You are on page 1of 11

SPE 113215

Application of a New Fully-Coupled Thermal Multiphase Wellbore Flow


Model
S. Livescu, SPE, L.J. Durlofsky, SPE, K. Aziz, SPE, Stanford University, and J.C. Ginestra, SPE, Shell Oil
Company
Copyright 2008, Society of Petroleum Engineers

This paper was prepared for presentation at the 2008 SPE/DOE Improved Oil Recovery Symposium held in Tulsa, Oklahoma, U.S.A., 1923April2008.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.


Abstract
Thermal recovery processes are widely applied for the production of heavy oil and oil sands. Thermal reservoir simulation
models, however, often lack a comprehensive well modeling capability. Such a capability is required to capture the detailed
thermal effects that occur in the wellbore. These effects can be important as they impact wellbore pressure and temperature
and thus production and injection. We recently developed a fully-coupled black-oil thermal multiphase wellbore flow model
and implemented it into Stanfords General Purpose Research Simulator (GPRS). The model computes pressure, temperature,
and oil, water and gas phase fractions along the wellbore as a function of time and includes treatments for slip between fluid
phases, heat losses to the reservoir, and general variations of fluid properties with temperature and pressure. The purpose of
this paper is to validate and test the coupled wellbore-reservoir model for challenging and realistic cases. The thermal wellbore
model is first validated through comparison to field data for three-phase flow in a long well with both vertical and inclined
sections. Close agreement between the model and field data is obtained. Complex wells containing multiple branches are then
simulated, including a steam-water case with vaporization and condensation. The general conclusion from this work is that the
new model is capable of simulating a wide variety of complex coupled reservoir-wellbore phenomena.

Introduction
Thermal recovery processes such as steam flooding, steam assisted gravity drainage (SAGD), and downhole electrical heating
are essential for the production of heavy oil and oil sands, and they are also under investigation for the in situ upgrading of oil
shales. Under any of these recovery processes, a key to efficient reservoir management and process optimization is the ability
to perform accurate reservoir simulations. Although the current generation of flow simulators is able to model thermal effects
within the reservoir, the wellbore flow models linked with these simulators often lack comprehensive thermal capabilities.
This may limit simulation accuracy for many cases. Thus, there is a need for the development, validation and testing of
comprehensive thermal wellbore flow models that are coupled to reservoir simulators. This is the overall goal of the work
described in this paper.
Over the last few decades, there have been many analytical and numerical models presented for nonisothermal wellbore
flow. This literature is reviewed in detail by Livescu et al. (2008) so our discussion here will be brief. The analytical models
include, among others, those of Ramey (1962), Satter (1965), Hasan and Kabir (1994, 1999, 2002, 2007) and Livescu et al.
(2008). These models are useful in many contexts though they typically include many simplifications, such as assuming steady
state flow or neglecting spatially varying inflow, which render them unsuitable for use in general simulators. Numerical
models are much more general as they potentially allow for transient effects, spatial variability, slip between phases, general
property variation, etc. Such models have been developed by Farouq Ali (1981), Farouq Ali and Abou-Kassem (1989), Stone
et al. (1989, 2002), Holmes et al. (1998), Pourafshary et al. (2007) and Livescu et al. (2008). These models entail one-
dimensional (axial) representations of the wellbore and include coupled conservation equations for multiple components (e.g.,
oil, water and gas), an energy equation, and a pressure drop relationship. Flow from the reservoir into the wellbore (in the case
of a production well) provides source terms for the wellbore flow model. Previous formulations have been developed for both
black-oil models (Stone et al., 1989; Livescu et al., 2008) and fully compositional models (Stone, 2002; Pourafshary et al.,
2007).
In recent work (Livescu et al., 2008), we presented a new thermal wellbore flow model developed within a black-oil
framework. This wellbore model was implemented into Stanfords General Purpose Research Simulator (GPRS), described in
detail by Cao (2002) and Jiang (2007). The model contains a number of features that distinguish it from earlier developments,
2 SPE 113215
including a three-phase drift-flux representation, which enables it to capture slip between the oil, water and gas phases. In
Livescu et al. (2008) we described the governing equations and the numerical implementation in detail. A new analytical
model, which generalizes previous such models, was also derived. We then presented a validation against field data for a
vertical well, comparisons between the numerical and analytical models, and an example involving a multilateral well. As our
emphasis in Livescu et al. (2008) was on model development, the examples involved very simple reservoir models and/or
wells. In that work we did not consider the application of the coupled reservoir-wellbore modeling procedure to realistic cases.
In this paper, we first briefly describe the thermal wellbore flow model. We then apply the general modeling procedure to a
number of realistic examples and thus establish its applicability for practical reservoir simulation. Specifically, we present a
validation of the model in which we compare model predictions to the field data of Hasan and Kabir (2007). These data
include pressure and temperature for a long offshore well with both vertical and inclined sections. Heat losses to the formation
and ocean are included. We achieve very close agreement between field data and simulation results without adjusting any
model parameters. We then consider the numerical performance of the wellbore model in a fully-coupled simulation involving
a multilateral well in a complex reservoir. The last example includes a two-phase flow problem with phase change
(vaporization and condensation) in the wellbore. Taken in total, the results presented here serve to validate the thermal
wellbore flow model and to demonstrate its range of applicability for realistic reservoir-wellbore models.

