See

discussions, stats, and author profiles for this publication at: http://www.researchgate.net/publication/237012358

Pressure transient analysis with exponential

and power-law boundary flux

ARTICLE in JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING · JUNE 2014

Impact Factor: 1.42 · DOI: 10.1016/j.petrol.2014.06.030

READS

170

3 AUTHORS, INCLUDING:

Mayra Núñez López Jorge X. Velasco-Hernandez

Metropolitan Autonomous University Universidad Nacional Autónoma de México
14 PUBLICATIONS 5 CITATIONS 84 PUBLICATIONS 1,113 CITATIONS

SEE PROFILE SEE PROFILE

Available from: Jorge X. Velasco-Hernandez

Retrieved on: 04 December 2015
 Journal of Petroleum Science and Engineering 121 (2014) 149–158

Contents lists available at ScienceDirect

Journal of Petroleum Science and Engineering

journal homepage: www.elsevier.com/locate/petrol

Pressure transient analysis with exponential and power law

boundary flux
Y.N. del Angel a, M. Núñez-López b,n, J.X. Velasco-Hernández c
a
Escuela Superior de Física y Matemáticas ESFM-IPN, Av. Luis Enrique Erro S/N, Unidad Profesional Adolfo López Mateos, Zacatenco,
Gustavo A. Madero, 07738 D.F., México
b
Universidad Autónoma Metropolitana Unidad Lerma, Av. Hidalgo Poniente 46, Col. La Estación, Lerma de Villada, 52006 Edo. de México, México
c
Instituto de Matemáticas, Universidad Nacional Autónoma de México Campus Juriquilla, Boulevard Juriquilla No. 3001, Juriquilla, Querétaro 76230, México

art ic l e i nf o a b s t r a c t

Article history: The widely used conceptual model set to study flow phenomena in an oil reservoir or groundwater
Received 26 April 2013 system has been for several decades the homogeneous cylindrical model with a fully penetrating well
Accepted 29 June 2014 located at its centre. For wells that produce at a constant rate in a bounded reservoir the boundary
Available online 10 July 2014
conditions are usually zero flux at the outer boundary. In this work we analyze the pressure response in
Keywords: the presence of flow at the external border, following an exponential or a power law boundary flux. Our
pressure transient analysis solutions could be used for the analysis and interpretation of the pressure response when it is dominated
variable flux at the boundary by the boundary effects or discontinuities.
well test & 2014 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
2. The basic theory for variable flux at the boundary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
3. Solution for a well in a reservoir with flux at the boundary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
4. Numerical and analytical approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
5. Characteristic behaviour of the pressure derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Appendix A. Analytical solution for IBVP1 and IBVP2 cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
A.1. Analytical solution for inverse Laplace of f ðrD; sÞ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
A.1.1. Case IBVP1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
A.2. Analytical approximation for f ðrD; tDÞ, case IBVP2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
A.3. Long time approximation for g2ðrD; tDÞ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

1. Introduction boundary effects. For homogeneous reservoirs the pressure response

A fully penetrating well located at the centre of a cylindrical
reservoir producing at a constant rate is the widely used model to
study flow phenomena. Despite this simplified geometry, many
complications in well-test interpretation arise due to inner and outer
n
Corresponding author. Tel.: þ 52 728 2827002; fax: þ 52 91757463.
E-mail address: m.nunez@correo.ler.uam.mx (M. Núñez-López).
influenced by the outer boundary effects such as leaky faults, no flow
boundaries, and constant pressure boundaries has been studied by
several authors (Acosta and Ambastha, 1994; Earlougher et al., 1974;
Ehlig-Economides and Ramey, 1981; Fuentes-Cruz et al., 2010).
With regard to heterogeneous reservoirs, there are models
available in the literature that address similar issues like compart-
mentalized reservoirs and geochoke with applications in different
areas such as extraction of crude oils, underground oil displacement,
well drilling, and aquifer contamination. There are models of flow in

