Finite-Element Studies of Near-Wellbore

Region During Cementing Operations:

Part I
K.E. Gray, SPE, E. Podnos, and E. Becker, University of Texas at Austin

Summary or may not be centered in the wellbore, such as, for example,
A wellbore cement sheath is expected to provide zonal isolation casing lying on the low side of the hole in an inclined well.
and borehole integrity during well construction and well life. At the inside boundary, conditions are well known. Pressures
Mechanical interactions of the cement sheath to existing and inside the casing are readily measured, and the mechanical beha-
operationally induced stresses, along with other elements in pro- vior of the casing is well documented. From the casing/cement
ximity to the wellbore, have increasingly large technical, econo- interface outward through the cement, mudcake, plastic zone, and
mic, and environmental ramifications. formation, the stresses, displacements, fluid transport, and other
Staged-finite-element procedures during well construction behavioral features often are not known. Very little information
consider sequentially the stress states and displacements at and is available for quantifying interactions between and among
near the wellbore. The model replicates complicated stress states the several components. Outer interface stresses, displacements,
arising from simultaneous action of far-field stresses, overburden fluid-transport parameters, and coupling conditions change con-
pressure, cement hardening and shrinkage, debonding at the inter- tinuously as the formation pore pressure changes with fluid pro-
faces, and plastic flow of cement sheath and rock formation. At duction or injection. These 3D and time-dependent changes are
present, temperature, flow, and poroelasticity effects are not manifested through all of the system components, interfaces, and
included. The technique tracks the time-dependent behavior of component couplings or interactions.
cement slurry, curing (with or without shrinkage), and hardened The cement sheath is the heart of any oil or gas well. It is
cement during the critical period after slurry placement. expected to provide mechanical support to the casing string in
Material models for casing, cement, and rock formation and general; but in particular, the cement sheath should provide both
failure criteria for cement, formation, and interface bonds were mechanical and hydraulic isolation for all production horizons
calibrated using published information and experimental data. during well construction and for the operational life of the well
Calculations were conducted for various loading and unloading (life-of-well). Cement sheaths interact mechanically with other
scenarios, geometric configurations, properties of rock forma- physical elements in the wellbore region in response to pre-
tions, and cement-slurry formulations. Results are discussed in existing stresses from geological processes and to stressed from
terms of field implication;, for example: (1) Interface microchan- operational activities. Quantifying these interacting physical
nels may or may not develop, depending upon shrinkage magni- components and processes has technical, economic, and environ-
tudes; and (2) simplifying modeling assumptions that are often mental implications of great and growing significance.
used, such as 2D stresses and/or deformations, may obscure criti- This paper is part of a series to quantify behaviorally coupled
cal casing, cement, and formation behavior in the wellbore region physical components and processes at, and in close proximity
and in the producing horizon. to, the well bore. Comparisons are made of analytical results
This paper, part of a series quantifying the interacting physical and field implications from simple to increasingly more realistic
components and processes at and near the wellbore region, initiates and complex assumptions (i.e., isotropic/directional-stress states;
useful comparisons of analytical results and field realities. The ser- isotropic/anisotropic casing, cement, and formation material
ies illustrates and compares results and practical implications from parameters; stress and temperature cycling; time-dependent para-
simple to increasingly complex, but more realistic, assumptions, meters; rock-mechanics/cement-testing standards; static and
such as isotropic/directional-stress states and isotropic/anisotropic dynamic poroelasticity; and quantitative results and methods
casing, cement, and formation-material parameters. using analytical and numerical simulators).

Introduction Literature Review

During well construction and subsequent wellbore, near-wellbore, The literature concerned with long-term mechanical durability of
and reservoir operations, several physical elements at and in the the wellbore cement sheath can be classified into two groups:
vicinity of the wellbore interact in response to downhole pro- (1) studies that use experimental techniques in the laboratory to
cesses and events. The life of any well—both its longevity and understand the behavior of the cement in the wellbore and
its productivity—reflects an integration of the various processes (2) those that use modeling methods to gain this understanding.
and events that have occurred during drilling, completion, and Of course, there is no clear separation between these two
production operations. approaches—the demanding loading and temperature conditions
Fig. 1 illustrates important interacting components that have created in the laboratory are based on analytical and computa-
physically and operationally coupled behaviors. From the inner- tional predictions of the wellbore stress state. The models need
most surface outward, the components are casing, cement (per- realistic, experimentally measured material properties of the for-
haps with some mud or gas channels), mudcake on all or part of the mation, slurry, cement, and casing.
formation face, a “plastic zone” (which may be unconsolidated Some modeling approaches are based on simplifying
sand or rock under such elevated stresses that it behaves inelasti- assumptions that allow for an analytical solution. For example,
cally), and finally, a main rock formation, which is assumed assumptions of linear elasticity and axisymmetric geometry and
to exhibit elastic behavior (linear or nonlinear). The casing may loading allow one to employ Lamé’s solution for a cylinder
under pressure, or assumptions of linear elasticity and plane
stress condition allow one to use Kirsch’s solution for an infi-
Copyright ã 2009 Society of Petroleum Engineers
nite plane with a circular hole (Obert and Duvall 1967). How-
This paper (SPE 106998) was accepted for presentation at the Production and Operations ever, increasingly, the modeling approach involves extensive
Symposium, Oklahoma City, Oklahoma, 31 March–3 April 2007, and revised for publication.
Original manuscript received for review 2 February 2007. Revised manuscript received for
use of computational techniques such as finite-element meth-
review 2 April 2008. Paper peer approved 28 April 2008. ods. This allows one to incorporate the mathematical theories