Model Formulation and Numerical Implementation
In this section we present the equations governing the nonisothermal multiphase wellbore flow. A comprehensive description
of this model can be found in our earlier paper (Livescu et al., 2008). The model involves time and one spatial dimension (the
axial direction, designated z). We consider a black-oil system involving oil, water and gas components (or more properly
pseudo-components) and three fluid phases, also referred to as oil, water and gas. In the current representation, the water
component can reside only in the water phase, while the other components can reside in the oil and gas phases. Other
representations are also possible. The governing equations include mass balance equations for each component, an energy
balance equation for the overall fluid, and a pressure drop equation. The unknowns are the gas and water phase volume
fractions, designated
g
and
w
, wellbore pressure, p
w
, mixture velocity (total volumetric flow rate divided by pipe area), V
m
,
and wellbore temperature, T
w
. The oil phase volume fraction is given by
o
= 1 -
g
-
w
. The well is discretized spatially into a
number of segments, as shown in Fig. 1, and the unknowns specified above are determined for each segment at each time step.
The pressure loss between adjacent segments accounts for hydrostatic, frictional and acceleration effects
a
w
f
w
h
w w
z
p
z
p
z
p
z
p

(1)
All variables are defined in the Nomenclature and the detailed terms are specified in Livescu et al. (2008). The mass
conservation equations for the gas, oil and water components are, respectively,

= = =
=

g o p
gp
g o p
gp sp p
g o p
gp p p
m V
z t
, , ,
(2)

= = =
=

g o p
op
g o p
op sp p
g o p
op p p
m V
z t
, , ,
(3)
( ) ( )
ww ww sw w ww w w
m V
z t
=

(4)
In the mass conservation equations,
p
and
p
represent the density and the volume fraction of phase p (p = oil, water, gas),
respectively,
cp
is the molar fraction of component c (c = oil, water, gas) in phase p, V
sp
is the superficial velocity of phase p
and m
cp
is the source/sink term, which is nonzero at well segments that are open to flow to or from the reservoir. All other
symbols are defined in Nomenclature. The m
cp
terms are represented in terms of the usual well index W and are written as
( )
w
p cp p p cp
p p W m = (5)
where
p
is the phase mobility.
The energy conservation equation is given by

+ =

p p
H loss sp p p p sp p
p
p p p p
m Q g V V H V
z
V U
t

2 2
2
1
2
1
(6)
where U
p
and H
p
are the internal energy and the enthalpy of phase p, respectively, and g is the gravitational component along
the well. The rate of work done on the fluid by viscous forces and the conductive heat transfer along the well are ignored. The
source term is analogous to Eq. (5) and is written as
( )

