SPE 79698

Coupled Analysis of Geomechanics and Fluid Flow in Reservoir Simulation

P. Samier, Atef Onaisi, TotalFinaElf, G Fontaine, TotalFinaElf Norge
Copyright 2003, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the SPE Reservoir Simulation Symposium held in
Houston, Texas, U.S.A., 35 February 2003.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
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
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
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
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
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.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
Generally, in classical reservoir studies, the geomechanical
behavior of the porous media is taken into account by the rock
compressibility. Inside the reservoir simulator, the rock
compressibility is assumed to be constant or to vary with the
pressure of the oil phase. It induces some changes in the
porosity field.
During the depletion phase or the cold water injection of HPHT reservoirs, the stress state in and around a reservoir can
change dramatically. This process might result in rock
movements such as compaction, induced fracturing,
enhancement of natural fractures and/or fault activation, which
continuously modify the reservoir properties such as the
permeabilities and the fault transmissibilities.
Modifications of such parameters strongly affects the flow
pattern in the reservoir and ultimately the recovery factor.
To capture the link between flow and in situ stresses, it
becomes essential to conduct coupled reservoirgeomechanical simulations.
This paper compares the use of 5 types of approach for the
reservoir simulations:
A classical approach with rock compressibility using only
a reservoir simulator,
A loose coupled approach between a reservoir simulator
(finite volumes) and a geomechanical simulator (finite
elements). At given user-defined steps, the hydrocarbons
pressures calculated by the reservoir simulator are
transmitted to the geomechanical tool which computes
the actual stresses and feeds back the modifications of the
petrophysical properties (porosities and permeabilities) to
the reservoir simulator.
A one way coupling: this approach is a simplification of
the loose coupled approach, the modifications are not fed
back to the reservoir simulator.

A simplified approach using permeability and porosity

multipliers inside a reservoir simulator. These multipliers
are user defined curves and vary with the pressure of the
oil phase. This approach uses only a reservoir simulator.
A fully coupled approach where the structural and the
flow unknows (displacement, pressure, saturations) are
solved simultaneously.
These approaches are compared for 2 field cases
described below.
Introduction
The importance of Geomechanics in problems such as
wellbore stability, hydraulic fracturing and subsidence is well
known. In recent years, there has been growing awareness of
the importance of the link between fluid flow and
geomechanics in the management of stress sensitive
reservoirs1-9. Standard reservoir simulation of compaction
drive accounts for nonlinear porosity changes determined from
uniaxial strain tests on cores. In many cases, laboratoryderived compressibility must be adjusted to match the
contribution of compaction to total hydrocarbon recovery.
Geomechanical effects such as stress arching and non-unique
stress path are among the causes of discrepancy between
laboratory-derived and field compressibility factors. If
compressibility varies linearly with the mean reservoir
pressure, then predictive reservoir modeling can be achieved
without coupling between stress and flow. However,
geomechanical effects are rarely linear for a number of
reasons. These include load variations due to modification of
pressure, temperature and saturation, change of the mechanism
of production, progressive activation of faults and fractures
that affect mechanisms such as stress arching and a non-linear
stress path. Unlike standard compaction drive simulation,
there is no simple linear method to account for the effects of
stress on permeability especially for fractured systems, where
the changes of permeability might be directional, localized and
strongly non-linear.
There are several ways to achieve the coupling between
flow and stress 10-17. The most rigorous is done with fully
coupled simulators, which not only solve the flow and the
mechanical equations simultaneously, but also allow for
anisotropy and non-linearity of the rock constitutive model.
The feasibility and accuracy of such simulators, as far as
complex and large scale reservoir systems are concerned, have
yet to be proved. Partial coupling on the other hand consists of
linking a flow simulator with a stress simulator allowing a
good compromise between feasibility and accuracy. A one
way link from flow to stress simulator is often used for

SPE 79698

subsidence forecasts. In this paper this one way link is

established to assess the faults behavior and the risk of fault
reactivation in a high pressure (HP) field. Explicit or implicit
two-way links are proposed to capture the effect of stress on
permeability. In the explicit process, a stress analysis is
performed at the end of given time steps and the permeability
is updated from the beginning of the next step in the flow
simulator according to pre-established relationships between
permeability and stress and/or strain. However, to solve the
compaction drive problem, explicit coupling is not sufficient.
To ensure the compatibility of pore volume calculations from
the flow and the stress simulators, iterations must be
performed within each stress analysis step before proceeding
to the next stress step with or without permeability changes.
Poro-thermo-mechanical equations
The theory of porous media was first presented by Biot18 and
then developed by Coussy19. In the latter reference, the porothermo-elastoplastic constitutive equations are presented
as follows:

ij = Dijkl kl pl
kl bPpij 3K d Tij

m
1
pl
= Pp + b kk pl
kk + kk flT

0fl

(1)
(2)

Eq. 2 assumes that the solid part of the rock is plastically

incompressible. The plastic strain is obtained with the plastic
flow rule
g
 ijpl = 
ij

(3)

and from the consistency equation

d F (inv , h ) = 0

(4)

F describes the yield surface and is function of stress

invariants and of the hardening parameters. In particular the
modified Cam-Clay yield function is
Q2
M2

+ P2 2PPc = 0

Pc = Pco * exp pl
kk

(5)

(6)
Compaction drive
In standard reservoir simulation, the equations of flow are
independent of stress and the contribution of rock
compressibility to production is expressed as a function of
pore pressure20,21. The assumptions underlying this separation
between stress and flow are twofold:
The incremental effective stress ratio, defined as the ratio
of horizontal to vertical effective stress increments, is
uniform in the reservoir.
One of the stresses (the overburden stress normally)
remains constant or changes uniformly everywhere in
the reservoir.

The former assumption implies that:

h
= K
v

(7)

Both assumptions provide a unique relation between

stresses, pressure and temperature as follows
v = 0

h = (K 1) bPp + 3K d T

(8)
The multiplying factor (K 1) b is generally referred to
as the reservoir stress path. Eqs. 7 and 8 are sufficient to
calculate the volumetric strain as a function of changes of
pressure and temperature. For an elastic rock for instance, the
volumetric strain is given by :
bPp

elkk = (1 + 2K )
+ T
3
K
d

(9)

The pore compressibility to be used in the flow simulator

becomes:
Cel
pp =

1 Vp 1 b
b2
(1 + 2K )
=
+
Vp Pp K s
3K d

(10)

