Professional Documents
Culture Documents
Three-Dimensional Black-Oil
Reservoir Simulation Problem
Aziz s. Odeh, SPE, Mobil Research and Development Corp.
Summary
A comparison of solutions to a three-dimensional 1. Stratification and reservoir properties are given in
black-oil reservoir simulation problem is presented. Fig. 2. The reservoir is initially undersaturated. A gas
The test of the problem and a brief description of the injection well is located at Grid Point (1, 1), and a
seven simulators used in the study are given. producing well is located at Grid Point (10, 10).
Pertinent data and constraints are given in Table 1.
Introduction PVT properties and relative permeabilities are given
Seven companies participated in a reservoir in Tables 2 and 3. The participants were asked to
simulation project to compare the results obtained by make the runs and report the results described below.
different black-oil simulators. The companies were
chosen to give a good cross section of the solution Runs To Be Made
methods used in the industry. The participants were Case 1
Amoco Production Co., Computer Modelling Group
of Calgary (CMG), Exxon Production Research Co., Let the bubble-point (saturation) pressure be con-
Intercomp Resource Development and Engineering stant with a value equal to the original value.
Inc., Mobil Research and Development Corp., Shell Case 2
Development Co., and Scientific Software Corp.
(SSC). The paper presents the text of the problem, a Let the saturation pressure vary with gas
comparison of results in graphical form, and a brief saturation - i.e., this is a variable saturation-pressure
description of each model. The descriptions were case. The PVT lines at pressures above the calculated
supplied by the participants. saturation pressures are parallel to the original line.
A variety of computers was used. Amoco used
IBM 3033, IBM 370/168, and Amdahl V/6. CMG Results To Be Reported
used Honeywell 6000 DPS, and Exxon used Amdahl The following results are to be reported.
470/V5 and IBM 370/168. Intercomp used Cray-l I.Plots of:
and Harris17. Mobil and SSC used CDC Cyber 175, a. Oil rate vs. time.
and Shell used Univac 111012C Level 36. The b. GOR vs. time.
number of time steps and the central processor times 2.Report annually and at abandonment:
varied considerably. Those interested in the actual a. The pressures of the cell where the injector
values should contact the individual companies. and producer are located. *
Except for Shell, all the participants used single- b. Gas saturation at Grid Points (1, 1, 1), (1, 1,
point upstream mobility weighting. Shell used two 2), (1, 1,3), (10, 1, 1), (10, 1,2), (10, 1,3), (10,
points upstream. Constraints and data are given in 10, 1), (10, 10,2), and (10, 10,3).
the text. 3.Report at the end of 8 years:
a. Tables of gas saturation.
Statement of the Problem b. Tables of cell pressures. *
Areal and cross-section views of the reservoir are c. Tables of saturation pressures for the variable
given in Figs. 1 and 2. The grid system is given in Fig. saturations-pressure case. *
0149·2136/81/0001·9723$00.25 'Report of all pressures referred to a depth of 8,325 It or at the center of the
Copyright 1981 Society of Petroleum Engineers of AIME respective blocks. If both are available, report both.
JANUARY 1981 13
GAS INJECTION OIL PRODUCTION
100 MM SCFI D WELL
.,. H. FT. KX
K
Ky (LINK) Sw So
8325 FT.
LAYER 1 .3 20 500 500 . 12 .88 8335 FT •
50
25
,.
Ci=±=ti=.t,=~~_~~~~ I : J: 1 2 3 4 5 8 7 8 II 10
8425 FT.
1000 FT.
18
-MOBIL
16 - SHELL
- AMOCO
14
- INTERCOMP
~ 12 - EXXON
IXI
....
(/) ----- SSC
(II) 10
0
....
W- 8
!i
a:
...J
0 6
EXXON AGREES WITH MOBIL BETWEEN 3-7 YEARS
4
EXXON AGREES WITH INTERCOMP BETWEEN 8-10 YEARS
AMOCO AGREES WITH SHELL AFTER 7 YEARS
2
CMG AGREES WITH MOBIL THROUGH 7 YEARS AND
WITH INTERCOMP BETWEEN 8-10 YEARS
0
1 2 3 4 5 6 7 8 9 10 TIME, YEARS
,.