=
p
w
p p p p H
p p H W m (7)
SPE 113215 3
The Q
loss
term in Eq. (6) describes the heat loss from the wellbore fluid to the surroundings. This is approximated using the
overall heat transfer coefficient, U
to
, which is computed as a sum of thermal resistances of all materials in the wellbore, i.e., the
tubing, annulus fluid, casing and cement. The heat loss term is then given by
( ) T T U r Q
w
to in loss
= 2 (8)
where r
in
is the internal radius of the well and T is the reservoir temperature outside the wellbore. Note that in Eq. (6) the
minus sign for the heat loss term corresponds to production wells. The sign is reversed for injection wells.
We apply a drift-flux model to represent the hydrodynamics inside the wellbore. Unlike the more complex (and potentially
more accurate) mechanistic models, drift-flux models are well suited for reservoir simulators as they are relatively simple,
continuous and differentiable. The drift-flux model was originally proposed by Zuber and Findlay (1965) for vertical two-
phase flow bubble flow. They expressed the gas velocity as a sum of two terms
d m g
V V C V + =
0
(9)
where C
0
is a profile parameter, V
d
is the drift velocity of gas relative to the liquid, and V
m
is the average mixture velocity (the
sum of the gas and liquid superficial velocities). The gas velocity is thus impacted by two different mechanisms, specifically
the higher concentration of gas near the center of the well, where the velocity is higher (this effect is captured by the term
C
0
V
m
) and the tendency of the gas to rise in the well due to buoyancy (captured by V
d
).
The functional forms for C
0
and V
d
(which depend on
p
, V
m
and the inclination angle of the well, ) that are used in our
model derive from the results of Shi et al. (2005 a, b). The representations developed by Shi et al. (2005 a, b) are based on the
experimental data of Oddie et al. (2003), who performed an extensive set of large-scale pipe flow experiments for one, two
and three-phase flows at various inclination angles. Shi et al. (2005 a, b) then applied an optimization procedure to determine
the parameters for C
0
and V
d
that best fitted the experimental results. For the nonisothermal case, we extended the drift-flux
model proposed by Shi et al. (2005 a, b) to include temperature effects. Specifically, the phase densities are now temperature
dependent, which in turn leads to temperature dependent phase velocities.
The well equations outlined above are solved simultaneously (fully implicitly) with the reservoir flow equations. For a
three-phase black-oil system, the wellbore model is represented by Eqs. (1) through (9). The thermal reservoir model consists
of three mass conservation equations and an energy conservation equation, as well as the capillary pressure relationship and
the saturation constraint. The thermal reservoir equations are discretized using a standard finite volume procedure (see, e.g.,
Aziz and Wong, 1989; Farouq Ali and Abou-Kassem, 1989, for detailed descriptions).
For the numerical discretization of the wellbore flow and energy equations, the well is discretized into N
s
segments (Fig.
1). The pressure drop equation is used for all segments except the first segment. A constraint equation is applied for this
segment, which can specify either the rate of any phase or a bottomhole pressure (BHP). Details on the discretized well
equations and their coupling to the reservoir equations are given in Livescu et al. (2008).

Results
We now apply our coupled numerical model to investigate several cases involving nonisothermal multiphase flow in
wellbores. In all cases, Stanfords General Purpose Research Simulator (GPRS) is used for the simulations.

Validation against field data. The first example represents a validation of the numerical model against field data published
by Hasan and Kabir (2007). Previously (Livescu et al., 2008), we reported a comparison of our numerical model against a
simpler field example (Hasan and Kabir, 1994). In that case, we compared the temperature profile along a vertical well 5355 ft
long perforated at the bottom. The reservoir temperature profile was geothermal. In the current example, we consider an
offshore system involving a 4 in wellbore with two sections. The first section extends vertically to a depth of 4250 ft, while the
second section is inclined at an angle of 39 from the vertical. The bottom of the well is situated at a depth of 14600 ft (Fig. 2).
The water depth is 2800 ft; the surface water temperature is 54F and the seabed temperature is assumed to be 30F. The
formation temperature at the bottom of the well is 164F. Thus, the exterior temperature is not monotonic.
Wellbore flow in this case involves all three phases. Specified data include the pressure at the bottom of the well (4201
psia), wellhead pressure (305 psia), wellhead temperature (70F), and superficial velocities (at the wellhead) of the gas, oil and
water phases (V
sg
= 7.157 ft/sec, V
so
= 1.613 ft/sec, V
sw
= 0.73 ft/sec). The overall heat transfer coefficient near the wellhead is
given by Hasan and Kabir (2007) as U
to
= 142.344 btu/(day ft
2
F). For the formation, we use an overall heat transfer
coefficient U
to
= 400 btu/(day ft
2
F). The well is discretized in 293 segments. Given the large gas flow rate, the drift-flux
model is expected to be more applicable than a homogeneous model with no slip between phases. The system is essentially at
steady state, which we were able to represent through appropriate specification of reservoir model properties and initial
reservoir saturation and pressure.
The comparisons between our numerical results using the drift-flux model (labeled GPRS), the field data (measured),
and the numerical results for the homogeneous model (HO) are shown in Fig. 3 and Fig. 4 for the pressure and temperature
profiles, respectively. In Fig. 4, the reservoir temperature (reservoir) is also shown. The pressure computed using the drift-
flux model is seen to closely match the field data. The homogeneous model does not provide as close a match for pressure.
However, the temperature profile agrees well with the field data for both models. The largest difference in temperature
predictions is close to the bottom of the well, where the pressure gradient difference is also largest. It is also of interest that in
4 SPE 113215
the case of the homogeneous model the superficial velocities obtained at the wellhead are different from the values listed by
Hasan and Kabir (all phases flow with the same velocity), while they are within 5% if the drift-flux model is used. These
results, along with the comparison in our previous paper, provide a degree of validation for our nonisothermal wellbore model
for cases of practical interest.

