Author's Accepted Manuscript

Numerical Solution of the Nonlinear Diffusiv-

ity Equation in Heterogeneous Reservoirs
with Wellbore Phase Redistribution
Kourosh Khadivi, Mohammad Soltanieh
PII: S0920-4105(14)00010-2
DOI: http://dx.doi.org/10.1016/j.petrol.2014.01.004
Reference: PETROL2584
To appear in: Journal of Petroleum Science and Engineering
Received date: 22 January 2013
Accepted date: 5 January 2014
Cite this article as: Kourosh Khadivi, Mohammad Soltanieh, Numerical
Solution of the Nonlinear Diffusivity Equation in Heterogeneous Reservoirs
with Wellbore Phase Redistribution, Journal of Petroleum Science and Engineering,
http://dx.doi.org/10.1016/j.petrol.2014.01.004
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal
pertain.
www.elsevier.com/locate/petrol


Numerical Solution of the Nonlinear Diffusivity Equation in Heterogeneous
Reservoirs with Wellbore Phase Redistribution

Kourosh Khadivi and Mohammad Soltanieh*
Department of Chemical and Petroleum Engineering
Sharif University of Technology, Azadi Ave., P.O. Box 11155-9465, Tehran, Iran.
* Corresponding author, Email: msoltanieh@sharif.edu

Abstract

We consider the application of the Finite Element Method (FEM) for numerical
pressure transient analysis under conditions where no reliable analytical solution is
available. Pressure transient analysis is normally based on various analytical solutions of
the linear one-dimensional diffusion equation under restrictive assumptions about the
formation and its boundaries. For example, the formation is either assumed isotropic or a
restrictive a priori assumption is made about its heterogeneity. The wellbore storage
effect is also often considered without regard to the possibility of phase redistribution. In
many practical situations such restrictions are not justified and analytical solutions do not
exist. Here we present a numerical solution of the nonlinear diffusion equation based on
the FEM that can be used without any restrictive a priori assumptions. Through the use
of the weak formulation of the FEM, solution can be obtained for a heterogeneous
medium with discontinuous or nonlinear properties. The weak formulation also enables
the handling of time dependent boundary conditions and hence problems involving
wellbore storage with significant phase redistribution. The speed and accuracy of the
numerical technique is first confirmed by comparison with simple test cases that admit an
analytical solution. The practical utility of the proposed method is then demonstrated for
a number of test cases that involve discontinuous and nonlinear formation properties
and/or wellbore storage with phase redistribution.

Keywords: Finite Element Method, Weak Formulation, Wellbore Storage, Phase
Redistribution, Heterogeneity, Pressure Transient Analysis.

1. Introduction

The pressure transient response of the reservoir is a widely used method for reservoir
characterization. However, the early time portion of such data sometimes is infected by
wellbore effects. To discriminate the reservoir portion from the wellbore portion and
extracting a true reservoir parameter, a wellbore model is required. Currently, there are
three types of flow models which can be used for modeling fluid flow in the wellbore;
empirical correlations, homogeneous models and mechanistic models (Shi et al. 2005a).
Empirical correlations are based on matching of experimental data and their range of
applicability are generally limited to the range of variables and the particular geometry
used in the experiments (Duns et al., 1963; Hagedorn and Brown, 1965; Orkiszewski,
1967; Aziz et al., 1972; and Beggs et al., 1973; Ansari et al., 1994). Homogeneous
models that sometimes are called drift-flux models, assume that a single-phase fluid
flows in the wellbore and the properties of this fluid is represented by a mixture of
properties (Shi et al. 2005a; Shi et al. 2005b). Both empirical correlations and
homogeneous models have been developed mainly for steady state multiphase flow in the
wellbore. However, during the well test, the wellbore condition is generally kept in
single-phase flow and pressure is always changing by time. Therefore, such wellbore
flow models are no longer applicable for pressure transient analysis. Mechanistic models;
however, are based on the fundamental conservation laws of mass and momentum
transfer e.g. Navier-Stokes differential equations. Although, they are accurate and reliable
models and can be used in transient condition; however, the computational burden is
significant to solve them and to get reliable solutions. They often encounter convergence
problems and the computation time is remarkably long. For instance, initializing Navier-
Stokes equations and establishing velocity field in the well, sometimes takes a longer
time than solving the diffusion equation in the reservoir. The computation time becomes
noticeable especially when the coupled wellbore-reservoir model should be executed for
many times to estimate reservoir parameters (Pourafshary et al. 2009; Khadivi et al.,
2013).
An alternative approach, which is often used for pressure transient analysis, is
considering the well as a boundary condition instead of a separate model. The accuracy
of this method in representing the wellbore dynamic depends on the equation that is used
as the boundary condition. This type of well modeling is fast and reliable to represent
wellbore dynamics in various well conditions. The overall goal of this work is developing
an efficient and fast numerical procedure to handle reservoir heterogeneity in the
presence of wellbore effects.
2. ProblemFormulation

In this section, we briefly review the governing flow equations in the reservoir and
give the associated weak form.
2.1 GoverningEquations

Most of the theoretical treatments of pressure transient analysis in well testing
consider a well situated in a porous medium of infinite radial extent and assume that the
fluids flow to a central cylinder (the well) that is normal to two parallel, impermeable
planar barriers. The theoretical analysis is based on simplified solutions of the continuity
equation in the porous medium:

t
(p) +v. (pu) = u (1)
where p is the density of the formation fluid, is porosity, u is the fluid velocity vector
and t is time. Flow of the fluid in the porous media is assumed to follow the form of
DArcys law,
u = -
k

7(p +pg7). (2)