16 -MOBIL
,. - - SHELL
_ AMOCO
...,m 12 - INTERCQMP
'".. 10 - - EXXON
<.> - .-. ~ sse
'"
"r£
0
•
" EXXON AGREES WITH MOBIL UP TO 3 YEARS
6 sse AGREES WITH MOBIL BETWEEN 2-7 YEARS
0
1 2 3
• 5 6 7
Fig. 4 - Case' -GOR vs. time.
• 9 10 TIME , YEARS
INTERCOMP,
SHEll,
5500 AMOCO
- MOBIL - INTERCOMP
- SHELL _ EXXON
- AMOCO _._ . sse
.'"
~
ui
sse AGREES WITH AMOCO BETWEEN 3-5 YEARS
II:
::>
'"'"
.
W
II:
3500
MOBil, INTERCOMI;P:::::::::::;::::~;·~-~.~-;;.~_~.;;_;:;.Z
30
- M O BIL
- SHELL, AMOCO
- INTERCOMP
-----SSC
I-
Z
w 20
~
o
a:
w
Q. INTERCOMP, SHELL, AMOCO
Z
o
~a:
::::l
ti:
(f) 10
EXXON AGREES WITH SHELL, AMOCO THROUGH 6 YEARS,
(f) AND WITH INTERCOMP BETWEEN 7-10 YEARS
<C
(!) CMG AGREES WITH MOBIL THROUGH 3 YEARS, AND
WITH INTERCOMP BETWEEN 4-10 YEARS
SSC AGREES WITH SHELL BETWEEN 4-9 YEARS
o+-----~----~----~----~--~~--~----~----~----_+----~
o 1 2 3 456 7 8 9 10
TIME, YEARS
Fig. 6 - Case 1 - gas saturation vs. time for producing wel l Cell 10, 10,3.
JANUARY 1981 17
7500
-MOBIL
- SHELL
- INTERCOMP
- EXXON
6500 _......... CMG
...0;
.
w
AMOCO AGREES WITH SHELL UP TO 6 YEARS,
AND WITH EXXON AFTER 6 YEARS
..
w
a: CMG AGREES WITH
MOBIL UP TO 4 YEARS
WITH MOBIL BETWEEN 3-4 YEARS, AND
WITH EXXON BETWEEN 6-10 YEARS
MOBIL-INTERCOMP~-...
4500
MOBIL-INTERCOMP::;~~~::;;:
3900±-__~__~__~__~__~__~__~__~__~__~.
o 1 2 3 4 5 6 7 8 9 10 TIME , YEARS
Fig. 7 -Case 1 - pressurevs. time for injection well Cell 1, 1, 1.
4500
_.
.-.
-. .
. _._. _.
.
4000 ._.
..'"
:!
W
-MOBIL
- SHELL
- .-.-
a: - AMOCO
:>
'"'"a:w
- INTERCOMP
. 3500
- EXXON
._._. SSC
.-.
CMG AGREES WITH AMOCO
I~M
3000
1,1 2,2 3 ,3 4 ,4 5,5 6,6 7,7 8,8 9,9 10,10
GRID POINT LOCATION
Fig. 8 -Case 1 - pressure vs. grid-poinllocation, time= 8 years, top layer.
. .1.... sf
1
+ M
S
A'I ' E= M -MOBIL
50 SSC = E
......t......" ...., .•..+ ,... 1=I S -
-
SHELL
AMOCO
SSC , E=A - INTERCOMP
- EXXON
....
z ·_ · _ · SSC
w ........... CMG
0
.
0:
W
Z E=I
0
~
0:
45 CMG~ A ·············:1
::> E= M
t(
<II
<II sse= A
50 I, E:A
CMG: M
40 -MOBIL
SSC = A - SHELL
- AMOCO
....
Z - INTERCOMP
W
o - EXXON
.