Multilateral well in a complex reservoir. The second example involves the simulation of a much more complex case. For the
reservoir, we use a portion of the Stanford VI geological model (Castro, 2007). The grid size is 751008 (a total of 60000
cells), and the cell dimensions are x = 80 ft, y = 80 ft and z = 15 ft. The permeability fields for the 8 layers are shown in
Fig. 5. The well is a producer and has three branches, as shown in Fig. 6. The first branch is vertical and has 8 segments. The
other two branches are horizontal and each is 6000 ft in length. Note that the three branches are in different planes. The well
has 4 perforations, one at each end of the two horizontal branches. The top of the reservoir is located at an average depth of
8800 ft. The initial pressure in the reservoir is 5080 psia at a depth of 7800 ft. There is no water in the system and initially
there is no free gas, though gas appears as the pressure drops below the bubble point.
For this example we consider the effect of the number of well segments on the simulation results. Thus we consider two
different cases, in which the horizontal branches have 150 and 300 segments each. The reservoir model is the same in both
cases. The flow is two-phase (oil-gas) and the drift-flux model is used. Note that, although only 4 segments are open to flow
from the reservoir, each segment has a thermal connection with a reservoir block because of the heat loss term in the energy
equation. We compare the pressure (Fig. 7), temperature (Fig. 8) and gas in-situ volume fraction (Fig. 9) along the well at t =
100 days for the two cases (which correspond to a total of 308 and 608 segments). Note that, for the 308 segment case, the
segment numbers are multiplied by 2 in order to show the two sets of results on the same plot. It is evident from the results that
the model gives very similar predictions for the two cases. This insensitivity to the number of well segments is encouraging
and suggests that the well model has essentially converged.

Two-phase (steam-water) with phase change in a complex well. The third example demonstrates the performance of the
well model in the case of water vaporization and condensation in the wellbore. We consider a two-phase steam-water system
with a complex well depicted in Fig. 10. The well has 5 branches and is represented using 100 segments. The first branch
(segments 1-20) is horizontal, the second branch (segments 21-40) is inclined at 45 from the vertical, the third branch
(segments 41-60) is vertical, the fourth branch (segments 61-80) is horizontal, and the fifth branch (segments 81-100) has both
horizontal and inclined (45) portions. The horizontal and vertical segments are 20 ft long and the inclined segments are 28.28
ft long. The wellbore diameter is 3 in. The well has 5 perforations, at segments 80 (the end of the lower horizontal branch), 85,
90, 95 and 100 (upper horizontal branch). The well produces under pressure control (with a constant well pressure of p
w
= 800
psia at segment 1) and the reservoir pressure decreases with time due to fluid withdrawal. The bubble point occurs at a
pressure of 1015 psia and a temperature of 545F. The overall heat transfer coefficient along the well is constant, U
to
= 500
btu/(day ft
2
F). The initial pressure at the top of the reservoir is 2000 psia and the gas-water contact is 200 ft below the top of
the reservoir. Thus, the perforated segments 95 and 100 are above the gas-water contact and all other perforated segments are
below this contact. Above the gas-water contact, the initial gas saturation is 0.1 and below it is 0.
Simulation results are shown for wellbore pressure (Fig. 11), temperature (Fig. 12) and volume fractions (Fig. 13) at t = 1
day. Hot water and steam enter the upper branch at different perforations. The fluid flows downward and the pressure
increases, reaching the bubble point pressure at segment 86. At this point the saturation curve is reached and condensation
occurs. The pressure then increases and the fluid enters the single-phase (water) region. In addition, hot water enters the well at
segment 80 of the lower horizontal branch. Due to decreasing wellbore pressure, this fluid (hot water) crosses the bubble point
at segment 28 and vaporizes. From this point to the first segment steam is also present.
In the two regions in which both phases are present (segments 1-28 and 86-100), the temperature is a function of pressure,
p
w
= p
sat
(T
w
), and can be interpolated from steam tables (Keenan et al., 1969). In the single-phase region the pressure and
temperature are independent. At later times, due to the decrease in reservoir pressure, the fluid entering the well is already
below the bubble point and both the condensation and vaporization points appear closer to the wellhead.

