Mecánica Computacional Vol XXIX, págs.

8767-8779 (artículo completo)

Eduardo Dvorkin, Marcela Goldschmit, Mario Storti (Eds.)
Buenos Aires, Argentina, 15-18 Noviembre 2010

SOLUTION OF A ONE-DIMENSIONAL THREE-PHASE FLOW IN

HORIZONTAL WELLS USING A DRIFT-FLUX MODEL

Arthur B. Soprano, Gustavo G. Ribeiro, António F. C. da Silva and Clovis R. Maliska

Computational Fluid Dynamics Lab - SINMEC, Federal University of Santa Catarina (UFSC), Campus
Universitário, Florianópolis-SC, Brasil, arthur@sinmec.ufsc.br, http:// www.sinmec.ufsc.br

Keywords: Reservoir-well coupling, drift-flux, multiphase flow, horizontal wells.

Abstract. To design wells and its equipments along the oil fields, one has to solve the flow that occurs
inside the oil reservoirs and its injecting and producing wells. This work presents a multiphase isothermal
flow model to the solution for horizontal wells using a drift-flux model (water, oil and gas) along a
one-dimensional domain. The problem is discretized with a Finite Volume Method and solved using
Newton’s Method. As a drift-flux model is a homogeneous model, it uses the mixture momentum and
continuity equations and it is also necessary to solve the gas and oil phase continuity equations in order
to solve the three-phase flow. Lateral mass inflow, due to the flow from reservoir to well, is considered
as a source/sink term in the one-dimensional continuity equations. Spatial and temporal interpolation
schemes are of first order, mostly upwind schemes. Results are compared with a CMG-IMEX black-
oil reservoir simulator. The algorithm is implemented using a C++ OOP programming language and
all the derivatives of the Jacobian Matrix are calculated numerically, making the code more generic,
allowing user to change property models and drift parameters according to the problem. The study of the
coupling between well and reservoirs is a state-of-art research activity. Most of the important petroleum
companies are developing proprietary softwares for modeling as much as possible the several phases of
the oil production chain.

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

8768 A. SOPRANO, G. RIBEIRO, A. SILVA, C. MALISKA

1 INTRODUCTION
For optimized petroleum exploitation, it is necessary to perform an intense sequence of stud-
ies and analysis. Reservoir simulation is a field that is in progressive growth since last decades
and it’s been always increasing its complexity to better represent the physical phenomena that
occurs during the process of oil extraction. Horizontal wells are today one of the alternatives to
maximize production in several situations.
Another requirement is to consider multiphase flow along the reservoir and wellbore do-
mains, for a better prediction of results and also different information like the total production
of each phase. A complete reservoir simulation procedure requires the knowledge of pressure
and saturation fields inside the reservoir subject to the boundary conditions that the injection
and production wells provide. In order to obtain these boundary conditions, it is also necessary
to determine the pressure and saturation fields along every well inside the reservoir (figure 1).
Hence, these two problems need to be coupled in some manner. If this goal is achieved, one can
use this tool to design and optimize well locations, well types, etc.
The solution of multiphase flow in horizontal wells can be done by solving the momentum
and continuity equations for each phase, applying the proper interphase and wall friction models
that may exist. This type of solution is know as multi-fluid models (Ishii and Hibiki, 2006). Yet,
it can also be solved considering a homogeneous flow and then apply a model to represent the
multiphase behavior, which are known as drift flux models. The main advantage of the latter
approach is that it simplifies the problem so that fewer equations need solved and good results
are still obtained.
Petroleum wells are usually represented as one-dimensional curves inside the reservoir with
lateral mass inflow or outflow, as its dimensions are much smaller than the ones for the reser-
voir. In the presented work, we model a one-dimensional three-phase flow (water, oil and gas)
corresponding to petroleum wells. The three-phase flow is modeled based on the approach pro-
posed by Shi et al. (2003), but using the one-dimensional drift-flux model proposed by Hibiki
and Ishii (2003).