Here, u is the superficial Darcy velocity vector, p is the reservoir pressure, p and
are the fluid viscosity and density, g is the magnitude of the acceleration due to gravity,
7 is a unit vector in the direction over which gravity acts and k

is the symmetric and

positive definite 3D permeability tensor, respectively. Substituting Eq. (2) into the
continuity equation yields,

t
(p) -7. _p
k

7(p +pg7)_ = u. (3)

Taking the fluid and rock compressibility as constant, c
]
=
1
p
[
p
p
= constont and
c
q
= -
1
q
[
q
p
= constont , leads to a diffusion equation that describes the temporal and
spatial pressure changes in the 3D reservoir,
pc
t
p
t
-7. _p
k

7(p +pg7)_ = u, (4)

where c
t
= c
]
+c
q
is the total compressibility.
It is assumed that the top and bottom surfaces of the reservoir layer are sealed and a
uniform (either constant pressure or no flow) boundary condition is imposed at the outer
limit of the reservoir. It is further assumed that there is no variation in 0 direction (radial
symmetry) and the reservoir is long enough to be assumed as a single thin layer reservoir.
Then with such boundary conditions, the problem geometrically and physically is a one-
dimensional symmetric problem.
c
t
b
op
ot
-
1
r
o
or
_
kb
p
r _
op
or
]_ = u (S)
The analysis of the measured pressure transient is based on the solution of Eq. (5)
subject to suitable boundary conditions. In the case of an isotropic reservoir the following
initial and boundary conditions are often adopted. As the initial condition, the pressure in
the reservoir is assumed to be uniform at a given initial value, p(r, t = u) = p

. For the
outer boundary condition, the pressure gradient at the outer limit of the reservoir is zero
at all times,
p(,t)

= u ot r
c
. This may mean that the pressure transient will not reach to
the end of the reservoir, the so-called infinite acting reservoir. Alternatively, for long
durations or small radius reservoirs it may mean that the reservoir is closed at its outer
limit, n. 7u = u (r = r
c
).
Depending on the assumptions made regarding behavior within the wellbore, three
major alternatives are used:
I) Negligible Wellbore Storage: accumulation of fluid within the wellbore is ignored all
together and the fluid issuing from the reservoir is withdrawn directly into the
wellbore, the so-called line source well
n _-
k
p
vp] +
qB
o
2nr
w
b
= u , (6)
where r
w
is the wellbore radius and n is the unit vector normal to the boundary. In the
simplest case of no wellbore storage, the diffusivity equation in a one-dimensional
homogeneous reservoir admits a closed form analytical solution based on exponential
integrals (Theis 1935; Crank 1975; Carslaw and J eager, 1959; Bourdet, 2002).

II) Wellbore Storage: accumulation of fluid in the wellbore is considered and the flux is
assumed to consist of that coming from the well itself and that issuing from the
reservoir,
n _-
k
p
vp] +_
1
2nr
w
b
] _qB
o
+C
s
Jp
w
Jt
] = u (7)
where p
w
(t) is the instantaneous bottom-hole flowing pressure and C
s
is the wellbore
storage coefficient. In case of wellbore storage with no phase redistribution, the
diffusivity equation remains linear and an analytical solution can be obtained in the
Laplace domain (Van Everdingen and Hurst 1949; Gringarten et al. 1979; Bourdet et
al. 1983 and Bourdet et al. 1989), which can be readily inverted with the aid of the
Stehfast algorithm (1970).

III) Wellbore Storage with Phase Redistribution: in many cases expansion and
contraction of the fluid in the wellbore is accompanied with redistribution of oil and
gas within the wellbore due to gravity effect. This highly nonlinear phase
redistribution effect occurs frequently in the build-up test because gas bubbles tend to
migrate upward and oil slugs tend to migrate downward under the action of gravity.
The inner-boundary condition in this case is taken as:
n _-
k
p
vp] +_
1
2nr
w
b
] _qB
o
+C
s
_
Jp
w
Jt
+
C
q

c
-t :
]_ = u (8)
Here C
q
is a measure of the potential pressure increase due to gas phase
redistribution and is the rise time of the gas bubbles (Fair, 1981). In case of
wellbore storage with phase redistribution, the diffusivity equation is nonlinear and
must first be (approximately) linearized to enable the use of Laplace transforms. The
procedure employed is described by Fair (1981) for positive values of C
q
that
correspond to an increase in pressure due to phase redistribution. Hegeman et al.
(1991) modified Fair's procedure for dealing with a negative value of C
q
for cases
exhibiting a pressure decrease as a result of phase redistribution.

To summarize, the combination of the diffusivity Eq. (5) and boundary conditions are
linearized where necessary and solved with the aid of the Laplace transform technique for
a wide range of homogeneous reservoirs. For the situations where the analytical inversion
of the Laplace transform is too complicated, Stehfest (1970) has presented a convenient
algorithm for numerical inversion. Such (semi) analytical solutions have been
parameterized and collected into a series of type curves (Gringarten et al. 1979 and
Bourdet et al. 1983) that can be used for identification of the type of flow regime and its
associated properties, for example the wellbore storage coefficient and skin parameters.
However, analytical solutions have a restriction, which can be used for infinite acting
reservoir only and there is no explicit solution for the closed boundary reservoirs. Even
more, available analytical solutions cannot handle fluid flow problems in spatially
heterogeneous reservoirs with significant phase redistribution effect in the wellbore.
Spatial variation of porosity and permeability in the reservoir can result in a highly
nonlinear diffusivity equation. The problem is further complicated by the need to account
for non-ideal wellbore storage effects and phase redistribution. However, this problem
can be solved numerically and leads to the more modern topic of numerical well
testing, which is capable of dealing with heterogeneous reservoirs directly.
2.2 FEMandWeakForm

