1 (Spring 2007): 26–43.
Moldoveanu N, Combee L, Egan M, Hampson G,
Sydora L and Abriel W: “Over/Under Towed-Streamer
Acquisition: A Method to Extend Seismic Bandwidth to
Both Higher and Lower Frequencies,” The Leading
Edge 26, no. 1 (January 2007): 41–58.
23. Özdemir H: “Unbiased Seismic Inversion: Less Model,
More Seismic,” presented at the Petroleum Exploration
Society of Great Britain, Geophysical Seminar, London,
January 30–31, 2008.
24. Özdemir H, Leathard M and Sansom J: “Lost Frequencies
Found—Almost: Inversion of Over/Under Data,” paper
D028, presented at the 69th EAGE Conference and
Exhibition, London, June 11–14, 2007.
for the marine environment is generating
considerable interest among geophysicists, and
these techniques may hold the key to detecting
properties that elude seismic surveys.
Another area of potential improvement lies in
enhancing the data content in seismic
recordings. Low frequencies not contained in
most seismic data have to be obtained or
modeled from log data for inversion to absolute
rock properties. However, in areas far from wells,
this step may introduce unwanted bias into the
results. For example, when lithologies thin,
thicken, disappear or appear between wells, data
from wells might not form an accurate basis for
seismic models.
A new seismic data-acquisition technique is
being evaluated as a means of supplying
necessary low-frequency information in the
absence of log data. Known as over/under

60 Oilfield Review
 Conventional Data Over/Under Data
2,000
Water bottom Water bottom
2,200

2,400

2,600

2,800

3,000

3,200
Top Breydon Top Breydon
3,400
Base Breydon Base Breydon
3,600

3,800
Two-way time, ms

4,000
Top Balder Top Balder
4,200 Top Flett Top Flett
Top basalt Top basalt
4,400

4,600

4,800

5,000

5,200

5,400

5,600

5,800

6,000

6,200
–65 Amplitude 65
6,400

6,600
1,030 3,261 1,030 3,261
CDP CDP
> Seismic sections from a conventional survey with deep source and receivers (left) and an over/under survey (right). The over/under survey shows
significant signal strength from deep reflectors below the basalt. In the conventional survey, the basalt blocks the penetration of seismic energy.

acquisition, the technology effectively eliminates 1.00

the gaps in seismic bandwidth that plague most

0.75
surveys.22 The additional low frequencies Over/under
Amplitude

contained in over/under data have been shown to Conventional

0.50 Acoustic impedance
improve imaging of deep reflectors (above). The CDP 5, raw
low frequencies, often below 6 Hz, are also useful Acoustic impedance
0.25 CDP 5, filtered
for enhancing inversion.23
Modeling has been used to study the impact 0
of these additional low frequencies on seismic 0 1 2 3 4 5 6 7 8 9 10 11 12

inversion.24 The starting point is a wedge- Frequency, Hz

shaped acoustic impedance model with
reservoir intervals of varying thickness
(previous page, bottom left). Two synthetic
seismic sections are constructed: one with a
wavelet extracted from a conventional survey
and the other with a wavelet extracted from an
over/under survey (right). In essence, the first
> Spectra of conventional and over/under wavelets. The over/under wavelet
(green) is richer in low frequencies, especially from 3 to 6 Hz, than the
conventional wavelet (dark blue). The frequency content of the synthetic
acoustic impedance log at CDP 5 of the wedge model is shown in brown. A
low-pass filtered version of this log (gold) formed the background model for
inversion of the synthetic seismic sections.

Spring 2008 61
 Conventional Data Over/Under Data
CDP CDP
1 5 11 21 31 41 1 5 11 21 31 41

Two-way time, ms 1,600

1,700

1,800

Acoustic impedance

4,750 m/s . g/cm3 7,750

1,900

> Inversion of synthetic conventional and over/under data. Both datasets were inverted using a
background model comprising a filtered version of the acoustic impedance log at CDP 5. The
over/under acoustic impedance section (right) delivers a wedge-shaped result that more closely
matches the well information than does the conventional acoustic impedance section (left). The
over/under version maps the low acoustic impedances (green) of the reservoir, which thickens to the
right, and also produces a better match with the high acoustic impedance zones (yellow and red)
below the reservoir, which also thicken to the right.

synthetic section has the frequency content of
a conventional survey, and the second synthetic
section has the enhanced frequency content of
an over/under survey.
These synthetic seismic sections were inverted
using a single low-frequency background model.
The background model was created by low-pass
filtering (between 0 and 3 to 4 Hz) the acoustic
impedance from one well. This simulates an
exploration setting in which data from only
one well are available for constraining the
inversion model.
Comparing inversion results of the conven-
tional and the over/under sections shows that the
acoustic impedances from the over/under section
correlated much better with acoustic imped-
ances “measured” at wells, and therefore
matched the actual model better than did the
results using the conventional data as input
(above). The addition of data in the range from 3
to 6 Hz, supplied by the over/under technique,
made a significant difference in the inversion,
returning reliable rock-property information
although log data were sparse.
Seismic data with a large bandwidth and high
positioning accuracy also allow detection and
measurement of minute stress effects in 3D and
4D seismic data.25 For example, the effects of
subsidence-induced stress have been seen in the
properties of S-wave velocities measured by a
multicomponent survey in the North Sea (next
page). Seismic inversion can potentially be used
to infer spatial and temporal stress changes in
the subsurface from seismic data.26
Schlumberger geophysicists envision the use of
seismic inversion for determining the triaxial