http://dx.doi.org/10.1016/j.petrol.2014.06.030
0920-4105/& 2014 Elsevier B.V. All rights reserved.
150 Y.N. del Angel et al. / Journal of Petroleum Science and Engineering 121 (2014) 149–158
Nomenclature
Field variables
ϕ porosity, dimensionless
μ fluid viscosity, M=LT
k formation permeability, L2
ct total system compressibility, LT 2 =R
rw wellbore radius, L
r radial distance, L
R reservoir drainage radius, L
p pressure, M=LT 2
pi initial reservoir pressure, M=LT 2
q oil flow rate, L3 =T
porous media that have considered moving boundary to describe the
transient flow of power law fluids (Pascal and Pascal, 1985). The
variable flow concept has also been studied for non-Newtonian fluids
in bounded and homogeneous domain (Ciriello et al., 2013; Ciriello
and Di Federico, 2012).
The model presented in this paper can be equally considered as
special cases of these compartmentalized reservoirs, the so-called
composite reservoirs consisting of two or more concentric regions
with a single well at the centre. The composite models have been
generalized in different ways, for example, including effects of
various trends of mobility and storativity variations to determine
the swept volume in thermal wells (Acosta and Ambastha, 1994).
Acosta and Ambastha (1994) present a general methodology for
analyzing pressure transient test for composite reservoirs includ-
ing a fractal region between the homogeneous circular region. On
the other hand Corbett et al. (2012) presented a new well testing
response considering vertical influx to reservoir.
These early studies have not considered that influx into the
reservoir is prescribed by the external boundary condition. Doub-
let and Blasingame studied the numerical pressure response and
the transient flow at the well due to the variable flow at the
boundary. In this work we propose a new temporal variable
condition at the external boundary (power law), numerical and
analytical solution is obtained for studying pressure response in
reservoir. The new well test responses are compared with the
Fig. 1. Agglomerations around the boundary (either shaly sand facies or shale
lenses) that can obstruct the flow channels at a radial distance R from the well
radius rw.
t time, T
h formation thickness, L
Dimensionless variables
pD dimensionless pressure
rD dimensionless radial distance
tD dimensionless time
RD dimensionless reservoir drainage radius
qRD dimensionless terminal (endpoint) boundary influx
τD dimensionless starting time for boundary influx
α scaling exponent (real number)
Hðt D Þ Heaviside function
pressure curves considering the exponential flux condition at
boundary as proposed by Doublet and Blasingame (1995).
Our physical model is a single well centred in a bounded
circular reservoir with two different conditions at the inner
boundary (constant rate production and constant pressure pro-
duction) for describing pressure behaviour. Additionally a pre-
scribed flux at the outer boundary is considered. This prescribed
flux or variable flow at the boundary means that the influx at the
outer boundary is initially zero changing after a time to a fixed
value following two different ways: exponential and power law.
It is hoped that by incorporating flux at the boundary we can
obtain a better approach to modelling homogeneous reservoirs
(mainly on those modelled as closed ones) with some hetero-
geneity at the external boundary. A closed reservoir is mainly
characterized by an impermeable boundary (no-flow condition).
Starting from this physical concept, the model analyzed here
describes a boundary that provides fluid to the system with a
different rate such that the boundary acts like a low-permeability
boundary. From geological point of view this can be associated
with reservoirs bounded by discontinuities due to agglomerations
of shaly sand facies or shale lenses that obstruct partially or totally
the flow channels as shown in Fig. 1. Discontinuities due to
variations in rock types are not precisely circular, but we assume
a uniform distribution around the boundary as an approximation
(Rosa, 1996). The rationale or motivation for our model is that in
these types of geometries, for example lenses of sand surrounded
by shale barriers or fractures, flow inside the lenses occurs in an
homogeneous rock but restrictions to flow are located at the
boundary where fractures or shale structures induce a subdiffusive
flow into the region under exploitation. This scenario is not purely
theoretic; it is a plausible model in reservoirs such as Chicontepec
(Comisión Nacional de Hidrocarburos, 2010) in Mexico.
In this work we first present the basic theory for a model with
variable flux at the boundary. Second we establish the conditions
under which each of the models with exponential or power law
boundary flux is applicable. Third we present and compare the
analytical and numerical approaches of the pressure response for
both models. Finally we analyze the characteristic behaviour of the
pressure derivative.
2. The basic theory for variable flux at the boundary
One of the basic models for describing fluid flow in cylindrical
porous media is described by the diffusion equation in radial
coordinates given by
∂2 p 1 ∂p ϕμct ∂p
þ ¼ ; ð1Þ
∂r 2 r ∂r k ∂t
 Y.N. del Angel et al. / Journal of Petroleum Science and Engineering 121 (2014) 149–158 151

in SI units, where pðr; tÞ represents the pressure at time t and exemplified in the introduction. Finally, let
location r, ϕ is the porosity, k is the permeability, ct is the total pD ðr D ; 0Þ ¼ 0; ð5Þ
compressibility and μ is the viscosity of the fluid. It is convenient
to express the diffusivity equation in terms of dimensionless indicates that at time zero the pressure at all points in the
variables: formation is constant, namely pi.
r
rD ¼ ;
rw
3. Solution for a well in a reservoir with flux at the boundary
kt
tD ¼ ;
ϕμct r2w In this section we solve Eq. (2) with the interior boundary
2π hk condition given by Eq. (4) (production at a constant rate) and
pD ¼ ðpi  pÞ ;
qμ exterior boundary conditions given either by


where rw is the well radius, q is the constant rate production and pi ∂p
rD D ¼  qRD ð1  e  tD =τD Þ or ð6Þ
is the initial pressure. With these variables, when r ¼ r w we obtain ∂r D rD ¼ RD
rD ¼ 1 and our main equation becomes


∂p
∂2 pD 1 ∂pD ∂pD rD D ¼  qRD Hðt D  1Þð1  t D α Þ: ð7Þ
þ ¼ : ð2Þ ∂r D rD ¼ RD
∂r 2D r D ∂r D ∂t D
The BVP given by Eqs. (2), (4) and (6) will be named the first
The diffusivity equation is equivalent to the standard heat con-
initial-boundary-value problem (IBVP1). The other case, where (6)
duction equation for which solutions with various boundary
is replaced by (7), will be named IBVP2 (Fig. 2). We denote the
conditions are well known. We can consider either of the follow-
generic flux term at the reservoir boundary as qDext ðt D Þ; α A ½0; 1Þ
ing inner boundary conditions at the well:
and τD are parameters that affect the pressure response at the
pD ð1; t D Þ ¼ 1 or ð3Þ inner radius of the well. Hðt D Þ is the Heaviside function. Eq. (7) can

 be represented as
∂pD
 (
rD ¼  1: ð4Þ 0; tD r 1
∂r D ∂p
rD ¼ 1 rD D ¼ α
∂r D rD ¼ RD  q RD ð1  t D Þ; t D 4 1:
The first condition represents production at constant pressure and
the second one represents production at constant rate for all time. Note that the negative sign in (6) and (7) is necessary because
Usually the condition for a finite reservoir is expressed as the flow comes through the boundary towards the well. This


∂p
rD D ¼ 0;
∂r D rD ¼ RD

where RD is the location of the outer boundary of the reservoir,

indicating that flow at the boundary is at equilibrium and that the
net flux in either direction vanishes. In oil or aquifer reservoirs this
is not quite the case. We postulate that the fluid depletes before
the fluxes equilibrate albeitly with the flux at the boundary very
small but non-zero. Given this, our outer boundary condition has
the following form:


∂p
qDext ðt D Þ ¼ r D D ;
∂r D rD ¼ RD

where qDext corresponds to the time varying flux at the outer

boundary, using the notation by Doublet and Blasingame (1995).
We postulate that this function at the boundary approximates Fig. 3. Behaviour of dimensionless boundary rate for IBVP2 case (log–log plot) with
the waiting times of a subdiffusive flow outside the sandstone as qRD fixed and α variable. The start of the pseudo-steady state is shown for α ¼0.66.

