Sei sulla pagina 1di 14

SPE 69718

A Coupled Reservoir-Geomechanics Model and Applications to Wellbore Stability and


Sand Prediction
Yarlong Wang, SPE, Petro-Geotech Inc. and Baoping Lu, China New Star Company

Copyright 2001, Society of Petroleum Engineers Inc.


and drawdonw, well stability and sanding risks are critically
This paper was prepared for presentation at the 2001 SPE International Thermal Operations controlled by (1) the solution-gas behavior, (2) the formation
and Heavy Oil Symposium held in Margarita, Venezuela, March 12-14, 2001
stiffness, and (3) the residual cohesion. The latter can be
This paper was selected for presentation by an SPE Program Committee following review of
senstitive to the wetting phase saturation. Considering
information contained in an abstract submitted by the author(s). Contents of the paper, as production enhancement and formation damage, the reservoir
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any porosity increase due to formation dilation is simulated, which
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
can generate a negative skin near a well and result in a
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper production enhancement. Our model can be used to calculate
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 both the enhanced production and the ranges of the enhanced
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
zone.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
Introduction
Abstract According to Shell, wellbore stability related problems cost
petroleum industry more than 500 million U.S. dollar each
Sand production and open hole collapse during production year1,2. Yet, such an estimate can well be in a conservative side
have been a great concern for production engineer, as the as many critical issues on production related stability are still
detrimental consequences are often associated to a desirable not well understood and adequately evaluated. For example,
high production. Sand production usually occurs under a high- we would be able to optimize production by minimizing the
pressure gradient in poorly cemented formations. Whereas well completion cost during underbalanced conditions if we
wellbore collapses could occur if a bottom hole pressure is could determine the onset of the stability accurately.
below a certain level or production reaches a certain rate, i.e. Concerning a wellbore integrity in the field, we often must
during underbalanced drilling, under a high rate production, or complete our wells with a casing or gravel packing strategy
when an aggressive well completion for high production is well above the regulated safety requirement, as we can not
used. To determine the critical flow rate or bottom hole quantify the onset of the instability in wellbore or perforation
pressure, a coupled reservoir-geomechanics model is cavity. An issue as such becomes extremely important when
developed to simulate the interaction between reservoir flow an off-shore production in a prone-sand formation is operated.
and deformation. The model is developed and implemented The cost in such operations are usually more costly than those
numerically in Finite Element method by fully coupling a on-shore, and yet sanding problem on an off-shore platform is
comprehensive geomechanics model to a three-phase reservoir usually strictly prohibited. Thus a design tool to accurately
behavior model. Gas component behavior is only considered determine the onset of sand production is absolutely essential
below a bubble point, and it remains inside the oil phase to balance the tolerance on sanding and wellbore stability and
before a critical gas saturation can be reached in our cost involved to achieve a certain stability factor. We include
simulations. The effects of multiphase behavior on near well both sand production and wellbore stability in our discussion
stresses and deformations, including high compressible gas as it is believed that both problems are associated to the
component in the solution gas phase (foamy oil) are analyzed. similar deforming and plastic yielding mechanisms. Wellbore
Both Mohr-Coulomb and Drucker-Prager criteria are collapse and solid production are believed to be caused by
introduced to outline the plastic yielding surface and to govern large in-situ stresses and resulted from an ultimate rock
the plastic flow. The effects of the stress-dependent formation formation damage near a well. In addition, drag forces induced
moduli and permeability changes are permissible. The final by the fluid in-flux (gradient or viscous drag) of reservoir fluid
stability and the onset of sand production are determined by a can also be a major driving force for sand production and
critical effective plastic strain and zero effective radial stress, wellbore stability in the underbalanced condition. It is highly
whichever condition occurs first. Our studies indicate that, desirable if we can cast the major mechanisms for rock
other than the well known factors such as bottomhole pressure collapse and sand production, and quantify such a process so
2 YARLONG WANG SPE 69718

