Académique Documents
Professionnel Documents
Culture Documents
~; + ()~·
Copyright 1995 Society of Petroleum Engineers
= 0................................. (11)
Original SPE manuscript received for review Oct. 31, 1990. Revised manuscript received
June 17, 1994. Paper accepted for publication July 25, 1994. Paper (SPE 21269) first pres-
ented at the 1990 SPE Eastern Regional Meeting held in Columbus. Oct. 31-Nov. 2. Substitution of Eq. 11 into Eq. 9 simplifies Eq. 9 further such that
;,. 8
azp a (;.g)
ap 1 ap
arz + 88 or B8 or + f';,.g or
psi, the uniqueness and independency of the proposed decline curves
to the pressure volume temperature (PVT) data become attenuated.
Fig. 1 displays the proposed decline curve generated at different ini-
tial formation pressures for Pwlp; = 0.5. It is clear that up to 2,000 psi
o2p a (;.w) ap 1 op op . ..... (15) the proposed decline curve displays unique characteristics and, as ini-
+ Aw arz + Bw or Bw or + rAw or = rpc,iii.
tial pressure increases, the uniqueness of the decline curve is increas-
Using the chain rule, one can write ingly compromised. This is expected because at relatively low pres-
sures the product of the gas viscosity and compressibility factor is
:r(~:)~~ =~(~:)(?,f. ········ ············· constant. As the pressure becomes higher, the constant nature of this
(16) functional group is compromised, and the uniqueness of the decline
curves start to deteriorate.
With the assumption of small pressure gradients and neglecting 1\vo separate production decline curves are generated for gas and
the squared terms, Eq. 15 reduces to water flow rates. Table 1 summarizes the parameters used in
constructing these production decline curves. To generate the pro-
a2p 1 op rpc,ap duction decline curves at different Pwlp; ratios, the initial pressure
or2 + r or = T;iii· (17)
was kept constant and the wellbore pressure was changed accord-
Eq. 17 can also be written as ingly. Fig. 2 displays the gas flow rate decline curves generated at
different Pwlp; ratios for different reservoir extent. Fig. 3 shows the
Discussion of Results
In Eqs. 17 and 18, ;., is defined as
Sensitivity Analysis. The performance of the proposed decline
A., = A8 + A.w. . .........•...................... (19) curves was investigated under a wide range of reservoir properties.
At this point, gas and water transport equations describing two- From different reservoir parameters, dimensionless gas flow rate and
dimensionless time groups were calculated at various times and were
phase flow conditions in a wet gas reservoir have been reduced to
marked over the proposed decline curves. As it is clearly seen in Figs.
a single expression, which is analogous to the single-phase equation
4 through 6, in every case a good match was achieved. These verifi-
used in well test analysis. Transforming Eq. 18 into dimensionless
cation tests were conducted at reo= 100 to save computer time. Fig.
form yields
7 shows different relative permeability characteristics used to investi-
1 a ( al:lp v) al:lp v (20)
rv orv rn orv = otn TABLE 1-RESERVOIR PARAMETERS USED TO GENERATE
PROPOSED PRODUCTION DECLINE CURVES*
with customary dimensionless groups defined as
Initial pressure, psi 1000
rn = :w' ...................................... (21) Bottomhole pressure, psi
Well radius. ft
500
0.5
(22) Permeability, md 1
Porosity,% 10
Initial gas saturation, % 40
(23) Formation thickness, ft 10
Gas gravity (air= 1.0) 0.6
Reservoir temperature, OR 530
..................... (24) Water viscosity, cp 0.9707
"For the relative permeability characteristics see Set A of Fig. 7.
100 1000 10000 1e+05 18+06 1e+07 1e+08 100 1000 10000 1e+05 1e+06 1e+07 11+08
Dimensionless nme Dimensionless nme
Fig. 4-Verification of the decline curves for a wet gas sand with Fig. 5-Verification of the decline curves for a wet gas sand with
following characteristics: SG=0.6, p,=1000 psi, Pw=500 psi, following characteristics: SG=0.6, p;=1000 psi, Pw=500 psi,
Sgt=40o/o, f= 10%, k=0.01 md, h= 10ft, rw= 0.5, re =50 ft. Sgt =50%, f=15o/o, k=0.15 md, h= 10ft, rw=0.5, r, =50 ft.