25. Olofsson B, Probert T, Kommedal JH and Barkved OI:
“Azimuthal Anisotropy from the Valhall 4C 3D Survey,”
The Leading Edge 22, no.12 (December 2003): 1228–1235.
Hatchell P and Bourne S: “Rocks Under Strain: Strain-
Induced Time-Lapse Time Shifts Are Observed for
Depleting Reservoirs,” The Leading Edge 24, no. 12
(December 2005): 1222–1225.
Herwanger J and Horne S: “Predicting Time-Lapse
Stress Effects in Seismic Data,” The Leading Edge 24,
no. 12 (December 2005): 1234–1242.
Herwanger J, Palmer E and Schiøtt CR: “Anisotropic Sayers CM: “Monitoring Production-Induced Stress
Velocity Changes in Seismic Time-Lapse Data,” presented Changes Using Seismic Waves,” presented at the 74th
at the 75th SEG Annual International Meeting and SEG Annual International Meeting and Exposition,
Exposition, San Antonio, Texas, September 23–28, 2007. Denver, October 10–15, 2004.
26. Sarkar D, Bakulin A and Kranz RL: “Anisotropic Inversion
of Seismic Data for Stressed Media: Theory and a
Physical Modeling Study on Berea Sandstone,”
Geophysics 68, no. 2 (March–April 2003): 690–704.

62 Oilfield Review
 Layer 1 Displacement Layer 1 Effective Stress Change
A W1
W2 W3
W4 B X C X
X
Y

W1 W4

W2 W3

Y
2,500 m 2,500 m

150 3,650 0
Depth, m Downward displacement, m 0.2

Modeled Shear-Wave Splitting Observed Shear-Wave Splitting

D X E

W1 W4
Y

W2 W3

Fast shear-wave
2,500 m 2,500 m polarization

0 4
Subsidence, m
> Modeled and observed subsidence-induced shear-wave splitting in the shallow subsurface. A 3D geomechanical model (A)
was constructed to investigate the effects of subsidence of a shallow layer (Layer 1, dark blue) caused by compaction of a
deeper reservoir (green intervals in Wells W1, W2, W3 and W4) under production. The resulting ground displacement in the
shallow subsurface causes a nearly circular subsidence bowl (B). Changes in effective stress associated with the modeled
deformation (C) are greatest in the center of the bowl. These stress changes give rise to elastic anisotropy, which in turn
causes shear-wave splitting, a phenomenon in which two orthogonally polarized shear waves propagate at different speeds.
The largest shear-wave splitting occurs at the flanks of the subsidence bowl (D), where the difference between the horizontal
stresses is largest. At the center of the subsidence bowl, where horizontal stress changes are large but isotropic, shear-wave
splitting is minimal. The azimuth of the bars shows the polarization direction of the fast shear wave, and the length of each
bar is proportional to the time lag between fast and slow shear waves. Observed shear-wave splitting in a subsidence bowl
over a compacting North Sea reservoir (E) follows a pattern similar to the modeled phenomenon.

stress state of the reservoir and overburden as a
function of time. This knowledge can be used to
plan well trajectories and anticipate wellbore
failure and rock damage. Characterizing
mechanical properties of the overburden and
monitoring stress changes over time open a new
field of application for seismic inversion.
Seismically derived rock and fluid properties
are playing an increasing role in characterization
of geological models, and therefore extend
naturally into the domain of the reservoir
production simulator. This characterization of
rock properties can be extended into the
overburden. The next steps in the progression of
seismic inversion will include increasing use of
reservoir and geomechanical simulation results
to generate starting models for inversion, and
vice versa. Closing this loop and operating in real
time on time-lapse data will take seismic
inversion far beyond reading between the lines to
reading between wells. —LS

