Professional Documents
Culture Documents
SPE
-ofmtma
SPE 19823
lis papar was prepared for presentation at the 64th Annual Technical Conference and ExhiblJon of the Society of Petrolebm Engineers hefd In Ban Antonio,TX, October S-1 1, 1SSS.
Iia papar was selected for preaen!ation by an SPE Program Committee following review of informationcontained in an abstract aubmlffad by the ink(s). Contents of the paper,
} preaentad, hava not bean reviewed by the Society of Petroleum Engineers ati are subject to correction by the author(a). The material, as presented, does not neceasarify reftesf
lY Positionof the ~etY of petroleum Engineers. its Officers.ormembers.Paparepresented at SPE mwfings are aubjact to publicationreview by Edifor161Cornmittwa of the Scdaty
PetrofeumEngineers.Permission10copy is restrictedto an ebalracfof notmorethan WI words. Illustrationsmay notbe coplad.The abarrasfahoutdc@ain conspicuousack~t
where and by whom the paper ia presented. Write Publications Manager, SPE, P.O. Box S3SS36, Richardson, TX 750SS-S9SS. Telex, 7S0SS9 SPEDAL.
In this paperl, two commercial simulators are utilized to Vogel [2] has proposed an empirical IPR equation that has
develop IPRs for horizontal wells producing from solution-gas been used in the industry successfully. Because flow into a hor-
drive reservoirs. The development parallele the work of Vogel izontal well is different from the flow into a vertical well, IPR
for vertical wells. First, a base case is considered with typi. equations developed for vertical wells should not be applied
cal fluid, rock and reservoir properties. Then, variations from to horizontal wells without verification, Since the analytical
the base case are investigated. Changes from the base case, calculations necessary to compute IPR.s from two phase flow
in fluid properties included variations in relative permeability theory are tedious, numerical simulation is used. Most com-
and PVT properties. Changes in reservoir properties included mercial black oil simulators do not include the feature of mak-
variations in drainage area, pay thickness, and absolute perme- ing complete IPR predictions for a reservoir.
ability. Changes in well propertied included variations in skin, In this paper, available commercial simulators are nev-
well location, and well length with respect to reservoir bound- ertheleae utilized to generate IPRs of a horizontal & pro-
itiea. The resulting IPRa “were made dlmenaionleee in order ducing from a solution-g~ drive reservoir. First, IPIts are
to compare their curvature, or the rate of change of oil pro- generated for a base case reservoir. Then, variation cf the
1---
duction rate with flowing bottom-hole pressure. These curves base case are examined. These -ariations cover a wide miige
were found to be sensitive to the stage of reservoir depletion. of fluid, reservoir and well characteristics. Fluid charact~:;rj
However, they were not affected significantly by chariges in the tics include crude oil PVT properties and relative perme.=%lity
fluid reservoir, or well properties. data. Reservoir characteristics include absolute yrmca.bfity,
An attempt wss made to develop a simple correlation to drainage area, and formation thickness. Wefl cL~racteristica
include skin, location, and length with respect to the reservoir
‘References and illustrations at end of paper
dimensions.
551
. .
INFLOW PERFORMANCE RELATIONSHIPS
2 FOR SOLUTION-GAS DRIVF Horizontal WELLS SPE 19823
—9.
90m.s
= 1.0 – 0.2~ - 0.8(~’)2 (1) I Horizontal Well Formulation
The connection between the horkkd well and the reservoir
where go is the oil flow rate, qoma=is the maximum flow rate block containing tke weti is based on th, assumptions that
corresponding to 100?!fopressure drawdown, pw~iB ?he flowing the flow is steady-state aiid radial, the fluid is in a single in-
bottom-hole pressure, and PR is the Bhut-in pressure or average compressible phase, and that the pressure head due to height
reservoir pressure. Another IPR equation was proposed by difference is negligible. The well productivity factor ia then
Fetkovich [3]. In dimensionless form, it can be written as: calculated aa follows:
(%)2 ‘
I
2rfkLu
—=
90m.*
90
(1.0-
) (2) WPs 0.00 H271n(%) +s (3)
ther oil rate or wellbore pressure to intersect with the expected ther production, dimensionless IPRs grow more curvature and
IPR curve. The repetition of the same procedure with different reach a shape like Vogel’s reference curve.
recovery factors results in dlffercr.t IPRa. Dimensionless IPR IPR curves of a horizontal well are nearly straight at the
curves are constructed by dividiltg the pressure coordinate of begivning, increase in curvature up to 8% recovery and then
each point on an IPR curve by the shut-in or average reservoir slightiv decrease in curvature at later stages of depletion.
pressure, and the oi! rate coordinate by the maximum oil rate
corresponding to 100~0 pressure drawdown. IPRs at Various Conditions
T’wo types of simulation runs are examined. In the first,
the well is constrained by a constant flowing bottom-hole pres- IPRs were produced for several variations of the base case
mre. In the second, a constant oil production rate is specified. conditions, These conditions concerned fluid, reservoir, and
F’or the same number of simulation runs, constant pressure well characteristics.
runs result in better IPR curve resolution that constant oil
rate runs. For this reason, all subsequent runs are done at a Effect of Fluid Characteristics
constant wellbore pressure constraint.