gate the performance of the proposed decline curves when relative house numerical simulator. To test the uniqueness and applicability
permeability characteristics of the reservoir are changed. The results of the proposed decline curves further, more tests were conducted
of this investigation are shown in Fig. 8. The data points shown in Fig. using two different sets of data in two different numerical models
8 correspond to the same real times. From Fig. 8 it is clear that the that were not instrumental in the construction of the decline curves.
characteristic nature of the decline curve is preserved successfully These tests were designed to discover whether the decline curves
against a variety of relative permeability characteristics. presented in this paper were biased toward the numerical model
used in their construction. The numerical model used in the first test
Tests Against Single-Phase Decline Curves. To demonstrate ex- (Test l) was a commercial simulator that uses IMPES procedure to
plicitly the need for a set of two-phase decline curves for conven- solve the multiphase flow problems. The in-house numerical model
tional gas reservoirs, a test was conducted in which the performance used in the generation of the decline curves is a fully implicit model.
of a single-phase production decline curve is compared with that of Furthermore, the commercial numerical model uses Cartesian coor-
a two-phase production decline curve in predicting the gas flow dinates, while the in-house model uses radial coordinates. Table 2
rates from a particular formation. The parameters of Table l were represents the relative permeability data used in this test. Formation
used to construct these curves, and Fig. 9 displays the results of this characteristics, which were input during the same test, are presented
comparison. As expected, the single-phase production decline in Table 3. Fig. 10 shows the agreement between the simulated data
curve predicts a higher gas flow rate than the two-phase production and predictions from the two-phase decline curves for the gas pro-
decline curve. The difference in the two predictions will vary ac- duction rate. Fig. 11 shows, for the same test, the agreement be-
cording to the relative permeability characteristics of the formation. tween the simulated data and the predicted water flow rates. Note
Obviously, as the mobility of the water in the formation increases, that the anomalous behavior in the simulated data in Fig. ll is
the prediction of the single-phase production decline curve becomes caused by the relatively long timesteps used during the simulation.
more unrealistic. The gas composition and calculation of dimen- In a separate test (Test 2), a new set of data was used on another
sionless groups are shown in Appendix A. commercial model. 7 Relative permeability data as well as formation
characteristics for this test are shown in Tables 2 and 3, respectively.
Tests Against Different Numerical Simulators. The proposed de-
cline curves were generated and subsequently tested using an in- In this test, two separate approaches were examined to demonstrate
the different usages of the proposed curves. First a type-curve
matching approach is tested. Using this technique, one may be able
to use the production data for reservoir characterization purposes.
Plotting the gas production data on the same scale as the type curves
and overlapping for a match, a match point was obtained (Fig. 12).
Usingthismatchpoint(qv8 =0.3, tv= 1 x 108 andq8 =0.7 MMscf/
D, t = 20,000 days) the permeability of the reservoir is calculated,
with Eq. 25, to be 0.74 md, while the actual permeability that was
put into the simulator was 0.75 md. The porosity calculated from the
type-curve matching procedure using Eq. 23 is 27 .8%, and the input
porosity to the simulator was 25%. This shows the ability of pro-
100 1000 10000 1e+OS 18+06 1e+07 1e+08
posed curves in accurately characterizing gas reservoirs producing
Dimensionless nme under two-phase (gas and water) flow conditions.
On a second approach in Test 2, it was attempted to predict the
Fig. &-Verification of the decline curves for a wet gas sand with performance of a well producing from the center of the reservoir
following characteristics: SG = 0.6, PI= 2000 psi, Pw = 1000 psi, with known reservoir chracteristics. Fig. 13 shows production data
Sgt = 60%, f= 30%, k= 1.0 md, h =40ft, rw = 0.5, r1 =50 ft. simulated by the simulator vs. those of decline curve prediction.
...
...
0.7
... ...
lla.t
... u
u
u
... •..
.. Set D
Set B
Fig. 7-Relative permeability characteristics used in testing the performance of the proposed decline curves.
_ _...,_(SIIB)
,__. -lnl C:U.W (Sol A)
•• ,.._............,_(SIIC)
_ __,_(SliD)
e 40ooo • /Single-Phase Decline CuMI Prediction, [Caner (1981))
0
~ ~
.. -e---·~·. . .---.. . . -.. . ~
•• _ 8 l
a:
'---.