Spring 2008 63

Frazer Barclay, Consulting Services Manager for
Schlumberger Data & Consulting Services (DCS), is
based in Perth, Western Australia, Australia. He began
his career in 1998 as a geophysicist with Western
Geophysical in London. He joined Ødegaard in 2000,
completing reservoir characterization projects for vari-
ous oil companies. In 2003, Frazer was appointed pro-
ject manager with Ødegaard in Kuala Lumpur and
transferred to Perth in 2005 as business manager.
When Ødegaard was acquired by WesternGeco, Frazer
assumed the position of Reservoir Seismic Services
manager. He has a BS degree in geology from
University of Glasgow, Scotland.
Mario Bellabarba works at the Schlumberger Riboud
Product Center (SRPC) in Clamart, France, where he
is the Product Champion for FUTUR* active, set-
cement technology. He began his career with
Schlumberger in 1998 as a cementing field engineer in
Venezuela. Prior to his current posting in France, he
held positions as an in-house engineer with Shell in
The Netherlands, cementing technical engineer in
western Siberia and cementing manager for eastern
Russia. Mario earned a BS degree in civil engineering
at the University of Waterloo, Ontario, Canada.
Rashmi Bhavsar has been Advisor and Materials
Métier Manager at the Schlumberger Reservoir
Completions Technology Center (SRC) in Rosharon,
Texas, USA, since 2007. His main research interests
include material selection, corrosion and erosion resis-
tance, failure analysis, welding development, swellable
elastomers and high-pressure, high-temperature mate-
rials. He joined Schlumberger in 1996 through the
Camco acquisition as director of materials and welding
engineering for Schlumberger Completions Systems,
Houston, and then served as principal engineer and
manager in materials engineering from 2000 to 2007.
Before joining Schlumberger, he worked as a metallur-
gical and quality assurance engineer in several
Houston companies. Rashmi holds an MS degree in
metallurgical engineering from The University of Texas
at El Paso, and a BS degree in metallurgical engineer-
ing from the University of Baroda, Gujarat, India.
Kenneth Brown is Principal Consultant with
Schlumberger DCS in Pittsburgh, Pennsylvania, USA,
where he manages the Gas Storage Technology prac-
tice. He supervises projects in underground gas-stor-
age (UGS) feasibility studies; full-field simulation
studies to evaluate UGS conversions; design, develop-
ment and optimization of UGS field operations; inven-
tory analyses; and assessment and modeling of UGS
deliverability. He also works on research projects
funded by industry consortia and governmental agen-
cies and provides expert witness testimony and litiga-
tion support involving UGS fields. In 1997, he joined
S.A. Holditch & Associates (acquired in 1997 by
Schlumberger) as a gas-storage engineer. Previously,
he was with Michigan Consolidated Gas Company.
Kenneth has also held positions as an independent
consultant, a reservoir engineer for Shell Oil
Company, Inc. (Bakersfield, California, USA) and a
64
Contributors
completion and production engineer for Marathon Oil
Company (Lafayette, Louisiana, USA). He received
BS and MS degrees in petroleum and natural gas
engineering from The Pennsylvania State University
in University Park.
Anders Bruun is Schlumberger Reservoir Seismic
Services Business Manager for Scandinavia, based in
Copenhagen, Denmark. He is responsible for sales
and marketing of reservoir characterization inver-
sion projects, and also lends his inversion expertise
to ongoing projects. Before this position, he was
Ødegaard business manager for Norway. He began
his career in 2000 as a processing geophysicist with
WesternGeco in Stavanger. Anders has an MS degree
in geophysics from University of Aarhus, Denmark,
and a diploma in business administration from
Copenhagen Business School.
Hélène Bulte-Loyer is a Development Engineer with
the Well Integrity Technology/Cementing group at
Schlumberger Riboud Product Center (SRPC) in
Clamart. She joined Schlumberger in 2003 as a chemi-
cal engineer at the Schlumberger Integrated
Productivity and Conveyance Center, Sugar Land,
Texas. She began in the Engineering Application
Department, working on Virtual Lab* geochemical
simulation software. Hélène moved to SRPC in 2004 to
be part of the FUTUR team and developed the techni-
cal side of the FUTUR technology. She obtained an MS
degree in chemistry from Ecole Nationale Supérieure
de Chimie de Lille, and an MS degree in materials
engineering from the Université de Nice-Sophia
Antipolis and Ecole Nationale Supérieure des Mines de
Paris, all in France.
Jose Camara Alfaro is Geophysical Interpreter,
Coordination of Design for Exploitation, at Pemex in
Tampico, Tamaulipas, Mexico. He earned an engineer-
ing degree in geophysical logging at the Instituto
Tecnológico de Ciudad Madero, Tamaulipas. Jose has
27 years of experience in geophysical data acquisition,
interpretation and subsurface characterization.
Keith W. Chandler is Senior Vice President, Earth
Sciences and Reservoir Operations for Falcon Gas
Storage Company in Houston. His current assignment,
which began in 2003, includes planning for reservoir
development of gas-storage fields for UGS operations.
A geologist for 52 years, he began his career in 1956
with Stanolind Oil & Gas. He also spent 10 years on
mine development of iron, copper, gold, water and
mineral exploration in Australia, the western USA and
Canada. Keith has been involved in exploration, devel-
opment and evaluation, working as an independent in
several depositional basins in the USA and interna-
tionally. He holds a BS degree in geology from
Oklahoma State University in Stillwater, USA.
Anthony Cooke, Business Manager for Schlumberger
Reservoir Seismic Services in the UK, is based in
Aberdeen. He is responsible for business development
and marketing of reservoir seismic opportunities in
the UK and Ireland. He joined WesternGeco in 2000 as
a reservoir geoscientist based in London, where he
worked on a variety of reservoir characterization stud-
ies for reservoirs in the North Sea, North Africa,
Middle East, North America and the Gulf of Mexico. In
2004, he moved to WesternGeco Reservoir Seismic
Services in Stavanger. Anthony received his BS degree
in geology from the University of Durham, England,
and his MS degree in petroleum geology from Imperial
College of Science, Technology and Medicine,
University of London.
Dennis Cooke is the Chief Geophysicist for Santos
