You are on page 1of 20

Rocks Matter: Ground Truth in Geomechanics

Stress and pressure act upon every reservoir, wellbore and completion. Drilling,
production and injection processes modify these stresses and pressures, sometimes
to the operators detriment. Through advances in geomechanical measurements,
modeling and monitoring, E&P companies are now able to predict and mitigate the
effects of stress and pressure as they change throughout the life of their fields
from appraisal to abandonment.

John Cook
Cambridge, England
Ren A. Frederiksen
Klaus Hasbo
Hess Denmark ApS
Copenhagen, Denmark
Sidney Green
Arnis Judzis
J. Wesley Martin
Roberto Suarez-Rivera
Salt Lake City, Utah, USA
Jorg Herwanger
Patrick Hooyman
Don Lee
Sheila Noeth
Colin Sayers
Houston, Texas, USA
Nick Koutsabeloulis
Robert Marsden
Bracknell, England
Morten G. Stage
DONG Energy
Hrsholm, Denmark
Chee Phuat Tan
Kuala Lumpur, Malaysia
For help in preparation of this article, thanks to Ben Elbel,
Dallas; Ian Walton, Rosharon, Texas; and Smaine Zeroug,
Clamart, France. Thanks also to Hess Denmark ApS, DONG
Exploration and Production A/S, Noreco ASA, and Danoil for
contributing their North Sea case study.
ECLIPSE, Petrel, TerraTek, UBI (Ultrasonic Borehole Imager)
and VISAGE are marks of Schlumberger.


Change the stress on a rock and it deforms,

altering its volume and geometry, as well as the
paths of fluid flow within. The stress regime of a
formation can be impacted by multiple factors,
including rock type, depositional settings, regional
tectonics, episodes of erosion or uplift, local
seismic disturbances and even tidal variations.
The influence of such stressors is further
complicated by differences in rock fabric.
The manner in which formations react to
changing stress is becoming a focus of increasing
interest to E&P companies. In-situ reservoir
stresses, having reached equilibrium over
geologic time, are altered by the process of
drilling, production and injection. If drilling- or
production-induced changes in stress are not
anticipated, the challenges and costs of managing
a prospect can far exceed an operators initial
expectations. To characterize stress, strain and
deformation in their reservoirs, E&P companies
employ the discipline of geomechanics. This
wide-ranging field applies solid and fluid
mechanics, engineering, geology and physics to
determine how rocks and the fluids they contain
respond to force or to changes in stress, pressure
and temperature caused by drilling, completion
and production.
In the past, most drilling and production
departments were not particularly attuned to
formation stresses and geomechanics; many
reservoirs were deemed technically straightforward and had undergone only limited

depletion. But declining resource volumes and

favorable oil prices are prompting operators to
drill deeper, more intricate well trajectories, at
the same time that new technologies are
extending the lives of mature fields. Operators
therefore are becoming more mindful of
geomechanics as they assess drilling and
production difficultiesespecially those who
endeavor to protect their investments in
expensive completions installed in highpressure, high-temperature, tectonically active
or ultradeepwater prospects.
Failure to appreciate the importance of geomechanics may have severe consequences.
Excessive mud loss, wellbore instability, casing
compression or shearing, reservoir compaction,
surface subsidence, sand production, fault
reactivation and loss of reservoir seal may all
be manifestations of stress changes within
a formation.
Some operators are forced to react to changes
in stress or rock fabric as they drill and produce
their wells. Others are more proactive. Through
core testing and geomechanical modeling of rock
strength, deformation and stress behavior, they
are engineering better wells and fields. These
efforts have recently been aided by newly
established centers of excellence for
geomechanics in Bracknell, England, and in
Houston and Salt Lake City, Utah, USA.
This article describes advances in geomechanics laboratory testing techniques,
stress-dependent reservoir simulations and

Oilfield Review

stress 1

Uniaxial compressive strength

Tensile strength

Minor effective stress 3

monitoring. Field studies performed at the

Schlumberger Geomechanics Laboratory Center
of Excellence and the Schlumberger Reservoir
Geomechanics Center of Excellence show how
this science is helping E&P companies optimize
drilling and production in increasingly
challenging reservoirs.

Autumn 2007

Stress in the Subsurface

The stresses acting on a formation can vary in
origin, magnitude and direction. Natural, in-situ
vertical stresses stem primarily from the weight
of overburden. Horizontal stresses also have a
gravitational component that may be enhanced
by tectonics, thermal effects and geological
structure. However, other factors such as litho-

logy, pore pressure and temperature influence

stress magnitude and orientation as well as the
degree to which rock responds to stress.
Stress, a measure of force acting on a given
area, is made up of normal and shear
components. Normal stress () is that which is
applied perpendicular to a plane or rock surface.
Shear stress () is applied along the face of the


> In-situ stresses and principal stresses. Stresses on a cube of material
buried in the earth are given the designation V, H and h, where V indicates
vertical, H indicates the direction of the larger horizontal stress, and h that of
the smaller horizontal stress. For simplicity, it is often assumed that these are
the principal stress directions, but the principal directions of stress can be
rotated significantly from these three axes. The principal stresses are
generally indicated as 1, 2 and 3, in decreasing order of magnitude. When
the principal stress directions do not coincide with the vertical and horizontal
directions, there will also be shear stresses on the cube faces in the
orientation shown.

Fracture point
Elastic field
Yield point


Ductile field


> Stress-strain diagram. Rocks that undergo elastic deformation store strain
energy as their volume changes. When the applied boundary stresses are
removed, the rock returns to its original state of deformation while the strain
energy returns to its original value. With application of greater stress, rocks
undergo inelastic deformation as nonrecoverable, internal structural changes
occur (starting at the yield point), such as tensile microcracking, grain crushing
or slippage at grain boundaries. These changes result in permanent volumetric
deformation, often referred to as plastic deformation. Higher stresses
eventually cause the rock to fail (fracture point), as exemplified by crushing or
fracturing of constituent grains and cement or by mineral dissolution.


plane. Mathematically, there is one orientation of

orthogonal axes defining the stress directions for
which the shear stresses are zero. That
orientation defines the principal axes of stress,
wherein applied stresses are strictly normal.
In situ, these orthogonal principal axes are
often assumed to be oriented vertically and
horizontally (left); however, this condition is
often not the case. The magnitude and
orientation of stresses in the earth change with
the structural dip of the formation, which can
rotate the orientation of principal stresses from
the vertical and horizontal orientations, as can
the presence of faults, salt diapirs, mountains or
other complex structures.1
In the earth, where deformation is restricted,
the three stress components are linked, and any
change of stress in one direction is accompanied
by changes in stress along the orthogonal axes.
For example, when continued deposition brings
about greater burial depths, the resulting
increase in overburden vertical stress can
generate changes in horizontal stress, depending
on the degree to which the formations are able to
spread out laterally. This response is generally
constrained by the presence of adjacent
formations that confine the rock deformation.
Differences in formation properties also impose
contrasts in stresses between adjacent lithologies. Furthermore, formation anisotropy can
result in greater lateral stress in one direction
than in another.
A body of rock responds to applied stress
through various modes of strain, or deformation,
causing changes in volume and shape, often
accompanied by changes in rock properties
(left). The spectrum of deformation ranges from
reversible, or elastic deformation, to permanent,
or plastic deformation, before eventually ending
in failure of the rock. Deformation caused by
compression, tension or shear can result in
compaction, extension, translation or rotation,
eventually ending in failure by shearing,
fracturing or faulting. In addition to the
magnitude of stress applied, a rocks response to
stress depends largely on rock type, cementation,
porosity and burial depth. In sandstones, the
size, shape and area of contact points between
individual rock grains influence deformation. In
limestones, the shape and strength of the
skeletal rock framework influence deformation.2
Small increases in stress generally cause a
small strain from which the rock may recover.

Oilfield Review

1. Addis MA: The Stress-Depletion Response of

Reservoirs, paper SPE 38720, presented at the SPE
Annual Technical Conference and Exhibition, San
Antonio, Texas, October 58, 1997.
2. Geertsma J: Land Subsidence Above Compacting Oil
and Gas Reservoirs, paper SPE 3730, presented at
SPE-AIME European Spring Meeting, Amsterdam,
May 1618, 1972.
3. For more on stress paths: Crawford BR and Yale DP:
Constitutive Modeling of Deformation and Permeability:
Relationships between Critical State and
Micromechanics, paper SPE/ISRM 78189, presented at
the SPE/ISRM Rock Mechanics Conference, Irving,
Texas, October 2023, 2002.
Rhett DW and Teufel LW: Effect of Reservoir Stress
Path on Compressibility and Permeability of
Sandstones, paper SPE 24756, presented at the SPE
Annual Technical Conference and Exhibition,
Washington, DC, October 47, 1992.