March 2009 SPE Drilling & Completion 127

 Researchers (Sabins and Sutton 1986; Le Roy-Delage et al.
2000; di Lullo and Rae 2000; Thiercelin et al. 1998) provide
various estimates for the tensile strength of cement as functions
of chemical composition, neatness of slurry, setting time, and
environmental conditions, among other things. The reported
values are usually correlated to compressive strength, flexural
modulus, or flexural modulus of rupture.
While some researchers develop computational techniques
based on simplifications such as assumptions of linear elasticity,
axisymmetric geometry, or lack of cement shrinkage (Mueller
et al. 2004; Boukhelifa et al. 2005), most of the proposed compu-
tational methods recognize the necessity to encompass interaction
of all components of the well and to include the entire loading
history to obtain valid information about the state of stress in the
cement sheath at a given time and given location. Some investi-
gators choose to develop or employ analysis techniques that lack
the detailed analysis of the stress state of casing, cement, and
formation, but concentrate more on the loading history, such
as the stress-analysis model (SAM) described and applied in
Thiercelin et al. (1998), di Lullo and Rae (2000), Dean and Torres
(2002), and Stiles and Hollies (2002) or the system-response-
curve (SRC) technique as proposed in Fourmaintraux et al.
(2005). While such a method is less demanding computationally,
the simplifying assumptions such as linear elasticity and axisym-
metric geometry of SAMs, or a one-parameter description of the
stress state of each wellbore component as required by the SRC
model, are too limiting for accurate analysis of complicated stress
states arising from action of nonaxisymmetric far-field stresses,
overburden pressure, cement hardening and shrinkage, debonding
at the interfaces, and history-dependent plastic flow of cement
sheath and rock formation.
Other researchers prefer to concentrate on the stress state at a
particular stage of life-of-well, such as pressure or temperature
Fig. 1—Physically interacting components during well con- loading of the set cement, and apply finite-element-analysis tech-
struction (from Life-of-Well Rock, Fluid, and Stress Systems, niques to model this stage (Shahri et al. 2005; Rodriguez et al.
University of Texas at Austin, April 21, 2006).
2003; Fleckenstein et al. 2000). However, without consideration
of the previous loading and history of deformation, it is not possi-
ble to specify the initial state of stress, amount of plasticity and
of solid mechanics such as elasticity, plasticity, poroelasticity, damage, interface conditions, and other parameters necessary to
viscoelasticity, and sophisticated constitutive models and fail- accurately set up the problem to be solved.
ure theories of wellbore components, which are calibrated on During the last decade, significant progress was made in the
the basis of experimental data. In contrast to an analytical development of robust nonlinear finite-element-method techni-
approach where the accuracy of the solution is by and large ques, particularly suitable to applications in geomechanics, soil
limited by validity and applicability of the underlying assump- mechanics, and rock mechanics. This development is reflected in
tions, the accuracy of a carefully developed computational the publication of papers that successfully use a staged approach to
model is limited by accessibility of experimental data and finite-element analysis of wells and examine short- and long-term
availability of computational resources. Because most compo- behavior of the cement sheath of the wellbore. The staged
nents in a wellbore exhibit strong nonlinearity and history approach imitates construction of the well, following all or some
dependence, the stress- and thermal-loading history during all of the development stages such as drilling, casing, cementing,
stages of a wellbore life should be taken into account. completion, hydraulic fracturing, and production. By use of the
Most of the experimental research is concerned with the mea- finite-element method, the stress state at each stage is modeled,
surement and improvement of cement mechanical properties such and state variables such as the amount of damage and plasticity,
as elastic moduli, coefficient of thermal expansion, fracture along with loading and boundary conditions, are transmitted to the
strength, and bulk shrinkage and expansion properties. James and model for the next stage. This technique eliminates the need to
Boukhelifa (2008) emphasize the importance of consistent mate- guess the initial state of stress, amount of plasticity, and damage.
rial properties to be used in computational models concerned with In particular, the approach developed by Bosma et al. (1999), Ravi
long-term mechanical durability of the cement sheath. et al. (2002a), and Ravi et al. (2002b), which is implemented in the
For the aims of the present study, which is concerned with DIANA software package (TNO DIANA User Manuals), is not
cement-sheath integrity during life-of-well—and in particular, only used to improve the understanding of sealant behavior, but is
with the modeling of the cement-sheath behavior during slurry also used successfully to evaluate the integrity of the
setting and hardening—cement shrinkage and tensile-strength cement sheath in offshore wells (Ravi et al. 2003) and to redesign
properties are of extreme importance. the casing and cement sheath to match extreme loading conditions
The work of API TR 10TR2 (1997), Baumgarte et al. (1999), in previously failed wellbores (Heathman and Beck 2006). Pattillo
Parcevaux and Sault (1984), Chenevert and Shrestha (1991), and Kristiansen (2002) implemented a staged-finite-element
Backe et al. (1997), Chenevert and Jin (1989), and Sabins and model using the Abaqus software package (SIMULIA) and
Sutton (1991) report bulk shrinkage of slurry as a function of applied it successfully to the analysis of horizontal-casing integ-
cement brand, chemical composition, neatness of slurry, and rity. The approach described in the present paper is yet another
environmental conditions such as temperature and availability of implementation of the staged-finite-element approach to cement-
free water. The authors discuss various mechanisms responsible sheath modeling. While the overall approach is similar to the
for cement shrinkage such as chemical shrinkage, which contri- aforementioned papers in that the rigorous analysis of each stage
butes to both bulk and inner shrinkages, and filtration, and they of life-of-well is performed using the finite-element method and in
propose ways to reduce or even eliminate shrinkage. that analysis for each stage is based on that of the previous one, the