in Perth, Western Australia. He has been with the
company for nine years. Before joining Santos, he was
a development and exploration geophysicist for Arco
in Indonesia, Alaska and other areas of the USA. His
current technical interests are reservoir characteriza-
tion, imaging and stochastic modeling and inversion.
Dennis has a PhD degree from the Colorado School of
Mines, Golden, USA.
Benoit Froelich, who is an Advisor at the
Schlumberger Riboud Product Center in Clamart, is
currently involved in the development of an acoustic
communication system for downhole well testing.
Since 1971, he has been active in the development of
tools such as the Isolation Scanner* cement evaluation
service and other tools for measuring and evaluating
well structures and their stability. His research has
taken him to France, Japan and the USA. Prior to his
most recent appointment, he was assigned to Cased
Hole Products at SRPC from 2000 to 2007. Benoit
earned a PhD degree in physical chemistry at the
University of Paris and an undergraduate degree at
Ecole Supérieure de Physique et de Chimie
Industrielles (ESPCI), Paris.
Partha Ganguly is a Senior Research Scientist at
Schlumberger-Doll Research (SDR) in Cambridge,
Massachusetts, USA. His research interests focus on
the development and characterization of materials for
oilfield application. He joined Schlumberger in 2004
after a postdoctoral position at Massachusetts
Institute of Technology (MIT), Cambridge, working on
constitutive modeling of porous materials subjected to
chemical and mechanical loading. Partha holds PhD
and MS degrees in materials science and engineering
from the University of British Columbia, Vancouver,
Canada, and a BS degree in metallurgical engineering
from the Indian Institute of Technology in Kharagpur,
West Bengal, India.
Craig Gardner is a Consultant in Cementing and
Cement Team Leader at Chevron in Houston. After
receiving a BS degree in chemistry from the University
of Houston, he worked for a major drilling fluids com-
pany prior to joining Gulf Oil as a drilling supervisor in
1980. He is involved in Chevron’s worldwide cementing
operations through technical services, technology
development and training. Craig is a member of SPE,
API and ISO and is a former chairman of the API
Subcommittee on Well Cements.
Oilfield Review
Robert Godfrey, based in Gatwick, England, is cur-
rently WesternGeco Fracture Characterization
Champion, focusing on seismic applications for char-
acterizing naturally fractured reservoirs. Prior to this
position, he was inversion services manager, also
in Gatwick. Robert began his career in 1979 as a
geophysicist for Mobil in Dallas. In 1982, he became
a research geophysicist with Digicon in London. He
joined Geco in Calgary as geosupport manager in 1990,
and since then has held a variety of research and busi-
ness development positions. Robert has a BS degree
in geological engineering from University of British
Columbia, Vancouver, Canada, and a PhD degree in
geophysics from Stanford University, California.
Francisco Gonzalez Pineda is the Prospect
Coordinator for Activo Integral Burgos at Pemex,
Reynosa, Tamaulipas, Mexico. He has 24 years of expe-
rience in oil exploration, primarily in the Tampico
Misantla River basin. Francisco earned his geology and
engineering degrees at the University of San Luís
Potosí, Mexico.
Dominique Guillot, Schlumberger Well Services
Research Champion, specializes in defining new
cementing markets at SDR in Cambridge,
Massachusetts. He joined Dowell in 1981 in Saint-
Etienne, France, as development engineer and section
head and later became section head and product
team manager on projects related to well cementing.
In 1990, he became a cementing specialist in Houston,
working on introduction of new technology. He subse-
quently returned to Saint-Etienne as a cementing
engineering specialist, working on cement mixing and
cement job evaluation. He was section head of Process
and Software and Field Support at SRPC from 1994 to
1996. He has also served as cementing engineering
specialist in Clamart, as InTouch knowledge cham-
pion for the cementing segment and as well integrity
technology discipline manager. He is currently the
technical editor of the SPE Drilling & Completion
Editorial Review Committee. A civil engineer,
Dominique trained at the Ecole Nationale des Ponts
et Chaussées in Paris, and received a PhD degree in
geological engineering from Ecole Nationale
Supérieure des Mines de Paris.
James Hawkins, who is based in Midland, Texas, is
Borehole Geologist and Well Placement Engineer for
the Schlumberger DCS Group for United States Land
(West, Northeast and Central) GeoMarket* region.
James has worked in the Permian basin and Eastern
Shelf area for the past 18 years as production chemist,
environmental geologist, owner and president of a geo-
chemical oil exploration company and independent
prospecting geologist. He has BS and MS degrees in
geology from The University of Texas of the Permian
Basin in Odessa.
Jorg Herwanger, Senior Geophysicist at the
WesternGeco Houston Technology Center, specializes
in the analysis of field seismic data for stress-induced
time-lapse effects. His main focus is on providing
closer integration of time-lapse seismic imaging with
reservoir modeling and reservoir geomechanics. Jorg
began his career with WesternGeco in 2003 in Gatwick,
England, as a Marie Curie Research Fellow. He was a
senior geophysicist at Gatwick before moving to
Spring 2008
Houston in 2006. He received a PhD degree in geo-
physics from Imperial College, London, and an MS
degree in geophysics from Technische Universität
Clausthal, Germany.
John M. Hopper, President, CEO and cofounder of
Falcon Gas Storage in Houston, has significant experi-
ence in all facets of the energy business, including nat-
ural gas storage, gas transportation management and
optimization, storage project development, project
financing, trading and marketing, and energy law and
regulation. Before founding Falcon, he was the presi-
dent and CEO of Inventory Management and
Distribution Company, Inc. (IMDCI) from 1994 to
2000. Founded by John, IMDCI pioneered the commer-
cial optimization of gas storage and transportation
assets for local distribution companies during the
early phases of pipeline unbundling. He began his
career with Pennzoil Company in 1976 as a tax accoun-
tant. In 1980, as co-owner and cofounder of Ash
Exploration Company, he was active in generating oil
and gas drilling prospects and property acquisition