Autumn 2007

Shear stress (Q): 1 3





failure sile












Beyond a certain point, it will deform plastically

or fail. The mode of deformation and failure is
dictated by the relationship between changes in
maximum and minimum stresses (right). This
relationship is called a stress path.3 In petroleum
geomechanics, the stress path (K) is
conventionally the ratio of change in effective
minimum horizontal stress to the change in
effective vertical, or overburden, stress from
initial reservoir conditions during fluid
pressure drawdown, more simply expressed as
K = 3/1. This can also be expressed in terms
of changes of shear stress (Q) and changes in
mean stress (P' ), as shown on the P'- Q diagram.4
A sufficiently low stress-path value implies
that the rock will fail in shear, generating a shear
plane. Shear strength increases as lateral
confining stress on the rock increases. Where
larger stress-path values are seen, the rock
undergoes compaction or reduction in porosity.
This is most common in soft, high-porosity rocks
such as chalk, porous sands and diatomite.5
When subjected to differential stresses, other
rocks, such as salts, will tend to flow over time to
reduce shear stresses and move towards
hydrostatic stress states.
To manage reservoirs, oil and gas companies
must contend with a variety of downhole
stressorsnot all of which are caused by overburden or tectonics. Pore pressure, temperature
differences and chemical interactions can also
affect localized perturbations in stress orientation and magnitude.
Stress and pore pressure are intrinsically
linked.6 In formation pore spaces, stress is
transmitted to liquids or gases in the form of
pressure. The magnitude of pressure applied in
any one direction is the same for all directions. If
a fluid is compressed, it reacts by exerting an
equal and opposite pressure outwards. Under
pressure, pore fluids often take up some of the
stress imposed on a formation. Thus pore

Near-elastic region



Mean effective stress (P'): (1 + 2 + 3) / 3

> Distortion and failure. Distinct modes of distortion and failure can be plotted
as a function of shear stress (Q) and mean effective stress (P' ). At relatively
low P' and high Q, rock failure typically occurs as localized shear along a
plane oriented at an angle to the principal stress axes. At relatively high P'
and low Q, rocks may undergo compaction or pore collapse. (Adapted from
Scott, reference 3.)

pressure is an important component of the net

stress applied to a body of rock.
Temperature is yet another contributor to the
overall stress regime. Temperature differences
between drilling fluids and downhole formations
will result in heat transfer between the two
media. Given the low thermal conductivity of
most rocks, these temperature variations
generate large strain gradients that may cause
severe fracturing and stress realignments. Since
thermal expansion of water in the pore space is
much higher than that in the rock matrix, the
heat transferred into a formation by drilling
fluid will generate a larger volume expansion of
the pore fluid and a corresponding increase in
pore pressure.7
Thermal expansion of the rock matrix under
constrained conditions will generate further

stress. A reduction in effective mud support is

often associated with an increase in pore
pressure. This reduction, together with the
thermal matrix expansion, will lead to a less
stable wellbore condition. Conversely, cooling
the formation may result in a more stable
condition because of decreased pore pressure
and tangential stress. The reduction of tangential stress may also lead to a lower hydraulic
fracture gradient, and, in extreme cases, the
tangential stress will become negative and
initiate hydraulic fracture.
Stress and pore pressure can also be affected
by interactions between rock and drilling fluid.
Shales, which account for the majority of drilled
sections in most wells, are particularly sensitive
to drilling fluids. Somewhat porous and usually
saturated with formation water, these rocks may
be susceptible to chemical reactions with certain

Scott TE: The Effects of Stress Paths on Acoustic

Velocities and 4D Seismic Imaging, The Leading Edge 26,
no. 5 (May 2007): 602608.
Teufel LW, Rhett DW and Farrell HE: Effect of Reservoir
Depletion and Pore Pressure Drawdown on In-Situ
Stress and Deformation in the Ekofisk Field, North Sea,
Proceedings of the 32nd US Rock Mechanics
Symposium. Rotterdam, The Netherlands: A.A. Balkema
(1991): 6372.
4. A relationship exists between the stress path, shear
stress and mean stress. While the stress path (K) can
be expressed as K = 3/1, shear stress (Q) is
expressed as (Q = 1-3), and effective mean stress (P' )
is [P' = (1+2+3)/3]. In laboratory uniaxial stress tests,
in which the minimum and intermediate principal stresses
are considered equal (2 = 3), the slope in the P'-Q
plane corresponding to the stress path K is given by this
equation, following Crawford and Yale (reference 3):

5. Doornhof D, Kristiansen TG, Nagel NB, Pattillo PD and

Sayers C: Compaction and Subsidence, Oilfield
Review 18, no. 3 (Autumn 2006): 5068.
6. Addis, reference 1.
7. Choi SK and Tan CP: Modeling of Effects of Drilling Fluid
Temperature on Wellbore Stability, Proceedings,
SPE/ISRM Rock Mechanics in Petroleum Engineering
Symposium, Trondheim, Norway (July 810, 1998):
Li X, Cui L and Roegiers J: Thermoporoelastic Analysis
for Inclined Borehole Stability, Proceedings, SPE/ISRM
Rock Mechanics in Petroleum Engineering Symposium,
Trondheim, Norway (July 810, 1998): 443452.


h = 2,000 psi

H = 3,000 psi

H = 3,000 psi

h = 2,000 psi
Hoop stress, psi






> Plan view of hoop stresses surrounding a vertical wellbore. In this model,
pore pressure and wellbore pressure are equal, while maximum and minimum
effective stresses within the formation equal 2,000 psi and 3,000 psi [13.8 and
20.7 MPa], respectively. However, hoop stress, which varies as a function of
radius and azimuth, is strongly compressive along the azimuth aligned with
minimum horizontal stress (h) (red shading above and below the wellbore),
where it reaches almost 7,000 psi [48.3 MPa]. Wellbore failure will be more
likely to occur along this axis. (Adapted from Sayers et al, reference 9.)

drilling fluids. When a formation is drilled with

an incompatible fluid, the invading filtrate may
cause the shale to swell, which can lead to
weakening of the rock and wellbore instability.
Shales may also be susceptible to timedependent changes in effective mud support
caused by differences between the mud pressure
and pore-fluid pressure, or between drilling fluid
salinity and formation salinity.8 Furthermore,
volume changes in shales arising from interactions between shale and drilling fluid can
locally disturb the stress orientation and
magnitude in a borehole.
Thus, while local and regional tectonic
stresses play a major role in rock deformation,
other downhole factors, such as pore pressure,
mud weight and downhole pressure fluctuations,
temperature and chemistry must also be
considered for their distinctive contributions to
the local stress-deformation continuum. Their
effects may also be tempered by textural properties unique to the local lithology, such as the size
and distribution of framework grains and pores,
mineralogy and the composition of diagenetic
cements. Given the variety of reactions to stress,
it is crucial for an operator to know as much as
possible about the rock surrounding a wellbore
and the conditions to which it will be subjected.


Changes in Stress
Drilling and production activities affect local
stress regimes. Problems encountered during
drilling may portend difficulties encountered
subsequently during the production phase.
Changes in stress may result in rock failure that
causes wellbore instability during drilling. These
changes may later lead to sand production once
the well has been completed. Other activities
during the life of a field can cause pore pressure
and temperature changes, which can modify
stresses acting farther from the wellbore. Stress
changes affect not only the reservoir but also
adjacent formations.
Drilling activity perturbs the initial equilibrium of stresses in the near-wellbore region. As
a cylindrical volume of rock is excavated
through drilling, the stresses formerly exerted
on that volume must instead be transferred to
the surrounding formation. This process creates
tangential, or hoop stresses, which must be
borne by the rock surrounding the borehole.
These wellbore stresses are a function of mud
weight, wellbore inclination, formation dip
angle and azimuth, and the magnitude and
orientation of far-field stresses (V, H and h).
Hoop stress varies strongly as a function of

borehole radius and azimuth.9 Furthermore, it

can far exceed H (left).
In most conventional drilling operations,
drillers use hydraulic pressure from drilling fluid
as a substitute for the mechanical support that is
lost through the cylindrical volume of rock
excavated while drilling a wellbore. They essentially replace a cylinder of rock with a cylinder of
drilling fluid. However, mud pressure is uniform
in all directions, and cannot balance against
oriented shear stresses in a formation. As stress
is redistributed around the wall of the wellbore,
shear stresses can exceed rock strength. When this
happens, the wellbore will deform or fail entirely.
Typical examples of geomechanics-related
drilling problems include wellbore instability and
fracturing of the formation. Ramifications include
financial loss resulting from lost circulation,
kicks, stuck pipe, additional casing strings,
sidetracks and even abandonment. To sustain
wellbore stability, operators must develop drilling
and well construction plans that consider stress
magnitude and direction, mud weight, trajectory
and pore pressure before, during and after a well
is drilled.
Drillers manage pressures imposed by mud
weight to avoid wellbore stability problems.
Their control of wellbore hydraulics reflects
a petroleum engineering approach to a
geomechanical problem. During drilling, wellbores can be compromised through a variety of
mud-induced modes of failure:10
Tensile failure occurs by increasing mud pressure until it causes the wellbore wall to go into
tension and eventually to exceed the rocks
tensile strength. This fractures the rock along
a plane perpendicular to the direction of minimum stress, often resulting in lost circulation.
Compressive failure may be caused by mud
weight that is too low or too high. In either
case, the formation caves in or spalls off, producing borehole damage and breakouts (next
page, top). Unless the wellbore is properly
cleaned out, the accumulation of breakout
debris can lead to stuck pipe as the borehole
packs off or collapses.
Shear displacement takes place when the mud
pressure is high enough to reopen existing
fractures that the wellbore has intersected. As
a fracture is opened, stresses along the opening are temporarily relieved, allowing opposing
faces of the fracture to shear, creating a small
but potentially dangerous dislocation along
the wellbore.
Wellbore stability is further affected by
structural factors, such as the interplay between
wellbore inclination, formation dip and