1 Vp b
= (1 + 2K )
Vp T

(11)

and
Cel
PT =

Equivalent relations can be established for elasto-plastic

rocks with a few differences insofar as the elasto-plastic
compressibility depends on the initial in situ stresses and
varies nonlinearly with pressure. Eqs. 10 and 11 illustrate how
the pore volume changes can be made independent of stress
and how rock compressibility for flow simulation can be
calculated from rock mechanical parameters, provided that the
stress path is known. In practice, petrophysical experiments
provide direct compressibility measurements by simulating the
depletion assuming uniform, uniaxial strain.
Generally, the compressibility derived from laboratory
measurements has to be adjusted to match the contribution of
compaction to fluids recovery. In fact, due to stress arching
and other geomechanical effects
The reservoir stress path is never uniform and does not
always follow a uniaxial strain condition;
The vertical stress changes during production and these
changes are not necessarily uniform.
Should the arching affect the compaction in a linear
manner, it would be possible to define an equivalent
compressibility to history match the compaction by
multiplying the laboratory-derived compressibility by a
constant factor
Ceq =

1
Vp

Vp
= Fc * C pp

lab
Pp field

{ }

(12)

SPE 79698

In complex situations however, the correction factor might

vary with time in relation with the evolution of pressure and
temperature patterns in the reservoir. Only coupled
simulations can then provide reliable forecast of
compaction drive.
Water weakening effects
Water injection in highly compacting North Sea chalk
reservoirs accelerated the compaction and subsidence rates,
contrary with the normal reaction to reservoir pressure
inflation. Chalk weakening by water is thought to be at the
source of compaction that overshadowed the effect of pressure
maintenance. Two main mechanisms of chalk weakening were
proposed. The first mechanism was inspired from the
principles of partially saturated soil and attributed the
compaction to the release of the capillary pressure upon
saturation, which in turn contributed to an increase of the
effective compressive stresses. The second mechanism related
the chalk weakening to a chemical reaction between the water
and the chalk, which reduced its yield stress. Experimental and
theoretical works have been conducted and it is now believed
that the latter is the dominant mechanism 22,24.
In standard flow simulation, the water weakening effects
are taken into account by defining a permeability and porosity
multiplier that is a function of the water saturation. The strong
effect on the field pore volume is depicted in example A.
Permeability versus stress
Changes of stress due to corresponding changes in pressure
and temperature induce deformation of the rock, which is
accompanied by changes in reservoir properties. Incremental
effective stress ratios K approaching unity lead to more
compaction, hence a reduction of the effective permeability.
Smaller values of K on the contrary might cause a dilation of
the rock and improve the effective permeability. Injection of
cold water enhances the effective permeability of naturally
fractured systems and promotes the creation of new fractures
that are labeled thermally induced fractures. Sealing faults
between reservoir compartments or within compartments
might become conductive because of slip driven by a pressure
differential across these faults.
In some flow simulators, empirical relationship between
permeability and pressure can be defined. Generally, these
relationships are obtained by performing permeability
measurements at various pressures during depletion tests.
Here, uniaxial strain paths are not necessarily representative of
the stress state in the whole reservoir as stated above. On the
other hand, there is no standard method for determining the
changes of fracture conductivity or transmissibility of initially
sealing faults.
Methods for modeling geomechanical effects in
reservoir simulation
A) Constant compressibility
This option is available in all reservoir simulators.
The pore volumes, only functions of the pore pressure, are
adjusted using

X 2
V P (P ) = V P Pref .1 + X +

where

( )

X = C. P Pref

C is the constant compressibility

V P Pref is the pore volume at reference pressure Pref

( )

B) Pseudo coupling 1: porosity and permeability

multipliers versus pore pressure
In this coupling, a conventional reservoir simulator itself
computes some geomechanics responses such as compaction
and horizontal stress changes through simple relations
between porosity and between permeability and stress
respectively. The porosity and the absolute permeability are
updated through an empirical model which is a function of
pressure. The empirical model is entered into the simulator as
tables of porosity and permeability versus pressure.
Since the set of curves can referred to several rock regions,
this option is more convenient even in the case of a reservoir
constituted of n different constant compressibilities.
This function can be nonlinear and can be divided into 2
regions: an elastic part (low variations of pore volume versus
pressure) and a plastic part (strong variations of pore volume
versus pressure) as shown in Fig 4.
This method is a significant and easily implemented
improvement compared to standard method A.
Pseudo coupling 2: porosity and permeability
versus pore pressure and water saturation
This method is similar to the previous method but the
empirical model depends also on the water saturation and
allows for water weakening effects.
C) Explicit coupling
This coupling is also called "one way coupling method" since
the information is transferred only in one way from a
simulator to a geomechanics module. The geomechanical
module we had access through this method had a specific fault
model computing relative displacement and stresses along
faults. In the case of a faulted geometry similar to case B, this
method was the most adequate for fault modeling compared to
the other methods available.
D) Iterative coupling
A unique relationship between pore volume change and pore
pressure is used to assess the pore volume variation in
reservoir simulation. In fact the pore volume variation depends
not only on the pore pressure but also on stress variations. And
the stress variations in the rock reservoir are function of: a) the
pore pressure and temperature variations b) the thermo-hydromechanical properties of the reservoir c) the geometry of the
geological structure: the computation of the mechanical
equilibrium of the structure during exploitation requires
modeling not only the reservoir but also its containment (or
surrounding rocks). This ensures that sufficient boundary
conditions at the peripheries of the model are imposed for
stress analysis purposes. The surrounding rocks consist in (see
Fig. 5):
The rock and soil lying between the seabed and reservoir
(overburden): its geometry, thickness and constitutive
properties determine the degree of transmission of reservoir

compaction to the surface as subsidence.

The rock layers adjacent to the reservoir (sideburden). The
sideburden influences dominantly the stress path and the
amount of compaction occurring in the reservoir: in case of
stiff sideburden, a part of the total overburden load is
transferred to the sideburden (arching effect). Consequently,
the vertical stress of the overburden is not entirely and
uniformly applied to the reservoir.
The rock lying below the reservoir (underburden): as the
sideburden is directly maintained by the underburden, a stiff
underburden allows greater arching than a less stiff
underburden and consequently influences both compaction
and subsidence.
The accurate stress field occurring in the structure -the
reservoir and its containment - can only be reproduced by a
geomechanics module.
Hence, to better characterize the pore volume variation in
reservoir simulation, one has to couple the standard reservoir
simulator with a geomechanics module. The two simulators
solve respectively, in a separate and sequential manner
reservoir flow variables and geomechanics variables, the pore
volume being the coupled term. Note that within the reservoir
simulation, the pore volume can be predicted by using either
the constant compressibility or the pseudo coupling. In this
explicit coupling frame, the iteration is performed on the pore
volume at selected users timesteps and controlled by a
convergence criterion.
E) Full coupling
The most adequate method would be a global simulation
solving simultaneously and implicitly the flow and the
structural analysis equations in the same linear system. We
had access to such a simulator but restricted in the reservoir
zone. It uses the same grid and the same numerical techniques
(finite volume method) for the reservoir and the
geomechanical. The main weakness of this fully coupled
software which is still at the development stage is that the
overburden, sideburden and underburden are not gridded.
Their effect is introduced using boundary conditions. In
particular, arching effects can not be accounted for.
These five methods were compared on examples A and B.
Examples
Field A : Compaction and subsidence
A simplified representative model of a large North Sea chalk
reservoir was constructed to evaluate the performance of
partial coupling in predicting compaction drive and
subsidence. The size of the original reservoir model has been
reduced from 49x69x12 (40572 cells) to a size of 29 x 19 x 5
(2755 cells) and simplified to a sugar box to be able to run
coupled simulations in a reasonable amount of time. The
central region extends over 3.81km by 2.286 km. The model is
shown in
Fig. 1. The production scenario consists in a
pure depletion phase during 15 years, followed by a water
injection program. The reservoir is populated by six producer
wells and two water injector wells and made of three