0:
W 30
Z
···T····
E= CMG
· _·_ · SSC
..........• CMG slc,
CMG=
o
~ GAS SAT. < .005 AT:
0: 1) 4 ,4- SHELL, EXXON, AMOCO
'c::> 20 2) 5 ,5 -ALL PARTICIPANTS
<II 3) 6 ,6- ALL PARTICIPANTS EXCEPT MOBIL,
<II CMG
4 ) 7 .7_SHELL, AMOCO
"" 10
5} 8,8- SHELL
A, E, CMG= M
I, Ef A
CjG= M E= CjG I r SSC, CMG: M
0 ~~~~'~~"'~
1,1
" ""~"~'~1~~
2,2 3 ,3 4,4 5,5 6,6 7,7 8,8 9 ,9 10,10
GRID POINT LOCATION
Fig. 10 - Case 1 _ gas saturation vs. grid-point location, time=8 years, middle layer.
20
EXXON,AMOCO
18
SHELL = MOBIL - M O BIL
- SHELL
16
- AMOCO
_._ . SSC
14
0
"-
a:I
~
(JJ 12
(\')
... 10
0
w
~
II:
8
-I
(5
6
INTERCOMP AGREES WITH MOBIL UP T O 7 YEARS,
4 AND WITH AMOCO BETWEEN 7-10 YEARS
18
-MOBIL
- SHELL
16 - AMOCO
- EXXON
14 _. _ . SSC
III
t- 12
,
III
U.
(,)
III 10
'"
ri
0 8 AMOCO AGREES WITH SHELL BETWEEN 8 - 10 YEARS
" EXXON AGREES WITH SHELL UP TO 4 YEARS. AND
6
, WITH MOBil BETWEEN 5-8 YEARS
,,
4 AND WITH SHELL BETWEEN B-1 0 YEARS
6000
SHELL- -MOBil
AMOCO - SHELL
UP TO - AMOCO
4 YEARS
- EXXON
_. - . SSC
........... CMG
5000
~ INTERCOMP AGREES WITH SHELL UP TO
III
0. 4 YEARS, AND WITH MOBIL AFTER THAT
W CMG AGREES WITH MOBIL THROUGH
rr
::> 2 YEARS
III
III
W
rr
0.
4000
"-'-.
3000 +----+--~~--~--~----~---T--~~--~--~~~
o 1 2 3 4 5 6 7 8 9 10
TIME, YEARS
Fig. 13-Case 2 - pressure vs. time for producing well Cell 10, 10,3.
JANUARY 1981 21
which are designed specifically for optimal solution Mobil's All Purpose Reservoir Simulator (ALPURS)
of two-phase problems: one for water/oil (dissolved- ALPURS is a three-dimensional, three-phase,
gas content constant) problems and one for multiwell, black-oil reservoir simulator which uses a
gas/water (no oil phase) problems. BETA II has a strongly coupled, fully implicit method to solve
large variety of user-oriented features such as in- simultaneously for all unknowns. 5 The nonlinear
put/ output options and well control options. intercell flow equations and well-constraint
There is a one-equation implicit pressure for- equations are linearized and iterated to converge
mulation in which the equations are decoupled and using Newton-Raphson iteration. Linear equations
solved in this order: pressure, gas saturation, and are solved with block successive overrelaxation. A
water saturation. For the one-equation formulation, typical block is an x-z, y-z, or x-y reservoir slice,
there are options to solve for both saturations ex- which is solved by sparse elimination. The relaxation
plicitly (lMPES) and either or both saturations parameter is computed automatically using the
implicitly (sequential). There is an alternative two- power method and Rayleigh quotients. ALPURS
equation formulation in which the program solves accounts for reservoir heterogeneity, rock com-
implicitly for pressure and gas saturation and then pressibility, gravity, gas dissolved in both the oil and
solves for water saturation. Similar to the previous water phases, constant or variable bubble-point
formulation, water saturation can be treated either pressures, hysteresis in saturation-dependent data,
implicitly or explicitly. In any of the formulations, tubing string pressure drop, and flash surface
multiple outer iterations may be taken to account for separation. Modern concepts of well flow equations
the nonlinearity of the basic flow equations. are incorporated, including pseudo gas-potential
However, if no vaporization or resolution of gas is function, skin factor to account for damage or
occurring, only one or two iterations are required to improvement, non-Darcy flow effect, and flow
converge the nonlinearities adequately. restriction due to limited entry such as partial
The large systems of linear algebraic equations penetration.