Oilfield Review

8. Gazaniol D, Forsans T, Boisson MJF and Piau JM:

Wellbore Failure Mechanisms in Shales: Prediction
and Prevention, paper SPE 28851, presented at the
SPE European Petroleum Conference, London,
October 2527, 1994.
Mody FK and Hale AH: A Borehole Stability Model to
Couple the Mechanics and Chemistry of Drilling Fluid
Interaction, in Proceedings, SPE/IADC Drilling
Conference, Amsterdam (February 2225, 1993): 473490.
Tan CP, Rahman SS, Richards BG and Mody FK:
Integrated Approach to Drilling Fluid Optimization for
Efficient Shale Instability Management, paper SPE
48875, presented at the SPE International Oil and Gas
Conference and Exhibition, Beijing, November 26, 1998.
van Oort E, Hale AH and Mody FK: Manipulation of
Coupled Osmotic Flows for Stabilization of Shales
Exposed to Water-Based Drilling Fluids, paper SPE
30499, presented at the SPE Annual Technical
Conference and Exhibition, Dallas, October 2225, 1995.
9. Sayers CM, Kisra S, Tagbor K, Dahi Taleghani A and
Adachi J: Calibrating the Mechanical Properties and
In-Situ Stresses Using Acoustic Radial Profiles, paper
SPE 110089-PP, presented at the SPE Annual Technical
Conference and Exhibition, Anaheim, California, USA,
November 1114, 2007.
10. For more on wellbore stability problems: Addis T, Last N,
Boulter D, Roca-Ramisa L and Plumb D: The Quest for
Borehole Stability in the Cusiana Field, Colombia,
Oilfield Review 5, no. 2 & 3 (April/July 1993): 3343.
11. Aoki T, Tan CP and Bamford WE: Stability Analysis of
Inclined Wellbores in Saturated Anisotropic Shales,
in Siriwardane HJ and Zaman MM (eds): Computer
Methods and Advances in Geomechanics: Proceedings
of the Eighth International Conference on Computer
Methods and Advances in Geomechanics, Morgantown,
West Virginia, USA, May 2228, 1994. Rotterdam, The
Netherlands: A.A. Balkema (1994): 20252030.
Yamamoto K, Shioya Y, Matsunaga TY, Kikuchi S and
Tantawi I: A Mechanical Model of Shale Instability
Problems Offshore Abu Dhabi, paper SPE 78494,
presented at the 10th Abu Dhabi International
Petroleum Exhibition and Conference, Abu Dhabi,
UAE, October 1316, 2002.
12. Rock fabric is a term that loosely encompasses the
mineral content, size, shape, orientation and
cementation of component grains within a rock,
including their overall arrangement into microscopic
laminations or larger beds.

Autumn 2007




Depth, ft

directional variations in strength between and

along formation bedding planes (below right). It
is not unusual for some degree of wellbore failure
to occur in vertical wells that encounter steeply
dipping shales, or inclined wells that intersect
shale bedding planes at low angles. Such failures
are initiated by low shear and tensile strength
along planes of weakness in shales.11
The issue of strength, or a rocks capacity to
withstand stress, points to an important underlying influence on deformation and failure: that
of rock fabric.12 Rock fabric can dictate whether
a given amount of stress will cause a rock to
deform or to completely fail, and can influence
the extent and orientation of fractures or
breakouts in a wellbore. Thus, although borehole
breakout is typically assumed to be oriented
along the axis of least stress, the bedding,
cementation, mineralogy and grain size of a rock
may actually redirect the course of a breakout
along the rocks weakest points.
For help in anticipating and circumventing
problems such as those described above, some



Radius, in.


> Borehole breakout. Results from a UBI Ultrasonic Borehole Imager logging
tool show the extent of stress-related damage in a wellbore. In isotropic or
transversely isotropic rock, where rock properties do not change along the
plane of the wellbore, such damage is generally aligned along a plane of
least horizontal stress.

> Formation effects on wellbore stability. Structural and stratigraphic factors

can combine to cause well damage. Here, incompetent beds overlie a stronger
formation near the crest of a structure; relative movement results in damaged
cement and collapsed casing.



Well pressure, psi














Reservoir pressure, psi

> Stress changes induced by production. As a field depletes, the magnitude

of stresses may alter drastically. Under such conditions, a completion or
perforation originally oriented in the most stable direction at the onset of
production may subsequently become unstable and fail as production
proceeds. Here, the horizontal perforation will permit the greatest safe
drawdown (blue curve) and solids-free production. However, as the field
depletes and stresses change, this previously stable perforation will
collapse and the vertical perforation will assume a greater role in
production, though safe drawdown pressure has decreased (red curve).
(Adapted from Marsden, reference 18.)

operators are turning to geomechanics experts at

the Schlumberger Center of Excellence for Pore
Pressure Prediction and Wellbore Stability
Analysis. Located in Houston, the geomechanics
experts in this group have a global reach, and
support operators around the world. This
interdisciplinary team is actively involved in
helping clients mitigate risk in drilling,
completing and producing wells in difficult
geomechanical environments, such as deepwater
exploration, subsalt drilling, unconventional gas
and unconsolidated reservoirs.
Beyond the Wellbore
Geomechanical influences can extend past the
borehole, into the reservoir and beyondthough
their extent may not be recognized until a
reservoir is produced. The pressure sink created
by a well to induce production will result in lower
wellbore pressures than the pore pressure of the
surrounding formation, and this difference can
increase the risk of rock failure.13
With the withdrawal of reservoir fluids during
production, the overburden load borne by pore
fluids must be transferred to the rock framework
surrounding the pore space. Resulting changes in
pore pressure will prompt adjustments in total
stress and effective stress. Within the rock,
increased loading will cause varying degrees of
deformation or failure, evidenced by grain sliding


and rotation, plastic deformation, cement

breakage at grain contacts, or activation of
existing fractures.14
On a larger scale, production-induced stress
changes on the rock framework can lead to pore
collapse and compaction of the reservoir.15
(Compaction is not always a problem, however
compaction drive has helped to pressurize oil in
some reservoirs, thereby increasing production
rates and improving ultimate recovery.)16 As a
result, operators have had to contend with
surface subsidence problems, deformation or
shearing of wellbore tubulars and buckling of
completion components. Other effects range
from reduction of porosity and permeability to
fault reactivation, formation fracturing, sand
production or loss of reservoir seal.
The effects of geomechanics are especially
pronounced in gas storage operations, where the
cyclic process of injecting and withdrawing gas to
or from a reservoir provokes changes in fluid
pressures within reservoir pore spaces. These
pressures cushion the stresses acting on the rock
mass, but the pressures increase or decrease
with injection and withdrawal. The loads acting
on the rock matrix thereby decrease and
increase in response to these cycles. Although
total overburden stress may remain constant
throughout these cycles, the total horizontal
stresses acting throughout the reservoir can vary

with pressure, generally decreasing as the gas is

withdrawn. If induced stresses exceed the elastic
limits of the rock, porosity and permeability may
be permanently reduced, along with reductions
in overall storage capacity. Furthermore, as the
surrounding rock adjusts to the isostatic
imbalance caused by pressure cycling and stress
changes, nearby faults may be reactivated.17
Production-induced changes can also affect
the rock beyond the productive areas of a
reservoir. Even in producing formations, reservoir attributes such as porosity and permeability
can vary, giving rise to nonuniform drainage and
depletion. As a reservoir is produced, the rock
may eventually compact, leaving surrounding,
undrained areas of the formation to compensate
for changes in pressure and displacement of the
adjacent rock. Above the productive formation,
compaction will lead to changes in the overburden, as described later in this article.
Changes in stress imposed on a producing
horizon can put the rock out of equilibrium with
its surroundings. The result is a corresponding
transfer of stress between the depleting
reservoir or injection interval and the rock
immediately surrounding the reservoir. Ensuing
rock deformations may compromise the
integrity of existing completions within the
reservoir and overburden (above left). The
significance of production-induced stress
changes and their potential to adversely
influence field operations, production and
economics will depend on mechanical
properties of the rocks, natural fractures and
faults.18 To understand and anticipate such
changes in the wellbore and beyond, operators
are increasingly turning to advanced geomechanical testing and modeling techniques.
13. Cook J, Fuller J and Marsden JR: Geomechanics
Challenges in Gas Storage and Production, presented
at the United Nations Economic and Social Council:
Economic Commission for Europe: Working Party on
Gas: Proceedings of 3rd Workshop on Geodynamic
and Environmental Safety in the Development,
Storage and Transport of Gas, St. Petersburg, Russia,
June 2729, 2001.
14. Sayers CM and Schutjens PMTM: An Introduction to
Reservoir Geomechanics, The Leading Edge 26, no. 5
(May 2007): 597601.
15. Doornhof et al, reference 5.
Sayers C, den Boer L, Lee D, Hooyman P and
Lawrence R: Predicting Reservoir Compaction and
Casing Deformation in Deepwater Turbidites Using a 3D
Mechanical Earth Model, paper SPE 103926, presented
at the First International Oil Conference and Exhibition,
Cancun, Mexico, August 31September 2, 2006.
16. Andersen MA: Petroleum Research in North Sea Chalk,
Joint Chalk Research Monograph, RF-Rogaland
Research, Stavanger, 1995.
17. Cook et al, reference 13.
18. Marsden R: Geomechanics for Reservoir
Management, in Sonatrach-Schlumberger Well
Evaluation Conference Algeria 2007. Houston:
Schlumberger (2007): 4.864.91.