2 MODEL FORMULATION
2.1 Drift-Flux Model
Drift-Flux models are homogeneous models that considers slip between phases allowing dif-
ferent velocities to be calculated for each phase. A complete explanation about those models
and the calculation of its parameters can be found in Ishii and Hibiki (2006), where it presents
the basic equations and formulas for the two-phase flow model. For the one-dimensional case,
all properties in each control volume along the well should not vary in the cross-sectional direc-
tions. Therefore, it is necessary to integrate the properties along the cross-sectional area in order
to obtain the one-dimensional equations. Given a property φ, its average over the cross-sectional
area (A) is Z
1
hφi = φ dA (1)
A A
the average operator h i will be omitted from equations from now on to simplify notation. All the
one-dimensional equations are derived from the integral along the cross-sectional area. Other
properties should also be defined, such as volumetric fraction (α), which represents the ratio
between the volume occupied by the phase p over the total volume. Assuming that inside a
control volume the properties do not vary, the volumetric fraction can also be calculated as the
ratio of the cross-sectional area occupied by the phase over the total area. For a phase p we have

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXIX, págs. 8767-8779 (2010) 8769

Figure 1: Coupling scheme between reservoir and well.

Vp Ap
= . αp = (2)
V A
The sum of the volumetric fraction for each phase is
X
αp = 1. (3)
p

The drift-flux model is a two-phase formulation that involves the calculation of the velocity of
each phase with a given constituve equation that involves the mixture center-of-mass velocity
and the drift velocity of the dispersed phase (Hibiki and Ishii, 2003). The drift velocity (Vgj ) is
the velocity of the dispersed phase relative to the volume center of the mixture

Vgj ≡ vg − j (4)

where vg is the gas phase velocity and j is the total volumetric flux, calculated as

j = αl vl + αg vg . (5)

the velocity (vp ) of a phase p is defined as the ratio between the volumetric flux of the phase over
the cross-sectional area occupied by the same phase. Now it is possible to define the mean drift
velocity, which allows the formulation of the constitutive equations to calculate the velocities
of each phase. The mean drift velocity is given by

V̄gj = Vgj + (C0 − 1)j (6)

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

8770 A. SOPRANO, G. RIBEIRO, A. SILVA, C. MALISKA

Figure 2: Extension of the drift-flux model to a three-phase flow.

and C0 is called the profile parameter that can assume different values according to the distri-
bution profile of disperse phase. Expressions to calculate the velocities for each phase can also
be derived. For the two-phase (gas-liquid) flow model, we have
ρl
vg = vm + V̄gj ,
ρm
αg ρ g (7)
vl = vm − V̄gj
αl ρ m
and the total volumetric flux can be calculated as
αg (ρl − ρg )
j = vm + V̄gj (8)
ρm
where vm is mixture velocity, defined by
ρg αg vg + ρl αl vl
vm = (9)
ρm
To extend the two-phase drift-flux model to a three-phase flow, we now consider that the liquid
phase is a mixture of two fluids (oil and water) and apply again the model for these phases with
new drift-flux parameters (figure 2):
0
V̄ow = Vow + (C00 − 1)j (10)

and the oil and water velocities are calculated as

ρw
vo = vl + V̄ow ,
ρl
αo ρo (11)
vw = vl − V̄ow
αw ρ l
This extension of the two-phase multiphase model to a three-phase flow may not be as precise
as a three-fluid model. Although, as pointed by Shi et al. (2003), this approach can produce
reasonable results and the expected behaviour if the drift-flux parameters are well adjusted to
the flow patterns that may exist along the well.

2.2 Wellbore Governing Equations

In the present model, four equations need to be solved in order to calculate all the necessary
properties, three continuity equations and one momentum equation:
Mixture Continuity Equation
 
∂ρm ∂ (ρm vm ) ṁ
+ = (12)
∂t ∂s V total

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXIX, págs. 8767-8779 (2010) 8771

Gas Phase Continuity Equation

   
∂ (αg ρg ) ∂ (αg ρg vm ) ṁ ∂ αg ρg ρl
+ = − V̄gj (13)
∂t ∂s V g ∂s ρm

Oil Phase Continuity Equation

     
∂ (αo ρo ) ∂ (αo ρo vm ) ṁ ∂ αo ρ o ρ w ∂ αg ρg
+ = − V̄ow + αo ρo V̄gj (14)
∂t ∂s V o ∂s ρl ∂s (1 − αg ) ρm