The examples presented in this section demonstrate the applicability of our nonisothermal wellbore model for challenging
cases. As demonstrated by the examples, the model is quite comprehensive and handles multi-branched wells, three-phase
flow, phase change, and fluid and heat exchange between the well and the reservoir.

Concluding Remarks
In this work we applied a recently developed coupled thermal reservoir-multiphase wellbore flow simulator to a number of
realistic cases. The thermal multiphase wellbore flow model was developed within a black-oil framework and determines
pressure, temperature and phase fractions along the wellbore as a function of time by solving mass and energy conservation
equations along with pressure drop relationships. Slip between the three fluid phases is captured using a drift-flux model.
In order to establish the model applicability for practical reservoir simulation, we presented a series of numerical results,
including validation against field data from Hasan and Kabir (2007) for a long offshore well with both vertical and inclined
sections, simulation of a nonisothermal wellbore model in a fully-coupled simulation involving a multilateral well in a
SPE 113215 5
complex reservoir, and simulation of steam-water flow with vaporization and condensation in the wellbore. This series of
results clearly demonstrates the broad capabilities of the coupled modeling procedure.
In the future we plan to extend the current formulation to treat fully compositional systems and to consider the
implementation of a semi-analytical procedure, which makes use of our previously developed analytical solution for
temperature in the wellbore. This may provide a fairly straightforward means for extending isothermal wellbore models to
handle thermal effects.

Acknowledgments
We are grateful to Shell Oil Company for providing funding for this work. We thank M.A. Cardoso for providing us with the
reservoir model used in the second example.

Nomenclature
C
0
= distribution coefficient in the drift-flux model, dimensionless
g = gravity component along the well, ft/sec
2
H
p
= enthalpy of phase p, btu/lbm
m
cp
= source/sink inflow term of component c in phase p through each perforation, lbm/day
m
H
= thermal source/sink term through each perforation, btu/(day ft)
p
p
= reservoir pressure of phase p, psia
p
w
= wellbore pressure, psia
Q
loss
= heat loss to the surroundings, btu/(day ft)
r
in
= internal radius of the well, ft
t = time, days
T = reservoir temperature, F
T
w
= wellbore temperature, F
U
p
= internal energy of phase p, btu/lbm
U
to
= overall heat transfer coefficient, btu/(day ft
2
F)
V
d
= drift velocity of gas in liquid, ft/sec
V
m
= mixture velocity, ft/sec
V
p
= velocity of phase p, ft/sec
V
sp
= superficial velocity of phase p, ft/sec
W = well index, (cP rb)/(day psi)
z = wellbore curvilinear coordinate, ft

p
= in-situ volume fraction of phase p, dimensionless

cp
= molar fraction of component c in phase p, dimensionless

p
= mobility of phase p, ft
2
/(cP rb)

p
= density of phase p, lbm/ft
3


Subscripts
a = acceleration
c = component (gas, oil, water)
f = frictional
h = hydrostatic
p = phase (gas, oil, water)
s = superficial