Oilfield Review

Autumn 2007

and without a common reference for scale. Until

recently, there has been no framework to make
the process consistent for every stage. However,
the development of continuous property profiling
and multidimensional cluster analysis of well
logs now provides a uniform scale of reference
for incorporating heterogeneity during all
aspects of reservoir analysis and evaluation.
Continuous ProfilingScratch testing,
known formally as continuous profiling of
unconfined compressive strength, provides a
quantitative means of evaluating variability in
strength, texture and composition of core
samples. By association, this variability may be
related to other rock properties. Scratch testing
has become critical in correctly defining facies
and heterogeneities that would be difficult or
impossible to observe by geologic description or
log characteristics alone. Digital photographs of
the core, in conjunction with scratch testing,
allow visualization of textural heterogeneity and
associated strength heterogeneity (below).
When continuous-strength profiling is
combined with cluster analysis of well logs, it

Specialized geomechanics laboratory testing

provides crucial data for wellbore and completion design and for reservoir management. This
was not always the case. Traditional engineering
analysis of reservoir potential and productivity
tended to overlook the heterogeneity of reservoir
rock. Although heterogeneity may have been
captured in mud logs and core photographs, or
inferred from logs of various petrophysical
properties, these characteristics were not
reflected in simplified homogeneous systems
created for reservoir and geomechanical models.
Properties related to reservoir rock
mechanics were often characterized as uniform
throughout all locations and for all orientations
within a particular geological unit. This approach
inevitably led to underestimations of the role of
material properties in geomechanics. The
industry, however, is coming to realize that the
rock matters, and that its varying properties
cannot be ignored in geomechanical analysis.
Further complicating the evaluation process
is the fact that each stage of reservoir analysis
from predrill geological studies, through
exploration, to reservoir modeling and
productiontends to be evaluated in isolation,

(continued on page 48)

50,000 psi

0 psi
50,000 psi

0 psi
50,000 psi

0 psi
50,000 psi

2.0 ft











1.0 ft









0 psi

0 ft

Measuring Ground Truth

Despite years of geomechanical analysis, many
E&P companies continue to experience drillingor production-induced problems. However, the
field of geomechanics involves much more than
analysis of stress. Though changing stress fields
can wreak havoc on drilling and production
plans, the orientation or magnitude of stresses
and strains have little significance without
framing such measurements in the context of the
rock itself. And rocks are highly variable. Other
problems are caused, in part, by oversimplified
characterization of rock behavior, and by limited
modeling and analysis capabilities compounded
by a lack of comprehensive rock property data.
These issues are being addressed by the
TerraTek Geomechanics Laboratory Center of
Excellence in Salt Lake City, Utah. TerraTek, Inc.
was acquired by Schlumberger in July 2006 (see
Geomechanics Laboratory: Testing Under
Extreme Conditions, page 44). The modern
high-pressure testing systems and techniques
developed at the TerraTek facility evolved from
an effort to characterize and predict ground
motion and crater development in response to
nuclear tests. Evaluation of these tests could not
be performed without rock property measurements obtained under high pressure. Measuring
these properties was very difficult, and
spawned a number of technical breakthroughs
by TerraTek.
Highly accurate load-deformation measurements were essential, requiring measurements
inside test vessels under extreme pressures.
TerraTek scientists conducted research to
measure rock properties to pressures of
150,000 psi [1,034 MPa]. The TerraTek highpressure rock property data enabled analysis of
the magnitude of ground motions caused by a
nuclear event.
TerraTek researchers carried out tens of
thousands of tests on rocks under high pressure.
Their testing capabilities were subsequently
applied to other geomechanics investigations,
including geothermal energy recovery, coal
mining, deep geologic nuclear waste storage,
underground energy storage, as well as oil and
gas recovery. Today, the TerraTek Geomechanics
Laboratory Center of Excellence regularly
conducts rock tests for deep wells, achieving
pressures of 30,000 psi [207 MPa], or to higher
pressures of 50,000 to 60,000 psi [345 to 414 MPa]
when required for drilling-rock destruction or
perforation analysis. In addition to high-pressure
geomechanics testing capabilities, the TerraTek
facility conducts large-scale drilling and
completions performance testing.

> Overlay of a core photograph and scratch test results. A scratch test uses a sharp point that is pulled
along the core with a fixed force to press it into the cores surface. The depth of the scratch, as an
indicator of rock strength (red curves), can be correlated to mechanical properties of the rock. Cored
intervals exhibiting visually similar properties (same shades of gray, points A and A) may have different
strengths, while other intervals exhibiting different visual properties (lighter and darker shades of gray,
points B and B) have equal strengths. Variability in mechanical strength along the length of the core is
high, ranging from 8,000 psi to 23,000 psi [55 to 159 MPa] within just 8 contiguous feet [3.6 m] of core.


Geomechanics Laboratory: Testing Under Extreme Conditions

The TerraTek facility in Salt Lake City, designated as the Schlumberger Geomechanics
Laboratory Center of Excellence, investigates
the impact of geomechanics on a wide range
of exploration and production applications.
The range of applications also provides insight
into the kinds of problems that operators
must try to circumvent:
Well construction and completion: evaluate
wellbore stability and the potential for sand
production and perforation collapse; analyze
multilateral junctions and evaluate stability
of conventional and expandable liners.
Completion and stimulation design: determine optimal completion alternatives based
on rock mechanical and physical properties;
investigate options for delayed gravel packing and oriented perforating; optimize
stimulation treatment design.
Long-term production behavior: investigate
stress regimes contributing to reservoir
compaction during production; predict
surface subsidence and subsequent loss of
reservoir permeability; analyze fines generated during the compaction process, along
with associated skin damage; evaluate
potential for casing collapse.
Overburden rocks: test for compatibility
between drilling fluids and shales; optimize
selection of drilling fluids; evaluate potential for delayed shale failure caused by
mud-shale interactions; analyze thermal
effects that cause delayed shale failure.
Exploration and frontier drilling operations:
develop field and laboratory correlations for
predicting mechanical properties and in-situ
stress prior to and concurrent with
exploratory drilling activity.
Testing is conducted in different specialized
laboratories, depending on available test material, client specifications and research efforts.
Many large-scale tests are carried out in the
completions laboratory. One of the more
prominent features of this facility is its largeblock polyaxial stress frame. The stress frame
provides a controlled environment for monitoring rock responses during pseudostatic testing.


> Large-block polyaxial stress frame for simulating downhole conditions.

Here, a worker lowers a steel platen while preparing to seal the test chamber.

In this setting, researchers can measure deformation parameters while simultaneously

measuring dynamic responses of rock samples
to different load rates and magnitudes. The
large-block stress frame can be configured to
simulate a variety of downhole pressures and
conditions. Large-block testing applications
range from wellbore stability analysis to evaluating sanding potential, liner and screen
loading, perforating effectiveness and hydraulic
fracturing simulations.
Located inside a pit, the exterior of the
stress frame is formed by a series of steel
rings. These rings are stacked to encase an

internal chamber that can accommodate

blocks of rock measuring up to 30 x 30 x 36 in.
[76 x 76 x 91 cm]. The chamber is sealed with
steel platens that are bolted over 12 large tie
rods (above).
Pairs of bladder-like devices, called flatjacks,
are placed on opposite sides of the sample to
apply independent triaxial loading in each of
the three principal stress directions. The three
pairs of flatjacks are internally pressurized,
with one surface of the flatjack reacting against
the face of the rock, and the other surface
reacting against the wall of the internal chamber of the stress frame, or its platen.

Oilfield Review

> Polyaxial stress frame. This device can

accommodate rock samples measuring up to
12 x 12 x 16 in. [30 x 30 x 41 cm].