Mixture Momentum Equation

 
∂ (ρm vm ) ∂ (ρm vm vm ) ∂P f ∂ αg ρg ρl 2
+ =− − ρm g sin (θ) − ρm vm |vm | − V̄ (15)
∂t ∂s ∂s 2D ∂s αl ρm gj

where the terms (ṁ/V )p correspond to the source/sink of each phase associated with the inflow
or outflow of fluids between well and reservoir. The mixture equations are obtained through the
sum of the equations for each phase. This problem is solved for pressure (P ), mixture velocity
(vm ), gas volumetric fraction (αg ) and oil volumetric fraction (αo ).
The friction factor f adopted in this work is the Churchill’s correlation:
 12 ! 121
8 1
f =8 + (16)
Re (A + B)1.5
and
 0.9 !!16
7 ε
A= −2.457 ln + 0.27
Re D
 16
37530
B=
Re

where ε is the pipe’s rugosity, D the internal diameter and Re the Reynolds number, calculated
as
ρm V D
Re = (17)
µm
where V is a characteristic velocity of the flow (vm in this case).
Even though the expression above provides a way to calculate the friction factor for laminar
and turbulent flow, it is only valid for single-phase flow and without lateral mass influx. There-
fore, it’s necessary to find a better way of calculating the fricton factor that better predicts the
pressure drop for the given conditions. Some alternatives are proposed by Ouyang (1998). The
calculation of the friction factor will not be discussed in this work.

2.3 Reservoir Governing Equation

The equations that represent the reservoir’s behavior, for a three-phase flow are the conser-
vation of mass to each phase involved. Usually we use the formulation by saturation, however,
this formulation presents computational problems when the gas component is fully contained in
oil phase, i.e., S g = 0. For resolve this situation we use the mass fraction formulation to express
the mass conservation for each phase (standard black-oil model) (Aziz and Settari, 1979).

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

8772 A. SOPRANO, G. RIBEIRO, A. SILVA, C. MALISKA

2.3.1 Continuity Equations

This section will be exposed the mass conservation as a function of mass fractions of three
phases (water, oil and gas). The black-oil model consider the three components in the mixture,
and three phases where the oil phase can be compound with two components, oil and gas. This
equations are
Water Conservation
 w w 
∂ m w kr ρ ¯ w
(φρ Z ) = ∇ · K̄.∇Φ − m̄w , (18)
∂t µw
Oil Conservation
o o
 
∂ m o oo kr ρ ¯
(φρ Z ) = ∇ · X o
K̄.∇Φ − X oo m̄w ,
o
(19)
∂t µ
Gas Conservation
krg ρg ¯ o o
 
∂ oo kr ρ ¯ .∇Φo
(φρm Z g ) = ∇ · K̄ .∇Φ g
+ (1 − X ) , K̄ − m̄g − (1 − X oo )m̄o (20)
∂t µg µo
where ρm is medium mass density defined to
nphases
X
ρp S p , (21)
p

and φ is the porosity of the porous media. Z p , krp , ρp and µp are the mass fraction, relative
permeability, mass density and viscosity of the phases, respectively. X cp is the mass fraction of
¯ is the absolute permeability tensor.
the componente c in the phase p and K̄
The variable Φp is the potential of the phase p and represent the relation between the oil
pressure and the water and gas pressure. This relationship is the capillary pressure as

Φo = po − ρo gz,
Φw = po − pcow −ρw gz,
| {z }
pw (22)
g o cog g
Φ = p − p −ρ gz.
| {z }
pg

Finally m̄p is the mass per volume entering or leaving the domain, to each phase. Adding to
the system of equations has been the global mass conservation equation given by
Z w + Z o + Z g = 1, (23)
this form we have the total system of equations for reservoir simulator formed by equations (18)
to (20) and eq. (23).