128 March 2009 SPE Drilling & Completion

assumptions, interpretations, and techniques used are different. Shrinkage. During the shrinkage stage, the cement reduces its
The specifics of the proposed method are explained in the follow- volume, which leads to deformation of the cement. While bonds
ing subsections. between cement and casing and cement and formation hold, the
deformation of the cement leads to the deformation of the casing
Stress State In and Near Borehole During Life-Of-Well. and formation, which might lead to plasticity and ultimately to
Unperturbed State. In its unperturbed state, the formation is, in damage and failure. Plasticity and failure might also occur within
the most general case, under a 3D state of stress—overburden the cement. The amount of cement shrinkage depends on tempera-
stress caused by the total weight of formation lying above the area ture and chemical composition of the cement. The amount of
of interest and horizontal stresses because of formation being deformation experienced by the cement, formation, and casing
constrained from unlimited expansion in horizontal dimensions depends on the far-field and overburden loading; hydrostatic state
and because of the presence of various geological conditions. of stress of the cement; material properties of the cement, forma-
The state of the wellbore at some particular depth is of con- tion, and casing; and history of plastic deformations experienced
cern. A horizontal section of an arbitrary, but small, thickness is by all of these components.
considered so that within the domain, there is no significant varia- If one or both of the bonds fail, the stress and deformation
tion of any of the quantities of interest in the vertical direction. distribution will be different, depending upon when failure occurs.
For the scope of this study, it is assumed that there are no shear Pressure Changes Caused by Completion Fluid, Hydraulic
stresses in any vertical plane. These stresses might arise because Fracturing, and Hydrocarbon Production. Effects caused by
of general anisotropy of formation material or because of nonver- completion fluid, hydraulic fracturing, and hydrocarbon produc-
ticality of the well itself. The treatment of these conditions will be tion are quite similar from a mechanics point of view. During
provided in subsequent papers. these stages, the fluid inside the casing applies hydrostatic pres-
Drilling. During the drilling stage, a cylindrical volume of the sure on the casing walls. This pressure is transmitted to the
formation is removed. This volume was under the state of stress cement and by the cement to the formation.
described above, and therefore was exerting traction on the sur- Change of the stress field leads to deformation. Depending on the
rounding formation material equal and opposite to that exerted by amount of this deformation, plasticity, damage, and failure of well
the formation. The removal of the material leads, of course, to components may occur. If hydraulic fracturing is performed, the
removal of the force exerted by this material on the surrounding pressure transmitted by the fracturing fluid is higher than that experi-
formation. Drilling fluid, which replaces the removed volume, enced during previous stages of a life-of-well. The production
exerts hydrostatic pressure on the surrounding-formation walls. “fluid” is either hydrocarbon-based fluid, which has a low density
Except for the special case of formation subjected originally to compared to that of slurry or mud, or pressurized natural gas. In any
equal far-field horizontal stresses, the pressure of drilling fluid case, the pressure at the casing walls at this stage is usually lower
cannot be sufficient to maintain the original equilibrium. The for- than that experienced during previous stages of life-of-well.
mation surrounding the removed volume of the well must deform. Both high pressure because of hydraulic fracturing and low
This deformation leads to creation of additional formation stresses, pressure because of well completion and hydrocarbon production
which together with hydrostatic pressure of the drilling fluid, are might lead to plastic deformation, damage, and debonding,
able to achieve equilibrium with the stresses induced by far-field depending on the near-field and far-field state of stress and the
horizontal loading and overburden vertical loading. In this aspect, history of deformations.
the “wellbore-stability” requirement defines the allowed range for
the hydrostatic pressure exerted by the drilling fluid, such that Important Aspects. Nonlinear Effects and Loading History.
deformation of the wellbore wall is within specified limits. Each deformation of the formation, casing, or cement might lead
Casing and Cementing. During the casing and cementing to plastic, and hence unrecoverable, deformation if yield criteria
stage, the steel casing is inserted in the wellbore. An effort is made are met. Plastic deformation in the formation might start as early
to align the casing in the center of the wall. However, because of as the drilling stage, and in the cement, during the curing of slurry.
deformation of the wellbore during the drilling stage described Accumulation of the plastic deformation might lead to damage,
above, and the general difficulty of placing a pipe many thousands which would result in cracking and/or creation of microchannels.
of feet long within fractions of an inch precision, the final casing Consequently, taking the loading history into consideration is very
location is not necessarily in the center of the removed cylinder of important. Modeling the wellbore state of stress by starting at
formation. At any given depth, the formation is still subjected to some particular time during life-of-well, and without taking into
far-field and overburden loading away from the well and to the account the loading history, could lead to erroneous conclusions.
hydrostatic pressure of the mud at the well. The casing is subjected For example, suppose that during the drilling stage, the combina-
to the hydrostatic pressure of the mud from inside and of slury tion of hydrostatic pressure applied by mud and far-field state of
from the annulus—the gap between the casing and the formation. stress led to plastic deformation and initiation of damage in the
These pressures are not always equal because there could be a formation near the well. The cracks located in this area might start
pressure gradient—for example, caused by mud friction. After the to grow and lead to the creation of undesirable microchannels
slurry has replaced the mud, it continues to apply hydrostatic during later stages of life-of-well, such as cementing, completion,
pressure on the casing and the formation. The value of this pres- or production. However, if modeling the problem were to begin at
sure is different from that of the mud pressure; it is defined by the one of these later stages, without accounting for history of pre-
slurry density, by the height of the cement top, and by the pressure vious damage, there would be no cracks to grow, and conclusions
maintained at the surface. The pressure inside the casing might be produced from such analysis would be erroneous, even dangerous.
the same or different, depending on cementing specifications. Mechanical Interpretation of Slurry to Hardened-Cement
Again, a change of the stress state in formation and casing will Phase Transformation. It is assumed that slurry does not “remem-
lead to deformations of near-well formation material and casing. ber” deformations that occurred while it was in a fluid phase. How-
Hardening and Shrinkage. There is a time period during ever, slurry in the wellbore is in a state of hydrostatic stress, which
which the cement slurry hardens and shrinks. Calculation of the must be inherited by hardened cement. Both the formation and the
stress state at the end of both processes is an objective of this casing were subjected to hydrostatic pressure exerted by slurry, and
work. For computational convenience, these two actions are sepa- according to Newton’s third law, they exert equal and opposite
rated during modeling (i.e., hardening and shrinkage are modeled pressure on the slurry, and later, on the hardened cement. Therefore,
as two separate sequential steps). it is assumed that the phase transformation from fluid to solid, which
Hardening. During the hardening stage, the fluid slurry happens in slurry, does not change the stress state of the surrounding
becomes solid cement. Because slurry as a fluid can be under only bodies (i.e., the formation and casing). In other words, if the cement
a hydrostatic state of stress, there will be no shear in the state of were to harden to be a stress-free material, it would need to deform
stress of solid cement. There is no deformation in the solid cement immediately to be able to sustain the pressure exerted by the sur-
by the end of this stage. rounding formation and casing, which is not the case.