In pressure transient analysis, we seek the pressure response such that it satisfies the
pressure diffusivity equation, e.g. A(p) = u in a domain (volume or radius), , together
with certain boundary conditions, e.g. B(p) = u, on the boundaries, , of the domain.
The FEM is one of the numerical approximations that seeks the solution in the
approximation form. In the FEM, a complex region defining a continuum is discretized
into simple geometric shapes called elements. The properties and the governing
relationships are assumed in terms of, usually fairly simple, approximating shape
functions over the individual elements and expressed mathematically in terms of the
desired unknown values at specific points in the element called nodes. An assembly
procedure is then used to link the individual elements within a given region. The
assembly procedure relates the local nodes of an element to the global node scheme and
relates the solution on neighboring elements together. Since the shape functions are
usually defined locally for elements or subdomains, therefore, the properties of discrete
systems can be recovered if the approximating equations are cast in an integral form.
Therefore, we seek to cast the equations in the integral form. The pressure transient
response of a 3D reservoir can be obtained by integrating A(p) over the reservoir volume
and B(p) across the boundaries, .
] A(p)J +] B(p). nJ
I
= u.

(9)
For a fixed (independent of t) volume and, under suitable conditions of smoothness
of the intervening quantities, we can apply the Gauss theorem:
] v B(p)J =

] B(p). nJ
I
(10)
to obtain,
]
|A(p) +v B(p)]J = u

(11)
For the integral expression to be zero for any volume , the integrand must be zero.
A(p) +v B(p) = u (12)
This implies an excessive smoothness of the true solution. For this reason, it is
sometimes called the strong or differential form of the equation. The familiar finite
difference method (FDM) uses the strong or differential form of the governing equations,
which requires strict adhesion to the smoothness property of the variables and is ill suited
to heterogeneous media. In contrast, the finite volume and finite element methods deal
with the integral form directly and obviate the need for strict smoothness and spatial
continuity of the variables. An alternative integral form can be obtained by the method of
weighted residuals. Multiplying Eq. (11) by a weight function (x) and integrating over
the reservoir volume we obtain
]
|A(p) +v B(p)](x)J = u

(13)
If Eq. (13) is satisfied for any weight function (x), then Eq. (13) is equivalent to the
differential form Eq. (12). The smoothness requirements on B(p) can be relaxed by
applying the Gauss theorem to Eq. (13) to obtain
]
|A(p)(x) -B(p) v(x)]J +] B(p). n(x)J
I
= u

(14)
This is known as the weak form [Peiro and Sherwin, 2005; Zienkiewicz et al., 2005]
of the diffusivity equation and does not demand the smoothness of the variables involved.
This is the essence of the finite element method and can be applied for any linear or
nonlinear partial differential equation (PDE), where the choice of the weight function
(x) defines the type of finite element scheme used.
2.3 ModelCharacteristics

A schematic of numerical model is presented in Figure 1. The reservoir is assumed to
be one dimensional (vertically homogeneous) but it can be heterogeneous in radial
direction as shown in Figure 1. This behavior is observed in typical damaged reservoirs
due to near wellbore skin [Van Everdingen, 1953; Hawkins, 1956]. The reservoir is
divided into two zones; near wellbore damaged zone and undamaged zone. The wellbore
radius is 0.1 m, the outer reservoir radius is 1,000 m and the interior boundary between
the zones is 1.0 m. In a homogeneous case, permeability as well as porosity is similar in
either zone; however, in a heterogeneous case, they are different. The weak formulation
of the FEM enables the handling of porous medium with sharp (discontinuous) variations
in rock properties in heterogeneous reservoirs. Accurate and stable numerical results are
obtained by using a sufficiently fine mesh close to the perforations. The performance of
numerical model is verified in the next section by comparing with the analytical or semi-
analytical solutions.
3. Verification

In this section, the performance of the new numerical solution is demonstrated in
three cases. In addition, to verify the performance of numerical solution, the results are
compared with analytical or semi-analytical solutions.

3.1 Case1:HomogeneousReservoirwithoutanyWellboreStorageandSkin
Effects

We consider a central well that is fully perforated over the reservoir layer with a
constant rate drawdown of 31.8 m
3
/day (200 stb/day). The reservoir is a very large
homogenous reservoir, where the pressure transient response does not reach the outer
boundary and acts as an infinite acting reservoir. The basic reservoir fluid and reservoir
characteristics for this model are summarized in Table 1. In order to validate the
numerical model, the numerical pressure transient response of a homogeneous reservoir
without any effects of wellbore storage and skin is compared with its analytical solution
as shown in Figure 2. Figure 2 compares the dimensionless pressure of a well in an
infinite acting reservoir. In this case, the numerical FEM solution using Eq. (6) at the
wellbore and the closed form analytical solution are virtually indistinguishable in middle
time and late time regions. There is a good match at the middle and late times; however,
pressure transient responses derived by analytical and numerical methods are different at
the early time region. The reason for this discrepancy is due to an unrealistic assumption
in deriving the analytical procedure, which assumes the well can be represented as a
point. The so-called well model is recognized as the line source. The line source
approximation assumes that the wellbore radius to be vanishingly small. Although
physically this is acceptable because the dimension of the wellbore is so small compared
to the radius of the volume drained by the well; however, in mathematical point of view,
it may not be true. The discrepancy of numerical and analytical solutions in the early time
is highlighted by using log-log plot of dimensionless pressure and dimensionless pressure
derivatives as shown in Figure 3. Mueller and Whitherspoon (1965) investigated the
validity of this solution in the case of finite wellbore radius and concluded that the line
source solution is an excellent approximation within one percent for all values of r

when
t
D

D
2
Su and for r

2u then
t
D

D
2
u.S. As we will see later, generally during this time
the pressure transient response of the well is mainly affected by the wellbore storage
effect. Furthermore, the computation time required for the FEM solution is not so
different from that required for evaluating the closed form analytical solution.