A maximum stress of 8,000 psi [55 MPa] can be

applied in all three directions, with a maximum
difference of 2,000 psi [13.8 MPa] between the
two horizontal stresses. Each stress can be controlled independently.
The stress frame also has the capability to
control pore pressure within a sample. In such
tests, the rock sample is encased in a thin
steel canister. Thick rubber sheets are placed
at the top and bottom surfaces of the rock to
act as pore-pressure fluid seals. A porous
proppant pack placed around the block establishes a constant pressure boundary
condition. Custom software controls each of
the three principal stresses, along with pore
pressure and wellbore pressure. The software
can be programmed to keep a constant effective stress on the sample block at all times.
Some experiments require a simulated
permeable zone bounded above and below by
impermeable formations. In these kinds of

Autumn 2007

> Instrumented sample for triaxial testing. This test-frame assembly is used to measure radial
and axial strains, along with compressional and shear wave velocities. In this configuration, both
pseudostatic and dynamic elastic properties are determined concurrently under simulated in-situ
stress conditions. Here, a core consisting of alternating light and dark layers of siltstone and
mudstone is subjected to ultrasonic pulses to test seismic responses in the rock. The sample is
sealed with a clear polyurethane jacket that prevents fluid communication between the confining
fluid pressure and the pore pressure. These test frames can also perform uniaxial strain
compaction testing, thick-walled cylinder testing and other specialized stress paths up to
temperatures of 200C [392F]. Axial force up to 1.5 x 106 lbf [6.7 MN] can be applied to samples
up to 6 in. [15 cm] in diameter. Confining pressure and pore pressure are monitored with
conventional pressure transducers with pressure limits of 30,000 psi [207 MPa]. Another system
in this laboratory can attain 60,000 psi [414 MPa].

tests, a servo-controlled injector is used to

supply the fluid at either a constant rate or a
constant pressure. The injected fluids can
range from brine to drilling mud to various
completion fluids. The injection can simulate
a scaled or actual-sized wellbore.
For smaller samples, a medium-sized
polyaxial stress frame is used (above left).
This device is often used for studying acid
fracturing and other stimulation techniques,
providing a wide range of testing capabilities.
Another unique testing facility is the rock
mechanics laboratory, where 14 stress frames
are used to test cylindrical samples with diam-

eters ranging from 0.5 inches [12.7 mm] to

6 inches [152.4 mm]. Testing on a smaller
scale can also provide valuable insights into
rock characteristics.1 A special triaxial test
frame has been designed to measure rock
strain as well as its effects on seismic velocities (above right). Ultrasonic velocities,
1. Laboratory capabilities include an extensive variety
of testsunconfined compression, uniaxial-strain
compression, triaxial compression, multistage triaxial
compression, controlled constant stress path, thickwalled cylinder (with and without radial fluid flow and
measurements of produced sand), and tensile strength
tests, as well as testing with concurrent ultrasonic
velocity and acoustic emissions measurements
along with many customized test programs and
research efforts.


obtained in combination with deformation

measurements of axial and radial strain, provide information on static and dynamic
mechanical properties that can be correlated
to well-log data.
The triaxial test frame holds a core sample
between polished, hardened-steel end-caps.
The sample, measuring 1 in. [2.5 cm] in diameter by 2 in. [5 cm] in length, is jacketed by an
impermeable membrane. Axial and radial cantilever-beam sets are mounted to the sample to
measure displacements when the sample is
subjected to stress and pressure. The axial
strain cantilever set is attached to the upper
end-cap and measures axial displacement
through deflection on the base cone attached
to the bottom end-cap. The radial-strain cantilever set consists of a ring with four
strain-gauge arms, which measure radial displacement at four points, forming two
perpendicular directions at the midpoint of the
sample. The bottom end-cap rests on an internal load cell, and the axial stress is calculated
from measurements of force on the internal
load cell. During testing, data are corrected for
elastic distortion of the end-caps and for
strains associated with the jacketing material.
The end-caps also contain ultrasonic transducers. Ultrasonic velocity measurements are
performed with piezoelectric transducers that
transform electrical pulses into mechanical
pulses and vice-versa. Compressional and
shear pulses are generated by a pulse generator that applies a high-voltage, short-duration
electrical pulse at an ultrasonic frequency to
one of the piezoelectric transducers. This
pulse is transmitted through the rock sample
in the form of an elastic wave. The receiving
transducer at the opposite end of the rock
sample transforms this elastic wave into an
electric signal, which is captured on a digital
oscilloscope. The P-wave and S-wave velocities
are calculated on the basis of the time required
for the compressional or shear pulses to travel
through the specimen.
The instrumented test sample is next
placed inside a pressure vessel. The pressure
vessel is then filled with either mineral spirits or oil to apply confining pressure. Axial
stress, axial strain, radial strain and confining
pressure are all measured and controlled


during each test. Depending on testing objectives, these tests may be performed with pore
fluids drained to atmospheric pressure, or
with pore fluids undrained. Temperatures can
also be increased to better approximate
actual in-situ conditions.
The triaxial test frame permits measurements to be taken at different orientations
with respect to bedding planes. Using these
measurements, the failure envelope of the
rock sample can be defined as a function of
stress orientation to bedding; in addition,
anisotropic properties of the rock can be
defined. This information is essential for predicting wellbore stability, evaluating in-situ
stress and designing hydraulic fracture programs for strongly anisotropic formations
such as those found in unconventional tight
gas shales.
Ultrasonic velocities, obtained in combination with deformation measurements of axial
and radial strain, provide information on
static and dynamic mechanical properties
that can be correlated to well-log data. Ultrasonic wave velocities in sandstones, particularly those that are poorly consolidated, are
strongly dependent on stress; thus, stress
changes can be calibrated to seismic velocity
measurements. Other, more consolidated
rocks, such as tight sands and tight shales,
exhibit an entirely different behavior. Wave
velocities in these rocks are virtually independent of stress, so changes in measured
seismic velocities can be attributed to other
phenomena such as anisotropy.
Early knowledge of rock behavior was based
on testing of homogeneous and isotropic
materials; early models reflected this simplicity. New opportunities, such as unconventional
hydrocarbon plays, are emerging, and call
attention to the true nature of the rocks in
which they are based. Platforms such as the
triaxial test frame provide data that are fundamental for developing new models to honor
the heterogeneous, anisotropic nature of complex formations.
The TerraTek facility is also called upon to
test new drilling, completion and stimulation
technologies, including evaluation of drilling
fluids and bits at high pressures. Although
capabilities exist for measuring individual rock

> TerraTek wellbore simulator. The full-scale

drilling rig and wellbore simulator can be
configured to test the performance, wear,
deviation and dynamics of full-size drill bits in
overbalanced or underbalanced conditions
and at simulated depths. A triplex mud pump,
equipped with a special high-pressure fluid
manifold, can produce wellbore pressures up
to 11,000 psi [75.8 MPa] to simulate highpressure drilling conditions. Effects of various
fluids on drilling performance, bit balling,
formation damage, coring and core invasion
are also investigated here.

properties or fluid properties at extreme temperatures and pressures, determining the

manner in which complex rock cutting and
breakage mechanisms interact in the presence
of drilling fluids at great depth is much more
difficult. To accommodate large-scale geomechanics testing, the drilling laboratory is
equipped with a wellbore simulator capable of
reproducing pressure conditions at reservoir
depth while also accommodating the flow rates
typically required to drill in extreme environments (above).

Oilfield Review

> Breakout simulation. With no drilling mud used to drill this

sandstone subjected to increasing confining pressure, this
simulated wellbore progressively broke down, producing a classic
borehole breakout pattern.

> Bottomhole drillbit patterns. Bottomhole impressions track

performance of a bit as it drills a borehole through high-strength
sandstone. In this case, a polycrystalline diamond compact bit was
drilling with a 16-lbm/gal (ppg) [1.9-g/cm3] oil-base mud at 10,000-psi
[68.9-MPa] wellbore pressure. The patterns on the bottom were
subsequently studied to determine how various drilling conditions
affected drilling performance. As depth of the rings decreases, so
does the cutting efficiency of the bit, and hence the ROP
decreases. With different drilling fluids, the patterns sometimes
disappear altogether.
2. Spurt loss is an instantaneous loss of a volume of
the liquid component of drilling fluid as it passes
through the borehole wall prior to deposition of
competent filtercake.
For more on ROP testing: Judzis A, Bland R, Curry D,
Black A, Robertson H, Meiners M and Grant T:

Autumn 2007

Optimization of Deep Drilling Performance;

Benchmark Testing Drives ROP Improvements for
Bits and Drilling Fluids, paper SPE/IADC 105885,
presented at the SPE/IADC Drilling Conference,
Amsterdam, February 2022, 2007.

The TerraTek wellbore simulator was central to a recent high-pressure drilling study
sponsored by the US Department of Energy
(DOE) joint industry program, called Deep
Trek. The facility was contracted to provide
full-scale laboratory tests of drill bits and prototype drilling fluids at 10,000-psi [68.9-MPa]
borehole pressuresubstantially higher pressures than any previously studied. Results
from these tests may influence the economics
of deep drilling.
The study demonstrated that drilling rates
of penetration (ROPs) can be increased in
deep-well applications using advanced bit and
drilling fluid designs. Although previous studies have shown that ROP usually falls with
increasing borehole pressure, these earlier
studies did not account for certain mechanisms that affect ROP at great depth, such as
type of drilling fluid, weighting material and
spurt loss.2
Another common wellbore stability problem
involves borehole breakouts. Although breakouts often occur during drilling, they can also
affect the completion process. In one breakout investigation, TerraTek engineers drilled
an 812-inch [21.6-cm] borehole in a large
sandstone core. The core was subjected to
increasing rates of confining pressure in the
laboratory. The resulting borehole breakout
was similar to that produced in actual wellbores when drilling fluid weights are too low
(above left).
The sample was subsequently used for an
expandable sand screen (ESS) mechanical
integrity test. The screen and basepipe assembly was compliantly expanded to the borehole
wall and partially into the breakout zone.
Results from this test showed how far the
screen could be expanded into the borehole
breakout, in addition to determining the collapse load resistance of the ESS product.
Other problems that adversely impact
drilling performance, such as vibration or
borehole spiraling, are identified through
examination of drilling patterns (left).
Through the aid of the borehole simulator,
researchers have an opportunity to closely
study bottomhole patterns that would otherwise not be accessible.