SPE 79698

materials. Table 1 summarizes the properties of the

five layers:

Layer 1 & 2

Layer 3

Porosity

40%

20%

Layer 4 & 5
38%

Kh

100mD

100mD

100mD

Kv

25mD

0,25mD

25mD

Table 1 -- Reservoir model properties

Five types of analysis are performed and compared:
1. A constant rock compressibility: this is the standard
reservoir modeling option. The constant rock
compressibility value is determined by linearizing the
compressibility curves obtained from uniaxial strain tests.
For a given material, two extreme values can be
exhibited, as seen in Fig. 2. One value (C = 25.10-7 PSI-1)
corresponds to an elastic behavior (low pore volume
loss with pressure decrease), the other (C = 350.10-7 PSI-1)
corresponds to a plastic behavior (high pore volume
loss with pressure decrease). Both behaviors are
simulated.
2. Pseudo-coupling 1: this option respects the non-linear
behavior of the compressibility curves and consequently
is a generalizing of Analysis 1. The non-linear behavior is
shown in Fig. 3 via the pore volume modifier parameter.
3. Pseudo-coupling 2: this case is the same as the previous
but now the porosity multipliers depend both on pressure
and on water saturation. The curves given the pore
volume loss as a function of the water saturation are
shown in Fig. 4.
4. Full coupling: a coupled flow and geomechanics analysis
is performed within a reservoir simulator. The overburden
is taken into account as a boundary condition and
modeled as constant load. In this case, the geomechanical
model and the reservoir model are identical.
5. Iterative coupling: with this loosely coupled analysis, the
flow and geomechanics equations are solved in separate
softwares. The geomechanical model is obtained by
modeling the overburden, sideburden and underburden
around the reservoir model (see definition in Fig.5).
Model 5, shown in Fig. 6 is built from the reservoir model
by adding 4 cells in the I+, I-, J+ and J- direction to create
the sideburden, 6 cells in the K+ direction for the
overburden and 3 cells in the K- direction for the
underburden. The lateral dimensions of the overall model
are 26.67 km by 16 km. The total simplified mesh
comprised 13986 8-node 3D-brick elements (37x27x14).
A total of 39312 degrees of freedom were solved. The
boundary conditions involve fully fixing the bottom
surface and the four side surfaces of the model in their
normal directions. The reservoir simulation is divided into
6 periods. At the end of each period, a stress calculation
is performed.
With the iterative coupling, it is possible to simulate the water
weakening via the reservoir simulator or via the geomechanics
modeler or via both simulators. In our example, two iterative
coupling simulations are performed. In one, the water
weakening is not activated. In the other, the water weakening

SPE 79698

is activated via the geomechanics modeler: the water

weakening effect is supposed to be initiated as soon as the
local variation of the water saturation exceeds 0,4 and leads to
a reduction of 75% of the consolidation pressure Pc introduced
in the Cam-Clay model.
The mechanical properties used in Case 4 and 5 and related to
the five layers and the overburden are given respectively in
Table 3 and in Table 4.
Constant compressibility and Pseudo coupling 1 It can be
shown that simulating the compaction with the constant
elastic compressibility approximates the pseudo coupling 1
simulation with a very good accuracy during the first 25 years
of production (see Fig.7 to Fig.10).
On the contrary, a simulation performed with the constant
plastic compressibility overestimates the loss of pore
volume. As a result, the pressure maintenance due to the
compaction is also overestimated.
Fig. 7 shows that the constant elastic compressibility and
pseudo coupling 1 cases predict a loss of pore volume lower
than 0.8% at the end of the field exploitation. This value is
insignificant compared to the pore volume loss predicted by
others methods and does not fit the measures made on a highly
porous chalk field. This indicates that the chalk
compressibility obtained by performing tests on samples (local
scale) is not directly applicable at a field scale.
Full coupling and Iterative coupling Although the full and
iterative couplings are performed using the same
geomechanical data, they give different results. It seems that
there is no benefit to perform the full coupling simulation
instead of the pseudo-coupled simulation: both types of
simulations predict the same evolution in term of pore
pressure and oil and gas production.
On the contrary, the geomechanics implemented in the
iterative-coupled simulation modifies the results obtained with
the pseudo-coupled simulation. For instance Fig. 9 and Fig.
10 show that the iterative coupling is responsible for a 20%
increase of produced oil and a 8% increase for produced gas.
The pseudo-coupled analysis predicts a high pressure drop
after 3 years of production. The geomechanics delays this
decrease and acts as a pressure maintenance according to Fig.
8. It is interesting to focus on the pressure and pore volume
profile drawn during the 6 first years of production to
understand how the iterative coupled simulation works. This
interval corresponds to the first coupling-period defined by the
user. The geomechanics simulator performs the stress
initialization in year 0. Then the reservoir simulator calculates
the pore pressure and pore volume changes until the next
stress analysis, scheduled in year 6. The pressure decreases
highly (Fig. 8) but not enough to result in a pore volume
reduction (
Fig. 7). But from the point of view of the
geomechanics modeler, such pressure drop should disturb the
structural equilibrium and lead to compaction. Thus a pore
volume decrease in year 6 is predicted by the geomechanics.
This lowered pore volume, used as an input data in the
reservoir simulator at the beginning of the second couplingperiod induces a pore pressure gain. This explains the peak
observed after 6 years of exploitation. Finally, the iterative