that we can determine the onset of such instability through phase, or not rigorously coupled for stability analyses, many
information observed on surface ground. aspects of the problem such as nonlinear fluid flow or
Concerning primarily with production, the focus in this study geomechanics behavior and coupling effects may not be
is on the stability and sand production in the underbalanced evaluated effectively for wellbore stability and sand
conditions, i.e. sand prediction during production and stability production. Thus a numerical model coupling a multi-phase
during either production or underbalanced drilling. A fully fluid flow to the elastoplastic geomechanical deformation is
coupled reservoir-geomechanics model by finite element developed. According to the proposed model, a corresponding
method is presented here. A black-oil model with possible plastic yielding zone (or a disturbed zone) initiates from the
foamy oil flow coupled to a comprehensive elastoplastic wellbore or cavity tip and propagates further into the
model is presented. Both the onset of plastic yielding, sand formation, because of the time-dependent fluid pressure
production, and wellbore collapse are defined based on a change and the transient effective stresses change near a
combined criteria in which stress concentration and strain are wellbore. A possible absolute permeability change inside the
calculated and compared to critical strength and strain based yielding zone is also considered, as the dilatant deformation
on experimental tests. The coupling to a multiphase flow developed may enhance the porosity in the plastic zone. As a
presented herein, rather than a single-phase flow used primary unknown, saturation is changing with the wetting
conventionally, is absolutely essential because a multiphase phase pressure. The relative permeability is updated by the
flow condition often exists in various underbalanced saturation, which in turn changes the responses of pore
conditions and in the potential pay zones during production. pressure and the skeleton deformation.
The effects of capillary pressure, wetting phase change, and Note that fully coupled formulations have been presented by
the relative permeability on pore pressure and pore volume several researchers on reservoir compaction and steam
changes can only be addressed appropriately through a injection14,15,16,17. Near wellbore deformation and their impact on
multiphase flow-geomechanics coupling as such. stability and sand production are not addressed. The purpose
The wellbore stability and the onset of sand production have in this study therefore is to investigate wellbore stability and
been evaluated by considering the critical stress concentration sand production in an environment both loosely cemented
3,4,5,6,7
. Despite we recognize that wellbore instability is reservoir and multiphase fluid flow exist. An elastoplastic
triggered by concentrated stresses against the existing strength, model is used to characterize the loosely cemented reservoirs
a complicated deformation/rupture/collapsing process and the sanding and stability are defined by a critical plastic
involving solid-fluid interaction makes a self-consistent strain, as those criteria based on peak stress/strength
coupled model difficult to obtain. Often a model was exclusively for stability proposed in the past are often too
developed to calculate stresses alone, and strains calculated conservative. Also it has been observed that the hole collapse
separately, which may not be used to determine a general and sand production often occur in a post plastic yielding
stability problem adequately, when both deformation and stage6. There are no adequate criteria, which can be used to
stresses are required for stability analysis. A self-consistent determine such a critical stress for the final collapse after a
model should represent the constitutive behavior of the plastic yielding, unless a sophisticated hardening and softening
formation rock simulated, and can be verified by a physical model may be used. The criterion using the effective critical
model through experiments, before it may be used with plastic strain was suggested10 to determine the onset of sand
confidence in the field applications8. Parallel to the production. Using such a criterion they have demonstrated
constitutive modeling, the fact of that the corresponding reasonably accurate predictions can be made to the field
strength used to evaluate stability depends on multiple results on sand production. Thus in the following, we use this
mechanical and geometrical factors simply adds further criterion for the analysis of wellbore stability and sand
complication to the stability problems. production in our discussions.
Investigations on sand production have also been extensive,
but primarily limited to the areas of single-phase flow. Bratli
Coupled Two-Phase Flow and Geomechanics
and Risines4 studied sand arching and production initiation
Formulation
from a cavity. A critical flow rate before sanding was defined
for a single-phase steady-state flow. Weingarten and Perkins7 Basic Formulation: In general, to define stability through
extended such a study to a single-phase gas reservoir, in which stresses, strains and pressure gradient, three primary variables
the gas density is characterized as a function of pressure. The are of interest: formation volumetric strain, wetting phase pore
critical condition for sanding is also defined based on an pressure, and wetting phase saturation. A fully coupled
effective stress and a multiphase flow is not considered in such formulation containing the three variables has been presented13
a study, including an extended study for a non-Darcy’s flow to simulate a formation compaction problem. The solid
case9 and some comprehensive study on sand production10. displacement and three pressures are identified as the primary
Several simplified uncoupled multiphase flow formulations variables, whereas the saturation is treated as intermediate
are also developed by coupling the porosity change to stresses, variables. Earlier, Lewis et al.14 also considered a coupled two-
iteratively11,12,13. Because all these aforementioned models are phase flow problem, assuming a linear elastic deformation.
somewhat restricted by the fact that they are analytical, single- Similarly Li et al.17 also studied a two-phase flow problem, in
which the primary variables are displacement, wetting phase
SPE 69718A COUPLED RESERVOIR-GEOMECHANICS MODEL AND APPLICATIONS TO WELLBORE STABILITY AND SAND PREDICTION 3

pressure, and saturation. Following Li et al.17, we also define saturation in phase j is denoted by Sj, the volume of the fluid
the skeleton displacement, wetting phase pressure, and wetting in phase j in the volume V is SjφV. The specific mass
phase saturation as the primary unknowns. To simplify the discharge is related to the displacement velocities as follows:
computational procedure, the calculation of saturation can be
decoupled from those pressure and deformation, but still q s = (1 − φ) v s (2)
coupled to the pressure and deformation. Such a simplification
is used so that we can take advantage of the well-defined qj = Sj φ vj
coupled formulation for single-phase flow. The computational
procedure may be summarized as following: The relative density θj of Phase j is defined as the mass of
phase j per unit total volume of material. Since the density ρj
I. we calculate the displacement and wetting phase pore
is defined as the mass of phase j per unit volume of that phase,
pressure with an initial saturation;
the relative densities are given by:
II. the saturation can be calculated using the newly
calculated pore pressure and displacements; θ s = (1 − φ ) ρ s (3)

III. the saturation-dependent reservoir properties such as θ j = S jφρ j


relative permeability are modified; and
IV. The pore pressure and the skeleton deformation are The basic principle, with which the specific mass discharges
calculated again with the new reservoir properties for have to satisfy, is the conservation of mass. In order to
the next time step. highlight the geomechanics/reservoir coupling, immiscible
flow without chemical reactions is considered. Hereafter, j =
The pore pressure changes respond to either pore volume w, o and g denote immiscible water, oil and gas components.
change or/and boundary loading conditions. The gas saturation The conservation of mass of each component requires that:
in the oil phase may not change initially above the bubble
point. Below a bubble point, free gases may be released from ∂θ s
the oil phase, forming the so-called solution gas with a higher + ∇ ⋅ ms + q s = 0 (4)
∂t
compressibility. The saturation of each component (oil or gas)
in the solution-gas phase is changing with reservoir pressure in ∂θ j
the foamy zone, which is presented in Table I for heavy oil16. + ∇ ⋅ mj + q j = 0
Considering a process in which the water phase in the near ∂t
well formation is the wetting phase, water saturation may or
gradually increase during production. The rising wetting phase
saturation reduces the relative permeability and mobility of the ∂[(1 − φ ) ρ s ]
oil phase, which requires a pressure gradient increase if the + ∇ ⋅ [(1 − φ ) ρ s v s ] + q s = 0 ; (5)
∂t
same level oil production be maintained. Such an increase in
pressure gradient, which participates the force balance in the ∂[ S j φρ j ]
mechanical equilibrium equations, may increase the risk on + ∇ ⋅ [ S j φρ j v j ] + q j = 0
sand production and wellbore instability. Such a pressure ∂t
gradient increase also, at the same time, provides an important with
driving energy for slurry sand to flow toward the production
well during cold production. Sg + S w + S o = 1 (6)
A mathematical description is required to define the physical
where
processes characterized above. The specific mass discharges
(the discharge of mass per unit area of a coordinate system φ = porosity
fixed in space) are denoted by ms and mj, where s represents ρ = density
solid phase, j = w, o, g for water, oil and gas phases, S = saturation
respectively. These specific mass discharges are related to the v = displacement phase velocity
specific volume discharges (the volume discharges per unit q = source/sink term
area of a coordinate system fixed in space) qs and qj by18,19:
The equation of motion of fluid has the following form:
ms = ρ s q s (1)
kk rj
u j = S j φ (v j − v s ) = − ∇p j (7)
mj = ρ j q j µj