Well 1



Gamma Ray

Bulk Density



gAPI 150


Neutron Porosity

15 0 ohm.m 1,000 0.45

vol/vol 0.15

Well 2


Depth, ft

Depth, ft

Depth, ft































Cluster Tag



> Cluster tagging between two wells. Color-coding of log responses from each well, combined with
analysis of compliance in the Error track, is useful in identifying changes in thickness and location of
previously defined cluster units between wells. Here, the red-blue-yellow sequences are significantly
higher and thicker in Well 1 than in Well 2. Three excursions above 40% error (red line) indicate
candidate zones for further sampling to better describe the range of facies encountered.

> Cluster analysis of well logs. A multidimensional

statistical algorithm is applied to the well-log
measurements to identify similar and dissimilar
combined log responses, enabling users to
identify rock units with similar and dissimilar
material properties. The output is displayed as a
color-coded representation of clusters for visual
interpretation of rock units with distinct properties
along the interval of interest (Track 4).

provides fundamental relationships for upscaling

or downscaling, and so is a powerful tool for corelog integration.
Cluster AnalysisCluster analysis defines
log-scale heterogeneity, based on multidimensional analysis of log responses (above left). This
technique uses detailed algorithms to distinguish
similar and dissimilar patterns of log responses.
Because this technique interprets the combined


effect on all measurements, it is able to

recognize small but consistent variations in
combined log responses. As applied to heterogeneous distributions of material properties,
cluster analysis also provides a relevant scale for
manipulating property variability in subsequent
evaluation steps throughout a project.
Cluster TaggingThe application of cluster
analysis can be extended to multiple wells,
providing comparisons between the cored, or
reference, well and other wells in a field. Details
obtained through cluster analysis of one well can
be used to recognize similar traits in adjacent
wells through a process known as cluster tagging.
Cluster tagging begins with log-response
clusters defined over discrete cored intervals in a
reference well, then compares these clusters
with log responses from a noncored well. Using
definitions established from core-log responses

in the reference well, the technique assigns

clusters to logs from the noncored well and then
outputs an error curve to help evaluate
compliance between two correlative zones.
Clusters showing poor compliance, where error
exceeds 40%, indicate a log response that is not
represented in the defined clusters, and thus a
new facies. These clusters are candidates for
detailed core sampling to provide new cluster
definitions and further characterize the range of
facies in a prospect (above right).
Cluster analysis is also used for optimal
selection of core samples. In reservoir studies,
both the strongest and weakest core samples
measured by continuous profiling must be tested
in proportion to their relative abundance in a
reservoir. Improper sampling in heterogeneous
or thinly interbedded formation cores can result
in biased representation of the formation.
Cluster analysis can help operators tie log

Oilfield Review

analysis identifies units by their material

properties and maps their distribution along the
length of a well. By relating laboratory measurements of these units to their combined log
responses, core-log relationships are developed
for each cluster. Since the method is unaffected
by variability in thickness or stacking arrangements of the various cluster units, it allows
prediction of properties along the length of the
logged section in a well.
Multiwell AnalysisFor basin-wide analysis,
cluster tags of multiple wells are tied to a single
reference model containing definitions of
material properties across the basin. The results
can be used for 3D visualization of lateral
variability in reservoir and nonreservoir units.

properties to core properties throughout the

reservoir, and thereby recognize which parts of a
core merit additional plug-sample analysis
(below). With cluster-analysis measurements of
log-scale heterogeneity and core-scale heterogeneity measurements obtained through scratch
testing, the operator can determine the location
and number of samples required to adequately
characterize the core.
Cluster-Level Property PredictionsSince
models are traditionally built around the
structure and stratigraphic layout of a basin, the
discontinuous and heterogeneous distribution of
reservoir and nonreservoir lithological units
within a single stratigraphic section is often
poorly represented across the basin. Cluster

Core-Scale Heterogeneity

1 ft

2 ft

2 ft

50 k

1 ft

50 k

2 0 ft

10 k

1 0 ft

10 k

50 k

10 k


Cluster tag analysis was instrumental in

generating a regional study for a client who was
pursuing an unconventional gas play. The goal
was to model the vertical and lateral discontinuity of principal reservoir units in a tight
gas-shale reservoir. These reservoirs are highly
heterogeneous both vertically and laterally, with
localized diagenetic alterations that create great
variability in material properties. As a result,
reservoir and mechanical properties change
significantly from location to location between
wells, and production performance often varies,
even between wells drilled in close proximity to
each other.
The client ordered a study to understand the
variability in permeability, gas-filled porosity and


50 k

10 k


3 0 ft

4 0 ft

1 ft

1 ft

2 ft

2 ft

2 in.

10 k

40 k

> Using rock heterogeneity to select laboratory samples. Log-scale heterogeneity, indicated by cluster colors (left), is compared against core-scale
heterogeneity data obtained through scratch testing (red curves) superimposed onto core photographs (middle). In the log-scale heterogeneity plot, color is
used to differentiate between zones of similar or dissimilar material properties as a function of unconfined compressive strength measurements. Here the
yellow clusters are the weakest units and brown clusters are strongest. Progressing from region 1 (yellow cluster), region 2 (yellow cluster transitioning to
dark blue), region 3 (dark blue transitioning to brown), and region 4 (brown cluster), the rock strength varies by more than 400%. Core photographs (middle)
show a corresponding transition in unconfined compressive strength from 10,000 psi [68.9 MPa] in the argillaceous mudstone (core section 1) to 40,000 psi
[275.8 MPa] in the basal carbonate (core section 4) within this 40-ft [12-m] interval. Sample plugs (right) are taken from the whole core for detailed analysis
and testing. This methodology helps operators ensure that their 2-in. sample plugs account for the variability present in the whole core.

Autumn 2007


Geologic Data
Regional tectonic framework
Structure depth maps
Lithostratigraphic column
Regional compaction trends
Basin analysis
Earthquake fault-plane solutions
Tiltmeter surveys
Core tests and descriptions
Rock composition and texture
Core-log integration
Heterogeneity and anisotropy
Petrophysical and mechanical

Seismic Data
3D seismic cube
2D seismic profiles
Tomographic velocity
Vertical seismic profiles and checkshot data
P-wave velocity profiles

Cluster tag
1 2 3 4 5 6 7 8 9 10 11 12

Formation Evaluation Data

Wireline and LWD logs
Gamma ray, resistivity, density, sonic, caliper
Acoustic scanning tool
Borehole imaging
Well test and production pressure measurements
Formation tests and drillstem tests

> Basin-wide multiwell cluster analysis. This presentation uses Petrel seismic-to-simulation software
to help operators visualize the cluster-analysis results and track reservoir quality throughout the field.
Different cluster units are associated with distinct reservoir qualities. They are also associated with
different values of fracture containment potential. Once the reservoir quality and fracture containment
potential are identified in detail by laboratory testing, they can be tracked laterally across the basin.
Surfaces identifying the intervals of best reservoir quality have been delineated. Cluster analysis in
this case identifies the heterogeneity inherent in any of these units that otherwise might be
considered homogeneous.

total organic content as they relate to reservoir

quality. It was also important to understand the
variability in the conditions of hydraulic fracture
containment along the various wells containing
units with best reservoir quality. For optimal well
productivity, reservoir quality must be coupled
with completion quality. In this field, reservoir
quality alone, without successful fracturing and
fracture-height containment, would result in
poor well productivity. By mapping locations
throughout the field where both conditions of
reservoir quality and completion quality exist
simultaneously, the client could identify sweet
spots in the reservoir (above left). The results of
this field study would also help improve
visualization of production distribution across
the basin.
TerraTek geoscientists used cluster analysis
and cluster tagging to evaluate the field.
Understanding the vertical stacking patterns of
cluster units on a well helped the client define
the location and thickness of clusters with the
best reservoir-quality properties.
19. McCann GD and Wilts CH: A Mathematical Analysis of
the Subsidence in the Long Beach-San Pedro Area,
technical report, California Institute of Technology,
Pasadena (November 1951), in Geertsma, reference 2.
20. Ali AHA, Brown T, Delgado R, Lee D, Plumb D,
Smirnov N, Marsden R, Prado-Velarde E, Ramsey L,
Spooner D, Stone T and Stouffer T: Watching Rocks
ChangeMechanical Earth Modeling, Oilfield
Review 15, no. 2 (Summer 2003): 2239.


Once these parameters were defined, the

client could select the best geometry of
horizontal wells and the best locations for
perforating. Understanding the properties of
cluster units immediately above and below the
best reservoir-quality units also helped identify
mechanical properties and conditions for
hydraulic fracture containment.
Modeling Geomechanical Properties
The interaction between geology, wellbore
orientation and stress changes caused by drilling
or production is a complex 3D process. This
interaction continually changes over time,
adding yet another dimension of complexity.
Over the life of any productive field, innumerable
events take place that alter the geomechanical
framework between the reservoir and the
surface. Exploration wells are drilled and tested;
additional wells are drilled and produced; some
may be turned into injectors, some are worked
over, while others are plugged and abandoned.
Each activity causes changes in stresssome
ephemeral, others more enduring. And these
changes can be costly, with potential to affect
formation integrity, porosity and permeability;
reservoir compaction and subsidence; and well
and completion integrity.

Drilling Data
Daily drilling reports
End of well reports
Mud weight profile
Leakoff tests, extended leakoff tests, formation
integrity tests, minifrac tests
Directional surveys
Mud logs

Calibration Data
Laboratory measurements on cores
In-situ stress measurements from hydrofracturing tests
Observed breakouts and stress-induced features
Field and production observations

> Array of input parameters for a mechanical

earth model.