coupling tends to maintain the field pressure during the early

years of production. The pressure drop computed by the
reservoir simulation is delayed by 4 years.
Water weakening effect. The effects of the water weakening
due to a large water injection after 15 years of production are
now investigated. In the Pseudo-coupling and the Iterative
coupling, without water weakening option, the water injection
leads to a pore volume maintenance (
Fig. 7). If one adds a
water saturation dependency on the chalk strength, the
reduction does not stabilize and goes on decreasing as
expected. With the Pseudo-coupling, the water weakening
effect is responsible for 75% of the total pore volume
decrease, while this effect is reduced to 13% with the Iterative
coupling. This last value is close to the 10% measured on the
field. As a result of the increase of compaction, the water
weakening leads to a reservoir repressurization (Fig. 8). In the
last years of production this pressure increase reaches 1465
PSI with the iterative coupling and 2200 PSI with the Pseudocoupling 2. It can be seen in Fig. 9 and Fig. 10 that the oil
production and in a less perceptible manner the gas production
are also affected by the water weakening effect.
The implementation of the water weakening effect in reservoir
simulation is too simplified and does not fit the reality. In
reservoir simulator, the water weakening is directly
responsible for a pore volume decrease. It means that a
waterflooded cells compacts as soon as its water saturation
exceeds a critical threshold. In the reality, the water
weakening affects first the geomechanics properties of the
rock. Then, the weakened rock is going to compact only if its
environment allows it.
Updating of Pseudo coupling 2 In standard reservoir
simulations, the integration of geomechanics is simplified: a
rock cell located in the reservoir is supposed to behave in
the same manner as a rock sample does during laboratory
tests. The properties of the structure in which this rock cell
is placed are not taken into account. In other words a cell
behaves mechanically independently of its neighbors.
On the contrary, in coupled simulation, a cell responds
mechanically to a stimulation (pressure drop, water saturation
variation for instance) by considering also its environment: is
this cell close to the sideburden? Is it connected to a fault? Are
its neighboring cells exposed to compaction? . A cell-bycell analysis based on local pore volume variations versus pore
pressure proves that these local effects are not negligible. The
plot in Fig. 11 compares the porosity multiplier implemented
in the standard reservoir simulation (thick line) to the one
predicted by the iterative-coupled simulation (thin lines). In
the legend, each (x,y) couple refers to the location of the
considered cell.
In reservoir simulation one curve is assigned to all cells made
of the same material. This is why only one thick line is drawn
on the plot. The iterative-coupling enhances this simplified
description of the mechanical behavior by distinguishing the
cells close to the sideburden (X=1,Y=1) from the others. All
the cells exhibit a similar mechanical behavior, excepted the
one located at the periphery of the reservoir. The cells
connected to the sideburden undergo a low compaction
compared to the cells situated in the middle of the reservoir.

SPE 79698

This lowering, reaching 3% at low pressure is observed on

compacted fields: their flanks are maintained by the
sideburden. This phenomena is known as the arching effect.
Another cell-by-cell study has been performed to compare
the pore volume variation due to water saturation implemented
in the reservoir simulation to the one predicted by the
iterative-coupled simulation. The results in Fig. 12 show that
the water weakening effect implemented in the reservoir
simulator needs to be reduced to fit the geomechanics
prediction. But in both cases, it is activated as soon as the
change in water saturation exceeds 0.10. One can be surprised
to see that a 38% porosity chalk is more compacted than a
40% porosity chalk. But here again, the effects of water
weakening depend on the environment of the considered
cell. In our case, the 38% porosity layers located below the
40% porosity layers undergo a higher load, and consequently
are more subject to the compaction.
Now, the idea is to use the results of the iterative-coupled
simulation presented in Fig. 11 and Fig. 12 to enhance the
implementation of geomechanics in a pseudo-coupling 2
simulation. Enhanced porosity-multiplier versus pressure
and water saturation sets are built: a local geomechanical
effect like the arching phenomena can be simulated by
assigning in each layer two different curves of porosity
multiplier induced by pressure variations. A mechanical
behavior similar to the one predicted by the iterative coupling
for the cell (X=1,Y=1) is assigned to the cells close to the
sideburden. The remains cells reproduce the mechanical
behavior of the cell (X=2,Y=2).
As expected, the update of the Pseudo coupling 2 leads to a
better matching with the iterative-coupled simulation in term
of oil production, field pressure and pore volume
(
Fig. 13). The oil production gap observed between
the iterative coupling and the pseudo-coupling is now reduced
from 17% to 6%. The case simulating the sideburden effect
calibrates the pore volume loss accurately. On the contrary,
the gas production does not seem to be influenced by
this update.
This first example proves the benefit of using a flow and
geomechanical coupling to better capture the effects of stress
variation within highly compacting reservoirs. However,
coupled simulations are still complex and time consuming.
Table 2 compares the computing time required to complete
the five analysis. The CPU of the constant plastic
compressibility case is a reference. The computation needed to
achieve an iterative-coupled simulation is around 200 times
longer than a classical reservoir simulation! The CPU time for
case 4 can increase much significantly in case of a
plastic behavior.
Case

CPU

1
C=25 C=350
E-7
E-7
1,13

1,16

1,39

5
No
WW

With
WW

235

200

Table 2 -- CPU comparison (WW: Water Weakening)

Furthermore, the integration of geomechanics in reservoir
simulation increases the number of input data. Then, the

number of parameters to be tuned during the matching phase

is also increasing.
Field B : Faults activation
Field B is an HP/HT field situated in the North Sea. The initial
reservoir pressure is around 1100 bar, yielding a pore pressure
gradient of 2.10 sg at the top reservoir. The field will be
depleted quite quickly by about 800 bar. Five major faults
separate the reservoir in 4 compartments and there is an
uncertainty about their transmissibilities.
Fig. 14 shows the 5 major faults. During the depletion, the
faults could be reactivated, and generate communication
between isolated regions.
The initial reservoir model assumes that all major faults
are closed. The high temperature effects may have sealed the
interfaces along faults.
A geomechanical study (explicit coupling or one way
coupling) was performed to predict the evolution of stresses,
and the displacements of faults.
The study has been divided into two steps:
First, the last significant tectonic event which is at the
origin of the current pressure and stress state in the field and
reservoir is simulated in order to obtain the actual in-situ state
of stress;
in the second part of the study, dealing with the depletion
phase itself, the validated results of this first step will be used
as initial data to calculate the variations in stress as a result of
pore pressure changes in the reservoir as determined in
reservoir simulations .
This approach, although less straightforward, has two main
advantages: i) by construction, the initial conditions are much
simplified and ii) the input parameters used for the second
phase are validated during the first phase. The validation of
the results is obtained by a close matching with the
distribution of measured pressures. The benefit of such a
procedure is particularly obvious when simulating such
complex geological settings.
Boundary and loading conditions
For the first step, the actual in-situ stresses were obtained
by prescribing a lateral displacement of 125m on the right
vertical boundary.
Boundary conditions of the model at step 1 are indicated in
Fig 15.
For step 2, the right lateral boundary is horizontally fixed to
the value obtained at the end of the phase 1, after a
compression of 125 m.
Geomechanical characteristics of materials
The mechanical behavior of the various lithologies is modeled
using an elasto-plastic constitutive law with a Drcker-Prager
- Van-Eekelen yield criterion25. All the rock layers are
assumed to have a coupled hydro-mechanical behavior. Table
5 gives some mechanical and hydraulic characteristics of
layers, used inside the geomechanical model during that first
phase. The fault behavior is assumed to follow a MohrCoulomb frictional criterion. For all the faults, penalty
coefficients Kp and K, characterizing the reversible part of
fault behavior, and the friction angle have been respectively
chosen to 1 x108 N/m3 , 8 x107 N/m3 and 30.