3.2 Case2:HomogeneousReservoirwithWellboreStorageandSkinEffects

Next we consider a homogeneous reservoir with ideal wellbore storage. In this case
the well is open or closed on the surface and there is the possibility of expansion or
contraction of the wellbore fluid but no phase redistribution. This characteristic flow
regime, called pure wellbore storage effect, can last from a few seconds to a few minutes
(Bourdet, 2002) and appears like a hump at early times of the derivative plot. This case
admits a solution in the Laplace domain that can be inverted using the Stehfest algorithm
(1970). The FEM solution is obtained using Eq. (7) as the inner boundary condition. The
results are compared in Figure 4 and are practically indistinguishable. Furthermore, the
FEM solution is delivered with a computation time faster than that required by the
Stehfest algorithm (1970). The wellbore storage effect is observed in the early portions
of the pressure response and appears as a hump on the log-log plot that gradually
vanishes at later times where the flow becomes purely radial. In addition, the
dimensionless pressure obtained using the Stehfest method and the numerical solution are
compared with the data reported by Agarwal et al. (1970) for wells with different storage
coefficients and skin factors in Figure 5. Again, there is excellent agreement between the
semi-analytical and FEM solutions and both methods fit the data well.

3.3 Case3:HomogeneousReservoirwithPhaseRedistributionandSkinEffects
We now turn to an example involving non-ideal wellbore storage effects. In this case
the well is shut in on the surface and a phase redistribution phenomenon is assumed to
occur during wellbore storage. The inner boundary condition is now taken as Eq. (8) that
allows for phase redistribution. Again, the solution of the diffusivity equation in the
Laplace domain is too complicated for analytical inversion. The inverse Laplace
transforms were calculated numerically using the Stehfest algorithm (1970). The results
obtained are compared with the FEM solution in Figures 6 and 7 that show excellent
agreement. Depending on the value of C
q
and C
q
, the wellbore storage hump and its
duration on the log-log plot vary in the early stages of the pressure response. A number of
data also for buildup tests in wells exhibiting phase redistribution have been reported by
Fair (1981), which are superimposed on Figures 6 and 7 for comparison purposes.
4. Application

The results presented above confirm the accuracy and reliability of the FEM solution.
In this section, the application of the new numerical solution in pressure transient analysis
of heterogeneous reservoirs is highlighted. It is well understood that wellbore storage and
phase redistribution effects have major influences on the early time portion of the
pressure transient response. Skin, for example, can change the shape of dimensionless
pressure and dimensionless pressure derivative as depicted on Figure 8. Skin in fact
delays the establishment of radial flow in undamaged zone. Dimensionless wellbore
storage (C
D
) influences the shape of hump on both dimensionless pressure and
dimensionless derivative plots as shown in Figure 9. Dimensionless phase redistribution
pressure parameter (C
q
) and, dimensionless phase redistribution time parameter (

)
have distinct influences on both dimensionless pressure and dimensionless derivative
plots as shown on Figures 10 and 11, respectively. At large values of dimensionless
phase redistribution pressure parameter (C
q
), a "gas hump" appears on dimensionless
pressure (Fair, 1981) and at small values; the phase redistribution diminishes (Figure 10).
In contrast, the dimensionless phase redistribution time parameter (

) dumps the phase

redistribution effect at large values and, induces oscillation and instability in the
dimensionless pressure at small values (Figure 11). As far as major differences are seen
on pressure transient responses by variations of the above parameters, the numerical
reservoir model is sensitive and such sensitivity can be used for parameter estimation.
Therefore, these parameters can be parameterized in the model and determined by
matching the pressure response of the numerical model with the experimental pressure
response using nonlinear regression technique [Gill et al., 1981].
Although the results presented above confirm the accuracy and reliability of the FEM
solution, they are obtained for simple and ideal cases, in which both the porosity and
permeability are spatially uniform (except that of skin case). Now let's see the FEM
performance in cases that there is no analytical solution for them.

4.1 Case4:ARadiallyHeterogeneousClosedSystemwithPhaseRedistribution
andSkinEffects

Consider an important practical case of radial heterogeneity in a closed boundary
reservoir. In such a heterogeneous reservoir, porosity is assumed to be vertically
homogeneous but exhibits radial deterioration along the reservoir radius in undamaged
zone. In addition, permeability in correlation with porosity deteriorates exponentially
along the reservoir radius. A schematic of such heterogeneity is illustrated in Figure 12.
This behavior can be observed in typical channel bars in deltaic environment [Nichols
Gary, 2009; Gastaldo and Huc, 1992; Reading, 1996]. In particular, the porosity is
assumed to deteriorate linearly with radial distance from well center, also the
permeability is assumed to be locally isotropic, k