References
1. Aziz, K. and Setari, A.: Petroleum Reservoir Simulation, Blitzprint Ltd., Calgary, 2002.
2. Aziz, K. and Wong, T.W.: Considerations in the Development of Multipurpose Reservoir Simulation Models, Proceedings of the 1st
and the 2nd International Forum on Reservoir Simulation, Alpbach, Austria, Sept. 12-16, 1988 and Sept. 4-8, 1989.
3. Cao, H.: Development of Techniques for General Purpose Simulators, Ph.D. dissertation, Stanford University, 2002.
4. Castro., S.A.: A Probabilistic Approach to Jointly Integrated 3D/4D Seismic Production Data and Geological Information for Building
Reservoir Models, Ph.D. dissertation, Stanford University, 2007.
5. Farouq Ali, S.M.: A Comprehensive Wellbore Steam/Water Flow Model for Steam Injection and Geothermal Applications, SPE
Paper 7966-PA, SPE J. 21 (5): 527-534, 1981.
6. Farouq Ali, S.M. and Abou-Kassem, J.: Simulation of Thermal Recovery Processes, Proceedings at the 1st and the 2nd International
Forum on Reservoir Simulation, Alpbach, Austria, Sept. 12-16, 1988 and Sept. 4-8, 1989.
7. Hasan, A.R. and Kabir, C.S.: Aspects of Wellbore Heat Transfer During Two-Phase Flow, SPE Paper 22948-PA, SPE Prod. & Fac. 9
(3): 211-216, Aug. 1994.
8. Hasan, A.R. and Kabir, C.S.: A Simplified Model for Oil/Water Flow in Vertical and Deviated Wellbores, SPE Paper 54131-PA, SPE
Prod. & Fac. 14 (1): 56-62, Feb. 1999.
6 SPE 113215
9. Hasan, A.R. and Kabir, C.S.: Fluid Flow and Heat Transfer in Wellbores, Society of Petroleum Engineers, Richardson, TX, 2002.
10. Hasan, A.R., Kabir, C.S. and Wang, X.: A Robust Steady-State Model for Flowing-Fluid Temperature in Complex Wells, SPE Paper
109765, SPE Annual Technical Conference and Exhibition, Nov. 11-14, 2007, Anaheim, California, U.S.A.
11. Holmes, J.A., Barkve, T. and Lund, O.: Application of a Multisegment Well Model to Simulate Flow in Advanced Wells, SPE Paper
50646-MS, European Petroleum Conference, Oct. 20-22, 1998, The Hague, Netherlands.
12. Jiang, Y.: A Flexible Computational Framework for Efficient Integrated Simulation of Advanced Wells and Unstructured Reservoir
Models, Ph.D. dissertation, Stanford University, 2007.
13. Keenan, J.H., Keyes, F.G., Hill, P.G. and Moore, J.G., Steam Tables, John Wiley & Sons, 1969.
14. Livescu, S., Durlofsky, L.J., Aziz, K. and Ginestra, J.-C.: A New Model for Simulating Nonisothermal Multiphase Flow in Wellbores
and Pipes, in review (available at http://pangea.stanford.edu/ERE/research/suprihw/publications/).
15. Oddie, G., Shi, H., Durlofsky, L.J., Aziz, K., Pfeffer B. and Holmes, J.A.: Experimental Study of Two and Three Phase Flows in Large
Diameter Inclined Pipes, Int. J. Multiphase Flow 29: 527-558, 2003.
16. Pourafshary, P., Varavei, A., Sepehrnoori, K. and Podio, A.L.: A Compositional Wellbore/Reservoir Simulator to Model Multiphase
Flow and Temperature Distribution, in review.
17. Ramey, Jr., H.J.: Wellbore Heat Transmission, SPE Paper 96-PA, Transactions of the Society of Petroleum Engineers of AIME 225
(4): 427-435, 1962.
18. Satter, A.: Heat Losses During Flow of Steam Down a Wellbore, SPE Paper 1071-PA, J. Pet. Tech. 17 (7): 845-&, 1965.
19. Sharma, Y., Shoham O. and Brill, J.P.: Simulation of Downhole Heater Phenomena in a Production of Wellbore Fluids, SPE Paper
16904-PA, SPE Prod. Eng. 4 (3): 309-312, Aug. 1989.
20. Shi, H., Holmes, J.A., Durlofsky, L.J., Aziz, K., Diaz, L.R., Alkaya B. and Oddie, G.: Drift-Flux Modeling of Two-Phase Flow in
Wellbores, SPE Paper 84228-PA, SPE J. 10 (1): 24-33, March 2005a.
21. Shi, H., Holmes, J.A., Diaz, L.R., Durlofsky L.J. and Aziz, K.: Drift-Flux Parameters for Three-Phase Steady-State Flow in
Wellbores, SPE Paper 89836-PA, SPE J. 10 (2): 130-137, June 2005b.
22. Stone, T.W., Edmunds, N.R. and Kristoff, B.J.: A Comprehensive Wellbore/Reservoir Simulator, SPE Paper 18419, SPE Symposium
on Reservoir Simulation, Feb. 6-8, 1989, Houston, Texas.
23. Stone, T.W., Bennett, J., Law, D.H.-S. and Holmes, J.A.: Thermal Simulation with Multisegment Wells, SPE Paper 78131-PA, SPE
Res. Eval. Eng. 5 (3): 206-218, June 2002.
24. Zuber N. and Findlay J.A.: Average Volumetric Concentration in 2-Phase Flow Systems, J. Heat Transfer 87 (4): 453-& 1965.


