IPRs can be generated from outputs of either the IMEX The effect of fluid characteristics include the effect of crude oil
skmdator or the ECLIPSE simulator. When the base case PVT properties aa a function of pressure, and relative perme-
IPR curves are produced from the two simulators, comparison ability data as a function of phase saturation.
between them reveals that the curves are ordy slightly dXferent The first run utilized a crude oil with oil viscosity about
in magnitude but essentially identical in shape. 3ecause The half of that of the base case and a solution gas.oil ratio about
Gould computer where IMEX is installed is much faster than twice as great. PVT properties corresponding to case b in
the Appolo work-station where ECLIPSE is installed, all runs Vogel’s paper. As expected, with a less viscous crude, pro-
were done with IMEX. ductivit y is much greater than in the base case. However,
progression of the shape of the dimensionless IPR curves is
quite similar to the previous case.
Interpretation of the Results A second run was made with s more viscous crude which
PVT properties correspond to case 9 (d) of Vogel’s paper. The
Base Case IHts minimum oil viscosity is now 3 cp compared with 1 cp in the
Examination of Figure 1 shows that IPRa of recovery factors base case. Dimensionless curves are plotted on Figure 4 Be-
between 0.1% and 2% deteriorate rapidly with initi~ deple- cause of the decrease in oil viscosity, oil productivity decreases.
tion. ‘His is equivalent to a rapid decrease of the productiv- As noted for a vertical weII, dimensionless IPR curves show
ity index (PI) with. depletion. The high initial flow rates are less curvature with a more viscous oiL These two runs indi-
caused by the large length of the well and by the sharp initial cate that an increase in oil viscosity increases the concavity of
pressure gradients. Productivity decreases rapjdly because oil the dimensionless IPR curves.
is depleted from the near well region at a higher rate than it A third run was made with a higher bubble point oil which
is replaced by oil from outer oil zones. PVT properties are those of case 9 (e) of Vogel’s paper, These
Further examination of the plot in the region where reser- properties are those of a less viscous oil sidar to those of
voir pressure is slightly less than initial pressure, shows that case 9 (b). Initial reservoir pressure was reset to the new bub-
there is a slight increase in PI with drawdown. At smidl draw- ble pressure of 3000 psig. Because of higher initial pressure,
downs, gas evolves from solution but stays nearly immobde lower viscosity, and higher solution gas-oil ratio, oil rates and
because of its low saturation. This increase in gas volume ultimate recovery have increased significantly. In addition, an
increases pressure in the oil phase. As a result, pressure gra- increase in the curvature of the dimensionless IPR curves is
clients become larger and productivity increases. This phe- - noted.
nomenon is reversed for larger drawdowns in which case greater A fourth run utilized a low solution gas-oil ratio which
gas saturation causes greater resistance to oil flow. characteristics are illustrated by case 9 (f) of Vogel’s data. In
At the later stages of depletion, Figure 2 shows that deteri- this run, oil viscosity did not change significantly from the base
oration of IPRs slows down. Because of an increaae in gas-oil case values. Although ultimate recovery decreases, dimension-
ratio, reservoir pressure diminishes at an equivalent rate to less IPR curves retain their shape,
maximum oil rates. In other words, the productivity index The effect of changes in relative permeabtity was also stud-
decreaaes only slightly with production. ied, A run utilized relative permeabfity data from case 10 (b)
Figure 2 also indicates that the maximum recovery is ap- of Vogel’s paper. Critical gas saturation increased to 8% and
proximately 15%. Inspection of the oil saturation distribution residual oil saturation reduced to 10.6Yo, More oil is recovered
at that stage of depletion reveals that there is a considerable as a result of a reduction in residual oil saturation. Therefore,
amount of movable oil left in the reservoir at a vertical level productivity increases at the later stages of depletion. The
below the well. Because of gravity segregation and absence same type of curvat ure progression is noted for the dimension-
of a significant pressure differential, the left over oil cannot be less IPR curves as seen in F~ure 5. After a recovery factor of
produced unless with some typ~ of secondary recovery process. about 12970, concavity reduces.