If the porosity of the reservoir is denoted by φ, then in an The phase pressures are related to capillary pressures through
elementary volume V the volume of the pore space is φV and the following equations:
the volume of the solid material is (1-φ)V. If the degree of
4 YARLONG WANG SPE 69718

Pcwo = p o − p w (8) Fully Coupled Two-Phase Flow: The solid velocity, which is
linked to the time-dependent volumetric deformation of the
Pcog = p g − p o reservoir formation, is usually not negligible near a wellbore
or perforation tunnels. Significant error either on the pore
In conventional reservoir engineering, where the solid medium pressure and deformation can be induced if such a component
is assumed stationary (vs = 0), uj corresponds to the Darcy is removed from the coupled formation20,21. If large
velocity. For deformable media where vs ≠ 0, a Darcy flow deformations are expected, such as those in poorly
describes fluid flow relative to the medium or solid phase flow consolidated formations, the contribution of the solid velocity
vs expressed in Equation (7). Thus when media deformation to the coupled flow system must be included. Thus in the
(i.e., effects of porosity or pore volume change) is considered, following, the solid velocity is included in the geomechanics
the fluid flow terms in the conservation equations depend also constitutive relationship and the derivation of the material
on the deformation through vs (and consequently porosity). balance equations, i.e. equation (10) will be used.
Equation (5) can be expanded into a form containing terms For an oil/water two-phase flow, an expansion of equation
involving the solid velocity, which can be related to the (10) gives:
porosity change. If the deformation is small, the solid velocity
can be neglected and simplified equations may be obtained. ∂φ ∂ε ∂( S o ρ o )
ρo So + ρ o S oφ v + φ
Using the mass conservation equations for a single fluid ∂t ∂t ∂t
component in phase j and the solid phase, i.e., Equations (5),  k ro 
where the terms φv and (1-φ)vs are equivalent to the fluid bulk = ∇ ρ o {k }(∇p + ρ o g ) + Ro
volumetric flux and the solid bulk volumetric flux,  µo 
respectively, the mass conservation for fluid phase in (13)
Equation (5) can be expanded into:
∂φ ∂ε ∂( S w ρ w )
∂ (φρ j S j ) kkrj ρwSw + ρ w S wφ v + φ
− ∇ ⋅ (ρ j ∇p j ) + v s ⋅ ∇( ρ jφS j ) ∂t ∂t ∂t
∂t µ
 k 
+ ρ jφS j ∇ ⋅ v s + q j = 0 (9) = ∇  ρ w rw {k }(∇p + ρ w g ) + Rw
 µw 
Dropping the third term in equation (9), which contains the A rigorous definition for porosity change has been defined 22:
solid velocity and gradient of the saturation, porosity, and
fluid density, assuming that it has a secondary effect on flow, dφ = (α − φ )[dε v + C s dp ]
the above equation can be simplied as:
and this expression can converge to the general simplified
∂ε ∂S j ∂p j kk rj expression commonly used in reservoir engineering:
S j v +φ + C j φS j − ∇ ⋅ (ρ j ∇p j ) + q j = 0 (10)
∂t ∂t ∂t µ dφ
= dε v
where 1−φ
1 d [(1 − φ ) ρ s ] ∂ ε v when the solid matrix compressibility is assumed zero, i.e. Cs
∇v s = − =
(1 − φ ) ρ s dt ∂t (11)
= 0.

∂ρ The fluid densities are normally defined as:


j ∂p
= C jρ
j
C j ( p − p0 )
∂t
j
∂t (12) ρ j = ρ je
Assuming the solid skeleton is stationary or vs ≈ 0, Equation where the subscript j denotes to w and o respectively if a
(9), in which the porosity change appears explicitly only in the water-oil two phase flow is considered. Substituting the above
accumulation term, is the standard mass conservation equation into (13) we have:
used in most reservoir simulators18,19. This effectively
decouples porosity changes from the fluid flow term, ∂ε v ∂p
facilitating an efficient but approximate treatment of the S wα + [C wφS w + (α − φ )C s S w ] w
∂t ∂t
coupling when the reservoir skeleton deformation is
 k rw  Rw
negligible. = ∇ {k}(∇p w + ρ w g ) +
In Equation (9), absolute permeability and porosity are  µw  ρw
functions of the deformation. The permeability is assumed to
change with porosity based Kozeny-Poiseuille correlation16.
SPE 69718A COUPLED RESERVOIR-GEOMECHANICS MODEL AND APPLICATIONS TO WELLBORE STABILITY AND SAND PREDICTION 5