= k
zz
= k, and taken as an
exponential function of the local porosity,
(r) = u.2u -(1.u1uE -4)r ; k(r) = S.u66 cxp(2u(r)) |m].
Skin factor is 1 and other properties of reservoir fluid and reservoir characteristics for
this example are summarized in Table 1. Consider a closed boundary reservoir with
heterogeneity mentioned in Figure 12, and the phase redistribution effect in the wellbore,
and then the reservoir is drawing down for 11.6 days at a constant rate of 31.8 m
3
/day
(200 bbl/day). Figures 13a and 13b show the semi-log and log-log plots of pressure and
pressure derivative of this well, respectively. They also show the sensitivity of pressure
and derivative of pressure with respect to variations of skin. Different shapes of pressure
and the derivative of pressure are observed by changing skin. Skin factor mainly affects
the early time region of pressure transient response of well; however, the middle time
region in this case, which is representation of the flow regime in undamaged zone does
not show a stable flow regime (Figure 13b). The pressure response of such a
heterogeneous reservoir is quite complicated and precludes the use of usual procedures
for detecting radial flow regions developed for homogeneous reservoirs. Evidently, the
pressure transient analysis in such heterogeneous reservoirs and the detection of the
spatial heterogeneity require the development of new techniques to solve highly
nonlinear inverse problems.
A numerical well test procedure capable of capturing the radial heterogeneity is
readily established. The porosity can be parameterized as, (r) =
0
+c
1
r, and the
permeability can be parameterized as, k(r) = k
0
exp (c
2
(r)).
0
andk
0
are representing
porosity and permeability at the wellbore so they can readily be found from either core or
log data. The sensitivities of the numerical model to variations of c
1
and c
2
are
demonstrated in Figures 14 and 15, respectively. Figures 14 and 15 show the influences
of c
1
and c
2
on pressure and pressure derivative response of the well, respectively. In
contrast to skin factor, which affects the early time region, c
1
and c
2
have major
influences on the middle time region of the well pressure transient response.
The parameters c
1
and c
2
are key elements in spatial variations of porosity and
permeability in this type of heterogeneous reservoir. Therefore, reservoir characterization
in such cases can proceed by searching for c
1
and c
2
. However, finding c
1
and c
2
is not
an easy task because of inherent nonlinearity of this problem due to porosity-permeability
interrelationship, but the sensitivity of pressure transient response of the well with respect
to skin, c
1
and c
2
may offer a test procedure for finding skin, c
1
and c
2
simultaneously.
Therefore, the constant parameters of skin, c
1
and c
2
can be determined by matching the
pressure response of the numerical FEM model with the experimental pressure response
using a nonlinear regression technique [Gill et al., 1981]. This technique is a reliable
approach for solving highly nonlinear and often ill-posed inverse problems [Kenneth
Levenberg, 1944; Donald Marquardt, 1963].
5. Conclusions

We presented a numerical solution for the nonlinear diffusion equation based on the
finite element method (FEM) that can be used without any restrictive a priori
assumptions. Through the use of the weak formulation of the FEM, the solution can be
obtained for heterogeneous media with discontinuous or nonlinear properties. The weak
formulation also enables the handling of time dependent boundary conditions and hence
problems involving wellbore storage with significant phase redistribution. The weak
formulation of the finite element method offers an accurate, reliable and fast procedure
for analyzing the pressure response of complex heterogeneous reservoirs. The response of
such reservoirs is very complex and cannot be analyzed using conventional procedures
developed for homogeneous reservoirs. The speed and accuracy of the numerical
technique is first confirmed by comparing with simple test cases that admit an analytical
solution. The practical applicability of the proposed method is then demonstrated for test
cases that involve discontinuous and nonlinear formation properties and/or wellbore
storage with phase redistribution.
We also developed a parametric numerical wellbore-reservoir model, which is a
reliable numerical well test procedure that can handle the phase redistribution in
heterogeneous reservoirs. This model is sufficiently sensitive with respect to the
parameters of phase redistribution and porosity-permeability variations. The model has
potential to be used to extract phase redistribution parameters in the wellbore and to
evaluate the porosity-permeability distributions in the radially heterogeneous reservoir,
which currently is either difficult or there is no analytical solution for them. This could be
achieved by coupling a nonlinear regression method to the numerical model. The
parameters can be extracted by matching the pressure response of the numerical model
with the actual pressure response. The pressure transient analysis of heterogeneous
reservoirs in the presence of phase redistribution is of great value in this development.
6. Acknowledgement
The authors, in particular K. Khadivi express sincere thanks, to his the late thesis co-
supervisor, Dr. F.A. Farhadpour, whose guidance and support made this research possible.
7. Nomenclature

C
D
Dimensionlesswellbore storage coefficient, C
D
= C
x
1
2ac
t
hr
w
2

C
D
Dimensionless phase redistribution pressure parameter, C
D
= C

kh
qB
u

Rock compressibility, Lt
2
/m, 1/Pa
c

Fluid compressibility, Lt
2
/m, 1/Pa
c
t
Total compressibility,c
t
= c

+c

, Lt
2
/m, 1/Pa
C
x
Wellbore storage coefficient, L
4
t
2
/m, m
3
/Pa
D Unit vector in gravity direction, L, m
g The acceleration of gravity, L/t
2
, m/s
2

h Reservoir thickness, L, m
k Reservoir permeability, L
2
, m
2

k

Symmetric and positive definite 3D permeability tensor

p Reservoir pressure, m/Lt
2
, Pa
p
w
Wellbore pressure, m/Lt
2
, Pa
p
D
Dimensionless pressure
p
D
Dimensionless phase redistribution pressure, p
D
= p

kh
qB
u

q Flow rate, L
3
/t, m
3
/s
r Radius L, m
r
D
Dimensionless radius, r
D
=
r
r
D

r
e
External radius of drainage area L, m
r
w
Wellbore radius L, m
S Skin factor
t Time, t, s
t
D
Dimensionless time, t
D
= t
k
c
t
r
w
2

u Darcy velocity (superficial) vector, L/t, m/s

Greek letters
Porosity, %
Viscosity, m/Lt, Pa.s
p

Fluid density, m/L

3
, kg/m
3
z Phase redistribution time parameter, s
z
D
Dimensionless phase redistribution time, z
D
= z
k
c
t
r
w
2