may be solved by a variety of methods, anyone of
which may offer significant speed advantages on a
given problem: (1) direct solution by reduced band- SSCModel
width Gaussian elimination, (2) several forms of one- SSC's black-oil model employs an Adaptive Implicit
line, two-line, and planar successive overrelaxation Method (AIM). This technique, which recently was
(SOR), or (3) SIP. developed at SSC, seeks to achieve an optimum with
30
-MOBIL
- SHELL, AMOCO
- INTERCOMP
I-
_ ._ . SSC
Z
w
o 20
a:
w
Q.
Z
o
fi
a:
::>
t( EXXON AGREES WITH INTERCOMP BETWEEN
UJ 10 3-10 YEARS
UJ
< CMG AGREES WITH INTERCOMP
CJ
O +----+----+----+----~--_r----~---r----~--~--~
o 1 2 3 4 5 6 7 8 9 10
TIME, YEARS
Fig. 14 - Case 2 - gas saturation vs. ti me for produci ng well Ceil 10, 10, 3.
-MOBIL
- SHELL
- AMOCO
- INTERCOMP
- EXXON
6500 _ . _ . SSC
.......... CMG
... MOBIL-SHELL-EXXON'.
\
EXXON AGREES WITH AMOCO
'" BETWEEN 6-10 YEARS
:i
::>
5500 SSC AGREES WITH MOBIL
AFTER 6 YEARS
'"'"a:w
.. CMG AGREES WITH EXXON
BETWEEN 6-10 YEARS
4500
3900' ~--+---~--~--~--~--~--~--~--~---:;
o 1 2 3 4 5 6 7 8 9 10 TIME, YEARS
Fig. 15 -Case 2 - pressure vs. time for injection well CellI, I , 1.
L
4500
...'" 4000
-MOBIL
- SHELL
ui - AMOCO
a:
::> - INTERCOMP
'"'"a:w
. - EXXON
_._" SSC
3500 .......... CMG
3000~~-+~~+-~~~~~~~~~-r~~~~-+~~~~~
1,1 2 ,2 3 ,3 4,4 5,5 6,6 7,7 8,8 9,9 10,10
GRID POINT LOCATION
Fig. 16 - Case 2 - pressure vs. grid-point location, time=8years, top layer.
JANUARY 1981 2l
respect to stability, truncation errors, and computer variable substitution. The simulator is a general
costs. Typically, only a small fraction of the total purpose package offering three-dimensional capabil-
number of grid blocks during simulation experience ity in both Cartesian and cylindrical coordinates.
sufficiently large surges in pressure and/or saturation
to justify implicit treatment. When it is needed, Shell Development Model
implicit treatment may not be required in all phases The Shell reservoir simulation system operates an
or for long periods of time. Moreover, those cells IMPES mode or an implicit mode. There are three
requiring implicit treatment will change as the pseudocomponents: water, stock-tank oil, and
simulation proceeds. Consequently, a model offering separator gas. There are three phases: aqueous,
a fixed degree of implicitness to all cells for all time hydrocarbon liquid, and hydrocarbon vapor. The
steps is not always the most desirable. For example, a aqueous phase contains water. The hydrocarbon
fully implicit model, while ensuring stable answers, liquid and vapor phases can contain both oil and gas.
amounts to overkill in most of the cells most of the A fourth component is also available for modeling
time, while an IMPES model can cause under kill. polymer or carbon dioxide. There are several indirect
With AIM there is no problem of over- or under kill. and direct solution methods as a user option. Ad-
Various degrees of implicitness are invoked ditionally, two-point upstream weighting is used to
regionally or individually cell by cell- i.e; the calculate phase mobilities.
solution is advanced with adjacent cells having
different degrees of implicitness. As the calculations References
proceed, the degrees of implicitness locally and 1. Behie, A. and Vinsome, P.K.W.: "Block Iterative Methods for
dynamically shift as needed - all automatically. The Fully Implicit Reservoir Simulation," paper SPE 9303
whole idea is to apply the right amount of im- presented at the SPE 55th Annual Technical Conference and
Exhibition, Dallas, Sept. 21-24, 1980.