∂ε v ∂p Equations (16) and (18) can be solved simultaneously for


S oα + [C oφS o + (α − φ )C s S o ] o skeleton deformation and pore pressure at a given saturation.
∂t ∂t Within each time increment, the saturation can be solved
k  R through (17). Thus a complete solution can be obtained.
= ∇  ro {k }(∇p o + ρ o g ) + o We emphasize here that, equation (16) has a similar format as
 µo  ρo those in a coupled single-phase flow, which is plausible as we
could use those well established algorithms for a single phase
here equation (11) is used and flow to achieve efficient solutions, while saturation treated as
d ∂ a known variable. One assumption made above is that the
= + vs∇ (14) capillary pressure is very small comparing to the total pore
dt ∂t pressures, which allows us to treat pressure in the oil and
For two-phase oil/water flow, we can define: water phase equal. This usually can only be applied when
reservoir pressure is high, but can trigger discrepancy near a
S = Sw ; So = 1 − S w = 1 − S (15) wellbore when the bottom hole pressure is extremely low.
Coupling Parameters: The fluid compressibility term C for
so that the above mass balance equations can be simplified as:
single phase has been defined as21:
∂ε v ∂p R R 9 µ (ν u − ν )(1 − 2ν u )
α +C − ∇[λ {k }(∇p + ρg )] = o + w (16) C = = φC f + (α − φ )C s
∂t ∂t ρo ρw 2GB 2 k (1 − 2ν )(1 + ν u )
where p can be considered as an average pore pressure, i.e. Considering a mixed fluid including three phases of water, oil,
and gas, the average fluid compressibility under such a
p = p w S w + po S o circumstance can further be defined as:
or the wetting phase pressure assuming the capillary pressure
is negligible. The other parameters are defined as: [ ]
C = C w S w + (C o S o + C g S g ) φ + (α − φ )C s

C = [C w S w + C o S o ]φ + (α − φ )C s where S o = S g + S o .
The gas saturation is defined as part of the foamy oil
k ro k
ρ = λ w ρ w + λo ρ o = ρ o + ρ w rw saturation, which affects foamy oil compressibility and oil
µo µw density (but not the foamy oil relative permeability before the
critical gas saturation). This is important during heavy oil
k ro k production when dissolved gases existing inside the oil diffuse
λ= + rw out of the oil into free gas below bubble point pressure23.
µo µw These gases remain in a discontinuous phase, forming a
unique type of oil - foamy oil, which makes the mixed oil
k ro k rw
λo λ = λw λ = λ w + λo = 1 phase more compressible. The introduction of the above
µo ; µw ; expression does not change the remaining process of the fluid
flow as the entrained free gas is assumed to be discontinuous
Further rearranging equation (13):
before a critical gas saturation can be reached.
∂ ( S w ρ w ) 1 ∂φ ∂ε v
+[ + ]( S w ρ w ) = Fs (17) In viscous heavy oil, a much larger critical gas saturation is
∂t φ ∂t ∂t required before a continuous gas phase can be established.
where Fs denotes to the defined reservoir properties and unknowns in This can justify the procedure used here to simulate the three-
the previous time step: phase system by the two-phase flow system. Once the critical
gas saturation is exceeded, the correlation between saturation
 k 
Fs = ∇  ρ w rw {k }(∇p + ρ w g ) + Rw and relative permeability must be defined. Table II displays
 µw
such a correlation for oil sands16, which may alternatively be

expressed in a continuous function in some case24:
and combing with equation (15), saturation may be solved in
an explicit manner to provide a complete solution for wetting k ro = 0.76( S o* )1.988
phase pressure, solid displacement and saturation.
k rg = 0.61(1 − S o* ) 2.782
The equilibrium equation for solid deformation can be written
as21: where
S o − S or
3 1 − 2ν 1  2 1 ∂ε v   C s  ∂p S o* =
∇ u+ = 1 −  1 − S wi − S or − S gc
1 − 2ν ∂t   C b
(18)
2 1 + ν Cb   ∂t
6 YARLONG WANG SPE 69718

where subscript or, wi, gc denote to residual oil, interstitial - M s -C  u  2 F 


water, and critical gas saturation, respectively. A fully coupled
= T ∆t  0 +
three-phase flow following the similar procedure above is C M f + C  p 0   0 
required, which is beyond the scope of this article.  2 
In reality for geomaterials, the above formulation may only
Numerical Discretization: To solve the above equation represent the incremental behavior. A plastic yielding surface
system, numerical solutions seem to be the only option separates stress states, which give rise to both elastic and
available. Finite element method has been considered the most plastic (irrecoverable) strains. To take account of complicated
powerful tool to solve nonlinear rock mechanics problem, and processes like cyclic loading the yield surfaces may move in
it is therefore used to solve the coupled reservoir- stress space ‘kinematically’ but in our case only immovable
geomechanics problem herein. Using finite element method in surfaces are considered.
a two-dimensional plane strain condition, the continuous The saturation then can be subsequently obtained in an
displacements u and v, in two principal directions which are explicit procedure by considering (17):
perpendicular to each other, are approximated in terms of their
nodal values, u1, u2, v1, and v2, through simple functions of the [M ]∂St + [W]S
w
w = Fs
space variable called ‘shape function’. That is, for example, ∂
where
u = N 1u u1 + N 2u u 2
(19) M = ∫ N Tfφ N f dV
p = N p1 + N p 2
1
p
2
p
V

W = u& ∫ N fT N f m T BdV + p& ∫ N fTφ C w N f dV


where N is the shape function and superscript u and p denoting
V V
to shape function for the displacement and pore pressure, s
respectively. It may be advantageous to apply a shape function Rw T Rw
Fs = ∫ N T
dV + ∫ N f dS +
for pore pressure with a lower order for a smooth solution25. A ρw ρw
f
discretized coupled formulation is written as: V A