Conversion Factors

bbl 1.589 873 E01 =m
3

cp 1.0 E- 03 =Pas
ft 3.048 E01 =m
psi 6.894 757 E +3 =Pa
mD9.86927574528 E-16 =m
2
8. References

1. Agarwal, R.G., Rafi Al-Hussainy, and Ramey, H.J . J r., 1970. An Investigation of
the Well Bore Storage and Skin Effect in Unsteady Liquid Flow - I. Analytical
Treatment, - Soc. Pet. Eng. J . 279-290; Trans., AIME, 249, (1970)
2. Ansari, A.M., Sylvester, N.D., Sarica, C., Shoham, O., and Brill, j.p., 1994. A
comprehensive mechanistic model for upward two-phase flow in wellbores. SPE
Production and Facilities, Volume 9, Number 2, pp. 143-151, May 1994.
3. Aziz, K., Govicr, G. W. and Forgarasi , M., 1972. Pressure Drop in Wells
Producing Oil and Gas, J . Cdn. Pet. Tech., J uly-Sept., 1972.
4. Beggs, H. D. and Brill, J . P., 1973. A Study of Two-Phase Flow in Inclined Pipes,
J PT, May, 1973.
5. Bourdet, D.P., Whittle T.M., Douglas, A.A. and Pirard, Y.M., 1983. A New Set of
Type Curves Simplifies Well Test Analysis, World Oil, 95-106.
6. Bourdet D., Ayoub J .A. and Pirard Y.M., 1989. Use of Pressure Derivative in Well
Test Interpretation, SPEFE4 (2): 293-302. SPE-12777-P.A.
7. Bourdet, D., Well Test Analysis: The Use of Advanced Interpretation Models -
(Handbook of petroleum exploration and production), First Edition, Amsterdam,
Elsevier Science B.V., 2002.
8. Carslaw H.S. and J aeger, J .C. 1959. The conduction of Heat in Solids, Second
Edition, Oxford, Clarendon Press.
9. Crank J ., 1975. The Mathematics Of Diffusion, Second Edition, Uxbridge, Oxford,
Clarendon Press.
10. Theis C.V., 1935. The relationship between the lowering of the piezometric surface
and the rate and duration of discharge of a well using ground-water storage,
Transactions American Geophysical Union, vol. 16, 1935.
11. Duns, H. and Ros, N.C.J ., 1963. Vertical flow of gas and liquid mixtures in wells.
Proceedings of the 6th World Petroleum Congress, Frankfurt, Germany, 1963.
12. Fair J r., Walter B, 1981. Pressure Buildup Analysis With Wellbore Phase
Redistribution. Soc. Pet Eng. J . 259-270, (SPE 8206), (1981).
13. Gastaldo Robert A. and Huc Alain-Yves, 1992. Sediment Facies, Depositional
Environments, and Distribution of Phytoclasts in the Recent Mahakam River Delta,
Kalimantan, Indonesia, PALAIOS, 1992, V. &, p. 574-590.
14. Gill, P.E., W. Murray, and M.H. Wright, 1981. Practical Optimization, London,
Academic Press, 1981.
15. Gringarten A.C., Bourdet D.P., Landel P.A., and Kniazeff V.J . 1979. A Comparison
Between Different Skin and Wellbore Storage Type-Curves for Early-Time
Transient Analysis. Paper SPE 8205 presented at the SPE Annual Technical
Conference and Exhibition, Las Vegas, Nevada, 23-26 September 1979.
16. Hagedorn A.R., Brown K.E., 1965. Experimental study of pressure gradients
occurring during continuous two-phases flow in small diameter vertical conduits,
J ournal of Petroleum Technology, pp. 475-484, April, 1965.
17. Hawkins, M., 1956. A note on the skin factor. Trans. AIME, 207, 356357.
18. Hegeman, P.S., Hallford, D.L. and J oseph, J .A., 1991. Well Test Analysis with
Changing Wellbore Storage, Low Permeability Reservoir Symposium, Denver,
Colorado, SPE 21829, (1991).
19. Khadivi k., Soltanieh M., Farhadpour F. A., 2013. A Coupled Wellbore-Reservoir
Flow Model For Numerical Pressure Transient Analysis in Vertically
Heterogeneous Reservoirs, J ournal of Porous Media, Volume 16, 2013, Issue 5,
pages 395-409, DOI: 10.1615.
20. Levenberg Kenneth , 1944. A Method for the Solution of Certain Non-Linear
Problems in Least Squares. The Quarterly of Applied Mathematics 2: 164168.
21. Marquardt Donald, 1963. "An Algorithm for Least-Squares Estimation of Nonlinear
Parameters". SIAM J ournal on Applied Mathematics 11 (2): 431441. DOI:
10.1137/0111030.
22. Mueller, T.D and Witherspoon, P.A., 1965. Pressure Interference Effects Within
Reservoirs and Aquifers, J PT (April 1965) 471-474.
23. Nichols Gary, 2009. Sedimentology and Stratigraphy, Second Edition, J ohn
Wiley & Sons Ltd, West Sussex, UK.
24. Orkiszewski, J ., 1967. Predicting Two-Phase Pressure Drops in Vertical Pipe, J PT,
J une:, 1967.
25. Peiro, J oaquimand Sherwin, Spencer, 2005. Handbook of Materials Modeling,
Volume I: Methods and Models, 132, Springer, Netherlands 2005.
26. Pourafshary, P., Varavei, A., Sepehrnoori, K. and Podio, A. L., 2009. A
Compositional Wellbore/Reservoir Simulator to Model Multiphase Flow and
Temperature Distribution, J ournal of Petroleum Science and Engineering, 2009,
69(1-2), 40-52.
27. Reading H.G., 1996. Sedimentary Environments: Processes, Facies, Stratigraphy,
Third Edition, Chapter 3, Alluvial Sediment, p. 45.
28. Shi, H., Holmes, J .A., Durlofsky, L.J ., Aziz, K., Diaz, L.R., Alkaya, B., and Oddie,
G., 2005. Drift-Flux Modeling of Two-Phase Flow in Wellbores, Paper SPE J ournal
84228-PA, Volume 10, Number 1, pp. 24-33, March 2005.
29. Shi, H., Holmes, J .A. Diaz, L.R. Durlofsky, L.J . and Aziz, K., 2005. Drift-Flux
Parameters for Three-Phase Steady-State Flow in Wellbores, Paper SPE J ournal
89836-PA , Volume 10, Number 2, pp. 130-137, J une 2005.
30. Stehfest, H., 1970. Algorithm 368: Numerical inversion of Laplace transform.
Communication of the ACM, vol. 13 no. 1, p. 47-49 (1970).
31. Van Everdingen , A. F. and Hurst , W., 1949. The Application of the Laplace
Transformation to Flow Problems in Reservoirs, Trans AIME (1949), 186, 305-324.
32. Van Everdingen, A.F., 1953. The Skin Effect and Its Impediment to Fluid Flow into
a Wellbore. Trans. AIME, 198: 171-176.
33. Zienkiewicz O.C., Taylor R.L. and Zhu J .Z., 2005. The Finite Element Method: Its
Basis and Fundamentals, Sixth edition, Oxford, Elsevier Butterworth-Heinemann.
9. Table