March 2009 SPE Drilling & Completion 129

 Summarizing this discussion, it is assumed that at the end of Finite-Element Formulation
the hardening stage, the cement is strain free, has no shear stress, General Considerations. Modeling Goals. The first objective of
but is in a state of hydrostatic stress equal to that of the slurry. this paper is to introduce a general finite-element framework
Plasticity and Vertical Stress. Because overburden stress is that takes into account crucial aspects of the near-wellbore
usually quite uniform at a given depth, the 3D description of the stress state, such as the three dimensionality of the stress
problem is commonly replaced by a 2D description, such as plane- state and nonlinearity of the material properties, hardening and
strain or plane-stress formulation. However, even though the shrinkage of the cement, debonding of the cement, and failure
vertical stress component is eliminated from equilibrium and of the cement, formation, and casing. This framework
compatibility equations describing stress and strain state in the should allow for seamless inclusion of various additional con-
vicinity of the wellbore, it cannot be omitted from the constitutive ditions such as eccentricity and noncircular geometry of the
equations. For example, the majority of plasticity models have a well, general anisotropy of the formation material, sophisti-
nontrivial dependence on all three principal stresses. Inaccurate cated plasticity and failure models, poroelasticity, well inclina-
treatment of the plasticity will lead to an erroneous solution. Still, tion, and others.
2D models can be used effectively if they account for the vertical The second objective is to apply this modeling framework to
component of stress. One way of doing this is to define an initial the problem of cement shrinkage and to study cement/casing-bond
condition for the vertical compression in a plane-strain model. failure and follow the development of plastic regions in the for-
This would lead to a uniform value of the overburden stress across mation and cement.
the entire domain. Another, more-versatile, approach is to use a Example Problem. Using data from literature, an example
generalized plane-strain formulation that assumes that the model problem of a cementing job for a vertical well with a perfectly
lies between two rigid planes that can move with respect to each centered casing was defined and studied. Simple geometry of the
other. This approach allows a linear distribution of the overburden model problem facilitates explanation of the analysis technique.
strain. In cases where a large deformation of the borehole might However, the proposed method can be applied easily to a nonver-
lead to a significant nonlinear variation of the overburden-stress tical well with eccentric casing; these cases will be discussed in
field in the vicinity of the well, a full 3D approach is necessary. subsequent papers.
Load Control and Displacement Control. It is not always The amount of plasticity and debonding observed during the
obvious what is the best way to apply an overburden load on a subsequent analysis of the example problem appeared to be very
3D model of a wellbore. Typically, the overburden load is stated sensitive to far-field and in-situ loading. The loadings used in this
as a pressure distribution across the horizontal plane. Therefore, it example problem were selected to illustrate various nonlinear
seems natural to use this pressure as a traction-boundary condition effects that could occur in the near-wellbore region during
applied to the top part of the model. However, this boundary cementing operations:
condition does not impose any restrictions on the vertical displa- • True vertical depth of the well is 15,000 ft.
cement of the material. For example, a material that experiences • The top of the cement is 13,500 ft.
thermal expansion or chemical shrinkage and is free to deform • The hole is drilled with an 8½-in. bit, with an estimated
vertically might shrink or expand mostly in this direction, so that drilled hole size of 9½ in.
the strain, and therefore the stress, in the horizontal plane would • The casing outside diameter is 7.625 in. and inside diameter
be almost unperturbed. Clearly, this is not the case for an actual is 6.625 in., which is based on casing supplied by Hunting Oilfield
well where there is hardly any deformation in the vertical direc- Services (Schlumberger 2006).
tion, compared to that in the horizontal plane. Indeed, if one • Overburden-pressure gradient is 1 psi/ft.
considers a material point of cement, far from the top and bottom • Pore-pressure gradient is 0.62 psi/ft.
ends, there is no reason to conclude that, during the shrinkage of • Ratio of maximum far-field horizontal stress to overburden
the cement, this point would move up instead of down, or down stress is 0.78.
instead of up. Therefore this material point must remain in the • Ratio of maximum far-field horizontal stress to minimum
same horizontal plane. To model this condition accurately, the far-field horizontal stress is 1.18.
overburden load can be applied by means of a rigid plane that • Density of drilling fluid is 13 lbm/gal.
allows a free horizontal displacement of the underlying material, • Density of slurry is 15.8 lbm/gal.
but only a uniform vertical displacement and therefore a uniform • Density of completion fluid is 7.4 lbm/gal.
vertical strain. This loading mechanism is similar to that of the • Hydraulic fracturing is achieved by 9,000 psi increase in
generalized plane-strain assumption. However, this approach has pressure beyond completion-fluid pressure.
its own drawbacks, such as the fact that the vertical deformation • Hydrostatic pressure in the casing below the packer during
of the entire model in this case will be defined by the vertical production of hydrocarbons is 4,000 psi.
deformation of the stiffest element of the assembly. All material properties were obtained from open source litera-
The first approach can be seen as a load-control method, while ture. Elastic material properties for cement were based on mea-
the second approach is similar to a displacement-control method. surements reported in James and Boukhelifa (2008) for Class-G
The choice of the loading method requires a judgment call for cement obtained from four different sources. The Mohr-Coulomb
each particular situation. For example, analysis of the formation plastic parameters; internal friction angle, f; and cohesiveness, c;
layer lying below a layer of loose sand fits better to the load- are based on properties measured in Bosma et al. (1999) for Class-
control description, while analysis of a formation layer below a G cement. Cement-elastic and -plastic properties are reported in
basalt plate fits better to the displacement-control description. Table 1.

130 March 2009 SPE Drilling & Completion

 The casing is assumed to be linearly elastic, with Young’s to the top face of the domain. The far-field boundary condi-
modulus E = 3.0e7 psi and Poisson’s ratio v = 0.3. tions are applied to the two remaining faces.
The formation is assumed to be composed of sandstone and/or The size of the horizontal rectangle is selected such that it is
shale. Elastic and plastic properties of the rock are compiled from large enough to accurately represent the effect of the far field on
data provided in Thiercelin et al. (1998), Bosma et al. (1999), and the near-wellbore region. A short study was conducted to estimate
Obert and Duval (1967). Rock properties are labeled “R1c1” the required dimensions.
through “R3c3,” as reported in Table 2. Fig. 2 shows that in the vicinity of the borehole, the difference
Dimensionality. The current study is concerned with non- in maximum stress values between a domain with a size of 6
linear effects occurring near the wellbore during life-of-well. borehole diameters and one with a size of 5 borehole diameters is
As mentioned earlier, an accurate analysis of nonlinear effects, less than 1%. The selected horizontal dimension of the modeling
such as plasticity, depends critically on the ability to model a domain is 50  50 in., which is 5.26 times larger than the dia-
3D state of stress. Therefore, the analysis was restricted to the meter of the borehole.
3D models. The vertical dimension of the formation domain is 5 in., which
Modeling Software. The Abaqus finite-element package is 5% larger than the radius of the borehole. This height is suffi-
(SIMULIA) was selected for this study. Abaqus is a general- cient to eliminate the artifacts in stress distribution that result
purpose finite-element-analysis code specifically developed to from end effects.
solve nonlinear stress-analysis problems. It has full 3D capabi- Interface Bonds. For the purpose of this study, debonding is
lity, a variety of nonlinear solvers, an extended library of materi- allowed only at the casing/cement interface. Formation of a
als, and an ability to include user-defined materials. microannulus at the casing/cement interface is commonly given
as the cause of failure in zonal isolation (Nelson 1990). Interac-
Implementation. Geometry. The formation is modeled as a tion at this interface is modeled using a contact condition, which
parallelepiped-shaped domain. Because of the symmetry of the does not allow penetration of two materials into each other but
problem, only one-eighth of this domain needs to be modeled if allows for their separation. The model can be modified such that
symmetry boundary conditions are applied at the three planes there is no tension transmitted across the interface (no bonding
of symmetry. The overburden boundary conditions are applied strength), or some amount of tension can be transmitted before the

Fig. 2—Convergence studies. Left: Dependence of stress in the vicinity of the borehole on the size of the modeling domain.
Distribution of radial stress along direction of maximum horizontal stress for rectanglular domains 4, 5, and 6 diameters of the
borehole in size. Right: Distribution of radial stress along direction of maximum horizontal stress for regular and fine mesh.