SPE 79698

Young's Poisson's
Rock layers modulus
ratio
(GPa)
Overburden 1
Overburden 2
C
Rese
B
rvoir
A
Sides
Pentland
Trias
Basement 3

12.5
10
12.5
12.5
12.5
12.5
20
20
20

0.19
0.12
0.19
0.19
0.19
0.19
0.19
0.19
0.19

Cohesion Friction Porosity

angle
(MPa)
()
20
20
20
20
20
20
25
25
25

24
20
24
24
24
24
24
24
24

0.2
0.1
0.15
0.2
0.1
0.2
0.1
0.1
0.1

Table 5 -- Rock layers characteristics.

Hydraulic transverse transmissivity has been taken equal to

10-8 m/(s.Pa) for fault F1 and 5x10-16 m/(s.Pa) for the four
other faults. Longitudinal permeability and cohesion are nil for
the whole of the faults.
Results of the geomechanical simulation
To simulate the depletion, the first approach is to impose a
modification of the pore pressure at each gridblock. This
loading means that the evolution in time of the pore pressures
is imposed everywhere and consequently fluid flows exist but
are not considered. The values of pore pressure imposed at
each time step govern the values of effective stress that
influence the mechanical deformation. It is thus a partially
coupled behavior.
The values of pore pressure are obtained from a single
phase dynamic simulation of the reservoir production with a
finite volume reservoir simulator. An interface has been
developed to carry these pore pressures from the finite volume
grid to the 8-node finite element grid. For nodes outside to the
reservoir zone, pore pressures are set to the value obtained at
the end of the first phase.
The simulation of the reservoir during the depletion uses
the values of pore pressure at each month during the 6 first
months, and after every 6 months till 6.5 years of depletion.
Field stresses
After 6.5 years of depletion, the major principal stress
indicated in Fig. 16 is always quasi horizontal and Fig. 17
gives contours of the minor principal stress. The effective
stresses are globally increased as a consequence of the
depletion. Indeed, pore pressures decreasing and total stresses
remaining constant, effective stresses increase. All the
reservoir area is now in compression.
Fig. 18 shows the stress path, in p'-q invariant axes,
followed during the two phases by an element inside the
reservoir where the depletion is the most significant. The
evolution of the plasticity criterion (Van-Eekelen) is also
given on Fig 18, it can be noticed that the plasticity is reached
only at the end of the last tectonic event. Moreover, the
depletion during the second phase leads to a stress path
departing from plasticity.
Fig. 19 a and b respectively indicate the interface normal
stresses and the interface tangential stresses along the faults
after 6.5 years of production. Interface normal stresses are

always relatively high and homogeneous. For the faults F5, F3

and F2 in the reservoir, the interface normal stress increases
with regard to the initial state before depletion; Tangential
stresses apply in the same direction along the faults F5, F3 and
F2 located at the reservoir center, while they are in the
opposite direction along the faults F8 and F1 located at the
reservoir sides.
During the depletion, the faults F5, F3 and F2 continue to
close up (sign -) in the same manner as during the last tectonic
event; on the other hand, the faults F8 and F1 tend slightly to
open (sign +) (Fig. 20a, maximum closing displacement
is 0.54 m).
During the reservoir production, shear displacements
appear along the faults (Fig. 20b, maximum shear
displacement is 0.43 m). Faults have been reactivated. Shear
displacements along the faults F5, F3, F1 and F2 partly, during
the depletion are in the opposite direction with respect to those
observed during the last tectonic event, contrary to what
occurs at the fault F8.
Table 6 summarizes the faults displacements evolution
during the depletion.
Fault F1

Faults F3 F2 F5

Fault F8

Closing/opening
displacement

Opening (in
reservoir) and
closing elsewhere

Closing

Opening

Shear
displacement

direction

and directions

direction

Table 6 -- Faults behavior evolution during the depletion.

From this geomechanical study, we modify the reservoir

simulation model in increasing the transmissibilities for faults
F1 and F8. The 3 majors faults F3 F2 F5 in the central
reservoir area remain closed.
Two type of flow simulations were performed:
- a "fully" coupled simulation with the same software as
in case A. Since this simulation tool doesn't model the
surrounding rocks, only boundary conditions are given
(a constant stress load). Also inactive cells induce on
this software a zero boundary condition in
displacement, the initial simulation grid has been
updated in order to remove all inactive cells.
- a stand alone reservoir simulation with the pseudo
coupling facilities.
The iterative coupling could not be used on this model (12,700
gridblocks in the reservoir zone) in a reasonable amount
of time.
Six different simulations were compared:
1. Constant rock compressibility: the initial reservoir model
using the classical approach (linearized rock
compressibility 2.10-5 bar-1) and all major faults closed.
2. Pseudo-coupling 1: The nonlinear behavior of the
compressibility and the permeability is shown in Fig. 21
(reservoir): the reservoir model uses pore volume and
permeability multipliers versus pore pressure. All major
faults are closed.
3. Pseudo-coupling 2 with condensate banking at wells: the
permeability multiplier feature is used to model a

simplified condensate banking model at wells using a

permeability reduction versus pore pressure as indicated
in Fig. 22. The permeability reduction can reach a value of
0.37 at the end of the depletion.
4. One way coupling: the flow model with an updating of
faults transmissibilities according to the results of the
geomechanical study.
5. Full coupling - elastic : the coupled analysis is performed.
The surrounding rocks are replaced by a constant
overburden stress pressure of 1,250 bars , fixed horizontal
and free vertical displacement on the lateral boundaries.
The rock material is supposed to behave elastically.
6. Full coupling - elasto-plastic : the rock material is
supposed to obey a Mohr-Coulomb constitutive law as
indicated in Table 5.
Fig. 23 compares the evolution of field pore pressure between
the differents runs. Fig. 24 to Fig. 26 compare respectively the
value of the field gas production, field oil condensate
production and field water production at the end of the
simulation. The highest gas production is obtained by the full
coupled method and the hishest oil production by the constant
compressibility method.
Constant compressibility and pseudo-coupling
Use of a constant compressibility induces a reservoir
pressure 1% lower (359 bars versus 364 bars at the end of the
depletion). It overestimates by 1% the total oil production. The
gas and water field production are also overestimated by
respectively 2% and 12%. The difference in water production
is not significant since the total production of water is not
large (2% of the condensate volume). The lack of reduction of
permeability during the depletion with a classical reservoir
simulator explains these small increases in water and
gas production.
Condensate banking effects
The effects are smaller as expected. The reservoir pressure is
not affected. The change in oil and gas production is not
significant. Only the water production increases by 16%.
Pseudo coupling and fully coupling
Since condensate banking effects were not accounted for
in the coupled simulation, results are compared to the pseudo
coupling run without condensate banking effects.
The fully coupled method gives a lower pressure at the end
of depletion (316 bars versus 361 bars). In terms of
production, the differences are 2% in gas production and 1%
in oil condensate production. The main difference concerns
the water production which is 30% lower in the
coupled model.
Geomechanical behavior between full coupling and one
way coupling
The modifications of the transmissibilities for faults F1
and F8 decrease of 5% the gas production and have no effect
on the oil production. The main goal of the one way coupling
geomechanical study was a justification of the closure of the
faults in the central reservoir area. It indicates also a small
displacement (less than 1m) at the top of the reservoir and a
small relative displacement of the faults (less than 50 cm).