plays in Texas, Arkansas, Louisiana, Mississippi,
Michigan and New Mexico, USA. He was a member of
the Butler & Binion law firm in its Houston and
Washington, DC, offices from 1985 to 1989, specializing
in oil and gas transactions and energy regulation. He
held various executive positions with Tejas Power
Corporation (1989 to 1994) before he founded IMDCI.
John holds a JD degree from South Texas College of
Law in Houston, and a BBA degree from The University
of Texas at Austin.
Alan Humphreys, based at SDR in Cambridge,
Massachusetts, is a Senior Research Scientist in the
Materials and Mechanical Sciences Department. He
has a PhD degree in metallurgical engineering from
the University of Birmingham and an MA degree in
materials science from the University of Oxford, both
in England. He has worked as a materials surveyor at
Lloyd’s Register, researching the structural integrity of
ship steels, and also as a postdoctoral fellow in materi-
als at McGill University, Montreal, Quebec, Canada.
His research at Schlumberger focuses on tribology, or
the mechanisms of surface interactions (such as fric-
tion and wear), under downhole conditions. Alan has
been coleader of the Schlumberger Eureka Materials
community since 2006.
Sylvaine Le Roy-Delage is Schlumberger Project
Manager and Principal for Well Integrity Technology at
SRPC in Clamart. She has managed the FUTUR project
from concept to commercialization. She is now in
charge of extension of the FUTUR platform. She began
her career at Schlumberger in 1990 at Melun, France,
as a development engineer and later became a project
leader for reservoir fluid modeling. She was a develop-
ment engineer and project leader on projects at
Clamart including equipment for drilling fluids,
cement gas wells and cement sheath integrity
(FlexSTONE* and DuraSTONE* advanced cement
technologies). Sylvaine has a PhD degree in petroleum
science from the French Institute of Petroleum and
Ecole Nationale Supérieure du Pétrole et des Moteurs,
both in Rueil Malmaison, France, and also holds
degrees in chemical processing and chemical
engineering from École Nationale Supérieure
des Industries Chimiques in Nancy, France.
Dominic Lowden is the Reservoir Seismic Discipline
Manager, responsible for coordinating WesternGeco
Research & Engineering (R&E) efforts, and is based in
Gatwick, England. After receiving a BS degree in geol-
ogy from the University of Reading, England, and an
MS degree in geology from the University of Guelph,
Ontario, Canada, he worked as a petrophysicist with
EnTec Energy Consultants in London in 1984, later
becoming technical director of reservoir studies. He
joined Western Atlas in 1996, responsible for reservoir
seismic marketing, and then became business develop-
ment manager for WesternGeco in 2001. Since then,
Dominic has held several marketing and business
development management positions, and he was
recently responsible for the integration of the newly
acquired Ødegaard company into WesternGeco.
Taoufik Manai is Schlumberger Principal Reservoir
Engineer, responsible for deploying new reservoir
simulation and production technologies and for coordi-
nating service delivery for key and strategic accounts.
Based in Paris, his current role involves overseeing the
reservoir engineering practice and technical consul-
tancy on Avocet* Integrated Asset Modeler projects.
Taoufik has worked extensively on reservoir and
production engineering projects worldwide and has
contributed to the design of underground gas-storage
facilities. He has an MS degree in mathematics from
the Faculté des Sciences de Tunis in Tunisia, an MS
degree from Ecole des Mines de Nancy, France, and a
PhD degree in petroleum engineering from Université
Pierre et Marie Curie, Paris.
Steve McHugo is a Principal Geophysicist in the
WesternGeco Integrated Solutions Group in Gatwick,
England. In his present role as facilitator for Q* pro-
jects, he provides a link between acquisition, process-
ing and reservoir services groups to identify
opportunities for delivering integrated geophysical
solutions to clients. After obtaining a degree in applied
physics from Middlesex Polytechnic in England, Steve
joined Geophysical Services Inc. in 1975 as a process-
ing geophysicist. During his career, he has held various
positions relating to specialized processing of land and
marine seismic data. His main area of interest and
expertise is in developing strategies and workflows for
stratigraphic inversion of seismic data, leading to seis-
mically guided reservoir description.
Gareth H. McKinley is Professor of Mechanical
Engineering, Director of the Hatsopoulos Microfluidics
Laboratory and Director of the Program in Polymer
Science and Technology in the Department of
Mechanical Engineering at MIT. He is also executive
editor of Journal of Non-Newtonian Fluid Mechanics.
His main research interests are rheology, non-
Newtonian fluid dynamics, hydrodynamic instabilities
and extensional rheometry. He is currently working in
the Materials Group at SDR in Cambridge,
Massachusetts, as an advisory consultant during his
sabbatical from MIT. Gareth received a BA degree in
natural sciences and MEng and MA degrees in chemi-
cal engineering from the Downing College, University
of Cambridge, England, and a PhD degree in chemical
engineering from MIT.
65
Nevio Moroni is the Cementing Technical Leader in
the Drilling Technology Department of Eni E&P.
Based at Eni headquarters in Milan, Italy, he joined
the company in 1977 and is an authority on drilling
fluids and cementing within the company. Nevio holds
a degree in chemistry from the Technical Institute S.
Cannizzaro in Milan.
Andrea Murineddu is a Reservoir Geophysicist with
Schlumberger in Stavanger. He is currently the tech-
nical manager of the Reservoir Seismic Services team.
He has been involved in several inversion projects,
including the time-lapse prestack inversion of five vin-
tages (1992, 2001, 2003, 2004 and 2006) of Norne field
surveys. Before moving with Schlumberger to Gatwick
in 2003, Andrea was a geophysicist with EnTec Energy
Consultants and then with WesternGeco in Isleworth,
England. He has a BS degree in electrical engineering
from the University of Cagliari, Italy.
Vladimir Onderka is the Manager of Engineering and
Underground Gas Storage Project Development in
RWE Transgas Net, s.r.o. He also chairs Working
Committee 2 (WOC 2) for UGS of the International
Gas Union. He has more than 20 years of experience
as a reservoir engineer, starting with VUGI (Research
