Condensate-Bank Characterization From

Well-Test Data and Fluid PVT Properties

Manijeh Bozorgzadeh, SPE, and Alain C. Gringarten, SPE, Imperial College London

Summary in which condensation kills gas production completely (Engineer

Published well-test analyses in gas/condensate reservoirs in which 1985). Capillary number effects reduce the productivity loss
the pressure has dropped below the dewpoint are usually based on caused by the condensate bank, and their impact must be ac-
a two- or three-region radial composite well-test interpretation counted for when predicting well productivity.
model to represent condensate dropout around the wellbore and In a well test (Gringarten et al. 2000; Daungkaew et al. 2002),
initial gas in place away from the well. Gas/condensate-specific the three mobility zones around the well yield a three-zone radial
results from well-test analysis are the mobility and storativity ra- composite behavior when real-gas pseudopressures (Chu and
tios between the regions and the condensate-bank radius. For a Shank 1993) are used. Real-gas or single-phase pseudopressure
given region, however, well-test analysis cannot uncouple the stor- linearizes the diffusivity equation and treats gas as the dominant
ativity ratio from the region radius, and the storativity ratio must be phase and condensate as a fluid heterogeneity (whereas two-phase
estimated independently to obtain the correct bank radius. In most pseudopressure replaces the gas/condensate two-phase fluid by a
cases, the storativity ratio is calculated incorrectly, which explains single fluid equivalent, resulting in homogeneous well-test behav-
why condensate bank radii from well-test analysis often differ ior). The derivative pressure response displays three stabilizations
greatly from those obtained by numerical compositional simulation. at different levels, corresponding to the different zone mobilities
In this study, a new method is introduced to estimate the stor- (Fig. 1, Curve b).
ativity ratios between the different zones from buildup data when The first stabilization, caused by near-wellbore capillary num-
the saturation profile does not change during the buildup. Appli- ber effects, develops as soon as condensate becomes mobile in the
cation of the method is illustrated with the analysis of a transient- reservoir but disappears as production continues (because oil satu-
pressure test in a gas/condensate field in the North Sea. The analy- ration increases and gas effective permeability decreases faster in
sis uses single-phase pseudopressures and two- and three-zone the near-wellbore region than elsewhere in the reservoir), and it
radial composite well-test interpretation models to yield the con- eventually becomes less than in the condensate bank (see Appen-
densate-bank radius. The calculated condensate-bank radius is dix A). When this occurs, a two-zone radial composite behavior is
validated by verifying analytical well-test analyses with composi- obtained, with only two derivative stabilizations (Fig. 1, Curve a).
tional simulations that include capillary number and inertia effects. Interpretation of gas/condensate pressure-transient data that ex-
hibit a two- or a three-region radial composite behavior yields the
effective reservoir permeability, the wellbore skin effect, and the
Introduction and Background
mobility or permeability ratios between the various regions [using
When the bottomhole flowing pressure falls below the dewpoint in straight-line analysis (Brown 1985)—see Appendix B—or type-
a gas/condensate reservoir, retrograde condensation occurs, and a curve matching (Olarewaju and Lee 1989)]. The storativity ratio
bank of condensate builds up around the producing well. This between two adjacent regions, however, cannot be uncoupled from
process creates concentric zones with different liquid saturations the radius because analytical well-test-analysis methods can only
around the well (Fevang and Whitson 1996; Kniazeff and Nvaille account for one set of pressure/volume/temperature (PVT) data,
1965; Economides et al. 1987). The zone away from the well, yet both are needed to estimate the total skin effect. The storativity
where the reservoir pressure is still above the dewpoint, contains in the condensate dropout region must therefore be estimated in-
the original gas. The condensate bank around the wellbore contains dependently to obtain the bank radius. This is not easy because of
two phases, reservoir gas and liquid condensate, and has a reduced two-phase flow and compositional changes (Novosad 1996). Ex-
gas mobility, except in the immediate vicinity of the well at high perience from the senior author of the present paper indicates that
production rates, where the relative permeability to gas is greater the storativity ratio is almost always underestimated, resulting in
than in the bank because of capillary number effects (Danesh et al. an overestimation of the condensate-bank radius and, conse-
1994; Boom et al. 1995; Henderson et al. 1998; Mott et al. 1999).
The capillary number,

␯g Ⲑ o␮g Ⲑ o
NC = , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)
␴

represents the ratio of viscous to capillary forces. In Eq. 1, ␯g/o is

the superficial velocity of gas/oil, ␮g/o is the viscosity of gas/oil,
and ␴ is the gas/oil interfacial tension. Capillary number effects are
also called positive coupling or viscous stripping effects.
The condensate bank grows as the reservoir pressure declines
and progressively impedes the flow of gas to the well, causing a
loss of well productivity (Fussell 1973; Afidick et al. 1994). This
has been well documented in the literature, with numerous field
examples of gas/condensate reservoirs that experience a sharp drop
in gas-well deliverability (Barnum et al. 1995), with extreme cases

Copyright © 2006 Society of Petroleum Engineers

This paper (SPE 89904) was first presented at the 2004 SPE Annual Technical Conference
and Exhibition, Houston, 26–29 September, and revised for publication. Original manuscript
received for review 7 June 2004. Revised manuscript received 7 December 2005. Paper Fig. 1—Schematic of pressure and derivative composite behav-
peer approved 26 July 2006. ior (Gringarten et al. 2000).

596 October 2006 SPE Reservoir Evaluation & Engineering

quently, of the well potential productivity loss. In marginal fields,
this may lead to the abandonment of a project that actually could
be viable.
The main objective of this paper is to present a new method,
based on the determination of the saturation profile during build-
ups, to estimate the storativity ratios between the different zones
from well-test and PVT data so that the radii can be calculated
correctly. Confident evaluation of the radius of the condensate
bank helps engineers estimate and predict well productivity, plan
infill drilling, and select and design the best remedial actions, such
as fracturing (Aly et al. 2001), changing the reservoir rock wetta-
bility (Li and Firoozabadi 2000), or gas cycling (Allen and Roe
1950). The method is first illustrated with computer-simulated
buildup pressure-transient data and then applied to an actual well
test from a gas/condensate field in the U.K. North Sea.