2.3.2 Discretization
The finite volume discretization of the equations is performed by integrating the control
volume P in space and time. Thus, equations (18), (19) and (20) turn into

h i ∆V nf
P
X
m w t+∆t m w t
 w
λf Tf (Φw w w

(φρ Z ) − (φρ Z ) = N B − ΦP ) − ṁ , (24)
P ∆t
f =1

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXIX, págs. 8767-8779 (2010) 8773

h i ∆V nf
P
X
m o t+∆t m o t
 oo o
X λf Tf (ΦoN B − ΦoP ) − X oo ṁo ,

(φρ Z ) − (φρ Z ) = (25)
P ∆t
f =1

h i ∆V nf
P
X
g t+∆t g t
 g
m
(φρ Z ) − (φρ Z )m
= λf Tf (ΦgN B − ΦgP ) +
P ∆t (26)
f =1

+X λf Tf (ΦN B − ΦoP ) − ṁg − X go ṁo ,

go o o


where f are the faces of the discrete control volume, λpf is the mobility of phase p in the face f
given by
ρp k p
λpf = wr , (27)
µ
and Tf is the transmissibility of face, where for the east face of control volume, on isotropic
medium, is written as
∆y∆z
Te = K . (28)
∆x
Finally, the ṁp is the mass flow rate that is entering or leaving the control volume by the
well, which is determined from
ṁo = ±λo W I (PPo − Pwell ) ,
ṁw = ±λw W I PPo − Ppcow − Pwell ,


(29)
ṁg = ±λg W I PPo + Ppcog − Pwell ,

where W I is well index determined with the Peaceman’s wellmodel (Peaceman, 1983). This
system of equations is composed of three unknowns, P o , Z o and Z w because Z g can be deter-
mined by eq.(23).

3 NUMERICAL PROCEDURE
The discretized equations are solved using a Newton’s method, but reservoir equations and
wellbore equations are solved separately. Therefore it is necessary to define a computational
algorithm that couples both solutions and guarantee the convergence of the problem. The pres-
sure along the well is a boundary condition for the solution of the reservoir flow and, once the
pressure and saturations fields are calculated, it is possible to calculate the lateral mass inflow
along the well. Therefore, the communication between reservoir and well is made through
source/sink terms. So it is necessary to solve iteratively wellbore and reservoir equations until
the fields are converged, always updating the boundary conditions of each domain. Figure 3
shows a simplified diagram of this procedure.
All the derivatives used to compute the jacobian matrix are calculated numerically, as follow

∂R R(X + ∆X) − R(X)

= (30)
∂X ∆X
where R is one of the residual equations obtained after the discretization of the PDE’s and
X one of the independent variables of the problem. This procedure is used for both reservoir
and wellbore equations and provides a way of make the algorithm more general, allowing the
inclusion of different property models in future studies, such as the ones proposed in Beggs
(2003).

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

8774 A. SOPRANO, G. RIBEIRO, A. SILVA, C. MALISKA

Figure 3: Simplified computational diagram of the numerical procedure.

4 RESULTS
For the numerical tests, we created a synthetic reservoir with dimensions of 550m × 550m ×
30m with the geometry given in figure 4. The reservoir is perforated with one vertical injector
well and one horizontal producer well. It also has initially certain amounts of gas, oil and
connate water. The simulation runs for 500 days and results are compared with the black-
oil reservoir simulator IMEX-CMG and the simulation parameters are given on the appendix.
Figures 5 and 6 show the total production of oil and gas and well as the flow rates of these
phases, respectivelly. The drift-flux parameters inside the well were considered constant values
and a no-slip condition for the oil-water model was imposed, so that oil and water have the same
velocity.
During the initial part of the simulation, a great amount of gas is produced, chiefly during
the first 100 days. After that, as the reservoir runs out of gas, the production of oil increases,
but with a decreasing rate.
We did not found information about the drift-flux parameters and friction fator used on the
CMG-IMEX well simulator. That could be one of the reasons why there’s a slight difference
between the results of gas and oil production. Different drift-flux parameters and friction factor
can result in a different pressure drop along the well, thus affecting the total production.

5 CONCLUSIONS
This paper presented a procedure for solving a three-phase flow along the well with a cou-
pled framework with the reservoir. The model is based on a general algorithm that allows the

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXIX, págs. 8767-8779 (2010) 8775

Figure 4: Scheme of the problem.

3.5E+07 1.0E+05

9.0E+04
3.0E+07
8.0E+04
Gas Production [m3]

Oil Production [m3]

2.5E+07 7.0E+04

6.0E+04
2.0E+07
5.0E+04
1.5E+07 IMEX IMEX
4.0E+04
Present
1.0E+07 3.0E+04 Present

2.0E+04
5.0E+06
1.0E+04

0.0E+00 0.0E+00
0 100 200 300 400 500 0 100 200 300 400 500

Time [days] Time [days]

(a) Results for total gas production. (b) Results for total oil production.
1.2E+00

1.0E+00

8.0E-01
Water Cut

6.0E-01

4.0E-01 IMEX

Present
2.0E-01

0.0E+00
0 100 200 300 400 500

Time [days]

(c) Results for water cut.