Institute of Geological Engineering) in Brno, Czech
Republic. Vladimir was later the technical director of
GeoGas, a.s., and Duke Engineering & Services. Since
the beginning of his career he has been involved in
UGS development including reservoir engineering,
well testing, simulation and expert systems on UGS
for RWE in the Czech Republic, reactive modeling
of flow and transport for Diamo uranium mining
company and other major projects. Vladimir is a
graduate of Charles University, Prague, with an
MS degree in geochemistry and a PhD degree in
geochemistry and applied geophysics.
Hüseyin Özdemir is Principal Reservoir Geophysicist
with Schlumberger Reservoir Seismic Services in
Gatwick, England, promoting prestack inversion to
rock and reservoir properties for all marine and land
seismic projects. He joined Schlumberger in 1985 as a
division geophysicist in Kuwait and the UAE, and five
years later moved to the Geco-Prakla reservoir char-
acterization group in Norway. Before assuming his
current position, he held various senior scientist and
leadership positions within WesternGeco and
Schlumberger. Prior to working for Schlumberger,
Hüseyin was an associate professor in applied geo-
physics at the Istanbul Technical University in Turkey,
and a consultant to the Turkish Petroleum
Corporation. He has a PhD degree in geophysics from
Imperial College of Science, Technology and
Medicine, University of London. He also holds a BS
degree from the University of Istanbul and an MS
degree from University of Birmingham, England, both
in geophysics.
Slavo Pastor is Schlumberger Oilfield Services
GeoMarket Manager for south Russia, based in
Tyumen. He began his career with Dowell
Schlumberger in 1993 as a field engineer trainee. He
66
has held managerial positions within Integrated
Project Management, Oilfield Services and
Schlumberger Information Solutions groups in Turkey,
Romania, Bulgaria, Poland, Czech Republic, Germany
and the UK. Slavo received an MBA degree from
Erasmus University, Rotterdam, The Netherlands. He
also holds an MS degree in computer engineering and
a BS degree in mining engineering, both from the
Technical University of Košice, Slovakia.
Frederic Pauchet has been an Engineer in the
Mechanical Technology Group at Schlumberger
Riboud Product Center (SRPC) in Clamart, France,
since 1994. He joined Schlumberger in 1988 and held
various engineering positions at SRPC. His research
interests include materials science, mechanics and
chemistry. Frederic has an engineering diploma from
École Nationale Supérieure de Chimie de Paris, France.
Stephen Pickering, Project Manager, Schlumberger
Career Management Initiative, and Geophysics
Domain Career Leader in Gatwick, England, focuses
on the use of seismic studies to enhance reservoir
management. He began his career as a seismic data
analyst with Western Geophysical. In 1981, he joined
Hamilton Oil as seismic interpreter on North Sea
acreage, including the Bruce field appraisal. From
1989 to 1995, he was UK and Europe exploration man-
ager for Hamilton Oil. After a move to BHP Petroleum,
he became manager of exploration technology with
primary responsibility for prospect evaluation and
portfolio management. He rejoined Western
Geophysical in 1999. In 2007, Stephen was president
of the Petroleum Exploration Society of Great Britain;
he was technical chairman of the society’s biennial
PETEX-2004 Conference and Exhibition. Stephen
received a BS degree in geology and an MS degree in
stratigraphy from University of London, and an MBA
degree from Open University in Milton, England.
Andreas Rasmussen, who is based in Stavanger,
is a Schlumberger Reservoir Seismic Services
Geophysicist. Recently, he worked on a 3D amplitude
variation with offset (AVO) multicomponent inver-
sion project for StatoilHydro. In 2006, he joined
Ødegaard in Copenhagen, where his main responsi-
bilities included seismic inversion of 2D, 3D and 4D
data using the ISIS* reservoir characterization soft-
ware. Andreas has a BS degree in geology and an
MS degree in earth sciences, both from University of
Aarhus, Denmark.
Klaus Bolding Rasmussen is Research Manager for
the WesternGeco Reservoir Seismic R&E group in
Copenhagen. As the developer of all the Ødegaard
inversion-related algorithms, he continues to support
commercial inversion projects performed by Reservoir
Seismic Services using ISIS software. He began devel-
oping algorithms for Ødegaard in 1991. Klaus has MS
and PhD degrees in electronics engineering from the
Technical University of Denmark in Lyngby. He also
has a bachelor’s degree in business administration
from Copenhagen Business School.
Ron Roberts is a Senior Geophysical Advisor for
Apache Canada, based in Calgary. Prior to this assign-
ment, he was the geophysical manager for Apache
Egypt and was based in Cairo for six years. He has also
been a geophysical interpreter for both Amoco
Production Company and Texaco North America, hold-
ing positions in New Orleans, Houston, London and
Denver. Ron has been championing the use of inver-
sion technology for the last 10 years. He received a BS
degree in geophysics from University of Delaware, an
MS degree in applied physics from the University of
New Orleans and an MBA degree from the University
of Denver.
Agathe Robisson is a Senior Research Scientist at
SDR in Cambridge, Massachusetts, where she is devel-
oping and modeling new polymeric materials for high-
temperature applications. She joined Schlumberger in
2000 as a mechanical and materials engineer at the
Schlumberger Riboud Product Center in Clamart.
Agathe earned a PhD degree in materials at Ecole des
Mines de Paris, France, and an engineering diploma
from Institut National des Sciences Appliquées in
Lyon, France.
Darren Salter, Senior Geophysicist with Santos, is
currently working on Australian offshore development
projects. His previous experience includes exploration
and development onshore Australia. Darren has been
with Santos for 10 years and has a Bachelor of Applied
Science degree in geology from the University of
South Australia, Mawson Lakes.
Lowell Thronson is the Vice President of Engineering
and Construction for Falcon Gas Storage Company. He
has more than 35 years of experience in engineering
and senior-level management for upstream and mid-
stream oil and gas projects, providing expertise, lead-
ership and direction for engineering and design
services; project management services; development
planning and marketing of engineering and construc-