March 2009 SPE Drilling & Completion 131

surfaces are separated. On the basis of the data from the literature 4. Shrinkage. On the basis of the data from the literature (API
(Sabins and Sutton 1991; Le Roy-Delage et al. 2000; di Lullo and TR 10TR2 1997; Baumgarte et al. 1999; Parcevaux and Sault 1984;
Rae 2000; Thiercelin et al. 1998), a bond strength of 1,000 psi is Chenevert and Shrestha 1991; Backe et al. 1997), conventional
assumed for the second case. oilwell cement shrinks anywhere from 0.5 to 7% in volume. A 5%
The contact condition is a source of nonlinearity in the model, cement-volume shrinkage was assumed for the purpose of this
in addition to that resulting from nonlinear-elastic material prop- study. Throughout the duration of this step, the cement elements
erties of the constituents. are prescribed with a volume shrinkage that is scaled linearly from
Loading Steps. The analysis starts with the formation at the 0% at the beginning of the step to 5% at the end of the step. While
initial state of stress in equilibrium with far-field horizontal stres- linear evolution of shrinkage is chosen for simplicity only, the
ses and overburden. The far-field and overburden loading remains method does not impose any restrictions on shrinkage history.
without any changes through the sequential steps. It is assumed that the top of cement is far away, and, therefore,
There is a total of seven loading steps in the model. no vertical displacement is allowed. The shrinkage of the cement
1. Drilling. At the beginning of the step, the cylindrical volume inflicts high stresses on the formation and casing and might lead
of the formation at the drill-hole location is removed. This step to the development of a plastic zone in the cement and formation
uses the “model change” feature of Abaqus (Stiles and Hollies and to debonding at the cement/casing interface.
2002). Using this feature, the elements corresponding to the 5. Completion. During this step, the hydrostatic pressure of the
drilled-out volume are deactivated. The slurry pressure is applied completion fluid acts on the inner surface of the casing. Because
at the surface of the wellbore. Because the remaining formation this pressure is lower than the previous pressure of the slurry, the
material is under the initial state of 3D stress, the formation near casing stretched under the previous loading will contract, impos-
the wellbore is not in equilibrium at the beginning of the step. By ing additional deformation on cement and formation.
the end of the step, the equilibrium has been reached, but the 6. Hydraulic Fracturing. During this step, the hydrostatic
original circular cross section of the well is deformed. pressure acting on the inner surface of the casing is increased by
The proposed approach is different from that chosen in Pattillo 9,000 psi beyond completion-fluid pressure. Under this load
and Kristiansen (2002), where the hole is defined in a stress-free the casing expands, forcing further compression of the cement
reference configuration and therefore, to obtain the proper shape and formation.
of the deformed borehole, modifications of the hole shape are 7. Production. The hydrostatic pressure acting on the casing
required. Because in the proposed approach the undisturbed con- below the packer is reduced during the production stage of life-of-
figuration in equilibrium in the initial triaxial state of stress is well. Similar to the result of the Completion-Step loading, the
assumed as the reference configuration, the removal of borehole casing contracts, thus reducing the compressive stress in the
material disturbs the equilibrium and automatically leads to a cement and formation.
change of shape of the borehole, which is similar to what happens The sequence of Completion, Fracturing, and Production Steps
in the field during the drilling of the formation. Therefore, after might appear to be redundant—pressure loading inside the casing
completion of this step, no additional iterations are required to is decreased, increased, and decreased again. However, because of
obtain the actual shape of the borehole. the nonlinearity of the problem—contact interaction between the
2. Cementing. At this stage, the pressure load corresponding to casing and cement, and accumulation of plastic deformation in
the hydrostatic pressure of the slurry is applied to the wall of the the cement and formation—the cumulative effect of this loading
borehole. Even though the applied pressure is uniform, the state sequence does not necessarily have a zero net effect; instead,
of stress near the wellbore is not. Therefore, deformation because it may lead to the failure of zonal isolation in the wellbore.
of the action of the slurry pressure may lead to the development of Loading and unloading because of perforation are not addres-
plastic regions in the formation. sed in this paper.
3. Hardening. Using the model-change feature of Abaqus Discretization. The model is discretized using 3D linear ele-
(SIMULIA), the elements representing the cement and casing are ments with full integration. The finite-element mesh is shown in
introduced. These elements are activated with no initial deforma- Fig. 3. Because interface contact plays a major role in the analy-
tion, but under an initial state of stress, equal to the slurry hydro- sis, the expected solution to the problem is not smooth, and, there-
static pressure. At the same time, the condition of contact is fore, the use of higher-order elements will not lead to faster
defined between the inner surface of the cement and the outer convergence.
surface of the casing. This paper does not discuss debonding at In the vicinity of the borehole, mesh size is selected such that
the formation/cement interface; therefore the outer cement surface there are at least 30 elements per 90 arc of the borehole and such
is “tied” to the surface of the hole (i.e., no separation is allowed). that all elements in the vicinity of the borehole have an aspect
The pressure loading because of the slurry, which was applied to ratio close to unity. A convergence study was conducted during
the walls of the borehole, is removed because the cement applies which the mesh was refined uniformly by defining twice as many
equivalent pressure. To maintain the equilibrium, a pressure equal
to the slurry pressure is applied to the inner surface of the casing.
In this step, the initial state of hydrostatic compression caused
by slurry pressure is imposed on the cement. Because the forma-
tion was already in equilibrium with the slurry at the end of the
previous step, no additional deformation of the cement or forma-
tion is required. As a result, by the end of this step, cement
elements are not deformed. The time-dependent behavior of
cement is not included in this paper.
The proposed technique allows for introduction of cement in
its reference state, without any history of previous deformations,
and under the state of hydrostatic compression inside of the well,
which is in the state of 3D stress.
The casing elements are activated simultaneously with the
cement elements. This is done for modeling convenience only.
The model accounts for small changes in the inner and outer
diameters of the casing, which are caused by hydrostatic compres-
sion. Replacement of the reference configuration from stress free Fig. 3—Finite-element mesh (green: formation; blue: cement;
to that of hydrostatically compressed does not affect the load red: casing; x-axis corresponds to direction of maximum hor-
response of the steel casing and, therefore, does not affect the izontal stress; y-axis corresponds to direction of minimum hor-
outcome of the solution. izontal stress; z-axis corresponds to vertical direction).

132 March 2009 SPE Drilling & Completion

Fig. 4—Equivalent-radial-stress distributions at the end of Cementing and Hardening Steps. Left: Casing and cement are deacti-
vated, uniform pressure applied to the wall of the borehole. Right: Cement and casing are in hydrostatic compression, which
is transmitted to the wall of the borehole.

elements in each direction. The results of this study confirm the Hardening. Figs. 4 and 5 illustrate the proposed method for the
described selection of the mesh size, as illustrated in the right- staged finite-element analysis. The left side of Fig. 4 shows the
hand side of Fig. 2. distribution of Von Mises stress by the end of the Cementing Step,
Solution Procedure. To implement the transformation from when the cement elements are still deactivated and the slurry is
slurry to solid accurately, “nonlinear-geometry analysis” was applying hydrostatic pressure to the formation. The right side of
assumed throughout the entire history of the loading. The problem Fig. 4 shows the same distribution by the end of the Hardening
is highly nonlinear because of the contact condition at the cement/ Step, when slurry hydrostatic pressure is replaced by pressure
casing interface and because of the presence of nonlinear-elastic applied to the formation by the cement elements. The distribution
materials; therefore, the Newton solver is used throughout the of stress in the formation does not change; the solid cement
analysis. At least 50 increments are requested during the Shrink- elements are activated and are free of shear strain.
age Step to capture the evolution of the stress state in the vicinity Stress distribution in the formation is not axially symmetric
of the wellbore accurately. because the minimum and maximum horizontal stresses are not
equal. Stress distribution in the slurry is hydrostatic and, therefore,
is axially symmetric, even for the studied case of asymmetric
Results loading.
Using the staged-finite-element-analysis model described, the The left side of Fig. 5 shows radial-stress distribution by the
stress states in the cement and in the formation in the vicinity of end of the Cementing Step. The right side shows radial-stress
the wellbore were analyzed for various combinations of cement distribution by the end of the Hardening Step. Stress distribution
and rock-material properties. The summary of the analysis is in the formation does not change, even though the radial load
reported in the following subsections. caused by slurry pressure is replaced by pressure applied by the

Fig. 5—Equivalent-Von Mises-stress distributions at the end of Cementing and Hardening Steps. Left: Casing and cement
are deactivated, uniform pressure applied to the wall of the borehole. Right: Cement and casing are in hydrostatic compression,
which is transmitted to the wall of the borehole.

March 2009 SPE Drilling & Completion 133

solid elements. The elements of cement and casing are under
initial conditions of hydrostatic stress in the slurry.

Debonding at the Casing/Cement Interface. Even though the

cement is shrinking, debonding at the casing/cement interface
does not necessarily occur. The lost volume can be accommodated
by the elastic expansion of the formation, cement, and casing, and
the elastic and plastic deformations of the cement and formation.
Debonding at the casing/cement interface was observed for
“stiff-material systems” that involve rock with a high value
of Young’s modulus and a high value of cohesive strength. In
“compliant-material systems,” rock deformed elastically (case of
low value of Young’s modulus) and/or plastically (case of low will develop in the vicinity of the borehole. Table 3 reports the
value of cohesive strength) to compensate for cement shrinkage. maximum amount of plastic strain observed in the cement, and
Another observation is that debonding does not necessarily Table 4 reports that for the formation.
start at the deepest location of the well. For example, for the The results show a strong dependence of the development of
material system with rock-type R3c3, cement-type A, and 1,000- plastic zones on the direction of the minimum horizontal stress.
lbf bond strength, the debonding did not occur at the maximum The plastic zone is mostly located at the side of the borehole with
well depth of 15,000 ft but was observed when the analysis was a normal collinear to the direction of the minimum principal
rerun for the shallower depth of 12,500 ft. stress, as illustrated in Fig. 7. If the amount of plasticity is small,
If debonding does happen, it is not necessarily observed first the plastic zone is localized. For the situation where there is a
by the end of the Shrinkage Step. For the material system given as great amount of plasticity, the plastic zone encircles the entire
rock-type R3c3, cement-type A, and zero bonding strength, Fig. 6 borehole. The width of the plastic ring is not uniform, with the
illustrates the creation of a microannulus at the casing/cement widest cross section again in the direction of the minimum hor-
interface by the end of the Production Step, after the wellbore izontal stress. This strong dependence on the orientation of the
has experienced loading not only because of cement shrinkage, minimum horizontal stress resembles the similar dependence of
but also because of the unloading/loading/unloading cycle of the development of debonding.
Completion/Hydraulic-Fracturing/Production Steps. Similar to debonding, the presence and size of the plastic
Even though the surface separation in this case is very small, zone depends strongly on the entire history of loading and defor-
0.00054 in. on average, a microannulus of 0.001 in. is sufficient to mation. For example, for the material system with rock-type
allow gas a flow path (Seidel and Greene 1985). Because fluctua- R1c1, cement-type A, and zero bonding strength, the plasticity
tions of the pressure in the casing during the production stage of starts not during the Shrinkage or the Hydraulic-Fracturing Step,
life-of-well might lead to further reduction of the diameter of the but only at the end of the Production Step (Fig. 7).
casing, the long-term zonal isolation is not achieved in this case.
The gap between the casing and cement is not uniform, as
shown in Fig. 6. The interface distance is smaller in the direction Conclusions
of the maximum horizontal stress and is larger in the direction of A review of literature concerned with wellbore cementing for
the minimum horizontal stress. This effect is probably based on zonal isolation is conducted, and the need for a careful, nonlinear,
the same mechanical principles as the conclusions of Hubbert and 3D analysis for vicinity of the wellbore during all stages of life-of-
Willis (1957) that hydraulic fractures should occur in the plane well is emphasized.
perpendicular to the least principal stress. A staged finite-element approach is proposed and discussed in
detail. This approach allows one to follow development of stress
Development of Plastic Zones. The high value of plastic strain and strain state during all stages of the life-of-well. The proposed
may lead to material damage, such as the creation of voids approach is specially designed to accommodate further extensions
or cracks, and consequently may lead to the failure of zonal that are listed in the Future Work section.
isolation. The method is illustrated using the example of a simple pro-
The lower the Young’s modulus and the lower the cohesive blem that involves cementing of a vertical wellbore for different
strength of the rock, the higher the probability that a plastic zone material properties of the formation. The behavior of the wellbore
components during life-of-well is observed, and the possibility of
zonal-isolation failure because of both debonding at the casing/
cement interfaces and material failure is discussed.
The results confirm the need for taking into account all speci-
fics of the complicated stress state near the wellbore, such as
triaxiality of stress, plasticity of the formation and cement, the
possibility of bond failures, along with the entire history of load-
ing and deformation, for an accurate analysis and prediction
of zonal isolation over the entire life-of-well.
This paper establishes a framework for a nonlinear, 3D ana-
lysis of the wellbore under various conditions, which will be
presented in the subsequent papers.