Figure 5: Comparisons with CMG-IMEX.

1.4E+06 9.0E+02

1.2E+06 IMEX 8.0E+02

IMEX
Gas Flow Rate [m3/day]

Oil Flow Rate [m3/day]

Present 7.0E+02
1.0E+06 Present
6.0E+02
8.0E+05 5.0E+02
6.0E+05 4.0E+02
3.0E+02
4.0E+05
2.0E+02
2.0E+05 1.0E+02
0.0E+00 0.0E+00
0 100 200 300 400 500 0 100 200 300 400 500
Time [days] Time [days]

(a) Gas flow rate. (b) Oil flow rate.

Figure 6: Comparisons of total rate with CMG-IMEX.

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

8776 A. SOPRANO, G. RIBEIRO, A. SILVA, C. MALISKA

inclusion of different models for the coefficients in each equations, e.g., different density, vis-
cosity and friction factor equations could be used. Also, the drift-flux parameters can be chosen
according to each flow pattern that will occur inside the well, that depends on the volumetric
fractions of each phase and its velocities (Provenzano, 2007). The design of an algorithm that
does not depend on the fluid models and constitutive equations was one of the main goals of the
presented work. Now it is possible to perform an indefinite sequence of studies of the correct
parameters to use on each problem and obtaining results that better represent and predict the
real situation of oil production.
As next steps, the inclusion of differents drift-flux parameters according to each flow pattern,
thus allowing the algorithm to handle most of the flow regimes. Also, the solution of the energy
equation would allow the model to handle not only the horizontal part of the well, but also the
vertical part (until the surface). The energy equation is necessary in the vertical part because
the temperature gradients are much greater on that region, thus affecting the fluids properties.

REFERENCES
Aziz K. and Settari A. Petroleum Reservoir Simulation. Applied Science Publishers, 1st edition,
1979.
Beggs H.D. Production Optimization, Using NODALTM Analysis. OGCI and Petroskills Publi-
cations, 2003.
Hibiki T. and Ishii M. One-dimensional drift-flux model and constitutive equations for relative
motion between phases in various two-phase flow regimes. International Journal of Heat
and Mass Transfer, 2003.
Ishii M. and Hibiki T. Thermo-Fluid Dynamics of Two-Phase Flow. Springer Science, 2006.
Ouyang L.B. Single Phase and Multiphase Fluid Flow in Horizontal Wells. Ph.D. thesis,
Stanford University, 1998.
Peaceman D.W. Interpretation of well-block pressures in numerical reservoir simulation with
nonsquare grid blocks and anisotropic permeability. SPE Journal, 23:531–543, 1983.
Provenzano C.E.C. Previsão Numérica de Escoamento Bifásico em Tubulações Utilizando o
Modelo de Deslizamento. Master’s Thesis, PUC - Rio de Janeiro, 2007.
Shi H., Holmes J., Durlofsky L., Aziz K., Diaz L., Alkaya B., and Oddie G. Drift-flux modeling
of multiphase flow in wellbores. SPE - Society of Petroleum Engineers, (SPE 84228), 2003.

APPENDIX - PROPERTIES AND SIMULATION INFORMATION

Medium physical properties:

• Porosity (φ): 0.2

• Rock Compressibility: 5.8015 × 10−10 P1a

• Reference Pressure Pore: 1 × 105 P a

• Absolute permeability on the x direction: kxx = 9.869−14 m2

• Absolute permeability on the y direction: kyy = 9.869−14 m2

• Absolute permeability on the z direction: kzz = 9.869−15 m2

Fluids physical properties:

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXIX, págs. 8767-8779 (2010) 8777

Pressure [Pa] Rs [fr] Bo [fr] eg [fr] µo [Pa.s] µg [Pa.s]