The movement to understand such changes

was spurred in part by recognition that
subsidence in certain fields was directly related
to production. Basic mathematical models were
developed by the early 1950s to understand and
predict subsidence in Wilmington field,
California.19 Later, subsidence of the North Sea
Ekofisk field, discovered in the early 1980s,
prompted development of more comprehensive
computer models, based on finite-element
analysis. These models linked hydrocarbon
production to changes in reservoir properties and
deformation and, in turn, to seabed movement
and faulting in the overburden.
E&P companies became interested in
learning how stress evolves as reservoirs become

Oilfield Review

Importing from ECLIPSE

or Petrel software, or both

Importing fault surfaces

Embedding in overburden,
underburden and sideburden

p, T
Data and results utilized
in engineering designs
and planning



kij, Vpore

Initialization and coupled

simulation (parallelization)

Population with properties and

assign behavioral models

> Workflow for 4D coupled reservoir geomechanics modeling. Formation and structural data form the framework for the initial reservoir model, then
characteristics from surrounding rock bodies are added. Stress and strain are modeled throughout the reservoir and adjacent rock to understand changes
over time.

depleted. If stress changes could be modeled

over the life of a field, operators could predict
problems during the life of a well or anticipate
the need for infill drilling. With a steady growth
of computational capacity, geomechanics
programs acquired increasingly sophisticated
modeling capabilities. Rock mechanical models
developed to analyze stress changes in reservoirs
included the VISAGE stress analysis simulator.
This advanced geomechanics modeling system
emerged from waterflood directionality studies
in the North Sea and elsewhere.
Developed in 1993 by V.I.P.S. (Vector
International Processing Systems) of Bracknell,
England, VISAGE geomechanics software solves
equations that relate rock stress and pore
pressure to deformation and reservoir properties.
By integrating geomechanics and rock mechanics
with reservoir engineering, V.I.P.S. developed the
worlds first coupled geomechanics stressdependent reservoir simulator. With acquisition
of V.I.P.S. by Schlumberger in April 2007, the
Bracknell facility was designated as the Reservoir
Geomechanics Center of Excellence.

Autumn 2007

Finite-element modeling is widely used for

stress analysis in both conventional engineering
and geomechanics. Finite-difference modeling is
used to analyze fluid flow. The advantage of the
VISAGE simulator is its capability to describe
and simulate the coupled nature of geomechanical
stresses and fluid flow as they change over time
by linking these two analyses. This capability is
key to development of 3D and time-sequenced 4D
mechanical earth models.
Unlike reservoir production models,
mechanical earth models (MEMs) must take into
account not only the reservoir, but also the
overburden, seabed, the underburden, or rock
beneath the reservoir, and sideburden, or
adjacent rock, which often provides stress
boundary conditions.20 The MEMs are usually
much larger than ordinary reservoir models. As
such, they have substantial data requirements
that may be difficult to satisfy.
Complex rock behavior, varying rock
properties and large-scale simulations require
better software and better data, especially with

regard to cores. Basic models of the past enabled

the industry to opt for simplified assumptions,
using homogeneous formation properties
throughout their models. Todays sophisticated
numerical simulators inevitably dictate a wider
array of data. The MEM is built to honor this wide
array of data (previous page, top right).
A geomechanical simulation might begin with
construction of a 3D structural model. Next, the
model is populated with mechanical properties
of each formation and fault. The properties are
derived from seismic data, logs, cores, geostatistical projections and inversion of breakout and
drilling data for individual wells. Boundary
conditions, simulating the expected stress
profiles at the sides of the model, are then added.
This populated model is imported into the
VISAGE system to calculate the evolution of
stresses throughout the model (above).
The driving mechanism of the modeling is
mainly pressure changes induced by fluid
extraction from the reservoir, or by injection into




> Three-dimensional view of a reservoir. The uppermost horizon of an anticlinal

reservoir is intersected by numerous faults (inclined planes of semitransparent
purple, red, green and blue). The axis of the anticline is aligned with the long
axis of this figure. Colors on the reservoir surface represent the computed state
of initial maximum principal stress acting on this horizon. In regions remote from
and unaffected by the presence of faults, the maximum principal stresses
(green) correspond closely to the magnitude of the vertical or overburden stress,
meaning that the principal stresses are near-horizontal and near-vertical. The
regions of reduced stress (blue) are the result of stress-arching in areas where
the structural geometry and the stiffness of overburden layers create an
incomplete transmission of overburden weight onto the underlying reservoir. The
high maximum stress concentrations (yellows and reds) near the faults coincide
with inclined principal stresses, causing the magnitudes of the maximum
principal stresses to exceed lithostatic or overburden stresses generated by
gravity and the weight of the overlying rock mass. The black box in the upper
quadrant represents the area of study shown in the following figure (next page).

the reservoir. Fluid flow is modeled using a

reservoir simulator, such as the ECLIPSE
reservoir simulation package. By accounting for
these pressure changes in the stress calculations
using VISAGE software, it is possible to accurately
predict subsurface deformations and stress
changes, and evaluate their influence on material
properties such as permeability and porosity.
The resulting model can be used as a source
of stress data for several key stages:
well planningwellbore stability and optimum
drilling azimuth
well completionssand management
formation stimulationhydraulic fracture
field managementpressure maintenance
and injection


well integritywell design to accommodate

compac tion and subsidence as the well
is produced.
This coupled approach was recently used in a
North Sea field study. The South Arne field,
located in the Danish sector of the North Sea,
produces from the Maastrichtian Tor and Danian
Ekofisk chalk formations. Oil production from
the low-permeability chalk is driven both by
water injection and by compaction of the chalk.
In 2006, a field study of the South Arne field
was conducted to quantify the effects of
production from 1999 to 2005, and to assess
outcomes of a proposed development plan. The
field study was carried out using a historymatched ECLIPSE model and the VISAGE
geomechanical simulator. The geomechanical
study comprised four phases.

The goal of the first phase was to enhance an

existing reservoir model by adding more rock
layers and structural detail. First, the reservoir
model was extended up to the seafloor, adding 20
new layers and eight horizons for optimal
description of the overburden sequence. Ten
layers were added below the reservoir layer to
serve as underburden, and eight cells were added
on each of the four vertical boundaries to serve
as sideburden. Next, 45 faults and two different
fracture sets were incorporated into the
embedded model. The mechanical properties
were determined based on laboratory tests, core
calibration and literature reviews. A 1D stress
calibration was determined from density log
integration, leakoff tests and pore-pressure
modeling based on wireline log data.
The second phase sought to characterize the
stress state prior to production operations. An
initial effective stress state was computed, based
on properties determined in the first phase. The
stress-state computations accounted for contrasts
in deformation and strength properties between
different rock layers, and also considered
discontinuity within the rock layers themselves
(left). The computed initial stress state was
checked to verify agreement with field data and
geological features relating to stress orientations,
stress magnitudes and fault orientation.
The goal of the third phase was to determine
the state of present-day stresses. The approach
called for both flow and stress modeling, starting
with the change in pressure predicted by the
ECLIPSE reservoir simulator. The changes in
stress and strain induced by production and
injection operations were then assessed using
the VISAGE geomechanical simulator. The
computed compaction at the top of the reservoir
was in good agreement with the estimated value
from 3D seismic inversion.
It was also important to assess the risk of well
failure. The coupled simulations demonstrated
that pore collapse within the reservoir layers
would cause compaction and subsidence, and that
differential pore collapse could result in localized
well failure (next page).
In the last phase, fluid-flow and stress
simulation was performed in which permeability
changed in accord with stress and strain
changes. After history-matching of production
and injection data, the geomechanical model
agreed with production history.

Oilfield Review



> Production-induced compaction. These figures correspond to the boxed area shown in the previous figure (page 52 ). Production-induced time shifts seen
from the 4D seismic response (left) closely match the pattern of computed plastic strains obtained through coupled numerical simulation (right). Maximum
compaction (red) follows the NW trend of horizontal wellbores (dark blue lines) in the upper part of this figure. As expected, the area of greatest compaction
corresponds to that part of the reservoir experiencing the greatest production and consequently the greatest depletion. The computed maximum compaction
of 1.45 m [4.76 ft] at the top of the reservoir was in good agreement with the estimated value of 1.4 m [4.59 ft] from 3D seismic inversion. The absence of 4D
seismic data (white zone) is caused by a gas cloud. Close agreement between the 4D seismic data and the numerical model reinforces confidence in model
results over the area where seismic data were not available.

Monitoring: Geomechanics and

4D Seismic Data
Once a field model is developed, it should be
periodically updated with data obtained through
monitoring. A variety of techniques have been
devised for monitoring field-scale geomechanical
effects. For example, global positioning systems,
bathymetry and borehole tiltmeter surveys have
been used to measure surface subsidence.
Reservoir compaction can be detected by monitoring casing collar movement, although this
method is not precise. Microseismic techniques
have been used to detect regions of movement
and rock failure during depletion, and are
particularly useful for identifying fault
movements and monitoring fracture creation
during injection and thermal recovery processes.21
Time-lapse, or 4D, seismic surveys are also being
used for geomechanical monitoring.22

Autumn 2007

Both seismic compressional and shear waves

are influenced by production-induced stress
changes inside and around a reservoir. Timelapse seismic surveys, which predominantly use
compressional waves, have long been used to
monitor reservoir changes caused by production.
Repeatedly surveying a reservoir over time
enables geophysicists to compare differences in
seismic attributes, such as reflection amplitudes
and traveltimes, between the initial baseline
survey and subsequent monitor surveys. These
differences are particularly useful in detecting
movements of gas/liquid contacts that occur as
reservoirs are produced. In recent years, 4D
seismic techniques have also been used to
monitor production-induced changes in reservoir
geomechanical properties.