Fig. 2. Dimensionless boundary rate. (a) IBVP1 case, (b) IBVP2 case.
152 Y.N. del Angel et al. / Journal of Petroleum Science and Engineering 121 (2014) 149–158

scenario involves, for instance, issues of water encroachment into that after inversion gives
reservoir. The basic model for this case, constant external pressure
2 
2 rD R2
with a well producing at constant rate, was provided by van pD ðr D ; t D Þ ¼ 2 þ tD  2 lnðr D Þ
ðRD 1Þ 4 ðRD  1Þ
Everdingen and Hurst and they showed that the steady-state flow
e  ui t D
2
is reached for long times. The constant external pressure indicates 3R4D  4R4D lnðRD Þ  2R2D  1 1
  π ∑ F 1 ðr D ; RD ; ui Þ
that the aquifer supplied fluid to system at the initial reservoir 4ðR2D  1Þ2 i¼1 ui
pressure. However, in practice, pressure at the interface reservoir– ð12Þ
aquifer (external boundary for reservoir) is not constant or uni-
form. Thus, Eqs. (6) and (7) consider reservoirs that may have where
a physical barrier such as a surface water body that provides J 21 ðRD ui ÞΦ1;0 ð1; r D ; ui Þ
recharge to the finite reservoir through its external boundary with F 1 ðr D ; RD ; ui Þ ¼ :
J 21 ðRD ui Þ  J 21 ðui Þ
a boundary flux uniformly distributed around the perimeter. This
influx at the external boundary is independent of production since Additionally ui satisfies the relation Φ1;1 ð1; RD ; ui Þ ¼ 0 with
qRD ranges from 0 to 1 independent of q, further in Eq. (6) τD Φm;n ðψ ; φ; uÞ ¼ Y m ðψ uÞJ n ðφuÞ  J m ðψ uÞY n ðφuÞ:
indicates that the influx starts at a time t 4 0: In Eq. (7), the
Jn and Yn are the Bessel functions of the first and second classes of
initiation of the influx is not strictly controlled. Rather what is
order n respectively. The solution given by Eq. (12) provides the
controlled is the beginning of the pseudo-steady state1 flow across
pressure drop at any point in the formation for the constant rate
the external boundary via α (see Fig. 3): the steady-state will
case in a finite reservoir. The pressure drop at the wellbore, r D ¼ 1, is
depend of α since when it is close to zero the boundary condition


is similar to that of a boundary with restricted flow. For both cases 2 1 3R4  4R2D lnðRD Þ  2R2D  1
pD ð1; t D Þ ¼ 2 þ tD  D
IBVP1 and IBVP2 the inflow takes place even if the well is not put ðRD  1Þ 4 4ðR2D  1Þ2
on production.
e  ui tD J 21 ðRD ui Þ
2
1
For petroleum production the reservoir–aquifer (water-drive) þ2 ∑ 2 2
; ð13Þ
i ¼ 1 ui ½J 1 ðRD ui Þ  J 1 ðui Þ
2
scenario at the external boundary may not be the only one. In
enhanced oil recovery various forms of drive may be used. The where we have used the relation Φ1;0 ð1; 1; ui Þ ¼  2=π ui (Carslaw
representation of the interface between reservoir fluid and gas and Jaeger, 1963). The infinite series in (13) requires the knowledge
phase (e.g. hydrogen-injection) at the outer boundary is com- of all roots ui of Φ1;1 ð1; RD ; uÞ in order to obtain the pressure at
pounded by the presence of multiphase flow. However, for r D ¼ 1.
practical purposes the single-phase simplifications can be used. Klins et al. (1998) concluded that only two terms in the infinite
If the fluid that encroaches into the reservoir does it at a higher summation are necessary to give a good accuracy, which is not
rate than that of fluid withdrawal, that it turn depends on many entirely correct if the analysis for short times is required. In
factors such as fluid properties and rock characteristics, then Fig. 4(b) we show the expansion of (13) with two and 15 positive
qRD 4 1 and pressure depletion overcomes pressure increase. roots, although for our study we consider only u1 and u2. The
The initial condition is given as infinite summation in (13) quickly converges for large tD and a
pD ðr D ; t D Þ ¼ 0; t D ¼ 0; r D 4 1: linear approximation to the solution is feasible. This implies that
pðr; tÞ decreases linearly with time which is a behaviour character-
In Laplace space Eq. (2) becomes istic of producer wells when the boundary effects are acting
pffiffi pffiffi (Bourdarot, 1996). Note that the pseudo-stationary state depends
p D ðr D ; sÞ ¼ AðsÞI 0 ðr D sÞ þ BðsÞK 0 ðr D sÞ; ð8Þ
on how far the boundary is located. Eq. (13) corresponds to
where A and B are determined from the boundary conditions the inverse transform of the first term of Eq. (9) so the problem
(Carslaw and Jaeger, 1963). Solving Eq. (8) for A and B we have is reduced to finding the inverse transform of the second term
pffiffi pffiffi given by
Ψ 0;1 ðrD ; RD ; sÞ q Dext ðsÞ Ψ 0;1 ðr D ; 1; sÞ pffiffi
p D ðr D ; sÞ ¼ 3=2 pffiffi þ pffiffi pffiffi ; ð9Þ q ðsÞ Ψ 0;1 ðr D ; 1; sÞ
s Ψ 1;1 ð1; RD ; sÞ RD sΨ 1;1 ð1; RD ; sÞ f ðr D ; sÞ ¼ Dext pffiffi pffiffi : ð14Þ
RD sΨ 1;1 ð1; RD ; sÞ
where

Ψ m;n ðψ ; ω; yÞ ¼ K m ðψ yÞIn ðωyÞ þð  1Þm þ n þ 1 Im ðψ yÞK n ðωyÞ; ð10Þ

and q Dext is either the Laplace Transform of (6) or (7) depending on
the case. Similar to Doublet and Blasingame (1995) we have that
the solution p D ðr D ; sÞ is represented as the sum of two terms:
p D ðr D ; sÞ ¼ “No Flow” Boundary Solution þ Boundary Flux Solution:
The case of no flow condition at the boundary (the so-called
“closed” reservoir) was studied by van Everdingen and Hurst
(1949) taking q Dext ¼ 0. In this instance the above equation reduces
to the classical solution
pffiffi
Ψ 0;1 ðrD ; RD ; sÞ
p D ðr D ; sÞ ¼ 3=2 pffiffi ; ð11Þ
s Ψ 1;1 ð1; RD ; sÞ
1
At production start up time, the resulting pressure wave generated at well
diffuses outwards in the reservoir as if the reservoir was infinite. This is the
transient or the unsteady-state regime. Then the disturbance hits the reservoir
boundary; after some transient time, the reservoir pressure declines at a constant
rate. This is the semi-steady-state flow regime (or pseudo-steady-state).
4. Numerical and analytical approaches
In this section the analytical solutions for IBVP1 and IBVP2 are
obtained and denoted by p1D ðr D ; t D Þ and p2D ðr D ; t D Þ respectively.
The pressure at the wellbore is obtained by evaluating p1D and p2D
at r D ¼ 1. The pressure behaviour as a function of the reservoir
radius and time is given by
Z tD
1
piD ðr D ; t D Þ ¼ P 0D ðr D ; t D Þ þ qDext ðt D  ςÞgðςÞ dς;
RD 0
where qDext is given by (6) or (7) for i¼ 1,2 respectively, gðr D ; ςÞ is
given by Eq. (A.8) in Appendix A and P 0D ðr D ; t D Þ is the solution for
a closed reservoir given by Eq. (12). For the IBVP1 the above
equation becomes
p1D ðr D ; t D Þ ¼ P 0D ðr D ; t D Þ þ f ðr D ; t D Þ;
where f ðr D ; t D Þ is given by Eq. (A.11). For the IBVP2 the analytical
approximated solution is
p2D ðr D ; t D Þ ¼ P 0D ðr D ; t D Þ þ g 1 ðr D ; t D Þ þ g 2 ðr D ; t D Þ;
 Y.N. del Angel et al. / Journal of Petroleum Science and Engineering 121 (2014) 149–158 153

Fig. 4. (a) Pressure behaviour (semi-log plot) using Stehfest algorithm for external radius RD ¼ 1000 and α ¼ 0:89. (b) Comparison between analytical (dashed lines) and
numerical (continuous line) solutions for RD ¼ 1000, α ¼ 0:89 and qRD ¼ 0:5. Analytical approach 1 considers two terms of summation given by Eq. (16) and Analytical
approach 2 considers 15 terms.

where g 1 ðr D ; t D Þ and g 2 ðr D ; t D Þ can be seen in Appendix. Finally, the

explicit expressions for the pressure behaviour at the wellbore are
given by
!
J 1 ðui ÞJ 1 ðRD ui Þ½e  ui t D  1
2
2qRD 1
p1D ð1; t D Þ ¼ P 0D ð1; t D Þ  tD þ ∑
R2D  1 i¼1 u2i ½J 21 ðRD ui Þ  J 21 ðui Þ
!
qR 2RD τD ½1  e  tD =τD  1 J ðu ÞJ ðR u Þ½e  ui t D  e  t D =τ D  τD
2
D
þ D 2 ∑ 1 i 1 2 i
RD R2D  1 i¼1 ½J 1 ðRD ui Þ  J 21 ðui Þ ð1  τD u2i Þ
ð15Þ
and


2qRD 1
p2D ð1; t D Þ ¼ P 0D ð1; t D Þ  þ tD
ðR2D 1Þ 4
qRD ðR4D þ 2R2D  4R2D lnðRD Þ  3Þ
þ
4ðR2D  1Þ2
Fig. 5. Pressure behaviour as a function of α (IBVP2 case) for external radius RD ¼30
2qRD 1 J 1 ðui ÞJ 1 ðRD ui Þ e  ui t D 2qR Γ ð1  αÞ 1  α
2
and qRD ¼ 0:4.
 ∑ þ 2 D tD : ð16Þ
RD i ¼ 1 ½J 21 ðRD ui Þ  J 21 ðui Þ u2i ðRD  1ÞΓ ð2  αÞ
Note that the pressure response is affected by qRD and the solution
for a closed reservoir is immediately recovered when qRD ¼ 0,
as we hope. The complexity of either equation for the pressure
response at the well is obvious. One of the difficulties lies mainly
on the fact that Φ1;1 ð1; RD ; uÞ ¼ 0, this implies that there is an
infinity of roots, and would be suitable to find a finite number of
roots to allow the convergence of the series. However, the solution
for the reservoir pressure behaviour as a function of time has been
obtained numerically by Stehfest (1970) such that the pressure
behaviour at the wellbore pD ð1; t D Þ is


ln 2 N ln 2
pD ð1; t D Þ ¼ ∑ V i p D 1; i ;
tD i ¼ 1 tD

where N is a even number and Vi are the appropriate weights.

Once N is fixed, Vi are given by
N=2 þ 1
minði;N=2Þ
k ð2kÞ! Fig. 6. Pressure behaviour for IBVP1 case (log–log plot) as a function of τD with
V i ¼ ð  1ÞN=2 þ i ∑ N  : external radius RD ¼ 30 and qRD ¼ 0:5.
k ¼ ½ði þ 1Þ=2 2 k !k!ðk  1Þ!ði  kÞ!ð2k  iÞ!