Determination of Condensate Saturation Profiles

in Buildup Tests
Well-test analysis is usually performed on buildup data because a maximum liquid dropout of 2.1 vol%, whereas Fluid B is a rich
drawdown data tend to be affected by rate fluctuations and well- gas with a maximum liquid dropout of 10.5 vol%. The composi-
bore dynamics. Hence, we need to be able to calculate the stora- tions of Fluids A and B are given in Table 3. The Whitson et al.
tivity ratios and the condensate-bank radius from buildup data. It (1983) and Coats and Smart (1986) techniques were used for split-
has been shown (Jones et al. 1989) that reservoir parameters during ting and lumping fluid components. Two equation-of-state (EOS)
shut-ins in gas/condensate reservoirs can be estimated if the oil- models for Fluids A and B were constructed using the Soave-
saturation profiles are known. We therefore need to determine the Redlich-Kong (SRK) EOS and the Peng-Robinson (PR) EOS, re-
condensate saturation after shut-in, which can reach 100% accord- spectively. The EOS parameters were tuned by regression to obtain
ing to some authors (Novosad 1996; Whitson et al. 1999). a good match between predicted values and observed data. The
In the present study, we used compositional simulation with variables that were regressed upon in the tuning process were the
capillary number and inertia effects under different production critical properties of the pseudocomponents because they were less
conditions to determine the various condensate saturation profiles well defined. Figs. 2 and 3 show a good match between the ob-
that may exist in the near-wellbore region at the end of a shut-in served liquid saturations in the constant composition expansion
and established a theoretical justification for the calculation of (CCE) and constant volume depletion (CVD) experiments and the
storativity ratios. In all cases, the initial reservoir pressure was set liquid saturations calculated with the SRK EOS for Fluid A and the
just above the dewpoint pressure, so that a liquid-phase condensate PR EOS for Fluid B, respectively.
formed at the start of production. Irreducible water was assumed to To investigate the impact of oil/gas relative permeability on the
be immobile. Capillary number and non-Darcy effects were ob- radius of the gas/condensate bank, three sets of imbibition relative
tained from correlations developed by Henderson et al. (1998, permeabilities (Models 1, 2, and 3) were generated with Corey’s
2000a, 2000b) and Geertsma (1974), respectively. formula (Liu et al. 2001). These models differed only in the critical
oil saturation, which was fixed at 0.1, 0.2, and 0.3, respectively
Grid Definition. A finely gridded 1D radial, fully compositional (Fig. 4). The relative permeabilities of gas and oil decrease as the
model was used, with no wellbore storage, mechanical skin effect, critical oil saturation increases.
or flow within the producing string. The well was located at the Although capillary number and inertia (non-Darcy) effects are
center, and 62 cells were used in the radial direction, with an always present, runs were also made with (NC) and without (nNC)
innermost grid size of 0.1 ft and four outer cells of 2,000 ft to capillary number and with (non-Darcy) and without inertia effects
eliminate any boundary effects (Table 1). to evaluate their individual impacts. The impact of a threshold or
This high-resolution grid arrangement combined with small base capillary number on saturation profiles at the end of the
timesteps yielded linear pressure gradients and smooth saturation buildup periods was also investigated.
profiles on a semilog scale. Parameters used for simulation are
listed in Table 2. Simulation Results. Pressure and rate histories for a typical simu-
lation run are shown in Fig. 5. There are four periods, alternately
Fluid Characterization and Relative Permeability Modeling. A
key component in modeling gas/condensate reservoirs is the de-
velopment of a representative fluid model. We used two commer-
cial fluid PVT packages to generate PVT models for two real
retrograde gases, referred to as A and B. Fluid A is a lean gas with

October 2006 SPE Reservoir Evaluation & Engineering 597

Fig. 2—Match of liquid saturation from the CCE experiment for
the lean-gas Fluid A.
Fig. 3—Match of liquid saturation from the CVD experiment for
the rich-gas Fluid B.
drawdowns and buildups, labeled 1DD, 2BU, 3DD, and 4BU. The
duration of each period is 10 days.
Laboratory PVT tests of CVD and CCE of gas condensate As can be seen in Figs. 6 and 7, this is independent of the oil
show that the condensate drops out from the gas as soon as the critical saturation and fluid type.
gas/condensate pressure falls below the dewpoint. Further pressure High Production Rates. During a buildup following a draw-
drop causes an increase in condensate saturation until a maximum down at high production rate and without capillary number effects,
value is reached. Any further reduction of the pressure leads to oil saturation increases because of liquid swelling from gas disso-
revaporization of the condensate. lution (Novosad 1996). Hence, the oil-saturation profile during the
One would expect that in a buildup, the condensate would buildup deviates from that in the preceding drawdown. This dis-
revaporize owing to the increase in reservoir pressure and that the crepancy increases with increasing oil saturation in the reservoir
oil saturation would decrease. In reality, as documented by previ- and is much higher for a rich gas than for a lean gas (Fluids B and
ous authors (Fevang and Whitson 1996; Fussell 1973; Novosad A, respectively, in Fig. 8). It also increases with increasing critical
1996; Wheaton and Zhang 2000; Vo et al. 1989) and verified in oil saturation (Fig. 9) and production rate (Fig. 10).
our simulations, revaporization only occurs near the outer edge of The difference decreases, however, when the capillary number
the two-phase region and only in buildups that follow drawdowns effect is incorporated in the simulation model (Fig. 11), especially
at high production rates. In buildups following low production-rate at low value of the base capillary number (LBNC). At high base
drawdowns, oil revaporization in the two phase-region is very low capillary number (HBNC), the saturation profile is the same as with
because (1) excessive accumulation of condensate near the well- no capillary number effect because, at a constant production rate,
bore and pressure gradients prevent it (hysteresis effect) (Econo- the radius of the capillary number effect zone is inversely propor-
mides et al. 1987; Fussell 1973) and (2) the heavier components tional to the base capillary number and is zero above a certain
condense and are left in the reservoir as production continues, value of the base capillary number.
whereas the lighter components are produced. The total concen- Adding inertia to the capillary effect increases the pressure
tration of the heavy components in the reservoir fluid mixture drop, especially in the immediate vicinity of the wellbore, where
increases, and, thus, the dewpoint pressure may become signifi- gas velocity is highest and inertia effects control the oil saturation.
cantly higher than in the original reservoir fluid mixture (compo- Consequently, oil saturation increases significantly within 1 ft
sitional changes) (Whitson et al. 1999; Wheaton and Zhang 2000; around the well. This yields a much different oil-saturation profile
Vo et al. 1989). This process occurs throughout the reservoir but is during the subsequent shut-in, particularly near the well, where
more significant in the near-wellbore region. compositional changes are such that gas can behave like a black oil
Low Production Rates. At low production rates (Fig. 6 for lean and dissolve into the oil (Fig. 12). In this case, the condensate
gas Fluid A and Fig. 7 for rich gas Fluid B), the condensate saturation increases much more near the well than elsewhere in the
saturation profiles during a buildup and the preceding drawdown reservoir, and more so at high critical oil saturation. This saturation
are approximately the same if one neglects capillary number ef- increase, however, cannot be detected on a pressure-transient log-
fects, and they are exactly the same if these effects are included. log derivative plot because of the small zone extent.
High rates, however, cannot be sustained because the bottom-
hole pressure quickly reaches its minimum permissible value, un-
less the capillary number parameters are such that the relative

Fig. 5—Example of rate and pressure histories for simulation

Fig. 4—Relative permeability curves. runs.

598 October 2006 SPE Reservoir Evaluation & Engineering

Fig. 6—Impact of capillary number effect on a lean gas: satura-
tion profiles at the end of a buildup and of the preceding draw- Fig. 7—Impact of capillary number effect on a rich gas: satura-
down at low production rates with relative permeability models tion profiles at the end of a buildup and of the preceding draw-
1 and 2, with (NC) and without (nNC) capillary number effect. down at low production rates with relative permeability models
1 and 2, with (NC) and without (nNC) capillary number effect.

permeability curves become near-miscible. In such a case, the

discrepancy between the oil-saturation profiles at the end of the The saturation at the desired pressure then can be estimated by
drawdown and in the subsequent shut-in is negligible, and the comparing the calculated relative permeability ratio with experi-
maximum near-well condensate saturation decreases (Fig. 13). mental oil and gas relative permeabilities data measured at differ-
The above results clearly show that the leaner the gas and the ent oil saturations.
lower the production time, production rate, and critical saturation, 3. Calculating the oil saturation at the desired pressure using
the smaller the saturation-profile difference between a drawdown the analytical expression by Bøe et al. (1989) and the fluid PVT
and the subsequent buildup. Under these conditions, the total com- properties as a function of pressure:

冉 冊
pressibility of the two-phase region and, therefore, the storativity
ratio between the two-phase and gas regions can be calculated at dSg Sg dBg ␭g −Sg dBg SoBg dRs So dBo SgBo dRv
= + + − + ,
the pressure and saturation at the time of shut-in to obtain the dp Bg dp ␭t Bg dp Bo dp Bo dp Bg dp
condensate-bank radius from buildup data. These criteria most . . . . . . . . . . . . . . . . . . . . . . . . . . . (3)
often exist for drillstem testing (DST). However, it should be
where the total mobility ␭t is given by:
checked for longer-duration testing (i.e., production testing).
␭t = ␭g + ␭o = krg Ⲑ ␮g + kro Ⲑ ␮o. . . . . . . . . . . . . . . . . . . . . . . . . . . (4)
Calculation of Condensate-Bank Inner and
Outer Radii All other parameters are the same as before. Bøe et al. (1989)
The saturation at the time of shut-in can be estimated with one of demonstrated the validity of their formula with computer-
the following three methods: simulated drawdown and buildup tests. In Eq. 4, the total mobility
1. Using the EOS (Whitson and Torp 1983) tuned with experi- can be assumed equal to the gas mobility (␭g/␭t=1) because ex-
mental data. periments show that the oil mobility is about one-hundredth of the
2. Calculating the gas/oil relative permeability ratio at the de- gas mobility in a gas/condensate reservoir, and its impact on this
sired pressure with the steady-state formula by Fetkovich et al. ratio is negligible (Gondouin et al. 1967).
(1986), where the fluid PVT properties are known as a function of The second method is more accurate than the other two for
pressure: evaluating the oil saturation, but it requires the relative permeabili-
ties of oil and gas to be known. The first method can be used only
krg
kro
共p兲 = 冉
Rp − Rs ␮gBgd
1 − rsRp ␮oBo 冊
, . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)
for a test of short duration in which the near-wellbore condensate
saturation is below the critical oil saturation. The initial fluid PVT
properties can be used in that case because fluid sample and pres-
where krg/o and ␮g/o⳱gas/oil relative permeabilities and viscosi- sure-vs.-time well-test data are acquired at the same time, and fluid
ties, respectively; B o / g d ⳱oil/dry-gas formation factors; composition does not vary over the short test duration. The third
Rp⳱producing gas/oil ratio (GOR); Rv⳱solution oil/gas ratio
(OGR); and Rs⳱solution GOR.

Fig. 9—Impact of critical oil saturation on a rich gas without

Fig. 8—Impact of fluid type: oil-saturation profiles at the end of capillary number effects: oil-saturation profiles at the end of a
a buildup and of the preceding drawdown at high production buildup and of the preceding drawdown at high production rate,
rate without capillary number effect. with relative permeability models 1 (Soc=0.1) and 2 (Soc=0.2).

October 2006 SPE Reservoir Evaluation & Engineering 599

 Fig. 11—Impact of production rate on a lean gas with capillary
number effects: oil-saturation profiles at the end of a buildup
and of the preceding drawdown at high production rate, without
Fig. 10—Impact of production rate on a lean gas without capil-
capillary number (nNC) and with low (LBNC) and high (HBNC)
lary number effects: oil-saturation profiles at the end of a
base capillary number effects.
buildup and of the preceding drawdown at low (1DD-LPR) and
high (3DD-HPR) production rates.
reservoir-fluid experimental data, using the critical properties of
method provides the best estimation of saturation when oil and gas the pseudocomponents as variables in the tuning process.
relative permeability data are not available and for tests of long 2. Generate live-oil and wet-gas PVT tables as a function of
durations. In the latter case, the fluid composition changes with pressure over the pressure range of the test, using the tuned EOS
time, and the near-wellbore condensate saturation is best estimated and Whitson and Torp’s procedures (1983).
from the PVT properties of the produced wellstream at the time of 3. Generate a single gas pseudopressure m(p) and the corre-
production. sponding log-log plots of ⌬m(p) and derivatives:
When the derivative pressure response exhibits a three-region pws
␳g
兰 ␮ dp. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (5)
radial composite behavior (Curve b in Fig. 1), two radii must be
estimated: one for the inner, capillary number-enhanced gas- m共p兲 = 2
g
mobility zone, and the other for the condensate bank (intermediate pwf,s
zone). Both require an estimate of the corresponding storativity 4. Calculate the total compressibility for the two-phase region
ratios (i.e., the ratios of the total compressibilities). at the pressure and time of shut-in, using the PVT table generated
As discussed in the previous section, calculating the conden- in Step 2:
sate-bank outer radius (i.e., where the two-phase region starts)

再 冋 冉 冊 册冎
冢
Sg −dBg dRv B0 − RsBg

冣
requires the total compressibility ratio between the two-phase and
the gas regions at pressure and saturation at the time of shut-in. +
Bg dp dp 1 − RsRv

再 冋 冉 冊 册冎
The inner-zone radius, on the other hand, can be calculated by Ctc = 共1 − Sw兲.
taking the total compressibility in the inner zone to be equal to that So −dBo dRs Bg − RvBo
+ +
in the two-phase region. This is based on simulation results for a Bo dp dp 1 − RsRv
lean gas under various production rates and durations, which show
that the difference between the oil saturations in the immediate + SwCw + Cr. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (6)
vicinity of the wellbore and in the condensate bank is too small (10 In Eq. 6, only the mass exchange between the reservoir oil and
to 15% only) to make a difference in the total compressibilities and vapor phases is considered.
viscosities (Figs. 14 and 15). 5. Calculate the total compressibility for the outer gas region
The following procedure is therefore proposed to characterize a (where the reservoir pressure is above the dewpoint pressure) at
condensate bank from pressure buildup data: the average reservoir pressure
1. Construct an EOS model to predict the actual reservoir-fluid 6. Calculate the storativity ratio using the total compressibili-
properties using a PVT package. Tune the EOS parameters with ties from Steps 4 and 5:

Fig. 13—Impact of non-Darcy flow on a lean gas: saturation

Fig. 12—Impact of non-Darcy flow on a lean gas: saturation profiles at the end of a buildup following a high-flow-rate draw-
profiles at the end of a buildup following a high-flow-rate draw- down for relative permeability model 2 (Soc=0.2): effect of mis-
down at different critical oil saturation (relative permeability cible (NC) and more miscible (Mis.NC) relative permeability
models 1, Soc=0.1 and 2, Soc=0.2) with capillary number effects. curves.

600 October 2006 SPE Reservoir Evaluation & Engineering

Fig. 14—Effect of production rate on the oil-saturation profile
after 1 day of production for the lean-gas Fluid A.
Fig. 15—Effect of the production time on the oil-saturation pro-
ct1 file for the lean-gas Fluid A with capillary number effect (15
共␾ h ct兲1 Ⲑ 2 = . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (7) MMscf/D).
ct2
7. Assume the same compressibility and viscosity ratios for the
inner and intermediate regions in the case of a three-zone radial set equal to unity (Step 7). The other parameters and, in particular,
composite behavior. the inner and outer radii of the condensate bank are then obtained
8. Calculate the two-phase viscosity using the PVT table gen- by matching the data with a three-region radial-flow composite
erated in Step 2 and the pressure at the time of shut-in: model (Table 4 and Fig. 16). With the calculated storativity ratios,
the inner and outer radii from well-test analysis compare well with
␮two-phase = So ␮o + 共1 − So兲 ␮g. . . . . . . . . . . . . . . . . . . . . . . . . . . (8) the actual values, which can be obtained from the condensate
9. Determine the mobility ratios from the derivative radial flow saturation profile in Fig. 17 (3 vs. 5 ft and 7 vs. 10 ft, respectively).
stabilization lines, The analysis of the buildup (BU2) following the drawdown at

冉冊
high production rate* is shown in Fig. 18. Only two stabilizations
kh 共k1 Ⲑ ␮1兲inner zone are visible, for the reasons given in the Introduction (the difference
= . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (9)
␮ 1Ⲑ2 共k2 Ⲑ ␮2兲two-phase in gas effective permeabilities in the velocity stripping zone and in

冉冊
the condensate bank decreases with increasing production time). A
kh 共k2 Ⲑ ␮2兲two-phase two-region composite model is therefore selected for interpreta-
and = , . . . . . . . . . . . . . . . . . . . . . . . . . . . (10)
␮ 2Ⲑ3 共k3 Ⲑ ␮3兲gd tion, and the corresponding parameters are obtained by matching
and calculate the corresponding effective permeabilities using the (Table 5). There is again good agreement between the condensate-
viscosity calculated in Step 8 for the inner and two-phase regions bank radius obtained with the calculated storativity ratios and the
and the viscosity at the average reservoir pressure for the dry-gas actual value determined from the condensate-saturation profile in
region. Fig. 19 (20 vs. 35 ft).
10. Determine the condensate-bank radius by matching the We also investigated a case at high production rate in which the
m(p) derivative data with a three- or two-zone radial composite oil-saturation profile changes during the buildup, using Fluid B
model, as appropriate, using the calculated storativity ratio from with capillary number and relative permeability model 2
Eq. 7 and the mobility ratios from Step 9. (Soc⳱0.2). Fig. 20 shows the analysis of the last buildup BU4.
Only two stabilizations are visible on the derivative pressure re-
Application to Computer-Generated Data sponse. A two-region composite model was therefore selected for
interpretation, resulting in the parameters listed in Table 6. The
The procedure described in the previous section is first illustrated
condensate-bank radii obtained with the calculated storativity ra-
with buildup data from the simulated DST shown in Fig. 5. The
tios and the actual value determined from the condensate satura-
analysis is performed with single-phase pseudopressures using a
tion profile are 15 and 70 ft, respectively (Fig. 21). Such a sig-
three- or two-region radial composite model, as dictated by the
nificant difference is, as explained before, caused by the change in
log-log derivative shapes.
saturation profile during the buildup compared to that in the pre-
Fig. 16 shows the analysis of the buildup following the draw-
ceding drawdown (Fig. 21). However, this error may be reduced if
down at low rate (10 MMscf/D). The derivative exhibits three
the PVT properties of the produced wellstream at the time of
radial-flow stabilizations, indicating a three-region radial-flow
production are used for calculation.
composite behavior.* The last stabilization yields the effective
reservoir permeability (10 md). The ratio of the first to the second
Well-Test Analysis of Field Data
stabilizations yields the mobility ratio (kh/␮)1/2 between the inner
zone controlled by the capillary number and the two-phase con- The interpretation procedure presented in this paper is now applied
densate bank. Similarly, the ratio of the second to the third stabi- to a DST from a large North Sea lean gas/condensate reservoir
lizations yields the mobility ratio (kh/␮)2/3 between the two-phase (Daungkaew 2002). The gas/condensate radius from conventional
condensate bank and the rest of the reservoir, where the pressure analysis is compared to that obtained from pressure history match-
is still above the dewpoint. The storativity ratios are calculated ing with a compositional simulator.
according to the procedure in the previous section; that is,
(␾hct)2/3 is obtained as per Steps 4 through 6, whereas (␾hct)1/2 is Well-Test Interpretation Procedure. The pressure and produc-
tion-rate histories of the selected test are shown in Fig. 22. Flow-

max max
*In Fig. 16, a different relative permeability model (Corey’s parameters: krg =krog =1,
sgc=0.05, sorg=0.2, swc=0.4, ng=no=2.5) and 6 days of production are used to increase the *In Fig. 18, capillary number parameters are selected so that a more miscible relative
contrast between stabilization levels and make them easier to see. permeability is obtained in the velocity stripping zone (n1g/o=0.001).

October 2006 SPE Reservoir Evaluation & Engineering 601

 The interpretation procedure is the same as for the computer-
generated DST. The last derivative stabilization yields the effec-
tive reservoir permeability (2.3 md). The ratio of the derivative
stabilizations yields the respective mobility ratios, (kh/␮)1/2 and
(kh/␮)2/3. The storativity ratio (␾hct)2/3 is obtained from Steps 4
through 6 of the procedure described in this paper, and (␾hct)1/2 is
set equal to unity. The other parameters and the inner and outer
radii of the condensate bank are then obtained by matching the
data with a three-region radial-flow composite model (Table 9
and Fig. 28).
Fig. 28 also includes the response with the fault at 1,090 ft,
to confirm that that boundary is too far to have been seen during
the DST.
The permeability and initial pressure in Table 6 are in good
agreement with the core permeability and the measured initial
reservoir pressure, respectively.
Table 10 emphasizes the dependency of the condensate-bank
radius on the storativity ratio and the need for a procedure such as
the one presented in this paper to calculate the storativity ratio.
The above interpretation and, in particular, the condensate-
bank radius is verified with compositional simulation in the fol-
lowing section.

Compositional Simulation Model. A Cartesian grid (37×45×1)

with radial local grid refinement (LGR) was selected to include the
Fig. 16—Analysis of Fluid A well-test data simulated without fault boundary (Fig. 29). The radial local grid consisted of 19 cells
non-Darcy effects and with capillary number effects, low pro- with logarithmically increasing sizes away from the well to permit
duction rate (10 MMscf/D), and relative permeability model 1 evaluation of the near-wellbore behavior in more detail. Wellbore
(Soc=0.1): Log-log match (a), Horner match (b), and pressure-
history match (c) with a three-region composite model.

rate data are listed in Table 7, whereas the basic input data for
well-test analysis and the sources and origins of the data are given
in Table 8.
Log-log plots of buildups and drawdowns, normalized to the
rate of BU6, are shown in Figs. 23 and 24, respectively. All
drawdown data are affected by phase redistribution or condensate
unloading in the wellbore and are not interpretable.
BU6 was selected for analysis because it has the longest dura-
tion. Three derivative stabilizations can be identified on the log-log
graph of Fig. 25, around 0.7 hour, 3 hours, and just after 30 hours.
They correspond to the mobilities of the capillary number-
controlled inner zone, the two-phase condensate bank, and the gas
outer zone, respectively. The last stabilization is consistent with
the arithmetic average permeability from cores, a frequent occur-
rence in sandstone reservoirs (Gringarten et al. 2000). No bound-
ary effects can be seen on the log-log derivative or on the super-
position graph (Fig. 26) owing to the short duration of the test,
although the structure map of Fig. 27 indicates a single fault of
limited extent 1,090 ft away. Fig. 17—Condensate-bank radius for the data from Fig. 16.

602 October 2006 SPE Reservoir Evaluation & Engineering

 The relative permeability curves were selected from a set used
in the actual full-field compositional simulation based on the re-
ported fluid connate-water saturation (30%). The endpoint of the
gas relative permeability was corrected to obtain a gas effective
permeability consistent with that from the actual test data (Fig.
31).
Non-Darcy and Capillary Number Effects. As discussed ear-
lier, capillary number and inertia effects must be included in the
simulation.
The Forchheimer (1901) parameter ␤ that defines inertia was
obtained from Geertsma’s correlation (1974):
0.005
␤≈ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (11)
k ␾5.5S5.5
0.5

Capillary numbers can be obtained from a number of correlations

(Henderson et al. 2000b; App and Mohanty 2002; Henderson et al.
2001; Jamiolahmady et al. 2003; Mott et al. 2000; Mott 2003). In
this study, we used those of Henderson et al. (1998, 2000a, 2000b),
which include eight coefficients (n1o/g, mo/g, Ncbg/o, n2o/g) that
must be determined experimentally (Table 12).
Because no experimental data were available, parameters were
obtained as follows:
1. Calculate krg at different krg/kro values between 1 and 50, for
different capillary numbers, using the correlation presented by
Whitson et al. (1999), with ␣⳱4000 and n⳱0.7 as recommended
for the North Sea. The base gas relative permeability krgI for the
selected krg/kro is estimated by interpolating the measured relative
permeability data:
Fig. 18—Analysis of Fluid A well-test data simulated with non- 1

冉 冊
Darcy and capillary number effects, high production rate (20 krgM = kmax
rg −1
, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (12)
krg
MMscf/D), and relative permeability model 1 (Soc=0.1): log-log 1+
match (a), Horner match (b), and pressure-history match (c) kro
with a two-region composite model.

storage was not included. High-resolution timesteps were used,

which provided linear pressure gradients and smooth saturation
profiles on a semilog scale. Table 11 summarizes the simulation
input data. Permeability was set at 2.3 md, which is the core
arithmetic average.
The composition of a representative fluid sample is given in
Table 8. The modified PR EOS with three parameters was used to
predict the actual reservoir-fluid properties. The PR EOS param-
eters were tuned by regressing on the critical properties of the plus
fraction until a good match was obtained between predicted values
and observed data, such as dewpoint pressure, fluid density and
viscosity, fluid volume and composition, and liquid shrinkage dur-
ing CVD and CCE experiments. These variables were selected for
regression because the properties of the plus fraction were less
accurate and not well defined. Fig. 30 shows the comparison be-
tween the observed liquid saturation in the CVD experiment and
the calculated liquid saturation with the tuned EOS. Equally good Fig. 19—Condensate-bank radius comparison for the data of
matches were also obtained for other PVT data. Fig. 18.

October 2006 SPE Reservoir Evaluation & Engineering 603

 et al. method under the same conditions [i.e., at the same capillary
number and krg/kro (Fig. 32)].
Once the capillary number parameters are obtained, well skins
must be adjusted to obtain a good match for all the drawdown
periods on the pressure-history plot. These skins were found by
trial and error. Simulation results after skin adjustments are pre-
sented in Fig. 33.
A log-log comparison of model and data for the drawdown
DD5 is shown in Fig. 34, and the GOR comparison is displayed in
Fig. 35.
The log-log comparison shows a good match at late times.
Because wellbore storage effects were not included in the simu-
lation model, no match is to be expected at early times on the
log-log plot.

Condensate-Bank Radius. The inner and outer radii of the con-

densate bank can be obtained from the simulated oil-saturation
profile for flow period BU6. Fig. 36 shows the three separate
zones with different oil saturations that have been described at the
beginning of this paper. In Zone I, 38 ft from the wellbore, the
reservoir pressure is greater than the dewpoint pressure, and the
gas permeability is equal to the reservoir effective permeability.
The oil saturation increases rapidly in the intermediate zone (II),
while the gas relative permeability decreases. Finally, oil satura-
tion decreases and gas relative permeability increases in the inner
zone (III) near the well because of capillary number effects. The
saturation profile in flow period BU6 is reasonably similar to the
saturation profiles in flow period DD5 (Fig. 36), which is consis-
tent with the assumption made for well-test analysis.
According to the method proposed in this paper, the final pres-
sure during DD5 must be used as a reference for predicting the
Fig. 20—Analysis of Fluid B well-test data simulated with cap-
illary number effects, high production rate (15 MMscf/D), and
relative permeability model 2 (Soc=0.2): Log-log match (a), Hor-
ner match (b), and pressure-history match (c) with a two-region
composite model.

1
f= , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (13)
共␣.Nc兲n + 1
and krg = f.krgI + 共1 − f兲.krgM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (14)
2. Set mo and mg to zero on the basis of relative permeability
data (no residual oil saturation and low critical gas saturation equal
to 0.05).
3. Estimate the base capillary number (6×10−7), using the low-
pressure gas/oil surface tension and gas viscosity from the CVD
experiment.
4. Estimate n1g and n1o, and n2g and n2o , respectively, by run-
ning a sensitivity analysis and using a Mott et al. (2000) spread-
sheet model such that, for the selected krg/kro, a reasonable match
is obtained between krg calculated at capillary numbers greater Fig. 21—Condensate-bank radius comparison for the data of
than the base capillary number and krg calculated by the Henderson Fig. 20.

604 October 2006 SPE Reservoir Evaluation & Engineering

 Fig. 22—Pressure and rate histories for a North Sea well.

average near-wellbore oil saturation and calculating the fluid com-

pressibility during flow period BU6. Table 13 shows the calcu-
lated fluid compressibilities at different pressures using CVD ex-
perimental data and Eq. 6. The storativity ratio during BU6 was
calculated at 1,414 psi and used as an input to calculate the con-
densate-bank outer radius from well-test analysis. This yields inner
and outer radii of 7 and 30 ft, respectively, which is a good agree-
ment with the results obtained from simulation, (5 and 38 ft, re-
spectively) (Table 14).

Conclusions
We have introduced a method for estimating the condensate-bank
radius from buildup data, when the saturation profile does not
change during the buildup. The method uses the dry-gas
pseudopressure and an independent determination of the storativity
ratio between the oil/gas region around the well and the original
gas away from the well.
Compositional simulation shows that the saturation profile dur-
ing shut-in can be assumed to be the same as the saturation profile Nomenclature
at the time of shut-in for most drill stem testing and should be A ⳱ conversion factor
checked for production testing. Consequently, when analyzing a
B ⳱ formation volume factor
buildup below the dewpoint pressure, the storativity ratio between
the condensate bank and the reservoir in the resulting radial com- cr ⳱ rock compressibility
posite behavior must be calculated from the last pressure in the ct ⳱ total compressibility
preceding drawdown. This storativity ratio is equal to the total cw ⳱ water compressibility
compressibility ratio between the two zones, taking into account D ⳱ diffusivity ratio
the mass exchange between the reservoir liquid and vapor phases f ⳱ interpolation function for relative permeability
at reservoir conditions. The mobility ratio between two zones, on h ⳱ formation thickness
the other hand, is derived from the effective permeabilities using k ⳱ permeability
the derivative stabilizations and PVT data (assuming the derivative l ⳱ length
stabilizations can be identified in the derivative data). m(p) ⳱ single gas pseudopressure
The procedure to calculate the storativity ratio was applied to
M ⳱ mobility ratio
actual well-test data from a gas/condensate reservoir in the North
Sea. Conventional interpretation and the resulting condensate-bank n ⳱ exponent in equation for immiscibility factor
radius were verified by comparison with results from composi- NC ⳱ capillary number
tional simulation with capillary number and non-Darcy flow effects. p ⳱ pressure
A method for predicting the capillary number parameters in the pwf,s ⳱ pressure at the time of shut-in
absence of experimental data was also presented. pws ⳱ shut-in pressure

October 2006 SPE Reservoir Evaluation & Engineering 605

 Fig. 24—Log-log diagnostic plot of the drawdown periods nor-
Fig. 23—Log-log diagnostic plots of the buildup periods nor- malized to BU6.
malized to BU6.
cial support for this project as a part of a Joint Industry Project
r1/2 ⳱ condensate-bank inner or outer radius funded by the U.K. Dept. of Trade and Industry, Anadarko, Bur-
rw ⳱ well radius lington Resources, BHP Billiton, Britannia Operator Ltd., Cono-
R ⳱ front radius coPhillips, Gaz de France, and Total.
Rp ⳱ producing GOR
References
Rs ⳱ solution GOR
Rv ⳱ solution OGR Afidick, D., Kaczorowski, N.J., and Bette, S. 1994. Production Perfor-
mance of a Retrograde Gas Reservoir: A Case Study of the Arun Field.
S ⳱ saturation
Paper SPE 28749 presented at the SPE Asia Pacific Oil and Gas Con-
tDe ⳱ theoretical dimensionless deviation time ference, Melbourne, Australia, 7–10 November. DOI: 10.2118/28749-
tDx ⳱ theoretical dimensionless intersection time MS.
␯ ⳱ velocity Allen, F.H. and Roe, R.P. 1950. Performance Characteristics of a Volu-
␣ ⳱ scaling parameter for Nc metric Condensate Reservoir. Trans., AIME 189: 83–90.
␤ ⳱ inertial flow resistance coefficient Aly, A.M., El-Banbi, A.H., Holditch, S.A. et al. 2001. Optimization of Gas
␭ ⳱ mobility Condensate Reservoir Development by Coupling Reservoir Modeling
␮ ⳱ viscosity and Hydraulic Fracturing Design. Paper SPE 68175 presented at the
␳ ⳱ density SPE Middle East Oil Show and Conference, Bahrain, 17–20 March.
␾ ⳱ porosity DOI: 10.2118/68175-MS.
Ambastha, A.K. 1988. Pressure Transient Analysis for Composite Systems.
PhD thesis, Stanford U., Stanford, California (October 1988) 193.
Subscripts and Superscripts
App, J.F. and Mohanty, M. 2002. Gas and Condensate Relative Perme-
g ⳱ gas ability at Near Critical Conditions: Capillary and Reynolds Number
gd ⳱ dry gas Dependence. J. of Pet. Sci. & Eng. 36 (16): 111–126. DOI: http://
i ⳱ immiscible dx.doi.org/10.1016/S0920-4105(02)00269-3.
I ⳱ initial Barnum, R.S., Brinkman, F.P., Richardson, T.W., and Spillette, A.G. 1995.
M ⳱ miscible Gas Condensate Reservoir Behavior: Productivity and Recovery Re-
o ⳱ oil duction Due to Condensation. Paper SPE 30767 presented at the SPE
r ⳱ relative Annual Technical Conference and Exhibition, Dallas, 22–25 October.
DOI: 10.2118/30767-MS.
t ⳱ total
Bøe, A. and Whitson, C.H. 1989. Two-Phase Pressure Test Analysis.
SPEFE 4 (4): 601–610; Trans., AIME, 287. SPE-10224-PA. DOI:
Acknowledgments
10.2118/10224-PA.
The authors would like to thank R. Mott and C.H. Whitson for Boom, W., Wit, K., Schulte, A.M., Oedai, S., Zeelenberg, J.P.W., and
their constructive comments. We gratefully knowledge the finan- Maas, J.G. 1995. Experimental Evidence For Improved Condensate

Fig. 25—Log-log derivative plot of Flow Period 6 (BU6). Fig. 26—Superposition plot of the buildup periods.

606 October 2006 SPE Reservoir Evaluation & Engineering

Fig. 27—Structural map of the selected well (Daungkaew 2002).

Mobility at Near-Wellbore Flow Conditions. Paper SPE 30766 pre-

sented at the SPE Annual Technical Conference and Exhibition, Dallas,
22–25 October. DOI: 10.2118/30766-MS.
Brown, L.P. 1985. Pressure Transient Behavior of the Composite Reser-
voir. Paper SPE 14316 presented at the SPE Annual Technical Con-
ference and Exhibition, Las Vegas, Nevada, 22–26 September. DOI:
10.2118/14316-MS.
Chu, W.C. and Shank, G.D. 1993. A New Model For a Fractured Well in
a Radial, Composite Reservoir. SPEFE 8 (3): 225–232. SPE-20579-
PA. DOI: 10.2118/20579-PA.
Coats, K.H. and Smart, G.T. 1986. Application of a Regression-Based EOS
PVT Program to Laboratory Data. SPERE 1 (3): 277–299. SPE-11197-
PA. DOI: 10.2118/11197-PA.
Danesh, A., Henderson, G.D., Tehrani, D.H., and Peden, J.M. 1994. Gas
condensate recovery studies. Presented at the DTI Improved Oil Re-
covery and Research Dissemination Seminar, London, 22 June.
Daungkaew, S. 2002. New Development in Well Test Analysis. PhD the-
sis, Centre for Petroleum Studies, Imperial College London (October
2002).
Daungkaew, S., Ross, F., and Gringarten, A.C. 2002. Well Test Investi-
gation of Condensate Drop-Out Behavior in a North Sea Lean Gas
Condensate Reservoir. Paper SPE 77548 presented at the SPE Annual
Technical Conference and Exhibition, San Antonio, Texas, 29 Septem-
ber–2 October. DOI: 10.2118/77548-MS.

Fig. 28—Interpretation plots for a three-zone radial composite

model.

Economides, M.J., Dehghani, K., Ogbe, D.O., and Ostermann, R.D. 1987.
Hysteresis Effects for Gas Condensate Wells Undergoing Build-up
Tests Below the Dew Point Pressure. Paper SPE 16748 presented at the
SPE Annual Technical Conference and Exhibition, Dallas, 27–30 Sep-
tember. DOI: 10.2118/16748-MS.

October 2006 SPE Reservoir Evaluation & Engineering 607

 Fig. 29—Schematic of a simulation model.
Henderson, G.D., Danesh, A., Tehrani, D.H., and Al-Kharusi, B. 2000b.
Engineer, R. 1985. Cal Canal Field, California: Case History of a Tight and Generating Reliable Gas Condensate Relative Permeability Data Used
Abnormally Pressured Gas Condensate Reservoir. Paper SPE 13650 to Develop a Correlation with Capillary Number. J. Pet. Sci. & Eng.
presented at the SPE California Regional Meeting, Bakersfield, Cali- 25: 79-91 (January). DOI: http://dx.doi.org/10.1016/S0920-
fornia, 27–29 March. DOI: 10.2118/13650-MS. 4105(00)00004-8.
Fetkovich, M.D., Guerrero, E.T., and Fetkovich, M.J. 1986. Oil and Gas Henderson, G.D., Danesh, A., Tehrani, D.H., and Al-Kharusi, B. 2001.
Relative Permeabilities Determined From Rate-Time Performance Effect of Positive Rate Sensitivity and Inertia on Gas Condensate
Data. Paper SPE 15431 presented at the SPE Annual Technical Con- Relative Permeability at High Velocity. Petroleum Geoscience 7 (6):
ference and Exhibition, New Orleans, 5–8 Oct. DOI: 10.2118/15431- 45–50.
MS. Jamiolahmady, M., Danesh, A., Henderson, G.D., Tehrani, D.H., and Al-
Fevang, Ø. and Whitson, C.H. 1996. Modeling Gas-Condensate Well De- Kharusi, B. 2003. Variations of Gas-Condensate Relative Permeability
liverability. SPERE 11 (4): 221–230. SPE-30714-PA. DOI: 10.2118/ With Production Rate at Near Wellbore Conditions: A General Corre-
30714-PA. lation. Paper SPE 83960 presented at the Offshore Europe Conference,
Forchheimer, P. 1901. Wasserbewegung durch Boden. ZVD1 45: 1781. Aberdeen, 2–5 September. DOI: 10.2118/83960-MS.
Fussell, D.D. 1973. Single-Well Performance Predictions For Gas Con-
densate Reservoirs. JPT 25 (7): 860–870. SPE-4072-PA. DOI: Jones, J.R., Vo, D.T., and Raghavan, R. 1989. Interpretation of Pressure-
10.2118/4072-PA. Buildup Responses in Gas-Condensate Wells. SPEFE 4 (1): 93–104.
Geertsma, J. 1974. Estimating the Coefficient of Inertial Resistance in SPE-15535-PA. DOI: 10.2118/15535-PA.
Fluid Flow Through Porous Media. SPEJ 14 (5): 445–450. SPE-4706- Kniazeff, V.J. and Nvaille, S.A. 1965. Two-Phase Flow of Volatile Hy-
PA. DOI: 10.2118/4706-PA. drocarbons. SPEJ 5 (1): 37–44; Trans., AIME, 234. SPE-962-PA. DOI:
Gondouin, M., Iffly, R., and Husson, J. 1967. An Attempt to Predict the 10.2118/962-PA.
Time Dependence of Well Deliverability in Gas Condensate Fields. Li, K. and Firoozabadi, A. 2000. Phenomenological Modeling of Critical
SPEJ 7 (2): 113–124; Trans., AIME, 240. SPE-1478-PA. DOI: Condensate Saturation and Relative Permeabilities in Gas/Condensate
10.2118/1478-PA. Systems. SPEJ 5 (2): 138–147. SPE-56014-PA. DOI: 10.2118/56014-PA.
Gringarten, A.C., Al-Lamki, A., Daungkaew, S., Mott, R., and Whittle,
Liu, J.S., Wilkins, J.R., Al-Qahtani, M.Y., and Al-Awami, A.A. 2001.
T.M. 2000. Well Test Analysis in Gas-Condensate Reservoirs. Paper
Modeling a Rich Gas Condensate Reservoir With Composition Grad-
SPE 62920 presented at the SPE Annual Technical Conference and
ing and Faults. Paper SPE 68178 presented at the SPE Middle East Oil
Exhibition, Dallas, 1–4 October. DOI: 10.2118/62920-MS.
Show, Bahrain, 17–20 March. DOI: 10.2118/68178-MS.
Henderson, G.D., Danesh, A., Tehrani, D.H., Al-Shaidi, S., and Peden,
J.M. 1998. Measurement and Correlation of Gas Condensate Relative Merrill, L.S. Jr., Kazemi, H., and Gogarty, W.B. 1974. Pressure Falloff
Permeability by the Steady-State Method. SPEREE 1 (2): 134–140. Analysis in Reservoirs With Fluid Banks. JPT 26 (7): 809–18; Trans.,
SPE-30770-PA. DOI: 10.2118/30770-PA. AIME, 257. SPE-4528-PA. DOI: 10.2118/4528-PA.
Henderson, G.D., Danesh, A., Tehrani, D.H., and Al-Kharusi, B. 2000a. Mott, R. 2003. Engineering Calculations of Gas-Condensate-Well Produc-
The Relative Significance of Positive Coupling and Inertial Effects on tivity. SPEREE 6 (5): 298–306. SPE-86298-PA. DOI: 10.2118/86298-PA.
Gas Condensate Relative Permeabilities at High Velocity. Paper SPE
62933 presented at the SPE Annual Technical Conference and Exhi-
bition, Dallas, 1–4 October. DOI: 10.2118/62933-MS.