Figure 3 shows that the shape of the dimensionless IPR Another run was made with relative permeability data
curves is sensitive to the recovery factor. At early stages of taken from case 10 (c). These data show higher oil perme-
depletion, dimensionless IPRs are nearly straight with a slight.. ability, and lower gas permeabUity. Critical gas saturation is
concavity developing at low dimensionless pressure. With fur- now 10%, and residual gas saturation is 30.6%. As expected,
with such favorable relative permeabtit y conditions, produc-
---
.
INFLOW PERFORMANCE RELATIONSHIPS
4 FOR SOLUTION-GAS DRIVE HORIZONTAL WELLS CD~ 3i100*
01 u 1?042.3
tivity at the later stages of depletion increases considerably, Effect of Well Characteristics
The ultimate recovery is now around 30% of the original oil in
place. However, dimensionless IPR curves retain their usual Well characteristics include skin, location with respect to top
shape. and bottom boundary, and length with respect to lateral reser-
C~mbinations of differing relative permeability and PVT voir dimensions.
properties were also investigated. In a run, relative permeabil- A run treated the effect of a positive skin, which could re-
ity data were replaced by those of case 10 (b) and crude oil suit from mud filtrate damage ‘or pertoration restrictions. The
\
PVT properti~ by those of c~e 9 (b). Ultimate recovery haa steady state infinitesimal skin factor is given a value of five.
increased to around 20Y0. There is not much change in the The resulting dimensionless IPR curves ar.>graphed on ~lgure
curvature progression of the dimensionless IPR curves. 10. As expected, oil productivity decreas ~s. Dimensionless
In another run, relative permeability characteristics of case IPR curves are initially straight, increase in concavity with
10 (b) were combined with PVT properties of case 9 (e). Initial depletion of up to 8%, then decreases in concavity.
pressure was reset to the bubble point pressure of 3000 psig. Two runs treated the effect of wells vertical placement with
Figure 6 is a plot of the resulthg dimensionles,~ IPR curves. An respect to top and bottom boundary. In the first run, the well
:.ncrease in both productivityy and ultimate recovery is noted. waa placed at a quarter of the height from the bottom bound-
Dimensionless IPR curves do not change very much in shape. ary, Results appear no different from those of the base case.
The next run included a combination of relative permeabil- This indicates that a change in the placement of the weIl within
ky of case 10 (c) and PVT properties of case 9 (b): Productiv- a quarter oft he height from the middle does not affect oil pro.
ky increases at the later recovery factors. As a consequence, ductivity significantly. In the second run, the well was lowered
ultimate recoverjj increases to 28%. As b~fore, the increase to the bottom of the reservoir. Corresponding dimensionless
in productivity or recovery does not change the shape of the IPR curves are plotted on Figure 11. Here, Initial IPR curves
dimensionless IPR.s significantly. show lower productivity and lower curvature. The lower pro-
In one run, fluid properties were approximated by straight ductivity is a result of the well’s placement near the no flow
lines given by case 9 (d) and 10 (d) of Vogel’s paper. Even bottom boundary. There does not seem to be a significant
with these extreme approximations, dimensionless IPR curves chang~ in ultimate recovery. Probably, the reservoir th~ckness
show the progressive increase in concavity shape. of 23,5 ft is not high enough to produce a significant amount
A run was made to see the effect of an initial pressure of oil by the ~ravity drainage process.
al. ;e the bubble point pressure. Wltid pressure WW set to A run waa made to see the effect of well length. The reser-
3040 psig, a value much above the bubble point pressure of voir was modeled in three dimensions with 9x6x 7 grid blocks.
2130 psig. The productivity index increases considerably when The well length was cut in half and the well was completed in
the flowing bottom-hole pressure drops from above to below the first 3 y-direction blocks. R.esultiag dimensionless IPR
the bubble point pressure. When average reservoir pressure is curves are plotted on Figure 12. Results indicate a decrease in
below the bubble point pressure, the contavity of IPR curve .. . . productivity by a factor of four. However, the @mensionless
regains its usual shape. IPR curves show the progressive concavity feature. There is
a slight indication that these curves are closer together than
before. It is important to point out that, due to computer
Effect of Reservoir Characteristics limitations, not enough grid blocks can be used to reproduce
Reservoirs characteristics include absolute permeability, drainage the three dimensional flow geometry.
area, and reservoir thickness. In summary, in almost atl cases dimensionless IPR curves
A run was made with an absolute permeability of 200 md, are approximately straight lines at initial recovery, but become
a value 10 times larger than that used in the base case. ”Results more concave with increasing cumulative production up to a
indicate a 10 times increase in oil production rate for a given certain recovery factor. At that time the concavity decreases
r-covery factor. In spite of that, there has been no significant unti[ final recovery is reached.
increase in ultimate recovery from a previous value of 1570.