Table 1: Basic reservoir and fluid characteristics.
Variable Value SI Units Description
B
o
1.200 std. m
3
/res m
3
Oil formation volume factor
c
t
7.252E-09 1/Pa Total compressibility (5.0E-5 psi
-1
)
b 15.00 m Formation thickness
p 0.010 Pa*s Viscosity
p 913.050 kg/m
3
Density
r
w
0.100 m Wellbore radius
r
s
1.000 m Damage radius
r
c
1000.000 m Reservoir radius
p
0
4.482E+07 Pa Initial reservoir pressure (6500 psia)
q 3.680E-04 m
3
/sec Production Rate (200 bbl/day)
0.200 [m^3/m^3] Porosity (20%)
k 7.200E-14 m
2
Permeability of undamaged formation (~73mD)
k
s
3.602E-14 m
2
Permeability of damaged formation (~36.5mD)
C
s
1.366E-07 [m^3/Pa] Wellbore storage coefficient
C
q
6.510E+06 Pa Phase redistribution pressure parameter
20.144 seconds Phase redistribution time parameter
C

100.000

Dimensionless wellbore storage coefficient,C

=
C
s
1
2nc
t
h
w
2

C
q
10.000

Dimensionless phase redistribution pressure
parameter,C
q
= C
q
kh
qB
c

S 2.300

Skin factor, S = j
k
k
sRn
-1[ ln [

skin

w
, Hawkins,
Formula

10.000

Dimensionless phase redistribution time,

=
k
c
t

w
2


10. Figures

Figure 1: Schematic of a 1D axial symmetry reservoir model with a central well.


0.1 1 10 100 1000
RadialDistance[m]
Inner
Boundary
Wellbore
Outer
Boundary
Reservoir
Interior
Boundary
Continuity
Damaged
Zone(ks)
UndamagedZone(k)
Figure 2: Comparison of pressure transient
response of the numerical FEM model with the
analytical solutions. It is shown on the semi-log
plot of dimensionless pressure versus
dimensionless time.
Figure 3: Comparison of the numerical FEM
model with the analytical solutions (Stehfest
Method). It is shown on the log-log plot of
dimensionless pressure and dimensionless
derivative of pressure versus dimensionless
time.




0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
DimenssionlessTime,t
D
Pressure(Analytical)
Pressure(Numerical)
Numerical
Analytical
1.0E04
1.0E03
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
,

D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e

D
e
r
i
v
a
t
i
v
e
,

p
D
'
DimenssionlessTime,t
D
Pressure(Analytical)
Pressure(Numerical)
Derivative(Analytical)
Derivative(Numerical)
Figure 4: Comparison of numerical FEM
model with analytical solutions (Stehfest
Method). It is shown on the log-log plot of
dimensionless pressure and dimensionless
derivative of pressure versus dimensionless
time.
Figure 5: Comparison of numerical and
analytical solutions with the real field data
(Agarwal et al., 1970). It is shown on the log-
log plot of dimensionless pressure versus
dimensionless time.


Figure 6: Comparison of numerical and
analytical solutions with the data quoted by
Fair (1981) for the case of Non-Ideal Wellbore
Figure 7: Comparison of numerical and
analytical solutions with the data quoted by
Fair (1981) for the case of Non-Ideal Wellbore
1.0E03
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
,

P
r
e
s
s
u
r
e

D
e
r
i
v
a
t
i
v
e
,

p
D
'
DimenssionlessTime,t
D
pDStehfestAlgorithm
pDNumericalFEM
pD'StehfestAlgorithm
pD'NumericalFEM
1.0E03
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E+02
1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
DimenssionlessTime,t
D
WeakFormulation
StehfestAlgorithm
Agarwaletal.Data
Skin=20
Skin=0
1.0E03
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E+02
1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
DimenssionlessTime,t
D
WeakFormulation
StehfestAlgorithm
WalterB.FairJr.
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E+02
1.0E+03
1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05 1.0E+06 1.0E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
DimenssionlessTime,t
D
WeakFormulation
StehfestAlgorithm
WalterB.FairJr.
D=100
Storage.