plicitness where and when needed and for only as 2. Au, A.D.K., Behie, A., Rubin, B., and Vinsome, P.K.W.:
long as needed. "Techniques for Fully Implicit Reservoir Simulation," paper
The simulator also provides a wide variety of user- SPE 9302 presented at the SPE 55th Annual Technical Con-
oriented features. For example, one can override ference and Exhibition, Dallas, Sept. 21-24, 1980.
AIM and operate in a fully implicit, partially im- 3. Spillette, A.G., Hillestad, J.G., and Stone, H.L.: "A High-
Stability Sequential Solution Approach to Reservoir
plicit, or an IMPES mode. Variable bubble-point Simulation," paper SPE 4542 presented at the SPE 48th
problems, such as that in Case 2, are handled by Annual Meeting, Las Vegas, Sept. 30-0ct. 3, 1973.
SSC:I
t
.............
A:~ 1~ S7 E
55
E: S .... ·1·- -MOBIL
CMG:A - SHELL
- AMOCO
I-
Z - INTERCOMP
W
0 - EXXON
IX:
W
Q.
Z 50
_....... ......
~.-
· _· _·SSC
··········· CMG
0
A:S
fi
IX:
::J
t:c
f/)
f/)
<
_ . .- _.. .,
CMG.L S
0 45 CMG:M
SSC : E
I . .- ssc: E
SSC, CMG: A 'f
1 I
CMG~ A + .............
. !
............
A, CMG : M
40
1, 1 2,2 3,3 4 ,4 5 ,5 6 ,6 7 ,7 8 ,8 9 ,9 10,1 0
GRID POINT LOCATION
Fig. 17 - Case 2 - gas saturation vs. grld·point location, time = 8 years, top layer.
50
CMG= A '.
CMG= E -MOBIL
- SHELL
... 40 E,CMG=M
zw - AMOCO
o - INTERCOMP
-.~.J
a:
w
Q.
- EXXON
_._.- SSC
Z
o 30 ........... CMG A=I
t
~
a:
::)
GAS SAT. < .005 AT:
ti
rJ) 1) 4,4 AND 5,5-SHELL, AMOCO,
rJ) 20 SSC,EXXON
ct
~ 2) 4,4-CMG
3) 6,6-SHELL, ssc
10 S, I, E, CMG= M
SSC,S =E rSSC,CMG=M
JANUARY 1981 25
Discussion of Comparison of Solutions (SPE 9741)
to a Three-Dimensional Black-Oil
Reservoir Simulation Problem
William Hurst, SPE, consultant
I read with interest the article by Aziz S. Odeh in the conformal mapping to simulate reservoir per-
Jan. 1981 issue of JPT (Pages 13-25). Although the formance. This was the first thing that occurred to
article is excellently illustrated with color-coded me to test the validity of the method.
curves for each of the contributors, I question what The results have been excellent for the transient
this article accomplishes. treatment of fluid flow for a well in an enclosed
The same physical data are given to seven circle, either at the center of the circle or offset in the
programming groups to see if they could reproduce enclosure. In Table 4-2 of the text, I start with the
the same results for the model under consideration. conformal rectangle, whereby I can reproduce
This applied even for their mesh grid arrangements Muskat's solution that applies to the offset well in a
shown in the paper in Figs. 1 and 2. circle.
I recommend to Odeh and other authors that they No one has made these comparisons, which should
reproduce in their programming procedure the be made available if we are to place validity on the
analytical solutions of Morris Muskat and mine that known methods of reservoir simulation published in
apply to reservoir performance. I have done this in the literature. '
my text, Reservoir Engineering and Conformal JPT
Mapping of Oil and Gas Fields in treating with
Original manuscript (SPE 9741) received in the Society of Petroleum 0149·2136/81/0003·9741$00.25
Engineers office Feb. 5,1981. Copyright 1981 Society of Petroleum Engineers of AIME