SPE 79698

The coupled run with elastic rock properties indicates

higher vertical displacements (1.20 m to 2.75 m at the top of
the reservoir). With an elasto-plastic material, the convergence
became critical (too much plasticity in a local zone at the base
of the reservoir) and the cohesion parameter needed to be
increased from 200 bars to 250 bars to carry on the simulation.
In this case, the vertical displacement varies from 0.8 to 2.5 m
at the top. Plasticity occur for elements close to Fault F1, F2
and F8.
Fig. 27 indicates the vertical displacement compared to the
evolution of the pore pressure. The differences in terms of
displacements and stresses between the one way coupling
model and the full coupled model could arise from
several causes:
the difference in the faults modeling: faults are not allowed
to slip in the coupled model and this constrained energy could
increase stresses; the surrounding rocks which are not
discretized in the coupled model.
Conclusions
The paper has presented several ways of introducing
geomechanical effects inside a conventional reservoir
simulator. These methods have been applied and compared on
two industrial field cases. Several simulations have been
performed increasing successively the different coupling taken
into account:
- porosity and permeability multipliers versus pore
pressure and water saturation,
- geomechanical study integrating the pressure maps
given by the flow model followed by a
reservoir simulation
- a fully coupled software where the geomechanical and
the reservoirs unknows (displacement, pressure,
saturations) are solved simultaneously. This software
is still in the development stage and the main
limitations are a geometry limited to the reservoir
zone and no relative displacements allowed for faults.
- an iterative coupling of a reservoir simulator (finite
volume technique) with a geomechanical simulator
(finite element technique) including the surrounding
rocks and a more sophisticated constitutive law.
For the first field case (chalk reservoir), the paper leaves no
doubt about the importance of using a flow and a
geomechanical coupling. Difference in hydrocarbon
production and reservoir pressure can reach 25 %.
For the second case example (HP/HT gas reservoir in a
faulted geometry), the difference in hydrocarbon production
remains acceptable (less than 5%) but this low difference is
mainly due to the good quality of the constitutive rock on this
specific field (high Young's modulus, good cohesion and
favourable behavior of the main faults). The geomechanical
effects could be more critical during depletion on another
field. The geomechanical study demonstrated that the 3 major
faults in the central area of the reservoir were not significantly
reactivated by the depletion and justified the use of the pseudo
coupling method for reservoir simulation.

SPE 79698

With the simulation softwares currently at our disposal,

only the pseudo coupling method can be used in practise for
each simulation run of a complete reservoir study (more than
50 runs). The one way coupling method can be used once at
the beginning of the study to justify some hypotheses. The
fully coupled method we tested doubles the CPU time and
could encounter convergence problems when large plasticity
zones occur. Nevertheless it can be used for sensitivity runs in
an industrial study with a realistic reservoir grid. However,
this method needs to be extended to discretize the
surrounding rocks.
The iterative coupling tool we tried is more adequate in
terms of physical behavior (overburden, sideburden,
underburden and arching effects can be accounted for) but it
seems that at the present time it can only be used on small
model (less than 5,000 blocks). There is much scope for
improvement in the modeling and flow effects in
hydrocarbon reservoirs.
Nomenclature
....Linear coefficient of thermal dilation

.....Cam-Clay
hardening parameter

....Tensor
of total strain

.....Tensor of elastic strain

el
.Tensor of plastic strain
pl
Porosity

Plastic multiplier

..Biot Modulus

Poissons ratio

Stress tensor

.Vertical effective stress

v
h
....Horizontal effective stress
b
Biots coefficient
Cppel
..Elastic pore compressibility due to pore pressure
CPTel .Elastic pore compressibility due to temperature
Cppep Elasto-plastic pore compressibility
D
.Tensor of elasticity
G
.Plastic potential function
Kd
.Elastic bulk incompressibility modulus
K
Incremental effective stress ratio
Ks
Elastic solid incompressibility modulus
..Relative
change of fluid density
m/Ofl
M
.Slope of critical state line
P
Effective average stress
Consolidation pressure in Cam-Clay
Pc
Pco
.Initial consolidation pressure in Cam-Clay
Pp
.Pore pressure
Q
Deviatoric stress
T
Temperature
Vp
Pore volume
Acknowledgements
The authors would like to thank TOTALFINAELF to publish
this paper. Special thanks are due to P. Gauer and S. Thibeau
(TFE-Norge) for their valuable contributions.
References
1.

Chen H-Y. and Teufel L.W.: Reservoir Stress Changes Induced

by Production/Injection, Paper SPE 71087 (2001).

2.

3.

4.
5.

6.
7.
8.
9.

10.

11.

12.

13.
14.
15.

16.
17.

18.
19.
20.
21.