Figure 7 shows that a change in absolute permeability does
Attempted Curve Correlation
not affect the shape of dimensionless IPR curves,
In another run, drainage area was doubled to 40 acres. Since the dimensionless curves exhibit a common curvature
the resulting dimensionless IPR curves are shown on Figure 8. progression, an attempt is made to correlate those changes
Results show no significant change in magnitude or shape from first with a modified form of Vogel’s curve, and then with
those of the base case. However, it is important to emphasize Fetkovich’s equation. The modified form of Vogel’s equation
that for a given recovery factor, the am~unt of cumulative oil is:
produced gets larger for a larger amount of oil originally in
place. —
90
= 1.0 -v: -(1 - v)(~)’ (5)
Two runs were made to see the effect of thickness. The first !70m.=
run used a thickness of 50 feet. The resulting dlmensiordess This equation is curve fit to the dimensionless IPR curves
IPR curves are plotted ~ Figure 9. The second run used a of the base case. Similarly, Fetkovich’s equation is curve fit to
thickness of 100 ft. Apparently, with increase in thickness, the the dimensionless IPR curves. Figure 13 shows a comparison
upward concavity seen at early recovery and at high flowing between the fitted curves by Equation 5 and the actual ones.
bottom-hole pressures disappears. It is not a perfect match because the parabola does not repro-
duce the initial upward concavity of initial IPR.s. For IPRs
corresponding to stages of depletion where the upward con-
-- .
m
SPE 19823 HICHEM BENDAKHLIA AND KHALID AZIZ 5
5. An attempt is made to correlate this common change of [5] Whitson, C. S.: ‘Reservoir Well Performance and Pre-
curvature. Both Vogel’s and Fetkovich’s equations are dicting Deliverabtity~ paper SPE 12518, Nov. 1983.
tried. They adequately fit especially to the later stage
of recovery curves with well behaved curvature, [6] Camacho, R. G., and Raghavan, R.: ‘%&XI Performance
Relationship for Solution-Gas Drive Reservoirs,n J. Pet.
6. A new two parameter equation resulting from the com- Tech., (May 1989), 541-550.
bination of Vogel’s and Fetkovich equations provides a
much closer fit to the generated IPR curves at all stages [7] Standing, M. B.: ‘Concerning the Calculation of Inflow
of depletion. Performance of Wells Producing from Solution-Gas Drive
Reservoirs,” J. Pet. Tech. (Sept. 1971), 1141-1142.
555
.,
INFLOW Performance RELATIONSHIPS
6 ~f)n
. --- SOT.TITTON.f2
., -”4,A C llnlilm UnDTCJAX1m AT TIXW
““~~~~ T c-+ C?nm * nanm
J2 4 V04
“b* . u sx”~~~a”~~ ~~~
.,..” .. W.. ”
or
‘[8] Diaz-Couto, L. E., and Golan M.: “General Inflow Per- [14] Plahn, S. V., Startzman, R. A., Wattenbarger, R, A,: ‘A
formance Relationship for Solution-Gaa Reservoir Wells,” Method for Predicting Horizontal Well Performance in
J. Pet. Tech. (Feb. 1982), 285-288.
Solution-GM Drive Reservoirs? SPE paper 16201, pre-
sented at the 1987 Production Operation Symposium,
[9] Kelkar, B. G., wd Cox, R.: “Unified Relationship to
Predict Future IPR Curves for Solution-Gas Drive Reser- Oklahoma City, Oklahoma, March 8-10.
voirs,n paper SPE 14239 presented at the 60th Annual [15] Kossack, C. A., and Kleppe, J.: “Oil Production horn
Technical Conference and Exhibition of the Society of the Troll Field: a Comparison of Horizontal and Verti-
Petroleum Engineers held in Laa Vegas, Neviida, Sep. 22- cal Wells,” paper SPE 16869 ‘presented at the 62nd An-
!25, 1985. nual Technical Conference and Exhibition of the Society
of Petroleum Engineers held in Daflaa, Texaa, Sep. 27-30,
[10] Joshi, S. D.: 4Augmentation of Well Productivity Using
Slant and Horizontal Wells,” SPE paper 15375, presented 1987.
at the 1986 Annual Meeting, New Orleans, Louisiana, Oc-
tober 5-8.