2.7579 × 106 29.3773260 1.0120 3.01767 × 101 1.17 × 10−3 1.30 × 10−5
5.5158 × 106 59.6448740 1.0255 6.03534 × 101 1.14 × 10−3 1.35 × 10−5
8.2737 × 106 89.0222000 1.0380 9.08380 × 101 1.11 × 10−3 1.40 × 10−5
1.1032 × 107 118.3995260 1.0510 1.21118 × 102 1.08 × 10−3 1.45 × 10−5
1.3790 × 107 147.4207632 1.0630 1.50885 × 102 1.06 × 10−3 1.50 × 10−5
1.6547 × 107 175.3737340 1.0750 1.81676 × 102 1.03 × 10−3 1.55 × 10−5
1.9305 × 107 201.1901720 1.0870 2.11961 × 102 1.00 × 10−3 1.60 × 10−5
2.2063 × 107 226.1163880 1.0985 2.40609 × 102 9.80 × 10−4 1.65 × 10−5
2.4821 × 107 247.4817160 1.1100 2.73921 × 102 9.50 × 10−4 1.70 × 10−5
2.7579 × 107 267.0666000 1.1200 3.01767 × 102 9.40 × 10−4 1.75 × 10−5
3.0337 × 107 284.8710400 1.1300 3.29720 × 102 9.20 × 10−4 1.80 × 10−5
3.3095 × 107 298.4024144 1.1400 3.63352 × 102 9.10 × 10−4 1.85 × 10−5
3.5853 × 107 311.5777000 1.1480 3.95649 × 102 9.00 × 10−4 1.90 × 10−5
3.8611 × 107 322.2603640 1.1550 4.23923 × 102 8.90 × 10−4 1.95 × 10−5
Table 1: PVT tables.

• Reference Water Density (ρw,ref ): 1000kg/m3

• Reference Oil Density (ρo,ref ): 800kg/m3

• Reference Gas Density (ρg,ref ): 1.1245kg/m3

• Water Reference Formation Volume Factor (Bw,ref ): 1.01420

• Water Compressibility (cw ): 4.3511 × 10−10 P1a

• Oil Compressibility (co ): 1.4504 × 10−9 P1a

• Reference Water Viscosity (µw,ref ): 10−3 P a · s

• Water Viscosibility: 0.0 PPa·s

a

• Oil Viscosibility: 0.0 PPa·s

a

• Reference Water Pressure: 105 P a

Rock-fluid Properties Section:

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

8778 A. SOPRANO, G. RIBEIRO, A. SILVA, C. MALISKA

(b) Gas-Liquid solubility ta-

(a) Water-Oil solubility table. ble.
Sw [fr] Pcow [Pa] Sl [fr] Pcog [Pa]
0.22 48263.299 0.22 26889.5523
0.3 27579.028 0.3 24131.6495
0.4 20684.271 0.4 20684.271
0.5 17236.8925 0.5 17236.8925
0.6 13789.514 0.6 13789.514
0.8 6894.757 0.7 10342.1355
0.9 3447.3785 0.8 6894.757
1.0 0.0 0.9 3447.3785
0.96 1378.9514
1 0
Table 2: Solubility tables.

Initial parameters:

• Bubble Point Pressure: 108 P a

• Reference Depth: 30m

• Reference Pressure Depth: 7 × 106 P a

Relative permeability curves:

• Water-Oil curves: Corey Correlation (1,1)

• Connate Water Saturation: 0.22

• Irreducible Oil WSaturation: 0

• Gas-liquid curves: Corey Correlation (1,1)

• Connate Gas Saturation: 0.0

• Irreducible Oil GSaturation: 0.0

Reservoir Simulation parameters:

• Timestep: 0.1 days

• Final Time: 500 days

Well physical parameters:

• Injection Vertical Well Bottom Hole Pressure: 3.86110 × 107 P a

• Production Horizontal Well Bottom Hole Pressure (Pheel ): 2.75790 × 106 P a

• Well Radius: 0.07m

• Well Rugosity (ε): 10−4 m

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXIX, págs. 8767-8779 (2010) 8779

Well Drift-Flux parameters:

• Gas-Liquid Drift Velocity Vgj : 0.05m/s

• Gas-Liquid Profile Parameter C0 : 1.2

• Oil-Water Drift Velocity Vow : 0.0m/s

• Oil-Water Profile Parameter C0 : 1.0

Copyright © 2010 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar