Well-Test Analysis for Solution-Gas-Drive

Reservoirs: Part 2-Buildup Analysis

Kelsen v. Serra, * Alvaro M.M. Peres, * SPE, and Albert C. Reynolds, SPE, U. of Tulsa

Summary. This work presents new analysis methods for pressure-buildup data from a well completed in a solution-gas-drive reservoir.
New procedures for estimating effective phase permeabilities as functions of pressure and saturation are presented.

Introduction
In Part 1, we showed that, for the infinite-acting drawdown period, Am=krm/Ilm, for m=o, g. Because we assume, as in Part I, that
reasonably good approximations of oil and gas effective permea- only oil and gas are mobile (i.e., water is immobile), the total mo-
bilities as functions of pressure or of oil saturation can be com- bility is given by Eq. 2 of Ref. I, and the system compressibility 7,8
puted directly from measured values of the flowing well bore is given by Eq. 3 of Ref. 1.
pressure if flow rates and initial saturations are known. We also The pseudopressure function is defined by
showed that drawdown-pressure data could be analyzed with a
j' Pi( -k - ) dp .
semilog plot of the flowing wellbore pressure squared vs. time to ro
Pp= ............................... (2)
yield an estimate of effective oil permeability as a function of P Bolla
pressure and an estimate of the mechanical skin factor. 1
In this work, we show that a generalized form of superposition The wellbore pseudopressure is given by
applies if the same oil saturation/pressure relation is used to evaluate
all pseudopressures. The generalized superposition equation allows
us to extend the methods of Ref. 1 to the buildup case. In particular,
r
Ppwf= J Pi(-k ro- ) dp, .............................. (3)
we' show that effective permeabilities as functions of pressure can Pwf Bolla
be estimated directly from pressure-buildup data. Further, because
and the dimensionless pseudopressure at the wellbore is defined as
oil saturation can be approximated as a function of pressure from
a material-balance eqlJation, we can estimate effective permeabil-
ities as functions of oil saturation. In addition, we show that buildup- khp pwf
khJ'Pi( ~
Pwf
k
Bolla
) dp
pressure data can be analyzed with a semilog plot of shut-in pressure PpwD= -=-----=-"'--. . ................ (4)
squared vs. an appropriate Horner time ratio. As in Part 1, all re- 141.2qo
sults assume that the reservoir is infinite-acting during the test. Part
3 considers the analysis of buildup data obtained after a long Note that to compute Ppwf directly from measured pressure data,
producing time. 2 Because this work (Part 2) represents a direct ex- the oil relative permeability as a function of pressure must be known.
tension of those results to the buildup case, the reader is strongly This is the major shortcoming of the analysis procedures suggested
advised to see Part I for pertinent background, historical infor- in Refs. 9 through II.
mation, and references. Results similar to some of those shown here In considering buildup solutions, we work in terms of two dimen-
and in Part I were obtained independently and concurrently by Al- sionless buildup pseudopressure functions, PpsD and PpsD, defined
Khalifah et al. 3,4 as

Mathematical Model k ro ) [Pi(

The problem considered assumes radial flow to a completely kh khJ - - dp
penetrating well located at the center of a cylindrical homogeneous Ppws Pws Bolla
PpsD=---= ................... (5)
reservoir of uniform thickness. The top, bottom, and outer reservoir 141.2qo 141.2qo
boundaries are sealed, and the well is produced at a constant surface
oil rate. The initial pressure is assumed to be uniform throughout
the reservoir and is equal to the bubblepoint pressure. After pro- khrws(~)dP
duction, the well is shut in to obtain pressure-buildup data. As is _ kh(ppws-Ppwf,s) Pwf,s Bolla
standard, the skin zone is modeled as a region concentric with the and PpsD = ---''----'-~- ---'~---"''-''--, ....... (6)
141.2qo 141.2qo
wellbore where the absolute permeability in the skin zone, k s , is
different from the absolute reservoir permeability, k, which is
respectively. Throughout our discussion of pressure buildup, t and
assumed constant. s, r s' k, and ks are related by Hawkins'S
equation. Gravity and capillary effects are assumed negligible. tJ.t are measured in days, dimensionless producing time is defined as
The buildup pressures considered here were obtained from the
tD =O.OO6328kAtithpctir2;.. ........................... (7)
same variable-bubblepoint black-oil simulator used in Part 1.
Although irreducible water is present, all results assume that the and dimensionless shut-in time is defined as
water phase is immobile and incompressible, so the water saturation
remains constant throughout the producing period. Refs. I and 6 tJ.tD =O.OO6328kAtitJ.t/<I>Ctir2w' ........................ (8)
discuss the simulator further.

Definitions Background
The in-situ (flowing) reservoir GOR is given by From the results of Refs. 1 and 9 through 11, we know that, for
the infinite-acting drawdown period, an excellent approximation
Rr=Rs + [(krg/kro)(lloBo/ IlgBg)]. . .................... (1) to PpwD is given by
The producing GOR, R, is given by the right side ofEq. I evaluated
PpwD = Ihln(4tD/e'Y) +s . ............................. (9)
at rw; i.e., at r=rw, Rr=R. The mobility of Phase m is given by
To the best of our knowledge, Eq. 9 was first presented by Fet-
Copyright 1990 Society of Petroleum Engineers
kovich,12 in a different form, and later. verified numerically by
'Now at Petrobras SA Raghavan 10 with a reservoir simulator.

SPE Formation Evaluation, June 1990 133

 .;; 4100
c:a.
:a
0

0
°i -41=0
~ 3900
bJ 0 CASE 3 0 ~3900 o 41=IDAY
II::
::J
(/)
qo=2oo STB/D 0 li
(/)
bJ
3700 5=0 i1l~ 3700
II::
0- 0 CASE 2 0 CASE 3
I- qo'100 STB/D IE qo= 200 STB/D
z 3500 5=5 ~ 3500 5=0
50- 0

bJ
0
rs/rw=2.59
ks /k=0.16
~o
..J 3300 ~ 3300
CD
CD
::J m
CD
~
310?0_1 10 103 10 31D?0-1 10 102
DIMENSIONLESS INVERSE BOLTZMAN VARIABLE, tD/r~ RADIUS, r, feet

Fig. 1-Varlatlon In Pb for Cases 2 and 3. Fig. 2-Change In reservoir P b during buildup for Case 3.

Following the ideas of Refs. 10 and 11, we showed 1,6 that pressure. As shown later, variations in the bubblepoint pressure
during production influences the buildup response. Results presented
were obtained from our variable-bubblepoint, finite-difference
acx) ab a{3 ]

l
2kQD (aa simulator. 1,6
dSo = --,;;;; cxa; -aa; +cxa; -aa;; ,
Behavior of Bubblepoint Pressure. Results regarding the vari-
dp 2kQD (-cx~ +a~ ) -cx~ +a~ rw ation in bubblepoint pressure are discussed briefly. Details can be
kscx aso as o aso aso found in Appendix A of Ref. 13; see also Refs. 6 and 14.
.................................... (10) During the infinite-acting period, pressure and oil saturation at all
points in the reservoir can be accurately correlated for zero-skin
where QD =5.615qot/271'cf>r~h . ........................ (11) cases in terms of the dimensionless Boltzmann variable, ZD =
rJltD' if tD lrJ>50 or rD>50, where rD=rlrw.l Because all
Eq. 10 can be numerically integrated to obtain So as a function other dependent variables are functions of pressure and oil satu-
of pressure. For the s = 0 case, Ref. 11 derived Eq. 10 by assuming ration, it follows that PVT properties, relative permeabilities, bub-
that So and P were unique functions of the Boltzmann variable. blepoint pressure, etc., can be correlated in terms of the Boltzmann
Aanonsen 9 derived an analog of Eq. 10, however, by assuming variable; see Ref. 13.
that So and P were monotonic functions of r and t. Note that Eq. The solid curve in Fig. 1 shows the bubblepoint pressure obtained
10 assumes that So is a unique function of p. In Eqs. 10 and 11, for Case 3 vs. the inverse of the dimensionless Boltzmann variable.
we assume that the oil flow rate, qo' is constant; however, a similar The circular data points on this curve represent the sandface value
expression can be derived if qo varies. (rD = 1) of Ph plotted vs. tD' Here Pb at the sandface is slightly
above its initial value (Pbi =4,000 psi) for a small interval around
Simulator Data tD = 102 , but it decreases rapidly at later times. The square data
The PVT properties and relative permeability curves used in this points represent the sandface value of Ph obtained for Case 2
work are identical to those used in Part 1 (see Table 1, Eqs. 26 plotted vs. tD' (For nonzero-skin cases, variables cannot be cor-
and 27, and Fig. 1 of Ref. 1). For these relative permeability data, related in terms of the Boltzmann variable. 1) Similar results were
Sgc=O.O; water is assumed to be incompressible and immobile obtained for Cases 2 and 3; specifically, Ph(r w) decreases rapidly
with Siw=0.3. below its initial value as time increases. 13 The rapid variation in
For all cases considered here, cf>=0.3, r w=0.328 ft, h= 15.547 Pb near the wellbore during production can be detrimental to the
ft, and re =660 ft. (Because we consider only data from the determination of Pbi or Pb from a bottomhole fluid sample. 14
infinite-acting period, re has no influence on the results presented.)
These data are identical to those used in Refs. 1,9, and 11. The Approximating the Oil Saturation/Pressure Relation. From the
initial and bubblepoint pressures were always specified as results of Ref. 15, it follows that the dissolved GOR at a gridblock
Pi =Pbi =4,000 psi, and all results presented are for cases where value of bubblepoint pressure is given by
k= 10 md, but the value of k has no influence on the validity of
the analysis procedures. In all cases, qo was held constant during Rsh =Rs + (SgBoISoBg) . ............................ (12)
pressure drawdown. Once Rsb is computed, the associated value of Pb can be obtained
from the Rs vs. P curve. Thus, Eq. 12 provides a relationship be-
Cases Considered tween saturation and pressure. Note that Eq. 12 can be applied at
In this work, we present results for three cases. Similar results for every gridblock or, equivalently, at every point in the reservoir.
other data sets are in Ref. 6. In all cases, t = 1 day, tD = 1.2 X 105 , Using Sg=I-So-Siw in Eq. 12 and rearranging the resulting
and all simulated buildup tests are I-day duration. equation, it is easy to show that
In Case 1, qo=l00 STBID, s=O, andpwf,s=3,366.5 psi.
In Case 2, qo=l00 STBID, s=5, Pwf,s=2,570.2 psi, rslrw= Bo(l-Siw)
- - - - - ............................. (13)
2.59, and kslk=0.16. Bo+(Rsb-Rs)Bg
In Case 3, qo=200 STBID, s=O, and PWj,s=2,411.2 psi.
With Eq. 13, we can compute the sandface So value directly from
Computing 011 Saturation the wellbore pressure provided Siw and the sandface Pb value are
Here, we discuss methods for computing oil saturation as a function known. Then, if we assume that the variation in Pb does not in-
of pressure directly from buildup data. A major objective of this troduce significant errors, Eq. 13 can be approximated by
work is to provide procedures to estimate effective oil and gas per-
Bo(l-Siw)
meabilities, kk ro and kk rg , as functions of pressure and of oil satu- ........................... (14)
ration, directly from the measured values of the wellbore shut-in

134 SPE Fonnation Evaluation, June 1990

 0.71 r---,----,.---r----r--r--,----,.---,
0.70 o
en CASE 2
0
en o MATERIAL BALANCE Z 0.69 qo=IOO STBID
Z " MARTIN'S EQUATION
o s=5
0 eiII: 0.67 's/'w=2.59
eiII: ;:) ks lk=0.16
;:) 0.62 ~
ti en 0.65 v
en
o
..J
o "v
..J SIMULATOR
0 0.58 lj 0.63
DRAWDOWN CASE :3
I&J
(,J qo=200 STB/D ~
~ saO
"o ~ 0.61
" DRAWDDWN. Ea. 10
0 0.54 o MARTIN'S EQUATION
Z " ~
~
o v RAGHAVAN ,.
o 0·~~'='OO::---.l.--2:-:8"=OO-:--.L----:3~2'='OO::-----'-""""36""OO:-:--..I..--40~OO
05~0~_~6~~~~~10~_~4~WL~~~10~_2~~L-~~~~WU
WELLBORE PRESSURE, Pwf OR Pws ' psi
TIME, t, DAYS
Fig. 4-Approxlmatlon of saturation/pressure relation for
Fig. 3-Approximatlon of· sandface So for Case 3. Case 2.

Thus, if Pbi is known, Eq. 14 can be used to approximate So as Note that, if the gas saturation remains close enough to zero so that
a function of P directly. Whether drawdown or buildup pressures Ao IA, ,." 1, we can develop an approximation for Eq. 16 that does
are used in Eq. 14 has no effect on the So value obtained. The not involve relative permeabilities.
difference between the correct value of So from Eq. 13 and that
from Eq. 14, however, depends on time because this difference dS o =(SoBg dRs _ Sg dB g ) . .................. (17)
is controlled by the variation in Rsb during pressure drawdown or
dpws Bo dpws Bg dpws
buildup. All our computations, however, indicate that the variation
in Pb during pressure buildup is insignificant, as shown in Fig. 2 In solving Eq. 16 or 17 for oil saturation, we use Sg=I-So-Siw
for Case 3 results. In Fig. 2, the solid curve is the reservoir profile to express Sg in terms of So.
of Pb at the end of production (t = 1 day, Ilt =0) and the circular Eqs. 14 through 17 can be used to estimate oil saturation as a
data points are the Pb profile after 1 day of pressure buildup. Note function of pressure with either drawdown or buildup pressures in
that, at all points in the reservoir, the Pb value at the end of 1 day the computation. The So values from Eq. 14 depend only on
of shut-in is essentially identical to its value at the end of production. pressure and will be independent of whether a given value of
Thus, the results indicate that Eq. 14 can provide accurate estimates pressure corresponds to drawdown or buildup. The So values from
of So as a function of pressure if and only if the variation in Pb Eqs. 15 through 17, however, will be different for the drawdown
during production is sufficiently small. and buildup cases because their numerical solutions require a starting
B~e et al. II showed that, at early times, Eq. 10 can be approxi- value of oil saturation of So(t =O)=Soi for drawdown and
mated by So(llt=O) for buildup. Thus, the drawdown So-vs.-wellbore-
dSo So dB o Ao pressure relation may be different from the one for buildup So. As
-=--+c,-, ............................ (15) established later, we need the drawdown So-vs.-p relation to es-
dp Bo dp A, timate effective permeabilities as functions of oil saturation.
which is Martin's8 equation. Eq. 15 can be derived easily from Raghavan 10 presented a method for determining absolute per-
Eq. 10 by setting QD =0 and rearranging the resultant equation. meability from a semilog plot of pseudopressure vs. shut-in time.
Because QD=O during pressure buildup, B~e et at. also used The method requires the determination of oil saturation as a function
QD=O in Eq. 10 (or, equivalently, Eq. 15) to compute the of pressure to compute Ppws' Assuming that oil and gas relative
sandface So as a function of shut-in pressure. Note that QD given permeability curves are available, a krglkro-vs.-So plot can be con-
by Eq. 11 is small if qot is small (i.e., if cumulative production structed. If we replace Rr in Eq. 1 with the producing GOR at the
is smaIl, which occurs if the production rate is small or the producing instant of shut-in, R(llt=O), and rearrange the resulting equation,
time is short), indicating that results from Eq. 15 are rate- we obtain
dependent. 9 When applying Eq. 15, to estimate So from pressure-
buildup data, we set P=Pws' krg
-=[R(llt=O)-R (J.l.gB g )
Fig. 3 compares the simulator value of So at the sandface (solid s] - - , ..................... (18)
kro J.l.oBo
curve) with the So value from Eq. 14 (square data points) and the
So value obtained by solving Martin's8 equation (Eq. 15, triangular which gives krglk ro as a function of pressure. 10 Because we can
data points). These results pertain to Case 3. As expected, Eqs. obtain krglk ro as a function of So from the relative permeability
14 and 15 yield approximately the same pressure/saturation relation curves, we can usually obtain So as a function of P by comparing
and both equations yield accurate results at high pressure values the two krglk ro expressions. This method, however, cannot de-
that correspond to early times for pressure drawdown. Specifically, termine So when Sg<Sgc, and if R s >R(llt=O), Eq. 18 gives a
if QD is small, Eq. 15 yields accurate results, and if the variation negative krglk ro value. 9 ,16,17 Therefore, in applying this method,
inPb at the sandface is small, Eq. 14 applied at the sand face yields we set So =Soi, Sg =0, and k ro =k roi whenever Eq. 18 gives a
accurate estimates of the sandface value of So' Note that Eq. 14 negative krglk ro value.
yields highly accurate values of So for t< 10- 2 days, tD < 1.2 X Fig. 4 presents Case 2 results of the sandface So calculated with
103 . The Case 3 results in Fig. 1 for tD < 1.2 X 10 3 indicate that the various methods for a I-day buildup test. Results obtained from
the change in the sand face Pb value is negligible. our simulator are shown by the solid (drawdown) and dashed
(buildup) curves. Circular, triangular, and inverted-triangular data
Calculating Oil Saturation During Buildup. Using the definition
points represent the various methods identified in the legend. As
of c, (Eq. 3, Ref. 1), we can rearrange Eq. 15 to obtain
in Ref. 9, the results of Fig. 4 indicate that the solution of Martin's
dSo ( AO) So dB o (SoBg dRs Sg dB g ) Ao equation (Eq. 15) reproduces the actual (simulator) buildup satu-
dpws = l-~ Bo dpws + ---;;: dpws - Bg dpws ~. ration/pressure relation almost perfectly. (Refs. 9 and 11 assumed
s=O in all their results.) Results from the modified version of
.................................... (16) Martin's equation (Eq. 17) are not shown but are essentially iden-

SPE Fonnation Evaluation, June 1990 135

 1~ 3400 3500 3600 3700 3800 3900 ~O
0.71
CASE 1 CASE 2 0
Ul CASE I
qo =100 STB/D qo • 100 STB/D " " " " " :A!""
saO s=5
160
~ 0.69 qo=IOO STB/D
s=O
c DRAWDOWN
'" BUILDUP
--DRAWDOWN
o BUILDUP
~
II:
~ 0.67
150 '"
Ul
.oJ
SIMULATOR. /",,---
o o Db. 0 00 0 0 BUILDUP--..,..... '"
0 0.65 ,a",
o Il. 6. l:J. 6
UJ /'"
,if'"
o 00 0 0 0 0 0 0 0 0 140 ~
0
4~
'" DRAWDOWN. Ea. 10
z 0.63 o MARTIN'S EQUATION
SIMULATOR
-10000000000
'"
Ul DRAWDOWN " RAGHAVAN

110 130 0.61.~--d:v:---:d;:::o--::;::l;;;::---::::!V:,---=~-:::::\:,",,"".....,~

2600 2800 3000 3200 3400 3600 3800 4000 3300 3400 3500 3600 3700 3800 3900 4000
WELLBORE PRESSURE, Pwf OR Pws ' psi WELLBORE PRESSURE, Pwf OR Pws ' psi

Fig. 5-Drawdown and buildup sandface GOR for Cases 1 Fig. 6-Approximation of saturation/pressure relation for
and 2. Case 1.

tical to those identified by the circular data points in Fig. 4. Note Following superposition results for the liquid case, we consider
that Raghavan's procedure (inverted triangles) reproduces the the analogs
drawdown saturation/pressure relation accurately until P =- 3,300
PpsD =PpwD(tD +lltD) -ppwD(lltD) ................... (19)
psi. As pressure increases, however, the procedure becomes less
accurate until, at p=-3,700 psi, Eq. 18 gives a negative krg/kro and PpsD =PpwD(tD) -PpwD(tD + lltD) +PpwD(lltD), ....... (20)
value, and the method fails.
where the PpwD terms represent drawdown solutions. Ref. 18 may
The Case 2 results in Fig. 5 show the well bore GOR calculated
prove to be the starting point for an approximate analytical deri-
by applying Eq. 1 at the sandface for both drawdown and buildup.
vation of Eqs. 19 and 20, but here we assume that they hold and
Note that the drawdown GOR is approximately constant and equal
examine the consequences.
to R(llt=O) for P < 3,300 psi, which corresponds to the pressure
Using the definitions of relevant dimensionless pseudopressure
range where the oil saturation/pressure relation obtained from Eq.
functions, we can rewrite Eq. 19 as
18 is essentially identical to the drawdown So-vs.-p relation; see
Fig. 4.
Fig. 6 presents the Case 1 results of Fig. 4. Results for Case
1 indicate that the simulated and computed saturation/pressure re- j.Pi( -kro- ) dp= j' Pi ( -k ro- ) dp- Pi j' (kro
- -) dp
Pws BoJl.o bu Pwj(t+lJ.t) BoJl.o dd Pwj(lJ.t) BoJl.o dd '
lations for buildup (simulator or Eq. 15) and drawdown (simulator
or Eq. 10) are similar. For p<3,750 psi, these relations are also · ................................... (21)
similar to the saturation/pressure relation obtained with Raghavan's where subscripts dd and bu refer to the drawdown and buildup
procedure. Thus, the B~e et al. 11 and Raghavan 10 buildup- kro/(BoJl.o)-vs.-p relationships, respectively. Eq. 21 can be re-
pseudopressure functions should be about the same. For sufficiently written as
low rates or short times, the terms involving the flow rate (i.e.,

j' Pi( -k ro- )

QD in Eq. 10) do not dominate the drawdown saturation/pressure
relationship. Because Eq. 15 results from setting QD=O in Eq. 10, dp=
j' Pwj(lJ.I) (kro
--
) dp.. .......... (22)
it follows that, when applied to buildup data, Eq. 15 should give Pws Bo Jl.o bu Pwj(t+ lJ.t) Bo Jl.o dd
approximately the same saturation/pressure relation that Eq. 10 gives A similar procedure starting from Eq. 20 gives
for drawdown data. In fact, it is the relative influence of the terms

j' pws(-kro- )
involving QD in Eq. 10 (more than the reported fact that different
rates give different profiles at the instant of shut-in 9 ) that makes dp= j' Pi ( -k ro- ) dp- j' pWj(lJ.I)( -k ro- ) dp.
the B~e et at. 11 procedure rate-dependent. If the drawdown rate Pwj.s Bo Jl.o bu Bo Jl.o dd
Pwj.s Pwj(t+lJ.t) Bo Jl.o dd
were lowered to qo=50 STBID, the B~e et al. procedure for · ................................... (23)
generating the buildup saturation/pressure relation would correlate
even better with the one for drawdown. Results for Case 1 indicate Using Eq. 22 in Eq. 23 gives
that during drawdown Pb>P;, and hence, Pb>P. 13 Thus, the
j' pws( -k ro- ) r r
wellbore pressure and pressures within the reservoir will always ro ro
dp= J Pi ( -k - ) dp- J Pi( -k -) dp.
be less than the bubblepoint pressure during a buildup test. This
Pwj.s Bo Jl.o bu Pwj.s Bo Jl.o dd Pws Bo Jl.o bu
condition is necessary if we wish to obtain a unique buildup
pressure/saturation relation that applies at all points in the reservoir · ................................... (24)
or if we wish to correlate buildup variables in terms of the Boltzmann
variable. If we rearrange Eq. 24, it follows that

Bulldup.Pseudopressure Behavior . Pi ( kro )

-- dp= Pi ( - j'
kro- ) dp- Pws ( Jr -k - )
ro
dp.
Here, we explain superposition for the solution-gas-drive system \
Pws Bo Jl.o bu Pwj.s Bo Jl.o dd Pwf.s Bo Jl.o bu
under consideration. Throughout our discussion of pressure buildup,
t is the producing time in days and tD is the dimensionless · ................................... (25)
producing time. Clearly, Eq. 25 is valid if and only if the kro/(BoJl.o)-vs.-p re-
lation (or equivalently, the saturation/pressure relation) chosen for
Superposition for Buildup. Although previous work9- 11 on buildup is the same as that determined by the drawdown-pressure
buildup did not explicitly deal with superposition, attempts to response. Thus, we have shown for the first time the following
correlate buildup data with the liquid buildup solution clearly rep- theorem: If the natural analogs of the liquid superposition equation
resent attempts to apply some form of superposition. hold for the solution-gas-drive-reservoir problem, then the same

136 SPE Formation Evaluation, June 1990

saturation/pressure relation should be used for both buildup and the sandface So will be equal to its initial value, SOi=0.7 (Sgi=O),
drawdown. (Unfortunately, to date, we have been unable to establish when the shut-in pressure at the wellbore reaches the value of the
the converse of this theorem theoretically.) Because the drawdown- bubblepoint pressure at the sandface [i.e., when Pws=Pb(r w)=
pseudopressure function can be correlated with the liquid solution, 3,183 psi]. At distances far from the well, however, free gas will
it seems preferable to use thedrawdown relation for kro/(BolLo) still exist (Sg > 0) when the buildup pressure reaches 3,183 psi. The
vs. p (i.e., the drawdown saturation/pressure relation) to evaluate B¢e et al. pseudopressure is correct in that it uses correct simu-
Ppws if we want to analyze buildup data using standard semilog lator values, but it fails to provide a valid analysis procedure be-
methods. If we do this for an infinite-acting reservoir so that both cause it fails to correlate with the single-phase liquid-buildup
PpwD functions in Eq. 19 can be approximated with the semilog solution. Buildup saturations cannot be conelated in terms of the
equation (Eq. 9), then Eq. 19 becomes Boltzmann variable. As explained previously, we believe that, to
use single-phase-flow liquid analogs to analyze buildup data, the
khJfPi( ~
k ) pseudopressure function should be defined so that a form of super-
dp position applies. Aanonsen 9 did not investigate superposition, but
ws _
_P_ BolLo
__ dd_ =1.151 log (t+!J.t)
-- . he plotted buildup data vs. the Horner time ratio, which seems to
PpsD imply that some form of superposition applies.
141.2qo !J.t
Here, we provide an alternative explanation of why the pseu-
.................................... (26) dopressure function of Ref. 10 approximately represents the
drawdown So-vs.-p relation. If the So-vs.-p relation obtained from
The preceding theoretical result on superposition yields additional Eq. 18 represents the drawdown So-vs.- p relation, then it follows
insight and im~roved interpretations of the imp::>rtant contributions that R(!J.t=O) is approximately equal to the actual drawdown
of Aanonsen, Raghavan, 10 and B~e et al. I The B~e et al. II producing GOR for the last several hours of the preceding flow
procedure for computing the buildup-pseudopressure function uses period; i.e., that the producing GOR is approximately constant for
the buildup saturation/pressure relation obtained by solving Eq. 15 most of the preceding drawdown. not that the GOR is constant
numerically. It follows from the theoretical result established above during buildup. The results of Figs. 4 through 6 illustrate the va-
that ~e et al. 's buildup-pseudopressure function can be correlated lidity of this conclusion. First, we consider the Case 1 results of
with the liquid solution if and only if the buildup-pressure response Figs. 5 and 6. Fig. 5 shows that R(!J.t) is approximately equal to
is such that the kro/(BolLo)-VS.-p relationship is approximately the the drawdown GOR at times corresponding to p<3,750 psi, and
same for both drawdown and buildUp. As shown by the results in Fig. 6 indicates that for p<3,750 psi, the So-vs.-p relation ob-
Fig. 6, this can occur in some situations. tained from Eq. 18 is approximately equal to the drawdown So-
Raghavan intuitively argued that the pseudopressure function used vS.-p relation. For p>3,750, there is more variation in the
for buildup analysis should be different from the one used for producing GOR (Fig. 5) and Eq. 18 fails to produce the drawdown
drawdown. Quoting from Ref. 10: So-vs.-p relation. In fact, for this case, Eq. 18 gives negative
The Pp function used to analyze buildup should reflect conditions values for p <!: 3,800 psi. From the Case 2 results of Figs. 4 and
occunng during buildup in the reservoir. Again, the GOR at the in- 5, we see that (1) the drawdown producing GOR is essentially con-
stant of shut-in is a representative measure of the conditions prevailing stant and equal to R(!J.t=O) for p<3,300 psi, but there is more
. in the area drained by the well. Thus. the second method for caL- variation in the producing GOR at higher pressures (earlier
culating the pseudopressure function for buildup assumes that the drawdown times) (see Fig. 5); and (2) and So-vs.-p relation ob-
GOR is constant during the buildup period and is equal to the
tained from Eq. 18 deviates more from the corresponding drawdown
producing GOR at the instant of shut-in. Basically. this implies that
pressure changes occur much faster in the reservoir than the time relation as the pressure increases above 3,300 psi. As shown in
interval in which large changes in saturation take place. Fig. 5, the buildup GOR in Case 2 is not constant untilp>3,454
psi, which corresponds to the bubblepoint pressure at the sandface.
Aanonsen 9 showed that the buildup GOR*R(!J.t=O) at all shut-in
(Results for Case 3 are very similar to those shown in Figs. 4 and
times; see Fig. 5. He also showed that, in a relative sense, pressure
5; see Ref. 6.)
changes do not occur faster than saturation changes. This also can
be seen from the results of Fig. I, which indicate that, during an Correlation With Liquid Solutions. The previous theoretical dis-
ensuing buildup test, all gas in the near-wellbore region of the cussion indicated that it is appropriate to use the drawdown
reservoir will dissolve long before Pws reaches p. More impor- pressure/saturation relation when constructing the buildup-
tantly, Aanonsen 9 showed that when applicable, the buildup pseu- pseudopressure function if we want to apply superposition to obtain
dopressure used in Ref. 10 (see Eq. 18) approximately represents analyzable buildup results. Using the semilog approximation (Eq.
the drawdown pseudopressure. Note that the results of Figs. 4 and 9) for all drawdown solutions on the right of Eq. 20 gives
6 reaffirm this conclusion. Our theoretical arguments on superpo-
sition also indicate that if Raghavan' s buildup-analysis procedure ppsD=1.151 log(4tD/e'Y)+s-1.151 10g(RHl ), ......... (27)
is applicable, then it should be based on the drawdown pseu- where RHI =(tD+!J.tD)/!J.tD=(t+!J.t)/!J.t, ............... (28)
dopressure.
Aanonsen 9 showed that the proposal lO for computing a and the right side ofEq. 27 represents the single-phase liquid-buildup
"buildup" pseudopressure using the So-vs.-Pws relation from Eq. solution. Based on the results of Refs. 9 and 19, the correlation
18 approximately uses the drawdown saturation/pressure relation between the liquid and solution-gas-drive solutions might be slightly
and the drawdown kro/(BolLo)-vS.-p relation. This conclusion was improved by replacing RHI with a Horner time ratio that incor-
based on the fact that during buildup the reservoir flow equation, porates Agarwal's 20 pseudotime or Scott's21 normalized time. The
written in terms of a general pseudopressure function, is approxi- Horner time ratio based on Scott's normalized shut-in time is de-
mately linear (quadratic gradient terms can be neglected) and that fined by
the pseudopressure calculated by Raghavan approximately satisfies RHZ=(tD+!J.tD)/!J.tnD, ............................ (29)
the correct initial and boundary conditions for buildup, whereas
~e et at. 's pseudopressure function does not satisfy the correct
where !J.tnD =0.006328kkroi!J.t/cJ>CtlLor~ . ............... (30)
initial condition. As shown in Figs. 4 and 6, however, B~e et al. 's Aanonsen's9 results on pseudotime (he did not consider nor-
method yields the correct buildup-saturation values and uses the malized time) indicate that At should be used in place of kro;llLo
correct buildup initial values of pressures and oil saturations. An in Eq. 30. Our computations, however, indicate that the simplified
amplified explanation 9 of the failure of the B~e et al. buildup- version of normalized shut-in time defined by Eq. 30 works just as
analysis procedure is that their buildup saturation/pressure relation well. Moreover, with the definition ofEq. 30, Eq. 29 is equivalent to
applies only at the wellbore; at other points in the reservoir, a
R HZ = [(t+!J.t )/(lLoiCti)] /[!J.t /(lLoC t )]; ................. (31)
different saturation/pressure relation holds during buildup. This is
clear from our results on variable bubblepoint. For example, if we thus, RHZ can be estimated without a priori knowledge of effective
consider Case 3 (qo =200 STB/D and s =0) (Fig. 1), it is clear that or relative permeabilities. Throughout, the values of ct and ILo

SPE Formation Evaluation, June 1990 137

 0.70 10'---'---~----r---~---r---'----.---,

18 o MATERIAL BALANCE - SIMULATOR, DRAW DOWN

Z 0.66 o COMPUTED FROM
0 DRAW DOWN 0
Eia::
::I
0.62
~
II)
0 CASE 3
..I
(5 SIMULATOR
0.58 qo =200 STBID
° CASE 3
ILl s:0
Co) 0
qo=200STB/D
~ s=O
0 0.54 °
Z
c( 0 A COMPUTED FROM BUILDUP, RHI
II)
o COMPUTED FROM BUILDUP, RH2
0.50°
2400 2800 3200 3600 4000 2~00~~~--~2~8~00~--~~32~00~--~~3600~~~--4~OOO~
WELLBORE PRESSURE. Pwf OR Pws' psi PRESSURE. P. psi

Fig. 7-Approximatlon of sandface saturation/pressure rela- Fig. 8-Effectlve 011 permeability as a function of pressure
tion for Case 3. for Case 3.

needed in Eqs. 29 through 31 are evaluated from the instantaneous relation. Because we assume that the effective or relative permea-
wellbore values of shut-in pressure and saturation, where the value bilities as functions of oil saturation are not known a priori,
of So at the sandface is estimated by solving Eq. 17. In practice, Raghavan'slO procedure cannot be used. Eq. 14 provides the sim-
the application of Eq. 17 requires that the sandface value of So at plest procedure we have found for estimating the required drawdown
the instant of shut-in be known; however, we may be able to es- saturation/pressure relation directly from pressure-buildup data.
timate this So value using drawdown-pressure data to solve Eq. The Case 3 results in Fig. 7 show that Eq. 14 gives a reasonably
10. I Appendix B in Ref. 13 shows that Eq. 27 is a good approxi- accurate estimate of the drawdown oil saturation/pressure relation
mation except at very early shut-in times. at later shut-in times (higher values of shut-in pressure). The solid
curve in Fig. 7 represents the drawdown saturation/pressure re-
Computing Effective Permeabilities lation obtained from the simulator, and circular data points rep-
resent the saturation/pressure relation from Eq. 14. Note that at
Using basic chain rules from calculus, we can differentiate Eq. 26
low pressure values, Eq. 14 does not yield extremely accurate So
with respect to flt and then rearrange the differentiated equation values, but at high pressure values it gives excellent estimates of
to get
So. Because the semilog straight line obtained for buildup analysis
kk ro = -162.6qoJLoBo/{h[dpwsld 10g(RHI)]}' .......... (32) always corresponds to high values of shut-in pressure (see Figs.
B-1 and B-2 in Ref. 13), Eq. 14 should provide reasonably accurate
If fltnD is used, then RH2 replaces RHI in Eq. 32. Based on the So values corresponding to the pressure range where Eqs. 32 and
results from correlating PpsD with the liquid solution (Ref. 13), we 33 yield accurate effective-permeability values. Results similar to
expect that using RH2 in Eq. 32 will yield the most accurate results. those shown in Fig. 7 were also obtained for Cases 1 and 2 (see
After kk ro has been computed from Eq. 32, the effective gas per- Ref. 6).
meability can be approximated by the following rearrangement of Fig. 8 shows Case 3 results for drawdown and buildup. The solid
Eq. 18:
curve is the value of effective oil permeability at the sandface ob-
kkrg =kkro[R(flt =0) - Rs ][(JLg/Bg)/(JLo IBo)). . ......... (33) tained from the simulator during pressure drawdown. Circular data
points represent the kk ro values from the Ref. 1 drawdown analog
Application of Eqs. 32 and 33 yields effective permeabilities as of Eq. 32. Square data points represent the values of effective oil
functions of pressure. If we can also estimate So as a function of permeability from Eq. 32 using R H2 . Triangular data points are
pressure, then we can combine the two results to obtain effective the values of effective oil permeability from Eq. 32 based on R HI .
permeabilities as functions of oil saturation. From our earlier su-
As expected, using RH2 in Eq. 32 yielded slightly more accurate
perposition arguments, the effective-permeability values computed
results than RH\. Note that the buildup and drawdown results are
from Eqs. 32 and 33 should correspond to drawdown values. To complementary; the drawdown results yield good estimates of kk ro
obtain effective permeabilities as functions of saturation, we must at low pressure values and the buildup computation yields good es-
be able to obtain the corresponding drawdown pressure/saturation
timates at high pressure values. Thus, by combining drawdown and
buildup test results, we should get a good definition of the effective
oil permeability as a function of pressure for the pressure range
10 observed during the buildup test.
~ - SIMULATOR, DRAW DOWN The Case 2 results in Fig. 9 are similar to the results in Fig. 8.
t: o COMPUTED FROM DRAWDOWN Here, use of Hawkins'S formula to incorporate the skin factor
..I 8 A COMPUTED FROM BUILDUP, RHI
iii causes inaccurate results during time periods when the change in
c( o COMPUTED FROM BUILDUP. RH2
ILl
CASE 2 H8 pseudopressure across the skin zone is not stabilized. Note, however,
~"O
a::e_
ILl 6 qo - 100 STBID that even in this case, drawdown and buildup computations give
Ho
H
_...
Q.
..I
0 ...
0
~

4 ~
0

" r:f
0

0
good estimates of the effective oil permeability at low and high
pressures, respectively.
UJ 0 0 0 0 The square data points in Fig. 10 represent kk ro values from Eq.
> 0
i=
"0 0 " 32 using RH2 plotted vs. the corresponding So value from Eq. 14.
Co)
UJ
u..
2
,.1,..-5-2.59 ,,0

"
0
0 0
0
0 '6
0
Circular data points are kk ro values from pressure-drawdown data
u..
ILl
k,/k-0.16 " " " (using the Ref. 1 analog of Eq. 32) plotted vs. So values obtained
0 by solving Eq. 10. The solid curve represents the actual kkro-vs.-
2400 2800 3200 3600 4000 So relation. Fig. 10 pertains to Case 3, but as shown in Fig. 11,
PRESSURE. P. psi
roughly similar results were obtained for Case 2. Note that from
a practical viewpoint, the results are quite good for both cases. Ex-
Fig. 9-Effectlve oil permeability as a function of pressure
for Case 2. cellent estimates are obtained, particularly at the lowest and highest
So values. Equally good estimates of effective gas permeability vs.

138 SPE Formation Evaluation, June 1990

 '1:J 7 '1:J 10
E ~
E ~

CASE 2
~
0
... CASE 3
qo = 200 STBID ~
0
~ 9
qo =100 STBID, s =5 0
~ .Ie
=0 8 rs/rw =2.59, ks/k =0.16 0
~ S ~
t:: 0 DRAW DOWN ~
..J
6 ::J 7 o DRAW DOWN
m
«
0 BUILDUP, RH2
0
as
« o BUILDUP, RH2
L&J 0 L&J 6
~ ~
a: 0 a:
L&J L&J 5
Q. Q.
...J ...J 0
0
0 5 0 4
I&J L&J 0 0
> > 3 0
~ ~ 0 0 0
(,) 0 0
L&J 0 LIJ 2 0 0
lA- lA-
lA-
o 0 0
lA-
L&J 4 LIJ 1
0.60 0.62 0.64 0.66 0.68 0.70 0.56 0.58 0.60 0.62 0.64 0.66 0.68 0.70
OIL SATURATION, So OIL SATURATION, So

Fig. 10-Effectlve 011 permeability as a function of 011 satura- Fig. 11-Effective 011 permeability as a function of oil satura-
tion for Case 3. tion for Case 2.

oil saturation can be constructed from Eqs. 33 and 14. Case 1 re- ration for the pressure and oil-saturation ranges observed during
sults are even better than the results shown in Figs. 10 and 11 (see pressure drawdown and buildup. Combining drawdown and buildup
Ref. 6). Tables C-l and C-2 in Ref. 13 give the buildup data plotted analysis yields a reasonably good definition of part of the curve
in Figs. 10 and 11. defining effective permeability as a function of oil saturation.
Given results like those in Figs. 10 and 11, we can construct a 2. A reasonable estimate of the skin factor can be obtained from
reasonable estimate of part of the effective oil permeability curve a semilog plot of the shut-in wellbore pressure squared vs. a Horner
by smoothing or fitting the data with effective or relative permea- time ratio (see Appendix D in Ref. 13 and Refs. 6 and 23).
bility correlations. As in the drawdown case, I an estimate of ab- 3. Arguments and numerical results presented indicate that su-
solute permeability could be obtained by assuming that perposition equations based on linear problems can be applied, at
kro "" So-i.e., by dividing a kk ro estimate by the corresponding So least approximately, to solution-gas-drive reservoirs If approxi-
estimate. From the buildup results in Fig. 10, we have kk ro =6.7 mately the same saturation/pressure relation is used in all pseu-
md at So=0.693. The ratio of these two values gives k""9.67 md, dopressure functions involved in the superposition equations.
whereas the correct value of absolute permeability is 10 md. The
same procedure applied to the buildup data of Case 2 (Fig. 11) gave Nomenclature
k=9.96 md, and for the results of Case 1 (not shown), we obtained a = RskrofILoBo+krgllLgBg, sef/RB-cp
k=9.87 md. Note that by dividing the effective permeabilities ob- b = RsSoIBo+SgIBg, sef/RB
tained by these associated absolute permeability values, we can B = FVF, RBlscf and RB/STB
obtain an approximation to part of the relative permeability curves. c( = total (system) compressibility, psi- I
h = total reservoir thickness, ft
Remarks
k = absolute permeability, md
The specific results presented in this paper pertain to radial-flow km = effective permeability of Phase m, md
problems for solution-gas-drive reservoirs; however, we also expect krm = relative permeability of Phase m, fraction
the methods to apply to gas-condensate radial-flow problems. ks = value of absolute permeability in skin zone, md
The results were restricted to infinite-acting systems primarily
P = pressure, psi
because we did not· want buildup results to be influenced by
Pb = bubblepoint pressure, psi
boundary effects. However, Refs. 2 and 6 indicate that these
Pp = pseudopressure function, STB-psi/RB-cp
procedures can also be applied in an approximate sense to buildup
PpsD = dimensionless shut-in pseudopressure defined by Eq. 5
data obtained after a long producing time.
PpsD = dimensionless shut-in pseudopressure defined by Eq. 6
If a semilog plot of Pws vs. RHi (classic liquid analysis) exhibits
a semilog straight line, then dpwsld 10g(RHl ) is constant. Eq. 32 PpwD = dimensionless wellbore pseudopressure
indicates that this derivative will be constant if and only if Ppwf = wellbore pseudopressure function defined by Eq. 3,
krol(BoJl.o) is constant, which generally is not true. The variation STB-psi/RB-cp
in this term may be small, however, if the pressure changes during Ppwf,s = dimensionless wellbore pseudopressure at instant of
a test are sufficiently small-how small depends on the specific rock shut-in
and PVT properties. Pwf = flowing wellbore pressure, psi
Finally, Evinger and Muskat 22 also noted that Eq. 18 gave Pwf,s = wellbore pressure at instant of shut-in, psi
krglk ro as a function of pressure, and they also used a pseu- qg = free-gas flow rate, sefID
dopressure function for solution-gas-drive reservoirs. qo = oil flow rate, STB/D
QD = dimensionless cumulative rate defined by Eq. 11
Conclusions r = radial coordinate or distance, ft
1. Given the oil and gas flow rates and the wellbore pressure as re = external reservoir radius, ft
functions of time, it is possible to estimate the effective oil and gas rs = radius of skin zone, ft
permeabilities as functions of pressure and as functions of oil satu- r w = well bore radius, ft
SPE Formation Evaluation, June 1990 139
 R = producing GOR, scf/STB 1987 SPE Annual Technical Conference and Exhibition, Dallas, Sept.
RHI = primary Horner time ratio defined by Eq. 28 27-30.
RH2 = secondary Horner time ratio defined by Eq. 29 5. Hawkins, M.F. Jr.: "A Note on the Skin Effect," Trans., AlME (1956)
207, 356-57.
Rr = in-situ GOR defined by Eq. 1, scf/STB 6. Serra, K.: "Well Testing for Solution Gas Drive Reservoirs," PhD
Rs = dissolved GOR, scf/STB dissertation, U. of Tulsa, Tulsa, OK (June 1988).
Rsb = dissolved GOR at Pb, scf/STB 7. Ramey, H.J. Jr.: "Rapid Methods for Estimating Reservoir Compres-
Rsbi = dissolved GOR at Pbi, scf/STB sibilities," JPT (April 1964) 447-54; Trans., AIME, 231.
s = skin factor from damage or stimulation 8. Martin, J.C.: "Simplified Equations of Flow in Gas Drive Reservoirs
S = saturation, fraction and the Theoretical Foundation of Multiphase Pressure Buildup
Analyses," Trans., AIME (1959) 216, 309-11.
Sgc = critical gas saturation, fraction
9. Aanonsen, S.: "Nonlinear Effects During Transient Fluid Flow in
Siw = irreducible water saturation, fraction Reservoirs as Encountered in Well Test Analysis," PhD dissertation,
t = time, days U. of Bergen, Norway (1985).
tD = dimensionless time defined by Eq. 7 10. Raghavan, R.: "Well Test Analysis: Wells Producing by Solution Gas
tit = shut-in time, days Drive," SPEl (Aug. 1976) 196-208.
titD = dimensionless shut-in time defined by Eq. 8 1 1. B~e, A., SIgaeveland, S.M., and whitson, C.H.: "Two-Phase Pressure
Test Analysis," SPEFE (Dec. 1989) 604-10; Trans., AIME, 287.
titnD = dimensionless normalized shut-in time defined by
12. Fetkovich, M.J.: "The Isochronal Testing of Oil Wells," paper SPE
Eq. 30 4529 presented at the 1973 SPE Annual Technical Conference and Ex-
ZD = dimensionless Boltzmann variable hibition, Las Vegas, Sept. 30-0ct. 3.
ex = krolp.oBo, STB/RB-cp 13. Serra, K.V., Peres, A.M.M., and Reynolds, A.C.: "Supplement to
(3 = SolBo' STB/RB SPE 17048, Well-Test Analysis for Solution-Gas-Drive Reservoirs: Part
'Y = Euler's constant, 0.57722 2-Buildup Analysis," paper SPE 20376 available from SPE Book Order
Dept., Richardson, TX.
A = mobility, cp - 1
14. Peres, A.M.M. et al.: "Well Conditioning Effects on Bubblepoint
Am = mobility of Phase m, cp-I Pressure of Fluid Samples From Solution-Gas-Drive Reservoirs," paper
At = total mobility, cp - 1 SPE 18530 presented at the 1988 SPE Eastern Regional Meeting,
P.m = viscosity of Phase m, cp Charleston, WV, Nov. 1-4.
c/> = porosity, fraction 15. Stright, D.H. Jr. et al.: "Carbon Dioxide Injection Into Bottom-Water,
Undersaturated Viscous Oil Reservoirs," JPT (Oct. 1977) 1248-58.
16. Whitson, C.H.: "Topics on Phase Behavior and Flow of Petroleum
Subscripts Reservoir Fluids," PhD dissertation, U. of Trondheim, Norway (1983).
bu = buildup 17. Whitson, C.H.: "Reservoir Well Performance and Predicting Deliver-
dd= drawdown ability," paper SPE 12518 available at SPE, Richardson, TX.
D = dimensionless 18. Peres, A.M., Serra, K.V., and Reynolds, A.C.: "Toward a Unified
g= gas Treatment of Well Testing for Nonlinear-Radial-Flow Problems With
i
= initial Applications to Interference Tests," SPEFE (June 1990) 151-60.
19. Reynolds, A.C., Bratvold, R.B., and Ding, W.: "Semilog Analysis
j = 1 or 2
of Gas Well Drawdown and Buildup Data," SPEFE (Dec. 1987) 657-70.
m = phase, 0 or g 20. Agarwal, R.G.: "Real Gas Pseudo-Time-A New Function for Buildup
0= oil Analysis of MHF Gas Wells," paper SPE 8279 presented at the 1979
ws = shut-in SPE Annual Technical Conference and Exhibition, Las Vegas, Sept.
23-26.
Superscript 21. Scott, J.O.: "Application ofa New Method for Determining Flow Char-
acteristics of Fractured Gas Wells in Tight Sands," paper SPE 7931
= average
presented at the 1979 SPE Symposium on Low-Permeability Gas
Reservoirs, Denver, May 20-22.
Acknowledgments 22. Evinger, H.H. and Muskat, M.: "Calculation of Theoretical Produc-
Financial support for the graduate studies of Kelsen Serra and Alvaro tivity Factor," Trans., AIME (1942) 146, 126-39.
Peres was provided by Petrobnis-Petroleum of Brazil. Financial
support was also provided by the Dept. of Petroleum Engineering, 51 Metric Conversion Factors
U. of Tulsa. bbl x 1.589873 E-OI m3
cp x 1.0* E-03 Pa's
References cp-I x 1.0* E+03 (Pa's)-I
ft x 3.048* E-OI m
I. Serra, K.V., Peres, A.M.M., and Reynolds, A.C.: "Well-Test Analysis
for Solution-Gas-Drive Reservoirs: Part I-Determination of Relative
ft3 x 2.831 685 E-02 m3
and Absolute Permeabilities," SPEFE (June 1990) 124-32. md x 9.869233 E-04 p.m 2
2. Serra, K.V., Peres, A.M.M., and Reynolds, A.C.: "Well-Test Analysis psi x 6.894757 E+OO kPa
for Solution-Gas-Drive Reservoirs: Part 3-A Unified Treatment of the psi-I x 1.450377 E-OI kPa- 1
Pressure-Squared Method," SPEFE (June 1990) 141-50.
3. AI-Khalifah, A.J., Home, R.N., and Aziz, K.: "In-Place Determination • Conversion factor is exact. SPEFE
of Reservoir Relative Permeability Using Well Test Analysis," paper Original SPE manuscript received for review Oct. 21, 1987. Paper accepted for publication
SPE 16774 presented at the 1987 SPE Annual Technical Conference Oct. 20, 1989. Revised manuscript received Sept. II, 1989. Paper (SPE 17048) first
and Exhibition, Dallas, Sept. 27-30. presented at the 1987 SPE Eastern Regional Meeting held in Pittsburgh, Oct. 21-23.
4. AI-Khalifah, A.J., Aziz, K., and Home, R.N.: "A New Approach to SPE 20376, "Supplement to SPE 17048, Well·Test Analysis for Solution-Gas·Drive
Multiphase Well Test Analysis," paper SPE 16743 presented at the Reservoirs: Part 2-Buildup Analysis," available from SPE Book Order Dept.

140 SPE Formation Evaluation, June 1990