Among the differences between baseline and

monitor surveys, geophysicists sometimes
observed shifts in seismic traveltimes to a
specific horizon. Initially, these discrepancies
were attributed to logistical problems associated
with repeating surveys over a reservoir: namely,
the difficulty in placing seismic sources and
receivers in exactly the same position for every
survey. The slightest mispositioning of sources or
receivers could lead to modified raypaths that
traveled through slightly different parts of the
subsurface, generating perturbations in observed
traveltimes. In the past, discrepancies in seismic
21. For more on microseismic applications: Bennett L,
Le Calvez J, Sarver DR, Tanner K, Birk WS, Waters G,
Drew J, Michaud G, Primiero P, Eisner L, Jones R,
Leslie D, Williams MJ, Govenlock J, Klem RC and
Tezuko K: The Source for Hydraulic Fracture
Characterization, Oilfield Review 17, no. 4 (Winter
2005/2006): 4257.
22. Doornhof et al, reference 5.


Vertical displacement, z

z, m







Distance, m



Change in vertical P-velocity, Vp





Depth, m


Vp, m/s

t, ms


Depth, m


Time-lapse time shift for vertical P-waves, t


Decrease in TWT Increase in TWT

Depth, m







Distance, m









Distance, m



> Changing seismic characteristics. Both change in geometry (top left) and change in seismic velocity (bottom left) influence seismic reflection traveltimes.
The seismic two-way traveltime (TWT) (right) gradually increases toward the top of the reservoir due to overburden stretching and associated velocity
decrease. The largest time shifts are observed around the producing wells. Inside the reservoir, the seismic velocity increases because of increased stress,
so the time shifts become smaller.

traveltimes were frequently attributed to

differences in acquisition geometry or to
processing artifacts.
However, seismic acquisition and processing
technology has steadily improved, so that sources
and receivers can now be repeatedly positioned

with high accuracy, allowing reliable measures of

traveltime changes as small as 1 millisecond.
With this level of accuracy, geophysicists are able
to use time-lapse seismic techniques for observing depletion-induced traveltime changes for a

23. Barkved O, Heavey P, Kleppan T and Kristiansen TG:

Valhall FieldStill on Plateau After 20 Years of
Production, paper SPE 83957, presented at SPE
Offshore Europe, Aberdeen, September 25, 2003.
24. Guilbot J and Smith B: 4-D Constrained Depth
Conversion for Reservoir Compaction Estimation:
Application to Ekofisk Field, The Leading Edge 21, no. 3
(March 2002): 302308.
Nickel M, Schlaf J and Snneland L: New Tools for 4D
Seismic Analysis in Compacting Reservoirs, Petroleum
Geoscience 9, no. 1 (2003): 5359.
Hall SA, MacBeth C, Barkved OI and Wild P: TimeLapse Seismic Monitoring of Compaction and
Subsidence at Valhall Through Cross-Matching and
Interpreted Warping of 3D Streamer and OBC Data,
presented at the 72nd SEG International Exposition and
Annual Meeting, Salt Lake City, Utah, October 611, 2002.
25. Hatchell PJ, van den Beukel A, Molenaar MM,
Maron KP, Kenter CJ, Stammeijer JGF, van der Velde JJ
and Sayers CM:Whole Earth 4D: Monitoring

Geomechanics, Expanded Abstracts, 73rd SEG Annual

International Meeting, Dallas (October 2631, 2003):
Hatchell P and Bourne S: Rocks Under Strain: StrainInduced Time-Lapse Time-Shifts Are Observed for
Depleting Reservoirs, The Leading Edge 24, no. 12
(December 2005): 12221225.
26. Hatchell et al, reference 25.
Hatchell and Bourne, reference 25.
Herwanger JV, Palmer E and Schitt CR: Field
Observations and Modeling Production-Induced
Time-Shifts in 4D Seismic Data at South Arne, Danish
North Sea, presented at the 69th EAGE Conference
and Exhibition, London, June 1114, 2007.
27. Herwanger et al, reference 26.
Sayers C: Monitoring Production Induced StressChanges Using Seismic Waves, presented at the SEG
International Exposition and 74th Annual Meeting,
Denver, October 1014, 2004.


growing number of fields. In the North Seas

Ekofisk and Valhall fields, combined observations
by reservoir engineers, geophysicists and geomechanics specialists have led them to conclude
that the soft chalk of the reservoir rock was
undergoing substantial reservoir compaction,
Herwanger JV and Horne SA: Linking Geomechanics
and Seismics: Stress Effects on Time-Lapse MultiComponent Seismic Data, presented at the 67th
EAGE Conference and Exhibition, Madrid, Spain,
June 1316, 2005.
Sayers CM: Asymmetry in the Time-Lapse Seismic
Response to Injection and Depletion, Geophysical
Prospecting 55 (September 2007): 699705.
Sayers CM: Sensitivity of Time-Lapse Seismic to
Reservoir Stress Path, Geophysical Prospecting 54
(September 2006): 369380.
Sayers CM: Sensitivity of Elastic Wave Velocities to
Reservoir Stress Changes Caused By Production,
paper ARMA/USRMS 06-1048, presented at the 41st US
Symposium on Rock Mechanics, Golden, Colorado,
June 1721, 2006.
Sayers CM: Sensitivity of Elastic-Wave Velocities to
Stress Changes in Sandstones, The Leading Edge 24,
no. 12 (December 2005): 12621267.

Oilfield Review

Top reservoir reflection

shifts toward later arrival
time and brightens

Bottom reservoir reflection

shifts toward later arrival
time and dims

Baseline survey
Monitor survey

> Monitoring compaction over time. A

comparison of traces using the same source and
receiver position between the baseline (green)
and monitor surveys (blue) shows the effect of
overburden stretching on the arrival time of the
seismic signal. Note the consistent shift toward
later arrival times of the monitor survey
compared with the baseline survey.

Autumn 2007

phenomenonthat of overburden stretching.23
Resulting traveltime changes are significant and
of a magnitude that could not be explained by
nonrepeatability of survey acquisition geometry.24
Seismic data confirmed that the reservoir
rock did not deform uniformly, and deformation
in the reservoir rock caused the surrounding rock
to deform. In this case, the differential deformation associated with reservoir compaction and an
arching effect in the overburden resulted in
compressive stress relaxation and corresponding
stretching in the overburden. Similar overburden
time shifts were subsequently reported above
high-pressure, high-temperature reservoirs and
certain deepwater-turbidite fields.25
The geomechanical implications of timelapse time shifts are evaluated with reservoir
geomechanical models to characterize
production-induced subsurface deformation and
to predict associated stress changes. Established
workflows allow geophysicists to compare
observed time-lapse time shifts against time
shifts predicted by the reservoir geomechanical
models.26 Both subsurface deformation and stress
changes influence the seismic traveltime, either
by changing the length of the path that a seismic
wave must travel or by altering the propagation
velocity of the seismic wave, respectively
(previous page). The workflows allow predictions
of traveltime changes to any point in a threedimensional subsurface model.
Traveltime changes can also be observed from
4D seismic field experiments (left). The
prediction and observation of 4D traveltime
changes may be used to validate and calibrate
reservoir geomechanical models and thereby
improve their capability to predict stress
changes for a variety of projected production
scenarios. Furthermore, laboratory measurements conducted on rock cores are helping E&P
companies learn more about changes in
ultrasonic velocities under various stress
conditions and saturation states. This allows
operators to better manage reservoir stress and
optimize the trade-off between compaction drive
of hydrocarbon production and unwanted
compaction problems such as wellbore failure
and reduced permeability.
At present, the observation of changes in
vertical traveltime is a common practice for
monitoring geomechanical changes such as
vertical stress and strain. This technique
provides useful information, and allows geophysicists to identify compacting and noncompacting

reservoir compartments. However, to understand

and predict other geomechanical factors, such as
wellbore stability or rock failure, the triaxial
stress state must be known. Recognizing this
need, Schlumberger and WesternGeco scientists
are exploring the use of surface-seismic 4D
measurements to characterize the change in
tensor-stress over time.27
Future Developments
The industry is striving to develop further
capabilities for integrating rock fabric into
geomechanics analysis, with the vision of
enabling operators to extrapolate information
from rock microstructure observations to the
core-sample scale, through the well-log scale and
eventually up to the seismic scale. This capability
will let operators track reservoir characteristics
along the extent of a play and beyond, to
locations where no well control exists. In doing
so, geomechanics may change not only the way
that fields are drilled and produced, but also the
way in which they are explored. To this end,
researchers at Schlumberger are actively
investigating new laboratory measurement
techniques, wellbore logging methods, seismic
measurements and modeling programs. Indeed,
computational capabilities already exist; it is the
actual rock, its fabric, and the relation of fabric
to rock behavior that must be further