1
• • • • • • •
Sgl=60% (Twool'hase)
• ••
0~----~-------L-------L-·----~------~
200 400 800 800 1000
llma(days)
Fig. 9-Gas flow rate predictions obtained by single-phase and
Fig. &-Sensitivity of the proposed decline curves to relative two-phase decline curves for a fixed set of reservoir parameters
permeability characteristics. but with different initial gas saturations.
. 0.67
0.69
0.734
0.8
0.074
0.012
0.425
0.485
0.245
0.24
Jo.a r . 0.92 0.95 0 0.875 0.018
1
1--·--1------l
1.000
Match Point
1•...,--'f-- : .•
10,000.0 100,000.0
Fig. 12-Reservoir characterization by type-curve matching procedure using the proposed type curves.
~
•
Decline Curve Predictions
-....;.. 300
.S!
~
~ 200
0
it
~ 100
(!)
O+--+--~-r~~4--+--+--+--~-r~--+--+--+--+--r--r-1--+-~
• •
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800 2,000
Time(Days)
Fig. 13-Predicted gas flow rates from a commercial simulatorS and from the proposed decline curves.
TABLE 4-RESERVOIR AND FLUID PROPERTIES USED FOR Sl Metric Conversion Factors
CALCULATIONS IN APPENDIX A bbl X 1.589873 E- Ol = m3
cp X 1.0* E + 00 = Pa · s
fJ/, psi 1991
°F (0 f - 32)/1.8 = °C
rw. ft 0.5
ft X3.048* E-01 =m
Sgi 0.3 ft3 X 2.831685 E-02 = m3
Pwf, psi 400 in. X2.54* E+OO =em
h,ft 85.6 psi x6.894757 E+OO =kPa
Gas gravity 0.6 0
R 0 R/1.8 = °K
T. OR 530
·conversion factor is exact. SPEFE
Jl.w. cp 1.0
Jl.g,CP 0.01622
Zj 0.7995 Shahab Mohaghegh is an assistant professor of petroleum and
k,g 0.1024 natural gas engineering at West Virginia U.• Morgantown. His re-
k,w 0.1296 search interests include reseNoir characterization, application
k,md 22.4 of petroleum engineering technology to environmental pro-
cesses. and application of machine intelligence to energy and
41 0.157 environmental sciences. He holds BS and MS degrees In natural
Cw. psi-1 5x1o-s gas engineering from Texas A&l U. and a PhD in petroleum and
natural gas engineering from Pennsylvania State U. H.llkln 811·
A = kkrg = (22.4)(0.1024) = gesu is an assistant professor of petroleum and natural gas engi-
g #g 0.01622 141 .415 ' neering at West Virginia U. Previously, he worked for the Turkish
Petroleum Corp. and taught at South Dakota School of Mines
and Technology. He holds a BS degree from Middle East Technl·
1 = kkrw = (22.4)(0.1296) = cal U. in petroleum engineering. an MS degree from the Colora-
1\.w J.lw 1.0 2·903 '
do School of Mines in chemical and petroleum refining engl·
nearing, and a PhD degree In petroleum and natural gas
engineering from Pennsylvania State U. He was a member of
the program committee for the 1993 and 1991 SPE Eastern Re-
gional Meetings. Turgay Ertekln is the Quentin E. and Louise L.
Wood Professor and chairman of the Petroleum & Natural Gas
Engineering Section at Pennsylvania State U., where he earned
and c1 = (0.30)(
19~ 1 ) + {0.7){5 X 10--{;) = 1.542 X 10-4 a PhD degree in petroleum and natural gas engineering. He
holds BS and MS degrees In petroleum engineering from Middle
.................... {A-1) East Technical U. Ertekin was 1993-94 Executive Editor for SPE
Using Eq. 25, dimensionless rate is calculated Formation Evaluation, a 199Q-93 ABET Program Evaluator,
a199Q-92 Review Chairman, and a 1984-89 Technical Editor.
(7.11 X 105)(10)(0.7995)(530)
qDg = {141.415){85.6){1991){1991-400) = 0 ·078 •
and dimensionless time is calculated using Bq. 23,
= {2.367 X 10- 4)(144.318)(0.9)(24) = 1351819
tv {0.157){1.542 X 10- 4){0.25) ·
{A-2) Mohaghegh Bilgesu Ertekln