M s u + Cp = F − ∫ ∇N Tf λλ w [k]ρ w gdV − p ∫ ∇N Tf λλ w [k]∇N f dV


V V
∂u ∂p where superscripts T and s denote to matrix transpose and
C T
−Mfp+C =0 (20)
∂t ∂t source term on the boundary surface.
where u is the displacement vector, p is the wetting phase pore
Elastoplastic Yielding and Flow: The volumetric strains or
pressure, M is the stiffness matrix, C the coupling matrix, and
displacements in the equilibrium equation, i.e. in equation (18)
C the matrix characterizing the liquid compressibility, should be used only in an incremental manner, as they are
subscripts f and s for fluid solid, respectively. To integrate often related to the external loading in a nonlinear manner, and
equations above with respect to time there are again many typically such a relation is loading history-dependent in porous
methods available, but we consider only the Crank-Nicolson geo-materials. This is because only a limited loading can be
method to further discretize the above equation as: sustained on most geomaterials before the irrecoverable
ω M s u1 + ω C p1 = (ω − 1) M s u 0 deformation is generated and unloading does not follow the
same loading path. Such a limited loading condition in
+ (ω − 1) C p 0 + F (21) geomaterials is often defined by a Mohr-Coulomb criterion.
The equation for the Mohr-Coulomb surface is expressed as
ωC u 1 − ω ∆t M f p 1 + ω C p 1 = ω C u 0
T 2 T
σ'1 + σ'3 σ' + σ'
+ ϖ C p 0 − ω (ω − 1) ∆t M f p 0 F = sin Φ − 1 3 − c cos Φ (23)
2 2
provided that the body force F is independent of time or the which, when rewritten in terms of invariance, becomes
above formulation is incremental. In the Crank-Nicolson type
J1 sin θ sin Φ
of approximation, ω, ranging between 0 to 1, is made equal to F= sin Φ + J 2 (cosθ −
½ in the above leading to the recurrence relation25: 3 3
M s C  u  − c • cos Φ ) − c • cos Φ (24)
 T ∆t  1
C − M f + C  p 1  where the Lode angle
 2  (22)
SPE 69718A COUPLED RESERVOIR-GEOMECHANICS MODEL AND APPLICATIONS TO WELLBORE STABILITY AND SAND PREDICTION 7

−3 3 J  π π
with the plastic increments being given by the flow rule:
arcsin  
1 (25)
θ = 3 − ≤ θ≤
3  2 J3/ 2  6 6 ∂Q
 2  dε p = κ (31)
∂σ '
where
The relationship between the total stress and strain is:
J1 = σ 11 + σ 22 + σ 33
dσ ' = M s dε (32)
1
J 2 = [(σ 11 − σ 22 ) 2 + (σ 33 − σ 22 ) 2 + (σ 11 − σ 33 ) 2 ] which can be expressed in terms of the elastic and plastic
6 components as
J 3 = σ 11σ 22σ 33 dσ ' = M s dε = (M e − M p )dε (33)
Plastic flow follows a similar equation as the Mohr-Coulomb where
yielding criterion: T
∂Q  ∂F 
J1 sin θ sinψ Ms   Ms
Q= sinψ + J 2 (cosθ − ∂σ '  ∂σ ' 
3 3 Mp = T
(34)
 ∂F  ∂Q
A+  Ms
− c • cosψ ) − c • cosψ (26)  ∂σ '  ∂σ '
where ψ is defined as the angle of dilation. In practice, ψ is and
considered equal to the internal friction angle, Φ (fully
associated flow), less than Φ (non-associated flow) or equal to 1 ∂Q
A= dγ (35)
zero (flow with no plastic volume change). κ ∂γ
Alternatively, Drucker-Prager criterion is also used as a which is equal to zero for the perfectly plastic case. Parameter
yielding function such as:
γ is a hardening parameter, κ a proportional constant, and F
the yield function, i.e. Equation (24) or (26). Note that a
2 4
F= s − (c • cos Φ − w( J 1 )σ sin Φ ) general expression for the plastic flow is written in Equation
3 3 − sin Φ (34), but the plastic potential function Q can be equal to a
plastic yielding function (i.e. Q = F), giving an associated
(27) plastic flow condition.
where the norm s is The final stability or onset of wellbore instability or sand
production is defined when the following criterion for the
s = s⋅s (28) effective plastic strain, εep , is satisfied:

and an empirical relationship is used to define w(J1) as:


ε ep = a0 + a1 J1 = [
2 p 2
(ε11 ) + (ε 22p ) 2 + (ε 33p ) 2 ] (36)
w( J1 ) = exp[− ξ ( J1 / pa )] 3
(29)
and typical data a0 = 0.02 and a1 = 0.008 have been suggested
where σ is the mean stress, s deviatoric stress, pa the
for sand production, if the compression is taken positive10. A
atmospheric pressure, ξ an empirical decay constant. The single value of 1.1062% was also used27.
Drucker-Prager is proposed to avoid possible non-smooth
“corner” defined by the Lode angle θ, and this criterion also Numerical Simulations and Discussion
has an advantage when the intermediate stress component
must be included for stability calculations. As mentioned Wellbore Stability and Sand Production: wellbore instability
above that deformation in porous geo-materials is often non- may be broadly defined as open hole collapse, stuck pipe, and
recoverable after the loading reaches a certain condition. This casing failure. Sand production on the other hand is defined
deformation stage is defined as the plastic deformation. when solid is produced from a well, which is typically
Because of the characteristics of nonlinear irreversibility and associated with perforated hole subject to underbalanced
path-dependency in plasticity, plastic deformation is conditions. From a geomechanics point of view, both stability
calculated incrementally. The incremental stress-strain and sand production are evaluated through the stresses and
relationship after yielding is: strain development, and both problems are defined when a
critical stress or strain condition is reached. A difficult issue
dε = dε e + dε p (30) has been to quantify such a critical condition. Traditionally
8 YARLONG WANG SPE 69718