tion services to domestic and foreign energy-related
companies. Before joining Falcon, he founded and was
principal engineer of TECORP International, PLLC,
which provided development planning and engineer-
ing for many natural gas–storage and oil and gas pro-
jects around the world including underground natural
gas-storage projects and related pipeline facilities.
From 1983 to 2001, Lowell was the founder, president
and chief executive officer of Thronson Engineering
Corporation (a Houston 100 Company) and Thronson
Internacional de Venezuela, C.A. (TIVENCA), where
he provided engineering, operations management,
and planning and project development for the domes-
tic and international oil and gas production, process-
ing and petrochemical industries. He is a registered
professional engineer in petroleum and natural gas
sciences in Texas. Lowell received his BS degree in
mechanical engineering from The University of Texas
at Austin.
An asterisk (*) is used to denote a mark of Schlumberger.
Casing Drilling® is a registered trademark of Tesco Corporation.
Oilfield Review
Huilin Tu is a Research Scientist at SDR in
Cambridge, Massachusetts, where she is developing
high-temperature polymers and swellable elastomers.
She joined Schlumberger in 2006. Huilin has BS and
MS degrees in chemistry from Peking University,
China, and a PhD degree in materials science and
engineering from the University of Illinois at Urbana-
Champaign, USA.
Nitin Y. Vaidya is a Group Leader for Polymers at the
Schlumberger Reservoir Completions Technology
Center (SRC) in Rosharon, Texas. He is an expert in
swellable polymer technologies, high-temperature
polymer applications, elastomer formulation and pro-
cessing (molding and extrusion), structure-property
relationships in polymers and thermal analysis of poly-
mers. He joined Schlumberger in 2000. Before moving
to SRC in 2004, he was a polymer chemist and senior
R&D engineer with the Schlumberger Lawrence
Product Center in Kansas, USA. Nitin holds a PhD
degree in polymer engineering from the University of
Akron, Ohio, USA; an MS degree in chemical engineer-
ing from the Indian Institute of Technology, Mumbai,
Maharashtra; and a BS degree in chemical engineering
from the University of Mumbai.
Robert van Kuijk, Program Manager for iCoil* optical
fiber installed CT string, is assigned to Schlumberger
Well Services at SRPC in Clamart. He joined
Schlumberger in 1993 as project engineer at Prakla
Seismos in Uetze, Hannover, Germany. Before assum-
ing his current management responsibilities in 2007,
he held posts as senior project engineer, section man-
ager and project leader with Geco-Prakla Oslo
Technology Center (OTC) in Norway and Geco-Prakla
in Sugar Land, Texas. Robert has also been project
manager for Wireline at WesternGeco OTC and project
manager for the Isolation Scanner tool at SRPC. He
has an MS degree in electrical engineering and com-
puter science from the University of Twente in
Enschede, The Netherlands, where he specialized in
mechatronics.
Stefano Volterrani is MultiMeasurement Reservoir
Description and Geology Product Champion for
WesternGeco in Houston. He began his career as a
senior processing analyst with Western Geophysical in
the UK in 1985. In 1988, he became a technical man-
ager at EnTec Energy Consultants. He moved to Egypt
in 2000 as technical supervisor of integrated services
and held various management and product develop-
ment positions until he assumed his current position
in 2006. Stefano holds a BS degree in geology from the
University of Pisa, Italy.
Joachim Wallbrecht joined BEB Erdgas und Erdöl 30
years ago as a petroleum engineer. He is in charge of
Asset Management for BEB’s underground gas storage
business. Based in Hannover, Germany, he has worked
in various oil and gas reservoir engineering projects, as
well as in gas demand and supply balancing. During
the last 17 years, Joachim has planned and developed
new UGS facilities and has been active in gas reservoir
and gas-storage management. He has an MS degree in
petroleum engineering from Technische Universität
Clausthal, Germany.
Spring 2008
Nathan Wicks is a Research Scientist in Mechanical
and Materials Sciences at SDR in Cambridge,
Massachusetts. His research interests are in drilling
dynamics software, deployable structures and
mechanical and materials sciences. He joined
Schlumberger in 2005. Nathan received BS and MS
degrees in mechanical engineering from Yale
University, New Haven, Connecticut, USA. He also
received an MS degree in applied mathematics and a
PhD degree in solid mechanics from Harvard
University, Cambridge, Massachusetts.
Augusto Zanchi is a Drilling and Completion
Engineer for Stogit, the Eni company responsible for
gas storage. Currently based in Crema, Italy, Augusto
joined Eni in 1990 and moved to Stogit when it was
created in 2001. In addition to his current posting in
Italy, he has worked for Eni in West Africa. Augusto
holds a degree in chemistry from the Technical
Institute G. Galilei in Milan.
Georg Zangl is Technical Manager of Schlumberger
Information Solutions (SIS) Technology Center in
Baden, Austria. He is responsible for the consulting
group of SIS Austria as well as the petroleum engineer-
ing development concepts of DECIDE!* data mining
based production optimization software. With more
than 16 years of experience, he has worked on various
aspects of reservoir engineering. Georg has spent
many years on the development and commercializa-
tion of reservoir simulation–related software and, since
1996, has been an advocate of data mining technologies
in the petroleum industry. In the last five years, he has
been involved in the development of reservoir surveil-
lance systems and applications. The results of his work
have been published in several international confer-
ences and printed media. Georg is the author of the
book Data Mining: Applications in the Petroleum
Industry. He earned an MS degree in petroleum engi-
neering at the University of Leoben, Austria.
Smaine Zeroug is Technology Center Marketing
Manager at the Schlumberger Riboud Product Center
in Clamart. Before this, he managed the production of
the book Well Evaluation Conference–Algeria 2007, a
two-year project carried out with Algeria’s national oil
company, Sonatrach. He moved to that project from
SDR in Cambridge, Massachusetts, where he began his
career in 1992. There, he held several positions includ-
ing senior scientist, principal scientist and program
manager of the Modeling and Inversion Applications