Gutierrez M. Tunbridge, L., Hansteen H., Makurat A. and

Barton N.: Modelling of the Compaction Behaviour of
Fractured Chalk, Eurock94, 803-810, Balkema (1994).
Gutierrez M. Makurat A. and Cuisiat F.: Coupled HTM
Modelling of Fractured Hydrocarbon Reservoirs During Cold
Water Injection, 8th Int. Soc. Rock Mech. Int. Congress Proc.,
V3, 1387-1390, A.A. Balkema.
Gutierrez M. and Lewis R.W.: The Role of Geomechanics in
Reservoir Simulation, Paper SPE/ISRM 47392 (1998).
Osorio J.G., Chen H-Y. and Teufel L.W.: Numerical
Simulation of the Impact of Flow-Induced Geomechanical
Response on the Productivity of Stress-Sensitive Reservoirs,
Paper SPE 51929 (1999).
Settari T. and Mourits F.M.: A Coupled Reservoir and
Geomechanical Simulation System, SPE Journal, (Sep. 1998).
Someville J. and Smart B.G.D.: Coupled Reservoir Simulation
Applied to the Management of Production Induced StressSensitivity, Paper SPE 64790
Stone T., Bowen G., Papanastasiou P. and Fuller J.: Fully
Coupled Geomechanics in a Commercial Reservoir Simulator,
Paper SPE 65107 (2000).
D. Tran, A. Settari, L. Nghiem: New Iterative Coupling
Between a Reservoir Simulator and a Geomechanics Module,
paper SPE 78192 presented at the 2002 Oil Rock conference,
Dallas, Texas, 20-23 October.
R. Charlier, D Fourmaintraux, P. Samier, J-P Radu, C. Guiducci:
Numerical Simulation of the Coupled behavior of Faults during
the depletion of a HP-HT Reservoir, paper SPE 78199
presented at the 2002 Oil Rock conference, Dallas, Texas, 20-23
October.
A. Onaisi, P. Samier, N. Koutsabeloulis, P. Longuemare:
Management of Stress Sensitive Reservoirs Using two coupled
Stress-Reservoir Simulation Tools, paper SPE 78512 presented
at the 2002 Oil Rock conference, Dallas, Texas, 20-23 October.
D. P. Yale: Coupled Geomechanics-Fluid Flow Modelling
Effects on Plasticity and Permeability Alteration, paper SPE
78202 presented at the 2002 Oil Rock conference, Dallas, Texas,
20-23 October.
Chen H-Y. and Teufel L.W.: Coupling Fluid-Flow and
Geomechanics in Dual-Porosity Modeling of Naturally
Fractured Reservoirs , Paper SPE 59043 (2000).
Koutsabeloulis N.C. and Hope S.A.: Coupled Stress/Fluid/Thermal
Multi-Phase
Reservoir
Simulation
Studies
Incorporating Rock Mechanics, Paper SPE 47393 (1998).
Lewis R.W. and Ghafouri H.R.: A Novel Finite Element
Double Porosity Model For Multiphase Flow Through
Deformable Fractured Porous Media, Int. J. Num. Meth.
Geom., 21, 789-716.
Settari A. and Walters D.A.: Advances in Coupled
Geomechanical and Reservoir Modeling With Applications to
Reservoir Compaction, Paper SPE 51927 (1999).
Mainguy M. and Longuemare P.: "Coupling Fluid Flow and
Rock Mechanics: Formulation of the Partial Coupling between
Reservoir and Geomecahnical Simulators", Oil and Gas Science
and Technology - La Revue de l'IFP, n4, (2002)
Biot M.A.: General Theory of Three-Dimensional
Consolidation, J. Appl. Phys. 12, 155-164.
Coussy O. : Mcanique des Milieux Poreux, Ed. Technip
(1991).
Bvillon D., Boutca M.J. and Longuemare, P.: Contribution of
the Coupled HydroMechanical Theory in theEstimate of
Reservoir Production, Paper SPE 65175 (2000).
Boutca M.J.: Elements of Poro-Elasticity for Reservoir
Engineering, Revue de lInstitut Franais du Ptrole Vol. 47,
N 4 479-49.

10

SPE 79698

1,05
-1

C = 35,0E-06 PSI

1,00

porosity multiplier

22. Mat C. "Etude experimentale et modlisation mcanique des

effets du balayage l'eau dans une craie sature d'huile", PhD
Thesis (in French), ENPC (2001)
23. Dormieux L., Mata C., Sarda J-P.: "Mechanical modelling of the
chalk/water interaction: from micromechanical to macroscopic
modelling", Symposium on "Mathematical Models in Soil
Mechanics" (2000)
24. Schroeder Ch., Bois A.P., Maury V., Halle G. "Water/Chalk (or
Collapsible Soil) Interaction: Part II. Results of Tests Performed
in Laboratory on Lixhe Chalk to calibrate Water/Chalk Models",
SPE/ISRM 47587, Eurock'98-Rock Mechanics in Petroleum
Engineering,
25. Van-Eekelen, H.A.M.: "Isotropic Yield Surfaces in 3
Dimensions for Use in Soil Mechanics", Int. Jnl. For Num. And
Anal. Meth. In Geomechanics, 4, 98-101 (1980).

-1

C = 2,5E-06 PSI
0,95

0,90

Non linear compressibility (from unixial strain test)

Linearization with low compressibility coefficient (C=2,5E-06)
Linearization with high compressibility coefficient (C=3,5E-06)
0,85
0

1000

2000

3000

4000

5000

6000

7000

pressure P

(PSI)

Fig. 2 -- Linearization of the compressibility curve obtained with

uniaxial strain test (for a 40% porous chalk sample)

Porosity multiplier

1,00

0,98

0,96

0,94
20%

0,92

38%
0,90

40%

0,88
0

1000

2000

3000

4000

5000

6000

7000

Pore pressure (PSI)

Fig. 3 -- Porosity multiplier versus pore pressure for the 3

materials
1,00

Fig. 1 --The simplified reservoir model

porosity multiplier

0,95

0,90

0,85

20 %
38 %
40 %

0,80

0,75
0

0,1

0,2

0,3

0,4

0,5

0,6

water sauration

0,7

0,8

0,9

Fig. 4 -- Porosity multiplier versus water saturation for the 3

materials

SPE 79698

11

Seafloor

Layer
Porosity
Youngs modulus of rock

Overburden

bar
bar/m

Soil effective unit weight

Side
burden

Reservoir

Side
burden

Underburden

Fig. 5 --The reservoir and the surrounding rocks

Poissons ratio of rock

Biots coefficient b
Max horizontal stress coef

dimensionless

1&2

20%

38%

40%

133842

42754

37662

0,148

0,115

0,111

0,33

0,24

0,23

dimensionless

dimensionless

0,75

Min horizontal stress coef

dimensionless

0,75

Vertical stress inclination

degrees
bar

Cohesion
Friction angle

Radius of ellipse Pc
Saturated Radius of Ellipse
factor
Critical saturation
Hardening parameter H

H =

90
357

96

83

degrees

69

38

35

bar

43

10

Tensile stress cut-off

Material fluidity

4&5

dimensionless
bar

1
1085

dimensionless

125

98

0,75

dimensionless

0,40

dimensionless

1 + e0

= Compressibility coefficient
= Swelling coefficient
e0 = void ratio

15,6

20,2

20,8

Table 3 --Geomechanical properties of the reservoir

Overburden thickness
Vertical stress gradient
Vertical stress on the top of the reservoir
Average Youngs modulus (from core)
Average Poissons ratio (from core)

2845 m
0.204bar/m
580 bar
24131 bar
0.43

93333 ft
0.9 ft/psi
8400 psi
350000 psi

Table 4 -- Geomechanical properties of the overburden

Fig. 6 -- The geomechanical model

12

SPE 79698

Billions

Billions

10,1
Constant compresibility (C=25E-7)
Constant compresibility (C=350E-7)
Pseudo coupling (no water weakening)
Pseudo coupling (with water weakening)
Iterative coupling (no water weakening)
Iterative coupling (with water weakening)

FGPT (MSCF)

9,9

FRPV (RB)

9,7

2,5

1,5

9,5

Constant compresibility (C=25E-7)