Analytical and numerical approaches have been compared obtain-

ing excellent agreement for long time values. Two terms are not
necessarily sufficient in (15) or (16) to provide good accuracy of
pD ð1; t D Þ at short times as suggested by Klins et al. (1998). Given
that piD, i¼1,2, at early times the analytical solutions require all
roots of equation Φ1;1 ð1; RD ; uÞ which are infinite. However, at
short times the producing well has not reached the boundary and
the reservoir behaves as if it were infinite. Based on this fact and
given that our purpose is to study the pressure response when the
boundary effect has been reached we have used the Stehfest
algorithm to generate the solution curves and some conclusions
are inferred from the analytical expressions as described below.
The pressure response for a well with constant flow rate at the
well with variable flow in the external border (IBVP2) is shown in
Fig. 4 for different values of qRD and α ¼0.89 fixed. Fig. 4(a) shows
the result of the numerical inversion of Eq. (9); in Fig. 4(b) dashed
lines show analytical solution given by (16) (with two terms of
154 Y.N. del Angel et al. / Journal of Petroleum Science and Engineering 121 (2014) 149–158
summation) and continuous line shows numerical solution of
Eq. (9). We can observe that these solutions are correct for long
time values but at short time there is a significant difference, since
only two roots of the polynomial Φ1;1 ð1; RD ; uÞ are considered. In
Fig. 5 the effect of α in the pressure response (IBVP2) is shown.
One can appreciate that the pseudo-steady state depends on the
value of α, a situation already observed in Fig. 3. Notice that as
α approaches unity the effect on the pressure curve becomes
negligible. However, there is a marked difference in the pressure
curve when α a 0. An appropriate estimation of this parameter is
thus very important since it influences the characteristic beha-
viour of the reservoir. The effect of τD for IBVP1 is shown in Fig. 6.
As observed by Doublet and Blasingame (1995), the model with
variable flux at the boundary behaves as if it were a closed dual
porosity reservoir. In a log–log or a log–linear plot, for IBVP1 and
IBVP2 some combinations of qRD and τD (or qRD and α) could
have effects on p1D (or p2D) resembling ω and λ in Warren and
Root's mode (Warren and Root, 1963).
Having a variable flux at the boundary for either IBVP1 or IBVP2
indicates that the period of flow dominated by the boundary flow
differs substantially from that of a closed reservoir. The idealized
model can be applied to simulate low permeability reservoirs at
boundary where the discontinuities are assumed to be uniformly
distributed around the boundary (Rosa, 1996), these discontinu-
ities can be caused by variations either in rock or fluid properties.
In a log–linear or a log–log plot of piD vs tD the difference between
the IBVP1 and the IBVP2 may not be evident and such disconti-
nuities could be wrongly interpreted. This is illustrated in Fig. 7
where both models were compared with different reservoir
characteristics. Both show a qualitatively similar behaviour. When
Fig. 7. Pressure response (log–log plot) for IBVP1 and IBVP2 cases.
Fig. 8. Pressure response for IBVP1 case with qRD o 1, qRD ¼ 1 and qRD 41.
this situation occur, a starting point for identifying the reservoir
behaviour is the log–log plot of the log ðdpD =d ln t D Þ. The plot of
log ðdpD =d ln t D Þ versus the logarithm of time is known as the
pressure derivative (widely used tool for the interpretation of
well-test). When qRD ¼ 1 the concept of variable flux at the
boundary leads to oil-production under steady-state conditions.
More specifically, the pressure at the well is constant for long
times. This approach is basically the water influx model provided
by van Everdingen and Hurst, i.e. constant external pressure with
a well producing at a constant rate. When qRD 41 the pressure
increases in the system, see Fig. 8, to ideally reach its initial
pressure and stay near that value. Note that for case qRD o1 the
curve indicates that pressure decreases and thereby the fluid in
the system is depleting despite injection into the reservoir.
5. Characteristic behaviour of the pressure derivative
The analytical solution of Eq. (9) can be used to investigate the
transient flow behaviour through the pressure derivative as was
discussed initially by Bourdet et al. (1983, 1984). However its
computation presents difficulties since the derivative must be
known at the roots of Φ1;1 ð1; RD ; uÞ. Using the Stehfest algorithm
(Ozkan and Raghavan, 1997) a log–log plot diagnostic of the
pressure change derivative at the well with production at a
constant rate is shown in Fig. 9. Fig. 9 shows that the radial flow
at the beginning is not affected by boundary effects. After the
infinite acting period, the curve goes through a transition period to
finally end with a typical behaviour for a closed reservoir. Different
plots were made to observe the effect of α on the pressure
derivative (Fig. 10). A transition period is also observed for the
IBVP1 (see Figs. 11 and 12).
Fig. 9. Log–log of the pressure derivative behaviour and variable flux at the
boundary with RD ¼ 1000 and α ¼ 0:89.
Fig. 10. Log–log of the α-effect in the pressure derivative behaviour with RD ¼ 1000
and qRD ¼ 0:89.
 Y.N. del Angel et al. / Journal of Petroleum Science and Engineering 121 (2014) 149–158
Fig. 11. Log–log plot of the derivative pressure behaviour for IBVP1 case with
RD ¼ 30 and τD ¼ 100; 000.
Fig. 12. Effect of τD on the transition period for RD ¼ 30.
Note that depending on the value of qRD , being far or close to
one, the transition period will be dominant. When the effect of qRD
is not dominant, the well behaviour can be interpreted as that of a
closed reservoir. Fig. 13 shows in detail this observation for IBVP1.
Nevertheless, there is a small difference that could be significant
for reservoir parameter estimation. Unfortunately, we do not
expect that this short transition period could always be distin-
guished because of uncertainty in pressure measurement data.
We also note that the concept of compressible zone or pressure
disturbance area, discussed in detail by Hsieh et al. (2007, 2008),
may be affected by the variable flow at the boundary. Therefore,
the radius of the pressure disturbance area (compressible zone),
also known as the radius of investigation or radius of drainage,
must be studied and reinterpreted. According to Hsieh, in a closed
reservoir, when the duration of the pressure test or well producing
period is long enough the no-flow boundary is reached by the
pressure front and rebounds back to the well. Then, a semi-log plot
of pD vs t D shows a deviation of the pressure curve corresponding
to an infinite reservoir at a specific time. The starting time of this
deviation is dependent on RD as can be seen in Fig. 14. We note this
deviation also for the case of variable flux at the boundary model,
however, in our case, it is dependent on qRD (Fig. 4(a)).
Note that as tD increases, t D 4 τD , the pressure at the well given
by Eq. (15) is approximated by
2t D 3 2qR t D 2qR 1 J 1 ðui ÞJ 1 ðRD ui Þ
p1D ð1; t D Þ ¼ þ ln RD   2 D þ 2 D ∑
R2D 4 RD  1 RD  1 i ¼ 1 u2 ½J 21 ðRD ui Þ  J 21 ðui Þ R2D  1
i
So, for long times pressure derivative is characterized by a straight
line with slope 1 on a log–log plot. The same case can be seen from
the analytical solution for IBVP2, Eq. (16). Therefore, when qRD o 1
the pseudo-steady-state regime is reached and the fluid in system
155
Fig. 13. Derivative pressure behaviour for a closed reservoir and IBVP1 case.
Fig. 14. Dimensionless boundary effect for a closed reservoir as a function of RD.
is depleted, similar to a closed reservoir, despite injection into
reservoir across the external boundary. Once the boundaries of
the drainage area have been reached the pressure at the well
decreases linearly with time. The decline is expressed by the above
equation and may be written as
!
2 2qRD
piD ð1; t D Þ ¼  t D þ b; i ¼ 1; 2
R2D R2D  1
which describes a line with slope m as a function of qRD . The b
value depends of approximation for IBVP1 or IBVP2 case. So,
estimating the drainage area A or the pore volume drained hAϕ
is affected by qRD . The pressure behaviour piD (i¼1,2) for qRD o 1,
qRD ¼ 1; and qRD 4 1 can be seen from pressure derivative of
Eqs. (15) and (16)):
!
∂p 2 2qRD
t D iD   t D ; R2D  1  R2D ; t D 4 0:25R2D :
∂t D R2D R2D
When qRD ¼ 1 we have ∂piD =∂t D ¼ 0, steady-state conditions and
when qRD o1 pseudo-steady-state is reached. For qRD 4 1 instabil-
ities are identified due to the negative slope of ∂piD =∂t D (see Fig. 8).
6. Conclusions
2qRD τ D
þ : We have developed approximate analytical solutions for a well
producing at a constant flow rate in a bounded reservoir with a
prescribed flux at the outer boundary. Our approach considers
two kinds of empirical functions representing variable flux beha-
viour at the boundary. The Stehfest algorithm for numerical
inversion of Laplace transform has been used to verify analytical
156 Y.N. del Angel et al. / Journal of Petroleum Science and Engineering 121 (2014) 149–158