Storage.




(a) Dimensionless pressure plot (b) Dimensionless pressure derivative plot
Figure 8: Pressure and pressure derivative plots of a heterogeneous reservoir for different skin
effects.


1.0E01
1.0E+00
1.0E+01
1.0E+02
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
DimenssionlessTime,t
D
1.0E+03
1.0E+04
1.0E+05
1.0E+06
1.0E+07
1.0E+08
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
P
r
e
s
s
u
r
e

D
e
r
i
v
a
t
i
v
e
,

p
'
DimenssionlessTime,t
D
S=5
S=2
S=0
S=2
(a) Dimensionless pressure plot (b) Pressure derivative plot
Figure 9: Pressure and pressure derivative plots of a homogeneous reservoir without skin with
different values of wellbore storage.


(a) Dimensionless pressure plot (b) Dimensionless pressure derivative plot
Figure 10: Pressure and pressure derivative plots of a homogeneous reservoir without skin for
different values of the dimensionless phase redistribution pressure parameter.

1.0E01
1.0E+00
1.0E+01
1.0E+02
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
DimenssionlessTime,t
D
CD=10 1001,000 10,000
1.0E+03
1.0E+04
1.0E+05
1.0E+06
1.0E+07
1.0E+08
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
P
r
e
s
s
u
r
e

D
e
r
i
v
a
t
i
v
e
,

p
'
DimenssionlessTime,t
D
CD=10100 1,00010,000
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E+02
1.0E+03
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
DimenssionlessTime,t
D
C
D
=100
C
D
=10
C
D
=1
1.0E+03
1.0E+04
1.0E+05
1.0E+06
1.0E+07
1.0E+08
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
P
r
e
s
s
u
r
e

D
e
r
i
v
a
t
i
v
e
,

p
'
DimenssionlessTime,t
D
C
D
=110100


(a) Dimensionless pressure plot (b) Dimensionless pressure derivative plot
Figure 11: Pressure and pressure derivative plots of a homogeneous reservoir without skin for
different values of the dimensionless phase redistribution time parameter.



Figure 12: Radial variations of porosity and permeability in the reservoir.
1.0E03
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E+02
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
D
i
m
e
n
s
s
i
o
n
l
e
s
s

P
r
e
s
s
u
r
e
,

p
D
DimenssionlessTime,t
D
1.0E+03
1.0E+04
1.0E+05
1.0E+06
1.0E+07
1.0E+08
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
P
r
e
s
s
u
r
e

D
e
r
i
v
a
t
i
v
e
,

p
'
DimenssionlessTime,t
D

D
=1
10
100
1000
0 20 40 60 80 100
0.0
50.0
100.0
150.0
200.0
250.0
300.0
10.0%
15.0%
20.0%
25.0%
0 20 40 60 80 100
P
e
r
m
e
a
b
i
l
i
t
y
,

m
d
P
o
r
o
s
i
t
y
,

f
r
a
c
t
i
o
n
RadialDistance,m
Porosity
Permeability

(a) Semi-log pressure plot (b) Log-log pressure derivative plot
Figure 13: Pressure and pressure derivative plots of the closed boundary heterogeneous reservoir
with phase redistribution effect for different values of the skin factors.


(a) Semi-log pressure plot (b) Log-log pressure derivative plot
Figure 14: Pressure and pressure derivative plots of the closed boundary heterogeneous reservoir
with phase redistribution effect for different values of the c
1
factor.

0.0
100.0
200.0
300.0
400.0
500.0
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
P
r
e
s
s
u
r
e
,

[
B
a
r
]
Time,[seconds]
S=2.0
S=1.0
S=5.0
1.0E03
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E+02
1.0E+03
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07

d
p
/
d
L
n
(
t
)
,

[
B
a
r
]
Time,[seconds]
S=2.0, 1.0,5.0
200
250
300
350
400
450
500
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
P
r
e
s
s
u
r
e
,

[
B
a
r
]
Time,[seconds]
C
1
=+1.010E03
C
1
=0.0
C
1
=1.010E03
1.0E04
1.0E03
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E+02
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07

d
p
/
d
L
n
(
t
)
,

[
B
a
r
]
Time,[seconds]

(a) Semi-log pressure plot (b) Log-log pressure derivative plot
Figure 15: Pressure and pressure derivative plots of the closed boundary heterogeneous reservoir
with phase redistribution effect for different values of the c
2
factor.


200
250
300
350
400
450
500
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07
P
r
e
s
s
u
r
e
,

[
B
a
r
]
Time,[seconds]
C
2
=20
C
2
=5
C
2
=0
1.0E03
1.0E02
1.0E01
1.0E+00
1.0E+01
1.0E+02
1.0E01 1.0E+01 1.0E+03 1.0E+05 1.0E+07

d
p
/
d
L
n
(
t
)
,

[
B
a
r
]
Time,[seconds]
C
2
=0
C
2
=5
C
2
=20
Conversion Factors

bbl 1.589 873 E01 =m
3

cp 1.0 E- 03 =Pas
ft 3.048 E01 =m
psi 6.894 757 E +3 =Pa
mD 9.86927574528 E-16 =m
2



Bighlights


A numerical model is developed to analyze pressure transient response of well.
The model is based on the finite element method and uses the weak formulation.
It is utilized under conditions where no reliable analytical solution is available.
It enables us to evaluate wellbore phase redistribution in heterogeneous reservoirs.
A numerical procedure is developed to capture porosity-permeability distributions.