SPE 113215 7



z
N
s
V
sp,i
, V
m,i
i
w
, T
i
w
p

p,i i+1
1
i1
i


Fig. 1 Schematic of well discretization.

(b)
o
F
54
o
F
30
o
F
39
o
water
2
8
0
0
1
4
5
0
1
0
3
5
0
z
T
(a)
164


Fig. 2 Wellbore schematics for the field example: (a) well and (b) reservoir temperature.

0
500
1000
1500
2000
2500
3000
3500
4000
4500
0 2000 4000 6000 8000 10000120001400016000
p
w
,

p
s
i
a
z, ft
GPRS
measured
HO


Fig. 3 Pressure variation along the well for the field example.

8 SPE 113215
20
40
60
80
100
120
140
160
180
0 2000 4000 6000 8000 10000 12000 14000 16000
T
w
,

F
z, ft
GPRS
measured
reservoir
HO


Fig. 4 Temperature profile along the well for the field example.



Fig. 5 Permeability field for each layer for the second example (extracted from the Stanford VI model).

(a)
z
y
x
vertical branch (1 8)
upper horizontal branch (9 308)
lower horizontal branch (309 608)
8 9 10 157 158 308 307 160 159
2 309 310 457 458 608 459 460 607
1
(b)


Fig. 6 Multilateral well schematics: (a) orientation of the well in 3D and (b) segment numbering.

SPE 113215 9
2000
2500
3000
3500
4000
4500
5000
5500
6000
6500
0 100 200 300 400 500 600 700
p
w
,

p
s
i
a
Segment number
p
bp
308 segments
608 segments


Fig. 7 Pressure variation along the multilateral well.

70
80
90
100
110
120
130
140
0 100 200 300 400 500 600 700
T
w
,

F
Segment number
308 segments
608 segments


Fig. 8 Temperature profile along the multilateral well.

10 SPE 113215
0.5
0.55
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 100 200 300 400 500 600 700

v
Segment number
308 segments
608 segments


Fig. 9 Gas volume fraction along the multilateral well.

96
45
o
45
o
z
x
1 20
21
22
39
40
41
60 61 80
81 85
86
100
95


Fig. 10 Well schematic for the phase change example.

800
850
900
950
1000
1050
1100
1150
1200
1250
0 10 20 30 40 50 60 70 80 90 100
p
w
,

p
s
i
a
Segment number
p
bp


Fig. 11 Pressure variation along the well for the phase change example.
SPE 113215 11

515
520
525
530
535
540
545
550
555
560
0 10 20 30 40 50 60 70 80 90 100
T
w
,

F
Segment number


Fig. 12 - Temperature profile along the well for the phase change example.

0
0.2
0.4
0.6
0.8
1
1.2
0 10 20 30 40 50 60 70 80 90 100

g
,

w
Segment number

w


Fig. 13 Gas and water volume fractions along the well for the phase change example.

You might also like