approximations. Unlike analytical solutions, the numerical recipes

to solve these kinds of problems are available in commercial
packages. Despite the complexity of the analytical solutions
obtained these could also be used to determine the pressure
response or the rate transient at any distance from the confined
(e.g., at RD). In particular, the pressure behaviour at the wellbore
radius has been analyzed for both cases of flux at the external
boundary. Both models exhibit in general, three flow regimes:
infinite-acting radial flow period, transition period and pseudo-
steady state flow period. However, for some specific values of α
and τD the pressure response resembles the flow behaviour in dual
porosity reservoirs which may lead to misinterpretation. The
model proposed may lead to a better understanding of typical
pressure responses for a wellbore with pressure maintained at the
external boundary, closed reservoirs, bounded systems with
water/oil two phase flow for underground oil displacement.
Although the proposed model is based upon homogeneous reser-
voirs, it can be extended to heterogeneous reservoirs as naturally
fractured reservoirs or radial composite models for analysis and
interpretation of the pressure response when it is dominated by
the effects of boundary.

Acknowledgments
Fig. 15. Integration contour in the complex plane for g 1 ðr D ; sÞ. Four singularities
are shown.
The authors are indebted with D. del-Castillo-Negrete, D. Her-
nández, A. Altamira and L. Velasquillo. J.X.V.-H. acknowledges
project Métodos Avanzados para la Visualización del Subsuelo
(CYTED 2011 – P711RT0278). This work was supported by SENER-
CONACYT Grant no. 143935. Now, we consider a simple closed curve Ω which contains all
singularities of gðr D ; zÞ. Then the residue theorem states that
Z γ þ i1 ztD pffiffiffi
1 e Ψ 0;1 ðr D ; 1; zÞ  
pffiffiffi pffiffiffi dz ¼ Res eztD gðr D ; zÞ; 0
2π i γ  i1 zΨ 1;1 ð1; RD ; zÞ
Appendix A. Analytical solution for IBVP1 and IBVP2 cases
1
þ ∑ ResfeztD gðr D ; zÞ; λk g; ðA:3Þ
To complete the inverse transform of p D ðr D ; sÞ, Eq. (9), the case k¼1
is reduced to finding the Laplace inverse transform of f ðr D ; sÞ, see pffiffiffi
Eq. (14). First, we obtain the Laplace inverse transform of gðr D ; sÞ where we take into account that Ψ 1;1 ð1; RD ; zÞ has infinite simple
given by roots on the negative real axis and therefore λk represents a pole of
pffiffi order one of the function gðr D ; zÞ. At the origin gðr D ; zÞ has a pole of
Ψ 0;1 ðrD ; 1; sÞ order one and it can be expressed as
gðr D ; sÞ ¼ pffiffi pffiffi : ðA:1Þ
sΨ 1;1 ð1; RD ; sÞ   2RD
Res eztD gðr D ; zÞ; 0 ¼ 2 ; ðA:4Þ
Using Bessel function properties it can be shown that the denomi- RD  1
nator of gðsÞ has no positive roots for s 4 0. Near s ¼0 consider
moreover, the other poles on the negative axis must be considered
K 0 ðzÞ and K 1 ðzÞ expanded to second order and we consider I 0 ðzÞ
in order to obtain the exact inverse transform. To do this we do
and I 1 ðzÞ expanded to third order to conclude that s¼ 0 is a pole of
know that
order one. So, the integration contour for this transform is a cut ( I pffiffiffi )
along the negative real axis as shown in Fig. 15. The inverse 1  zt  1 1 ezt D Ψ 0;1 ðr D ; 1; zÞ
∑ Res e gðr D ; zÞ; λk ¼ ∑
D
pffiffiffi pffiffiffi dz ;
k ¼ 1 2π i C λ zΨ 1;1 ð1; RD ; zÞ
Laplace transform to g ðsÞ seems to have no simple solution, such a k¼1 k
solution was found by Mellin's inversion formula and the residue
ðA:5Þ
theorem. First, we obtain the Laplace inverse transform of gðsÞ
according to Mellin's inversion formula as where C λk are the simple closed curves inside Ω. As van Everdingen
Z γ þ i1 ztD pffiffiffi and Hurst (1949) let z ¼ u2 eiπ and using the properties of the Bessel
1 e Ψ 0;1 ðr D ; 1; zÞ functions (Carslaw and Jaeger, 1963), we obtain the integral around
gðr D ; t D Þ ¼ pffiffiffi pffiffiffi dz; ðA:2Þ
2π i γ  i1 zΨ 1;1 ð1; RD ; zÞ the boundary C λk :
( I pffiffiffi )
where 1 1 ezt D Ψ 0;1 ðr D ; 1; zÞ
∑ pffiffiffi pffiffiffi dz
k ¼ 1 2π i C λk zΨ 1;1 ð1; RD ; zÞ
Ψ m;n ðψ ; φ; yÞ ¼ K m ðψ yÞIn ðφyÞ þ ð  1Þm þ n þ 1 Im ðψ yÞK n ðφyÞ; ( )
I
e  u t D Φ1;0 ð1; r D ; uÞ
2
1 1
and γ 4 0 is a constant larger than the real part of any of the ¼2 ∑  ; ðA:6Þ
i¼1 2π i C u Φ1;1 ðRD ; 1; uÞ
singularities of i