[11] Karcher, B. J., Giger, F. M., and Cornbe, J.: “Some Prac- S1 Metric Conversion Factors
tical Formulaa to Predict Horizontal Well Behavior,” SPE acres x 4.046873 E+03 = m2
paper 15430, presented at the 1986 Annual Meeting, New bbl x 1,589873 E-01 = m3
Orleans, Louisiana, October 5-8. Cp x 1.000000 E-03 = Pas
D X 8.640000 E+04 = s
[12] Babu, D. K., and Odeh, A. Z.: “Productivity of a Hori-
ft x 3.048000 E-01 = m
zontal Well,” paper SPE 18298, 1988. ft3 x 2.331685 E-o2 = m3
[13] Sherrard, D. W., Brice, B. W., and MacDonalu, D. G.: lbm x- 4.535924 E-01 = kg
‘Application of Horizontal Wells at Prudhoe Bay,” J. Pet. md x 9.689233 E–04 = pm~
Tech. (Nov. 1987), 1417-1425. psi x 6.894757 E+OO = kPa
2000
~
in
a
a
5 1590
w
In
w
k
~
2
k
1000
i
o
.~
it
g
u.
500
0
0 2000 4000 6000 8000
Oil flow rate (STB/D)
c-
m -.
.
..”
SW 19823
1.0
0,8
1? 0.6
z
ri
0.4
0.2
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Oil Flow rala (STB/D)
Dhnensionks oil rste
1s
1.0
0.6
0.8 -
: 0.6 -
z
.5! NP/N =
2 0.4, 0.1, t.o
z
.-
0
‘“4 14.0, 2.0, t2.o,
0.4 - 6.0, 10.0 %
0.2
0.2 -
tJ.o~..J~
0.0 0.2 0.4 00 3.8 1 00 ~.1
0.0 0.2 0.4 0.6 0.8
Dimensionlessoil ra!e
Dlmsnsbnbss oil tlow rate
Flguro 4. Dhwenalonleaa IPR cuwes wllh Flgura 5. Olnrenalonle$s IPR curvee with
PW proparlles from csss (d). rolotlvo parrrruablllfy from osae (b)
667
SPE 1982h “ ~“ -
1.0
NPR4 -
0.1, 0.4, 1.0,
2.0, 10.0, 6,0 %
0.2 -
0.0
0.0 0.2 0.4 0.6 0.6 1.0
-..
0.0
0.0 0.2 0.4 0.6 0.8 1..0 Llimensbnieesoil ftow rate
Dimensbnleea oil flow rate
1.0 ,-
0.6
0.6
0.4
0.2
0.0 L \ 0.0
0.0 0.2 0.4, 0.6 0.6 1.0 0.0 0.2 0.4 0.6 0.6 1.0
Dimensionleaaoil flow“rete
Dimensionlessoil flow rata
.
sf% 1982~
1.0
~
a
% 0.8
g
0.
NPIN -
0.4, 0.1,
f .0, 2.0,
10.0, 6.0 %
-~
0.0 0.2 0.4 0.6 0.0
0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
Dimenebnless oil fbw rate
DhmsnebnleaOHfbw rete
Figure 10. Dlmenslonless IPR c ues with poal!lve bkln of 5.
# FIOut* 11. Dimenslonleee IPR curwe wlfh well on fhe bo~om.
1A,
“*.
2 F/pm . {
0) 0.1, 0.4, NPIN =
&?0.4 f .0, 12.0, Of,
~
.7 2,0, 6.0, 2.0,
.: I 10.0 % 6.0 %
.-E
Q
0.2
t
.......
0.0 I
0,0 0.2
J Aclualcarves0! the beae r *e
0.4 0.6 0.6 1.0
Dimensionlessoil flCWreIe 0.0 ~ \
0.0 0.2 9.4 0.6
Figure 12. Dhnenelonleee IPR curves wifh helf well length. 0.6 1.0
669 . .
1.0 1.0
0.8 - 0.8 -
NP/N =
0,6 - 0.6 - 0.1,
2.0,
NPIN - 6.0 %
0.1,
2.0,
0.4 - 6.0 % %, 0.4 -
\
...
‘8
‘.,
0.2 - — Fittedcur.es with FetkodchIPR. 0.2 -— Fittedcwves withthe new equatbn.
0.0 0.0
-0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.s 0.8 t ,0
Figure 14, IPR curvoe tftted by Fafkovfch aquatforr Figure 1S. IPR curves ftfled wffh the new quatkon
compared to acfuaf OIWea. -w~ :0 aefual Curses.
0.4
0.2 -
0.01
0.02 0.04 0.06 0.06 0.10 0.12 0.$4