Constant compresibility (C=350E-7)
Pseudo coupling (no water weakening)
Pseudo coupling (with water weakening)
Fully coupling
Iterative coupling (no water weakening)
Iterative coupling (with water weakening)

0,5

9,3

0
0

9,1

10

15

20

25

30

years

35

Fig. 10 -- Evolution of the field gas production versus years

8,9
0

10

15

20

25

layer 2 (initial porosity = 40%)

30 years 35
1,01

Fig. 7 -- Evolution of the field pore volume versus years

7500

FPR (PSI)

6500

porosity multiplier f(P)

Constant compresibility (C=25E-7)

Constant compresibility (C=350E-7)
Pseudo coupling (no water weakening)
Pseudo coupling (with water weakening)
Fully coupling
Iterative coupling (no water weakening)
Iterative coupling (with water weakening)

0,99

0,98

Pseudo coupling 1
Iterative coupling (1,1)

0,97

Iterative coupling (2,2)

Iterative coupling (3,3)

0,96

5500

Iterative coupling (4,4)

Iterative coupling (5,5)

0,95

4500

Iterative coupling (6,6)

0,94
3500

4000

4500

5000

5500

6000

6500

7000

7500

pore pressure (PSI)

3500
0

10

15

20

25

30

years

layer 4 (initial porosity = 38%)

35

1,01

800

Constant compresibility (C=25E-7)

Constant compresibility (C=350E-7)
Pseudo coupling (no water weakening)
Pseudo coupling (with water weakening)
Fully coupling
Iterative coupling (no water weakening)
Iterative coupling (with water weakening)

700
600

FOPT (STB)

porosity multiplier f(P)

Millions

Fig. 8 --Evolution of the field pore pressure versus years

0,99

0,98

Pseudo coupling 1
Iterative coupling (1,1)

0,97

Iterative coupling (2,2)

Iterative coupling (3,3)

0,96

500
400

0,95

300

0,94
3500

Iterative coupling (4,4)

Iterative coupling (5,5)
Iterative coupling (6,6)

4000

4500

5000

5500

6000

6500

7000

7500

pore pressure (PSI)

200

Fig. 11 -- Porosity multiplier (pseudo coupling 1 and iterative

coupling) - arching effect

100
0
0

10

15

20

25

30

Fig. 9 -- Evolution of the field oil production versus years

years

35

13

pseudo coupling 2 - porosity 38%

iterative coupling - porosity 38%
pseudo coupling 2 - porosity 40%
iterative coupling - porosity 40%
average (iterative coupling - porosity 38%) dg/dsw = -0,0979
average (iterative coupling - porosity 40%) dg/dsw = -0,0657

1,01

700
Pseudo coupling (with water weakening)

600

Iterative coupling (with water weakening)

Updated Pseudo coupling (with sideburden effect)

500

0,99

FOPT (STB)

Porosity multiplier g(Sw)

1,00

Millions

SPE 79698

0,98

400
300

0,97

200
0,96

100

0,95

0
0

10

15

20

25

30

0,94
0,1

0,2
0,3
0,4
0,5
(Water Saturation)current - (Water Saturation)initial

0,6

0,7

FGPT (MSCF)

Fig. 12 -- Water weakening effect predicted by the pseudo

coupling 2 and the iterative coupling

7500
Pseudo coupling (with water weakening)

7000

Iterative coupling (with water weakening)

Updated Pseudo coupling (with sideburden effect)

FPR (PSI)

6500

Billions

2,5

1,5

1
Pseudo coupling (with water weakening)

0,5

6000

Iterative coupling (with water weakening)

Updated Pseudo coupling (with sideburden effect)

5500

0
5000

4500

10

15

20

25

Fig. 13 -- Pseudo-coupling 2 updating

4000
3500

Billions

10

15

20

25

30

years

35

9,9
Pseudo coupling (with water weakening)
Iterative coupling (with water weakening)

9,8

Updated Pseudo coupling (with sideburden effect)

9,7
FRPV (RB)

years

9,6

9,5

9,4

9,3
0

10

15

20

25

30 years 35

Fig 14 -- Field B: major faults

30

years

35

35

14

SPE 79698

Fig. 15 -- Field B: loading and boundary conditions

Fig. 19 -- Interface normal (a) and tangential (b) stresses along the
faults.
Fig. 16 -- Principal stresses directions in reservoir zone.

Fig. 17 -- Minor principal stress

initial -> end phase 1
end phase 1 -> end phase 2
VE criterion : initial state
VE criterion : end phase 1
VE criterion : end phase 2

(Pa)
8.E+07
6.E+07
4.E+07
2.E+07

p'

(Pa)

0.E+00
-2.5E+08 -2.0E+08 -1.5E+08 -1.0E+08 -5.0E+07 0.0E+00

Fig. 18 -- Stress path in the reservoir

Fig. 20 -- Closing displacement (a) and shear displacement (b)

along the faults.

SPE 79698

15

1,1

1,1

constant compressibility
pseudo coupling 2 with condensate banking
pseudo coupling 1
one way coupling

fully coupled elastic

Field gas production ratio

Multiplier

fully coupled elasto-plastic

0,9
0,8
Porosity

0,7

K/Ki

0,9

0,6
0

200

400

600

800

Pressure

1000

1200

1400

0,8
0

bars

10

15

20

25

30

35
years

Fig. 24 -- Comparison of the field gas production

Field Oil production ratio

Fig. 21 -- Porosity and permeability multipliers

1,1
1

constant compressibility

0,95

0,9
Multiplier

Pseudo coupling 2 with condensate

banking

0,8

Pseudo coupling 1

0,7

One way coupling

0,9

Porosity

0,6

full coupled elastic

full coupled elasto-plastic

K/Ki

0,5

0,85

0,4

10

15

20

25

30

35
years

0,3
0

500

1000

1500

Pressure - bars

Fig. 25 -- Comparison of the field oil condensate production

Fig. 22 -- Condensate banking
Porosity and permeability multipliers at wells

500

250000

constant compressibility
Pseudo coupling 2 with condensate banking
Pseudo coupling 1
450

200000

One way coupling

Field water production (m3)

full coupled elastic

full coupled elasto-plastic

FPR (bars)

400

350

150000
constant compressibility
Pseudo coupling 2 with condensate banking

100000
Pseudo coupling 1
One way coupling

50000

full coupled elastic

300
0

10

15

20

25

30

full coupled elasto-plastic

35

years

0
0

Fig. 23 -- Evolution of the field pore pressure versus years

10

15

20

25

30 years 35

Fig. 26 -- Comparison of the field water production

Mean reservoir pressure (bar)

16

SPE 79698

1200
1000
800
600
400
200
0
0

0,5

1
1,5
2
2,5
Max. vertical displacement (m)

Fig. 27 -- Evolution of the vertical displacement at the top of the

reservoir versus pore pressure