wellbore stability is defined by comparing the hook stress perfectly plastic model. Thus, a strain-oriented model is likely
concentration with the uniaxial compressive strength. Because to avoid such a difficulty associated with stress/strength
the classic linear elastic model has been long suggested to be criterion. In addition, depending on the purpose of a well is
too conservative, a pressure-dependent model and a linear drilled for and completion strategies used, stability can be
elastic model with scale dependent strength are proposed5,27. defined differently: a stuck pipe and hole collapse by either
Even though a model as such may offer improvement on the shear rupture or compression are defined as instability
previous model in predicting the onset of hole instability, the problem, so does casing failure by shear. Strengths under
definition of a Young’s modulus depending on the minimum different loading conditions may be used for stability analysis,
principal stress is questionable28, as such a model itself which is inefficient and costly in practice. Finally different
theoretically can be unstable29. Sand production is believed to reservoir-rock types may only be simulated by an individual
be induced by a different failure mechanism from the well constitutive model, which itself remains a great challenge,
bore collapse, i.e. compressive or shear failure for wellbore theoretically, experimentally, and numerically. In summary,
collapse and radial tensile failure for sand production. A stability problems can be triggered by different mechanisms,
critical condition when the effective radial stress reaches zero and stresses analysis and strength measurement ought to be
has been suggested for the onset of sand production4. performed accordingly. All these complicate the simulation
Seemingly reasonable, the criterion also needs to be verified and analysis of wellbore stability. Nevertheless, stress and
both experimentally and in the field8. Our experiences in deformation analysis seems to be the most effective approach
numerical simulations suggest that it is difficult to achieve a for the analysis of sand prediction and wellbore stability.
negative effective radial stress anywhere from a wellbore wall Using a criterion by considering the effective plastic strain
into the reservoir formation when a cylindrical well is may offer partial solution for this complicated field problem.
simulated, even under the extremely unstable condition. This One must, however, understand the physical mechanisms
seems to be consistent with those concluded by Bratli and related to the field problems considered to adequately define
Rines4, who found that no sanding condition may be obtained the stability problems.
by the zero-effective radial stress criterion for a cylindrical Stresses Vs strength: The wellbore stability has been
well case until the whole formation surrounding the cylindrical conventionally evaluated by comparing the calculated stresses
hole fails. This condition seems to be in contradiction with with the measured strength, tensile or compressive. Such an
those in the field when a wellbore can be jammed by sand if approach can be problematic as neither the stress calculated
no sand control strategy is implemented. Another factor nor the strengths measured can be obtained with great
making our task more difficult is that the onset of wellbore confidence. On one hand, the stress calculation involves
collapsing or sanding is usually vaguely defined. For complex constitutive modeling and results produced are
example, production may impose a monatomic deformation difficult to be verified as all the confirmation for stress
until a wellbore loses its function as well. This final condition calculated can only be achieved indirectly, i.e. by comparing
is often not related to formation collapse, nor peak stressing to the slef-troubled strength. On the other hand, the measured
stage and failure in the geomechanics sense. Figure 1 displays strength, often used to verify stress levels, can be a function of
a stress distribution simulating a drilling process. The many other factors such as geometry, saturation, grain size,
tangential stress is usually building up during wellbore and sample size. It is practically impossible to pin down these
pressure reduction, but declines along with the radial stress at factors accurately for the purpose of determining the strength
the plastic yielding point (Figure 1b). A typical stress quantitatively at the wellbore wall in the field. An extreme
distribution produced at the balanced condition, i.e. when the example for such a shortcoming by using stress to evaluate
wellbore pressure is equal to the reservoir pressure, is wellbore stability is a borehole surrounded by an elastoplastic
presented in Figure 1a from the elastoplastic model. The perfectly plastic material. Under such a circumstance, no stress
noticeable peak tangential stress is correlating to the plastic build up can be sustained once the plastic yielding takes place
yielding, but may not represent the hole collapse or sanding on the borehole wall. Yet instability is surely taking place
condition, particularly in poorly consolidated reservoirs. On ultimately. There have been suggestions to use the range of the
the other hand, the strain tensors are usually built up plastic radius to determine the onset of instability30, but such
continuously (Figure 2). A limited deformation is reached an approach is difficult to be quantified through either
when the hole may lose its function as an integrated well. laboratory tests or fundamental physics.
During such a deformation period, neither the maximum nor Deformation and the Critical Effective Plastic Strain: the
the minimum principal stresses changes significantly (Figure magnitude of deformation may be alternatively used to define
1c). Figure 3 displays a relationship between the calculated formation collapse or sanding condition. Figure 4 shows the
stresses and strains during balanced and underbalanced deformation patterns near a well, Figure 5 shows the finite
conditions, respectively. Both the plastic yielding and sanding element mesh used, and Figure 6 presents the plastic yielding
are suggested. The former is defined by the peak stress and the zone and the final collapse or sanding zone in response to
latter through the degree of deformation, i.e. through the drilling and production time. Note no sand production occurs
effective plastic strain, whereas stresses may not build up after even if the critical effective strain is reached before an
yielding. Such a phenomenon may reflect the restriction of a underbalanced condition is imposed. The bottom hole pressure
SPE 69718A COUPLED RESERVOIR-GEOMECHANICS MODEL AND APPLICATIONS TO WELLBORE STABILITY AND SAND PREDICTION 9