Program. His research activities have centered on the
development of new-generation wireline ultrasonic
imaging and borehole sonic logging technologies.
Smaine holds a BS degree in solid-state physics from
the University of Algiers (USTHB), Algeria, and MS
and PhD degrees in electrophysics from Polytechnic
University, New York City. He received a Schlumberger
in-company MBA degree from Erasmus University,
Rotterdam, The Netherlands. The author of numerous
technical papers and holder of several patents, Smaine
is a senior member of the IEEE and has served as asso-
ciate editor of the IEEE Transactions on UFFC
(Ultrasonics) since 2002.
An asterisk (*) is used to denote a mark of Schlumberger.
67
Coming in Oilfield Review
New Dimensions in Resistivity.
Resistivity, perhaps the most funda-
mental of petrophysical measure-
ments, has been used to identify
oil and gas deposits for more than
80 years. Triaxial resistivity, made
possible because of increased pro-
cessing power and sensor develop-
ments, is changing the way hydro-
carbon reservoirs are evaluated.
Case studies from around the world
will help demonstrate several impor-
tant applications of this new tech-
nology: accurate resistivity measure-
ments in highly dipping beds or in
deviated wells, laminated sand anal-
ysis using vertical and horizontal
resistivities, and structural formation
dip from induction measurements.
Multistage Hydraulic Fracturing.
Today’s new oil and gas production
is more likely to come from improved
recovery of proven reserves than
from development of virgin reservoirs.
For this reason, the upstream indus-
try is drilling an increasing number
of high-angle wells to reach remote
but previously uneconomical
deposits, to pierce thin reservoir
hydrocarbon-bearing sands and
shales and to improve reservoir
drainage. These same goals of
increased production and greater
ultimate reserve recovery have also
spurred renewed interest in hydraulic
fracturing. This article looks at the
confluence of these trends and the
role of coiled tubing in fracturing
older wells, openhole completions
and wells with multiple zones.
Groundwater Management.
Groundwater constitutes an over-
whelming share of our freshwater
resources. However, rapid population
growth, rising standards of living,
changing weather patterns and pol-
lution are imposing great demands
on groundwater supplies. Managing
these reserves is key to achieving a
sustainable supply of fresh water.
Advanced logging, sampling and
modeling techniques—some bor-
rowed or modified from established
oilfield applications—are proving
vital for evaluating and managing
this precious resource.
68
NEW BOOKS
The emphasis on storytelling over Super Volcano captures the
concepts, however, creates pedagogical essence, the excitement, and the deep
challenges. For instance, the book and far-reaching influence of the
is free of pictures and diagrams: a world's greatest heat anomaly.
puzzling omission.
Breining writes for the layperson
Similarly, many of the most with enthusiasm and informality, bring-
exciting aspects of general relativity ing the subjects to life with copious
get short shrift. quotes from naturalists, field leaders,
and volcanologists.… A rare read!…
Caveats aside, Girifalco is a fluid Highly recommended.
writer, and his stories are compelling.
This book about the force of gravity Grose TLT: Choice 45, no. 7 (March 2008): 1189.
The Universal Force: Gravity— has its feet firmly on the ground.
Creator of Worlds Caroll S: Nature 451, no. 7175
Louis A . Girifalco (January 10, 2008): 130.
Oxford University Press
198 Madison Avenue
New York, New York 10016 USA
2007. 320 pages. $46.95
ISBN 0-19-922896-5
Understanding gravity has challenged
generations of great scientists, from
Galileo to Einstein. This book describes
the achievements of those who studied
the nature of gravity, its origins and its
effects. The author weaves an interest- Notes from the Holocene:
ing narrative from the complex history A Brief History of the Future
of this field. Dorion Sagan
Chelsea Green Publishing Company
Contents: Super Volcano: The Ticking Time 85 North Main Street, Suite 120
Bomb Beneath W hite River Junction, Vermont 05001
• The Seeker
Yellowstone National Park USA
• The Giants Greg Breining 2007. 224 pages. $14.95
• The First Modern Giant Voyageur Press ISBN 1-933-39232-0
729 Prospect Avenue
• The Grid
P.O. Box 1 Operating on the precept that the uni-
• The Universal Force Osceola, Wisconsin 54020 USA verse is much stranger than we might
• The Laws 2007. 256 pages. $24.95 imagine, Sagan—son of noted scientists
• The System of the World
ISBN 0-7603-2925-7 Carl Sagan and Lynn Margulis—uses
his knowledge of philosophy, science
• Force and Mass Below Yellowstone National Park is and sleight-of-hand magic to explore
• Two More Giants one of the world’s five super volcanoes. some of the deepest questions of life.
Science writer Breining provides a solid He provides fresh insights on the nature
• Ether
introduction to modern volcanology of technology, the prognosis for human-
• The Genius in this study of super volcanoes in ity, the living nature of our planet and
• Time and Space general, and the Yellowstone caldera— an explanation on why our universe may
potentially the most deadly active be just one of an infinite number.
• It Really Is True
volcano on the planet—in particular. In
• The Space-Time Continuum addition to describing what is likely to Contents:
• Time Warps and Bent Space happen when Yellowstone erupts, this • Earth
book describes how volcanoes function
• It Stands Alone • Water
and includes a time line of famous
• This Too Is True volcanic eruptions throughout history. • Air
• Crunch • Fire
Contents:
• Beyond Existence • Afterword: Twelve Mysteries
• The Big Blast
• Absolute Space? • Index
• Yellowstone Today
• Infinity
• How Weird Can It Get?
• Natural Wonders
This whimsical book takes on the
• Evolving Geology
• Scientific Truth big questions of the universe.… In
• The Yellowstone Puzzle
• The Meaning of Why the end, Sagan does not necessarily
answer these big questions, but, more
• Distant Death
• Final Comments importantly, he encourages readers to
• Most-Super Volcanoes be thoughful and creative in their own
• The Deadliest Volcanoes explorations of truth.
For anyone interested in the more
• The Next Big Blast Oberle GD III: Choice 45, no. 7 (March 2008): 1190.
human side of science, this work is a
valuable contribution. • Glossary, References, Index
Oilfield Review