Fig. 6—Debonding along casing/cement interface at the end of

the Production Step. Angle is 0 at the direction of maximum
horizontal stress and 90 at the direction of minimum horizontal
stress. Field parameters: 15,000-ft depth. Material system: rock-
type R3c3, cement-type A, zero bonding strength.

134 March 2009 SPE Drilling & Completion

Fig. 7—Dependence of plastic-strain distribution on material system and loading conditions. Left: Rock-type R1c1, end of
Production Step; minimal plasticity. Right: Rock-type R3c1, end of Hydraulic-Fracturing Step; significant plasticity.

Future Work Oil- and Gas-Well Zonal Isolation in a Full-Scale Annular Geometry.
Based on the method discussed in the previous section, the follow- SPEDC 20 (1): 44–53. SPE-87195-PA. DOI: 10.2118/87195-PA.
ing modeling studies are ongoing, and results will be included in Chenevert, M.E. and Jin, L. 1989. Model for Predicting Wellbore Pres-
future publications: sures in Cement Columns. Paper SPE 19521 presented at the SPE
• Inclined wells and horizontal wells Annual Technical Conference and Exhibition, San Antonio, Texas,
• Eccentric casing USA, 8–11 October. DOI: 10.2118/19521-MS.
• High-temperature effects Chenevert, M.E. and Shrestha, B.K. 1991. Chemical Shrinkage Properties
• Anisotropic-material behavior of Oilfield Cements. SPEDE 6 (1): 37–43. SPE-16654-PA. DOI:
• Time-dependent material models 10.2118/16654-PA.
• Damage and failure Dean, G.D. and Torres, R.S. 2002. Novel Cement System for Improved
• Debonding at the formation/cement interface Zonal Isolation in Steam Injection Wells. Paper SPE 78995 presented
• Poroelasticity at the SPE International Thermal Operations and Heavy Oil Sym-
• Fluid flow posium and International Horizontal Well Technology Conference,
Subsequent publications will also take into account experi- Calgary, 4–7 November. DOI: 10.2118/78995-MS.
mental results from the life-of-well program on oilfield cements di Lullo, G. and Rae, P. 2000. Cements for Long Term Isolation—Design
regarding shrinkage; mechanical behavior during loading/ Optimization by Computer Modelling and Prediction. Paper SPE
unloading processes; stress/strain paths; uniaxial, biaxial, and 62745 presented at IADC/SPE Asia Pacific Drilling Technology,
triaxial strengths; and static and dynamic moduli. Kuala Lumpur, 11–13 September. DOI: 10.2118/62745-MS.
Fleckenstein, W.W., Eustes, A.W., and Miller, M.G. 2000. Burst Induced
Stresses in Cemented Wellbores. Paper SPE 62596 presented at the
Acknowledgments SPE/AAPG Western Regional Meeting, Long Beach, California, USA,
The authors would like to thank Ametek/Chandler Engineering, 19–23 June. DOI: 10.2118/62596-MS.
BJ Services, Chevron, ExxonMobil, and Schlumberger for support Fourmaintraux, D., Bois. A.-P., Franco, C., Fraboulet, B., and Brossollet, P.
of this work in “Casing, Cement, and Formation Interactions 2005. Efficient Wellbore Cement Sheath Design Using the SRC (Sys-
During Drilling, Completion, and Production Operations,” Life- tem Response Curve) Method. Paper SPE 94176 presented at the SPE
Of-Well Rock, Fluid, and Stress Systems Research Program at Europec/EAGE Annual Conference, Madrid, Spain, 13–16 June. DOI:
The University of Texas at Austin. 10.2118/94176-MS.
Heathman, J. and Beck, F.E. 2006. Finite Element Analysis Couples Cas-
ing and Cement Designs for HP/HT Wells in East Texas. Paper SPE
References 98869 presented at the IADC/SPE Drilling Conference, Miami,
API TR 10TR2, Shrinkage and Expansion in Oilwell Cements, first edition. Florida, USA, 21–23 February. DOI: 10.2118/98869-MS.
1997. Washington, DC: API. Hubbert, M.K. and Willis, D.G. 1957. Mechanics of hydraulic fracturing.
Backe, K.R., Lile, O.B., Lyomov, S.K., Elvebakk, H., and Skalle, P. 1997. In Petroleum Development and Technology. Transactions of the
Characterizing Curing Cement Slurries by Permeability, Tensile American Institute of Mining and Metallurgical Engineers, Vol. 210,
Strength and Shrinkage. Paper SPE 38267 presented at the SPE Wes- 153–168. Littleton, Colorado: AIME.
tern Regional Meeting, Long Beach, California, USA, 25–27 June. James, S. and Boukhelifa, L. 2008. Zonal Isolation Modeling and
DOI: 10.2118/38267-MS. Measurements—Past Myths and Today’s Realities. SPEDC 23 (1):
Baumgarte, C., Thiercelin, M., and Klaus, D. 1999. Case Studies of 68–75. SPE-101310-PA. DOI: 10.2118/101310-PA.
Expanding Cement to Prevent Microannular Formation. Paper SPE Le Roy-Delage, S., Baumgarte, C., Thiercelin, M., and Vidick, B. 2000.
56535 presented at the SPE Annual Technical Conference and Exhibi- New Cement Systems for Durable Zonal Isolation. Paper SPE 59132
tion, Houston, 3–6 October. DOI: 10.2118/56535-MS. presented at the IADC/SPE Drilling Conference, New Orleans, 23–25
Bosma, M., Ravi, K., van Driel, W., and Schreppers, G.J. 1999. Design February. DOI: 10.2118/59132-MS.
Approach to Sealant Selection for the Life of the Well. Paper SPE Mueller, D.T., GoBoncan, V., Dillenbeck, R.L., and Heinhold, T. 2004.
56536 presented at the SPE Annual Technical Conference and Exhibi- Characterizing Casing-Cement-Formation Interactions Under Stress
tion, Houston, 3–6 October. DOI: 10.2118/56536-MS. Conditions: Impact on Long-Term Zonal Isolation. Paper SPE 90450
Boukhelifa, L., Moroni, N., James, S.G., Le Roy-Delage, S., Thiercelin, presented at the SPE Annual Technical Conference and Exhibition,
M.J., and Lemaire, G. 2005. Evaluation of Cement Systems for Houston, 26–29 September. DOI: 10.2118/90450-MS.