is a continuous declining function of (drilling or production) Special thanks go to Dr. D.W. Towson for his support and
time. The critical effective plastic strain has been used for technical guideline of our work.
sand production prediction17,31. Although difficult to measure
such parameter in the lab, it can avoid some of the difficulties References
mentioned for the stress/strength approach. We extend such a
critical strain approach for stability analyses, which can be 1. Bol, G.M., Wong, S.W., Davison,C.J., and Woodland,
useful if calibrated in the field. D.C.,”Borehole stability in shale”, SPE 24975 presented
Reservoir mobility consists of reservoir permeability and fluid at SPE European Petroleum Conf., Cannes, 1992
viscosity. An increase in the mobility indicates that the 2. Mody, F.K. and Hale, A.H.,”A borehole stability model to
pressure gradient may drop subject to a constant bottom hole couple the mechanics and chemistry of drilling fluid shale
pressure. For loose sand reservoir a porosity or permeability interaction”, JPT, pp.1093, 1993
reduction may be expected near a well, whereas a permeability 3. Bradley, W.B., “Failure of inclined borehole”, ASME
increase may occur near a wellbore in the dense formations. Journal of Energy Resources Technology, 1979, 101, 232-
Depending on the field conditions, operators must design their 239
wells accordingly to maintain stability. 4. Bratli, R.K. and R. Risnes, “Stability and failure of sand
Bottomhole Pressure Decline Rate is also an important factor arches”, SPEJ, Trans. AIME, April 1981, pp. 236-248.
for stability and sand production. A rapid pressure drop 5. Santarelli F.J., E.T. Brown, and V. Maury, “Analysis of
corresponds to a condition with a sharp pressure gradient, borehole stresses using pressure-dependent, linear elasticity”,
making the near wellbore formation collapse easier during Int. J. Rock Mech. Min Sci. & Geomech. Abstr. Vol 23, 445-
production. One may also take advantage of such a mechanism 449,1986
by purposefully collapsing the near wellbore formations32.33. 6. Ewy, R.T. and N.G.W. Cook, “Deformation and fracture
Both sand production and stability during underbalanced around cylindrical openings in rock-I. Observations and
drilling are subject to an in-flux condition. Under such a analysis of deformation”, Int. J. Rock Mech. Min Sci. &
circumstance, the reservoir pressure in-flux gradient Geomech. Abstr. Vol 27(5), 387-407,1990
contributes to mechanical equilibrium state unfavorably. Both 7. Weingarten, J.S. and T.K. Perkins, “Prediction of sand
the fluid flow behaviors and the time required to imposing production in gas wells: methods and Gulf of Mexico case
certain borehole pressures become extremely important for studies”, JPT, July 1995, pp. 596-560.
stability. 8. Veeken, C.A.M., D.R. Davies, C.J. Kenter and A.P.
Capillary Pressure is yet to be considered in our numerical Kooijman, “Sand production prediction review:
model for deformation calculations, rather as a factor to developing an integrated approach.” SPE 22792, 66th
enhance the formation strength. In fine-grain reservoirs, Annual Tech. Conf. SPE, Dallas, Texas, 1992
capillary force may act as a bonding force to the formation, 9. Ong, S., R. Ramos, and Z. Zheng, “Sand Production
which is the key to prevent sand production from occurring as prediction in high rate, perforated and open hole gas wells”,
the near well formation in post-yielding stage possesses a SPE 58721 held in Lafayette, Louisiana, 23-24 Feb., 2000.
small residual cohesion. A detailed model including the 10. Morita, N., Whitfill, D.L. Massie, I., and Knudsen, T.W.,
capillary pressure profile is required to simulate its impact on “Realistic sand production prediction: numerical approach,
wellbore stability. SPE Production Eng. Feb., 15-24, 1989
11. Rodriguez, H., Fung, L.S.-K., Silva, R., Zerpa, L., and
Conclusion and Summary Wan, R.G., “Thermal simulation of horizontal wellbore
stability in unconsolidated heavy oil reservoir”, SPE
A coupled reservoir-geomechanics model is presented. The
37102 presented at Int. Conf. On Horizontal Well
model may be used for wellbore stability design and sand
Technology held in Calgary, 1996
production risk evaluation in a reservoir formation during
12. Osorio, J.G., H.Y. Chen, and L.W. Teufel, “Numerical
production or underbalanced drilling. A criterion based on a
simulation of coupled fluid-flow/geomechanical behavior of
critical effective plastic strain is used to determine the onset of
tight gas reservoirs with stress sensitive permeability”, SPE
hole collapse and sand prediction. Such a criterion is superior
39055, SPE 5th Latin American and Caribbean Petroleum
to the conventional stress/strength criterion as the peak stress
Engineering Conf., 1997
is often neither associated to the wellbore collapse nor sand
13. Settari, A. and F.M. Mourits, “A coupled reservoir and
production. A criterion based on strain can be practically
geomechnical simulation system”, SPE 50939, 1998
useful either for sand production and wellbore stability if
14. Lewis, R.W. and Y. Sukirman, “Finite element modeling
adequately calibrated in the field.
of three-phase flow in deforming saturated oil reservoirs”,
Acknowledgments Int. J. Num. Anal. Meth. Geomech., Vol. 17, 1993, pp.
577-598
This project is partially funded by the National Research 15. Lewis, R.W., P.J. Roberts, and B.A. Schrefler, “Finite
Council Canada. Their financial support is deeply appreciated. Element modeling of two-phase heat and fluid flow in
10 YARLONG WANG SPE 69718

deforming porous media”, Transport in Porous Media, wells in the San Juan Basin”, SPE 24906, 67th Annual
Vol. 4, 1989, pp. 319-334. Technical Conference and Exhibition of the SPE,
16. Tortike, S. W.: Numerical Simulation of Thermal, Washington, D.C., pp.501-516.
Multiphase Fluid Flow in an Elastoplastic Deforming Oil 33. Boone, T.J., R.J. Smith, and R. Galway, “Sand Packing Liner
Reservoir, Ph.D. Thesis, University of Alberta, 1991. Completions in Heavy Oil Reservoirs”, 6th One Day
17. Li, X., O.C. Zienkiewicz, and Y. M. Xie, “A numerical Conference on Horizontal Well Technology, Calgary,
model for immiscible two-phase fluid flow in a porous Alberta, November, 12, 1997
medium and its time domain solution”, Int. J. Num. Anal.
Meth. Geomech., Vol. 30, 1990, pp. 1195-1212
18. Verruijit, A.: “Elastic Storage of Aquifers,” Flow
Through Porous Media, R.J.M. De Wiest (ed.) Academic,
San Diego, California, 1969, 331-376.
19. Chen, H.-Y, L.W. Teufel, and R.L. Lee, Coupled fluid
flow and geomechanics in reservoir study-I theory and
governing equations”, SPE 30752, 1995
20. Biot, M.A., “General theory of three-dimensional
consolidation”, J. Appl. Phys. 12, 155-164 , 1941
21. Rice, J.R. and Cleary, M.P., “Some basic stress-diffusion
solutions for fluid saturated elastic porous media with
compressible constituents, Rev. Geophys. Space., 14, 227-
241,1976
22. Wang, Y. and Dusseault, M.B., 1991, “The effect of
quadratic gradient terms on the borehole solution in
poroelastic media”, Water Resources Research, 27 (12),
3215-3223, (1991)
23. Maini, B., H.K. Sarma, and A.E. George, “Significance of
Foamy-oil behaviour in primary production of heavy oils”,
JCPT, Vol. 32, (9), 1993, pp. 50-54.
24. Muqeen, M.,”Effect of Temperature on Three-Phase relative
Permeability”, Ph.D. Thesis, University of Alberta,
Edmonton, Canada, 1994
25. Smith, I.M. and D.V. Griffiths, Programming the Finite
Element Method, 2nd Edition, John Wiley & Sons, 1988
26. Papamichos, E., and M. Stavropoulou, “An erosion-
mechanical model for sand production rate prediction”,
Int. J. Rock Mech. Min Sci. & Geomech. Abstr. Vol 35(5),
531-532,1998
27. Van den Hoek, P.J., Hertogh G.M.M., Kooijman A.P., Ph.
De Bree, Kenter, C.J., and Papamichos E.,” A new
concept of sand production prediction: theory and
laboratory experiments”, SPE 36418, pp.19-33, 1996
28. Wang, Y., Kessler, N., and Santarelli, F.J.,”Stress-induced
anisotropy an dloading path effects on borehole responses”,
Int. J. Rock Mech. Min Sci. & Geomech. Abstr. Vol 30(7),
1367-1370,1993
29. Piau, J.-M, Personal Communication, 1989
30. Peden, J.M. and A.A.M. Yassin, “The determination of
optimum completion and production conditions for sand-free
oil production”, SPE 15406, 61st Anuual Tech. Conf., New
Orleans, Oct. 5-8, 1986
31. Kjorholt, H., H. Joranson, P. Markestad, A.M. Raaen, and R.
Viken, “Advanced sand production prediction in a user
friendly wrapping”, SPE/ISRM 47333, Eurock 98, July, 8-
10, 1992, Trondheim, Norway, pp.133-142
32. Palmer, I.D., M.J. Mavor, J.L. Spitler, J.P. Seidle, and R.F.
Volz, “Openhole Cavity completions in coalbed methance
SPE 69718A COUPLED RESERVOIR-GEOMECHANICS MODEL AND APPLICATIONS TO WELLBORE STABILITY AND SAND PREDICTION 11