pffiffiffi where
Ψ 0;1 ðr D ; 1; zÞ
gðr D ; zÞ ¼ pffiffiffi pffiffiffi :
zΨ 1;1 ð1; RD ; zÞ Φm;n ðψ ; φ; uÞ ¼ Y m ðψ uÞJ n ðφuÞ  J m ðψ uÞY n ðφuÞ;
 Y.N. del Angel et al. / Journal of Petroleum Science and Engineering 121 (2014) 149–158 157

and ui is the root of Φ1;1 ðRD ; 1; uÞ, i.e, ui satisfies Φ1;1 ðRD ; 1; ui Þ ¼ 0. an alternative approximation method which is based on the same
Now, the poles are represented on the negative real axis by λi ¼  u2i . way as was obtained Eq. (A.8). For this reason the details, shown in
The integral in the right-hand side (RHS) of Eq. (A.6) around the the following section, are not presented fully.
boundary C ui is obtained by considering the residue theorem and
that Φ1;1 ðRD ; 1; uÞ has infinite simple roots. Therefore
( ) A.2. Analytical approximation for f ðr D ; t D Þ, case IBVP2
I n o
e  u t D Φ1;0 ð1; r D ; uÞ
2
1 1 1
¼ 2 ∑ Res e  u t D GðuÞ; ui ;
2
2 ∑ 
i¼1 2 π i Cu Φ 1;1
i
ðR D ; 1; uÞ i¼1 Consider f ðr D ; sÞ defined as above but now we consider
ðA:7Þ f ðr D ; sÞ ¼ g 1 ðr D ; sÞ þ g 2 ðr D ; sÞ;
where
where
Φ1;0 ð1; r D ; uÞ pffiffi
Gðr D ; uÞ ¼  ; qRD Ψ 0;1 ðr D ; 1; sÞ
Φ1;1 ðRD ; 1; uÞ g 1 ðr D ; sÞ ¼  pffiffi ; ðA:12Þ
n o π 2 RD s3=2 Ψ 1;1 ð1; RD ; sÞ
Res e  u t D GðuÞ; ui ¼ e  ui t D ui F 2 ðr D ; RD ; ui Þ;
2

2 pffiffi
qRD sα Γ ð1  αÞΨ 0;1 ðr D ; 1; sÞ
g 2 ðr D ; sÞ ¼ p ffiffi : ðA:13Þ
RD s3=2 Ψ 1;1 ð1; RD ; sÞ
and
J 1 ðui ÞJ 1 ðRD ui ÞΦ1;0 ð1; r D ; ui Þ To complete the inverse transform of f ðsÞ the case is reduced to
F 2 ðr D ; RD ; ui Þ ¼ : finding the Laplace inverse transforms of g 1 ðsÞ and g 2 ðsÞ. First, we
J 21 ðRD ui Þ  J 21 ðui Þ
obtain the Laplace inverse transform of g 1 ðsÞ. On more time, from
Finally, Eq. (A.2) is expressed as Bessel function properties it can be shown that s ¼0 is a pole of
2RD 1 order two. So, we can use the integration contour given by Fig. 15.
þ π ∑ ui F 2 ðr D ; RD ; ui Þe  ui t D :
2
gðr D ; t D Þ ¼ ðA:8Þ It is necessary to invert the Laplace Transform, Eq. (A.12), by
R2D  1 i¼1
Mellin's inversion formula. So, it can be shown that


 2qR r 2D qR
g 1 ðr D ; t D Þ ¼ 2 D þ t D þ 2 D lnðr D Þ
A.1. Analytical solution for inverse Laplace of f ðr D ; sÞ ðRD 1Þ 4 ðRD  1Þ
qRD ðR4D þ 2R2D  4R2D lnðRD Þ  3Þ
A.1.1. Case IBVP1 þ ðA:14Þ
4ðR2D  1Þ2
As we have said, pD ðr D ; t D Þ is completed if the Laplace inverse
transform of f ðr D ; sÞ is obtained. For this purpose we just need to holds when the time tD is large. The other poles on the negative
get its Laplace inverse transform. Fortunately, the Laplace inverse axis are found by van Everdingen and Hurst, through the Fourier–
transform for this function can be obtained through the convolu- Mellin Theorem and the residue theorem to obtain finally
tion theorem. To do so, we must solve the integral