March 2009 SPE Drilling & Completion 135

Nelson, E.B. ed. 1990. Well Cementing. Oxford, UK: Developments in SIMULIA. Abaqus/CAE User’s Manual. http://www.simulia.com/support/
Petroleum Science series, Elsevier. documentation.html.
Obert, L. and Duvall, W.I. 1967. Rock Mechanics and the Design of Stiles, D. and Hollies, D. 2002. Implementation of Advanced Cementing
Structures in Rock. New York: John Wiley and Sons. Techniques to Improve Long Term Zonal Isolation in Steam Assisted
Parcevaux, P.A. and Sault, P.H. 1984. Cement Shrinkage and Elasticity: A Gravity Drainage Wells. Paper SPE 78950 presented at the SPE Inter-
New Approach to a Good Zonal Isolation. Paper SPE 13176 prepared national Thermal Operations and Heavy Oil Symposium and Interna-
for presentation at the SPE Annual Technical Conference and Exhibi- tional Horizontal Well Technology Conference, Calgary, 4–7
tion, Houston, 16–19 September. DOI: 10.2118/13176-MS. November. DOI: 10.2118/78950-MS.
Pattillo, P.D. and Kristiansen, T.G. 2002. Analysis of Horizontal Casing Thiercelin, M.J., Dargaud, B., Baret, J.F., and Rodriquez, W.J. 1998.
Integrity in the Valhall Field. Paper SPE 78204 presented at the SPE/ Cement Design Based on Cement Mechanical Response. SPEDC 13
ISRM Rock Mechanics Conference, Irving, Texas, USA, 20–23 Octo- (4): 266–273. SPE-52890-PA. DOI: 10.2118/52890-PA.
ber. DOI: 10.2118/78204-MS. TNO DIANA User Manuals. TNO. http://www.tnodiana.com/.
Ravi, K., Bosma, M., and Gastebled, O. 2002a. Improve the Economics of
Oil and Gas Wells by Reducing the Risk of Cement Failure. Paper SPE
74497 presented at the IADC/SPE Drilling Conference, Dallas, 26–28 SI Metric Conversion Factors
February. DOI: 10.2118/74497-MS.
Ravi, K., Bosma, M., and Gastebled, O. 2002b. Safe and Economic Gas
ft  3.048* E–01 = m
Wells Through Cement Design for Life of the Well. Paper SPE 75700
gal  3.785 412 E–03 = m3
presented at the SPE Gas Technology Symposium, Calgary, 30
in.  2.54* E+00 = cm
April–2 May. DOI: 10.2118/75700-MS.
lbf  4.448 222 E+00 = N
Ravi, K., Bosma, M., and Hunter, L. 2003. Optimizing the Cement Sheath
lbm  4.535 924 E–01 = kg
Design in HPHT Shearwater Field. Paper SPE 79905 presented at the
psi  6.894 757 E+00 = kPa
SPE/IADC Drilling Conference, Amsterdam, 19–21 February. DOI: *Conversion factor is exact.
10.2118/79905-MS.
Rodriguez, W.J., Fleckenstein, W.W., and Eustes, A.W. 2003. Simulation
of Collapse Loads on Cemented Casing Using Finite Element Analy-
K.E. Gray received BS and MS degrees from the University of
sis. Paper SPE 84566 presented at the SPE Annual Technical Confer- Tulsa and a PhD from the University of Texas, where he is a
ence and Exhibition, Denver, 5–8 October. DOI: 10.2118/84566-MS. professor of petroleum engineering. His teaching and
Sabins, F.L. and Sutton, D.L. 1986. The Relationship of Thickening Time, research activities are in drilling, rock mechanics, wellbore sta-
Gel Strength, and Compressive Strength of Oilwell Cements. SPEPE bility, and geomechanics applications to managed pressure
1 (2): 143–152. SPE-11205-PA. DOI: 10.2118/11205-PA. drilling. Gray is a Senior Member, Life Member, Distinguished
Sabins, F.L. and Sutton, D.L. 1991. Interrelationship Between Critical Member, and Legion of Honor Member of SPE. He has served
Cement Properties and Volume Changes During Cement Setting. on and chaired a variety of SPE and AIME committees, forums,
SPEDE 6 (2): 88–94; Trans., AIME, 291. SPE-20451-PA. DOI: and applied-technology workshops; chaired an SPE Annual
Fall Meeting Program; and served twice as an SPE Distin-
10.2118/20451-PA.
guished Lecturer. Evgeny Podnos is a research engineer
Schlumberger. 2006. Well Services Field Data Handbook (i-Handbook), at the University of Texas at Austin. His activities deal
version 1,0,4,2. http://www.slb.com/content/services/resources/software/ with finite-element simulations and analytical modeling of
ihandbook.asp?entry=oilfield/ihandbook& complex physical systems incorporating 3D, nonlinear, time-
Seidel, F.A. and Greene, T.G. 1985. Use of Expanding Cement Improves dependent, and hysteretic behavior of solids and porous
Bonding and Aids in Eliminating Annular Gas Migration in Hobbs media. Podnos holds a BS degree from Leningrad Polytechnic
Grayburg-San Andres Wells. Paper SPE 14434 presented at the SPE Institute and MS and PhD degrees from the University of Texas
Annual Technical Conference and Exhibition, Las Vegas, Nevada, at Austin. Eric B. Becker holds BS and MS degrees from the
USA, 22–25 September. DOI: 10.2118/14434-MS. University of Texas at Austin and a PhD from the University of
California at Berkeley. He is a professor emeritus of engineering
Shahri, M.A., Schubert, J.J., and Amani, M. 2005. Detecting and Model-
mechanics at the University of Texas at Austin, where he
ing Cement Failure in High-Pressure/High-Temperature (HP/HT) developed curricula and research programs in computational
Wells, Using Finite Element Method (FEM). Paper IPTC 10961 pre- mechanics. Becker is a cofounder of the Texas Institute for
sented at the International Petroleum Technology Conference, Doha, Computational Mechanics (TICAM) and a pioneer in finite-
21–23 November. DOI: 10.2523/10961-MS. element methods now used as standard procedures.

136 March 2009 SPE Drilling & Completion