Nomenclature

B,G Skempton’s constant and bulk shear modulus


C Compressibility, and subscript b,s,w,o denotes to
bulk, solid matrix, water and oil, respectively.
c Cohesion
Table I: Oil/water Two Phase Case
E,ν, νu Young's modulus, drained and undrained
Poisson's ratios SWT KRW KRO PCOW(MPa)
J1,J2,J3 First, second, and third stress invarinces ------------------------------------
k Absolute permeability 0.200 0.0000 1.0000 0
krj The relative permeability for phase j 0.250 0.0102 0.7690 0
Mj The stiffness matrix for j component, j =o,w,s 0.294 0.0168 0.7241 0
mj Fluid mass for j component, j =o,w 0.357 0.0275 0.6206 0
p Pore pressure 0.414 0.0424 0.5040 0
Rj Source/sink term for component j 0.490 0.0665 0.3714 0
Q Plastic Potential 0.557 0.0970 0.3029 0
Sj, Saturation for phase j. 0.630 0.1148 0.1555 0
u Displacement vector, 0.673 0.1259 0.0956 0
vj Average flow velocity in j phase, where j=w,o,s 0.719 0.1381 0.0576 0
α Biot constant 0.789 0.1636 0.0000 0
φ Porosity 1.000 0.2500 0.0000 0
σ1, σ3 The maximum and minimum principal stresses
κ,γ Plastic constant and hardening parameter
λ, λj Absolute and the relative mobility for fluid
phase j, where j = w,o
σr, σθ The radial and tangential stresses
p0,pw The external and internal pressures
σh The in-situ stress
Φ,ψ Friction and dilatant angles
ρj, ρs Fluid and solid skeleton’s density, where j=w,o
TABLE II:OIL/GAS RELATIVE PERMEABILITY
εr,εθ The total radial and tangential strains
εer,εeθ The elastic radial and tangential strains SL Krog Krg Pcgo(kPa)
-------------------------------------
εpr,εpθ The plastic radial and tangential strains
0.2900 0.0000 0.1700 10.342
µ Fluid viscosity
0.3950 0.0294 0.1120 7.281
ξ Empirical constant for formation stiffness 0.4330 0.0461 0.1022 6.178
0.5150 0.0883 0.0855 3.785
0.5690 0.1172 0.0761 2.213
0.6140 0.1433 0.0654 0.903
0.6630 0.1764 0.0500 -0.524
0.7190 0.2170 0.0372 -2.158
0.7500 0.2255 0.0285 -3.061
0.8050 0.2919 0.0195 -4.661
0.8500 0.3373 0.0121 -5.971
0.8990 0.5169 0.0026 -7.398
1.000 1.000 0.0000 -10.342
12 YARLONG WANG SPE 69718

Plastic Yielding Time

Figure 1b: Wellbore Wall Stresses During Drilling

Wellbore Collapse Time

Plastic Yielding Time

Figure 2a: Strains on Wellbore Wall During Drilling Figure 1c: Wellbore Wall Stress During Drawdown

Plastic Yielding Time

Wellbore Collapse Time (Strains)

Figure 2b: Strains on Wellbore Wall During Drawdown Figure 3a: Wellbore Strains-stress Correlation during Drilling
SPE 69718A COUPLED RESERVOIR-GEOMECHANICS MODEL AND APPLICATIONS TO WELLBORE STABILITY AND SAND PREDICTION 13

Wellbore Collapse Time

Plastic Yielding Time

Figure 3b: Wellbore Strain-Stress Correlation During Drilling Figure 3c: Wellbore Strain-Stress Correlation During Drawdown

Wellbore Collapse Time

Figure 3d: Wellbore Strain-Stress Correlation During Drawdown Figure 4a: Near Wellbore Displacement in X-direction

Plastic Yielding Time

Figure 4b: Near Wellbore Displacement in Y-direction Figure 5a: Plastic Yielding/Sanding Radii Change during Drilling
14 YARLONG WANG SPE 69718

Wellbore Collapse Time

Figure 5b: Plastic Yielding/Sanding Radii Change during Drawdown Figure 6: FE Mesh and Boundary Conditions for the Simulation

Potrebbero piacerti anche