Z tD  2qR r 2D qR
  1 g 1 ðr D ; t D Þ ¼ 2 D þ t D þ 2 D lnðr D Þ
L  1 f ðr D ; sÞ ¼ qDext ðt D  ςÞgðςÞ dς: ðA:9Þ ðRD 1Þ 4 ðRD  1Þ
RD 0
qRD ðR4D þ 2R2D  4R2D lnðRD Þ  3Þ
Simplifying the last equation we obtain þ
Z tD Z tD 4ðR2D  1Þ2
qR qR
f ðr D ; t D Þ ¼  D gðςÞ dς þ D e  ðtD  ςÞ=τD gðςÞ dς: ðA:10Þ
qRD π 1 e  ui t D
2
RD 0 RD 0
þ ∑ F 2 ðr D ; RD ; ui Þ : ðA:15Þ
RD i ¼ 1 ui
Developing the first term in RHS of Eq. (A.10) we have
Z tD
qR 2qR π qRD 1 F 2 ðrD ; RD ; ui Þ  u2 tD
 D gðςÞ dς ¼  2 D t D þ ∑ ðe i  1Þ
RD 0 RD  1 RD i ¼ 1 ui
A.3. Long time approximation for g 2 ðr D ; t D Þ
Developing the second term in RHS of Eq. (A.10) we have
Z As we have said, pD ðr D ; t D Þ is completed if the Laplace inverse
qRD tD  ðtD  ςÞ=τD qR 2RD τD transform of f ðsÞ is obtained. For this purpose we just need to get
e gðςÞ dς ¼ D ½1  e  tD =τD 
RD 0 RD R2D  1 the Laplace inverse transform of g 2 ðsÞ. Unfortunately, the Laplace
!
1 τD inverse transform for this function cannot be obtained as g 1 ðsÞ
þ π ∑ ui F 2 ðr D ; RD ; ui Þ ½e  ui t D  e  t D =τD 
2
since if we use third-order approximations for the Bessel functions
i¼1 ð1  τD u2i Þ
I0 and I1 and second-order approximations for the Bessel functions
Finally we have that K0 and K1 as in van Everdingen and Hurst (1949) we conclude that
π qR D 1 g 2 ðsÞ has a pole at the origin but of fractional order, namely 2  α.
2qRD F 2 ðr D ; RD ; ui Þ  u2 t D
f ðr D ; t D Þ ¼  tD þ ∑ ðe i 1Þ For this reason the residue theorem cannot be applied immedi-
R2D 1 RD i¼1 ui
! ately. Of course, this observation clearly warrants the further
qRD 2RD τD  t D =τ D
1 τD  t D =τ D
þ ½1 e  þ π ∑ u F ðr ; R ; u Þ ½e  u2i t D
e  : study. Our goal is to attempt to approximate g 2 ðr D ; t D Þ for large
ð1  τD u2i Þ
i 2 D D i
RD R2D 1 i¼1
values of tD, thus an approximation for pD ðr D ; t D Þ is feasible. In this
ðA:11Þ work we utilize the Bessel functions and their asymptotic expan-
For the case when qDext ðt D Þ is given by Eq. (7) we can obtain the sions for small arguments as the mechanism for generating the
analytical expression for real space of f ðr D ; t D Þ from Eq. (A.9). So, solution approximation of g 2 ðr D ; t D Þ. We find that the approxima-
Z tD tion for g 2 ðsÞ for small values of s is given by
1
f ðr D ; t D Þ ¼  qRD Hðς  1Þð1  ς  α Þgðt D  ςÞ dς;

RD 0 2qRD Γ ð1  αÞ t 1  α 1 t α
lim g 2 ðsÞ ¼ þ ðr 2D  2 ln r D  1Þ ;
s-0 2
RD  1 Γ ð2  αÞ 4 Γ ð1  αÞ
should be further simplified. Evaluating the integral seems not to
have a simple expression. Therefore, only for this case, we use ðA:16Þ
158 Y.N. del Angel et al. / Journal of Petroleum Science and Engineering 121 (2014) 149–158

and can be shown from tables that inversion of (A.16) is Doublet, L.E., Blasingame, T.A., 1995. Decline curve analysis using type curves:
water influx waterflood cases. SPE 30774.
2qRD Γ ð1  αÞ 1 ðr 2D 2 ln r D  1Þ
t 1D α þ qRD t D α ;
Earlougher Jr., Robert, C., Kersch, K.M., Kunzman, W.J., 1974. Some characteristics of
g 2 ðr D ; t D Þ ¼ 2 ðA:17Þ
ðRD  1ÞΓ ð2  αÞ 2 R2D  1 pressure buildup behavior in bounded multiple-layered reservoirs without
cross flow. SPE 4499.
which holds when tD is large. Ehlig-Economides, C.A., Ramey, Jr., H.J., 1981. Pressure buildup for wells produced at
a constant pressure. SPE 7985.
Fuentes-Cruz, G., Camacho Velázquez R., Vásquez-Cruz M. 2010. A unified approach
References for falloff and buildup test analysis following a short injection/production time.
SPE 133539.
Acosta, L.G., Ambastha, A.K., 1994. Thermal well test analysis using an analytical Hsieh, B.Z., Chilingar, G.V., Lin, Z.S., 2007. Propagation of radius of investigation
multi-region composite reservoir model. SPE 28422. from producing well. Energy Sources Part A.
Bourdarot, G., 1996. Well Testing: Interpretation Methods. Éditions Technip, Paris. Hsieh, B.Z., Chilingar, G.V., Lin, Z.S., 2008. Determination of the constant coefficient
Bourdet, D., Whittle, T.M., Douglas, A.A., Pirard, Y.M., 1983. A new set of type curves in pressure propagation equation. Energy Sources Part A.
simplifies well test analysis. Word Oil 6. Klins, M.A., Bouchard, A.J., Cable, C.L., 1998. A polynomial approach to the van
Bourdet, D., Ayoub, J.A., Pirard, Y.M., 1984. Use of pressure derivative in well test Everdingen–Hurst dimensionless variables for water encroachment. SPE 15433.
interpretation. SPE 12777. Ozkan, E., Raghavan, R.S., 1997. Some strategies to apply the Stehfest algorithm for a
Carslaw, H.S., Jaeger, J.C., 1963. Operational Methods in Applied Mathematics. tabulated set of numbers. SPE 30552.
Dover Publications, Inc.; New York. Pascal, H., Pascal, F., 1985. Flow of non-Newtonian fluid through porous media. Int.
Ciriello, V., Di Federico, V., 2012. Similarity solutions for flow of non-Newtonian J. Eng. Sci. 23 (5), 571–585.
fluids in porous media revisited under parameter uncertainty. Adv. Water Rosa, A.J., 1996. Pressure transient behaviour in reservoirs with an internal circular
Resour. 43, 38–51. discontinuity. SPE 26455.
Ciriello, V., Longo, S., Di Federico, V., 2013. On shear thinning fluid flow induced by Stehfest, H., 1970. Numerical inversion of Laplace transforms. Commun. ACM 13,
continuous mass injection in porous media with variable conductivity. Mech. 47–49.
Res. Commun. 52, 101–107. van Everdingen, A.F., Hurst, W., 1949. The application of the laplace transformation
Comisión Nacional de Hidrocarburos. 2010. Proyecto Aceite Terciario del Golfo: to flow problems in reservoirs. Trans. AIME. 186.
Primera revisión y recomendaciones. Secretaria de Energía. Estados Unidos Warren, J.E., Root, P.J., 1963. The behaviour of naturally fractured reservoirs.
Mexicanos. SPE 426.
Corbett, P.W.M., Hamdi, H., Gurav, H., 2012. Layered fluvial reservoirs with internal
fluid cross flow: a well-connected family of well test pressure transient
responses. Pet. Geosci. 18, 219–229.