Fig. 30—Match of liquid saturation of the CVD experiment for Fig. 31—Gas/oil relative permeability curves of the selected
the selected well. well.

608 October 2006 SPE Reservoir Evaluation & Engineering

Mott, R., Cable, A., and Spearing, M. 1999. A New Method of Measuring Vo, D.T., Jones, J.R., and Raghavan, R. 1989. Performance Predictions for
Relative Permeabilities for Calculating Gas-Condensate Well Deliver- Gas Condensate Reservoirs. SPEFE 4 (4): 576–584; Trans., AIME,
ability. Paper SPE 56484 presented at the SPE Annual Technical Con- 287. SPE-16984-PA. DOI: 10.2118/16984-PA.
ference and Exhibition, Houston, 3–6 October. DOI: 10.2118/56484- Wheaton, R.J. and Zhang, H.R. 2000. Condensate Banking Dynamics in
MS. Gas Condensate Fields: Compositional Changes and Condensate Ac-
Mott, R., Cable, A., and Spearing, M. 2000. Measurements and Simulation cumulation Around Production Wells. Paper SPE 62930 presented at
of Inertial and High Capillary Number Flow Phenomena in Gas- the SPE Annual Technical Conference and Exhibition, Dallas, 1–4
Condensate Relative Permeability. Paper SPE 62932 presented at the October. DOI: 10.2118/62930-MS.
SPE Annual Technical Conference and Exhibition, Dallas, 1–4 Octo- Whitson, C.H. and Torp, S.B. 1983. Evaluating Constant-Volume Deple-
ber. DOI: 10.2118/62932-MS. tion Data. JPT 35 (3): 610–620. SPE-10067-PA. DOI: 10.2118/10067-
Novosad, Z. 1996. Composition and Phase Changes in Testing and Pro- PA.
ducing Retrograde Gas Wells. SPERE 11 (4): 231–235. SPE-35645- Whitson, C.H., Fevang, Ø., and Saevareid, A. 1999. Gas Condensate Rela-
PA. DOI: 10.2118/35645-PA. tive Permeability for Well Calculations. Paper SPE 56476 presented at
Odeh, A.S. 1969. Flow Test Analysis for a Well With Radial Discontinuity. the SPE Annual Technical Conference and Exhibition, Houston, 3–6
JPT 21 (2): 207–210; Trans., AIME, 246. SPE-2157-PA. DOI: October. DOI: 10.2118/56476-MS.
10.2118/2157-PA. Whitson, C.H., Norvik, H., Hvidsten, J., and Austad, T. 1983. Practical
Olarewaju, J.S. and Lee, W.J. 1989. A Comprehensive Application of a Aspects of Characterizing Gas-Condensates. Paper presented at the
Composite Reservoir Model to Pressure-Transient Analysis. SPERE 4 Conference on North Sea Condensate Reservoirs and Their Develop-
(3): 325–331. SPE-16345-PA. DOI: 10.2118/16345-PA. ment, London, 24–25 May.
Olarewaju, J.S., Lee, W.J., and Lancaster, D.E. 1991. Type- and Decline-
Curve Analysis With Composite Models. SPEFE 6 (1): 79–85. SPE- Appendix A
17055-PA. DOI: 10.2118/17055-PA. Existence of the Capillary Number Effect Derivative Stabili-
Ramey, H.J. Jr. 1970. Approximate Solutions for Unsteady Liquid Flow in zation. The conditions of existence of the initial derivative radial-
Composite Reservoirs. J. Cdn. Pet. Tech. (January–March) 32–37. flow stabilization are illustrated in Fig. A-1. The first stabilization
van Poollen, H.K. 1964. Radius of Drainage and Stabilization Time Equa- disappears when the gas saturation in the near-wellbore region
tions. Oil & Gas J. (14 September) 139–147. decreases below the saturation at Point B if the capillary number is
van Poollen, H.K. 1965. Transient Tests Find Fire Front in an In-situ 10 times the base capillary number, or below the saturation at Point
Combustion Project. Oil & Gas J. (1 February) 78–80. C if the capillary number is greater than the base capillary number
NCb by five orders of magnitude.

Fig. 32—krg vs. Nc at fixed krg/kro. Fig. 33—Pressure history match.

October 2006 SPE Reservoir Evaluation & Engineering 609

 Fig. 35—Comparison between measured and simulated GOR.

Fig. 34—Log-log comparison between actual and simulated

data (DD5). A ⭈ tDX
R2D = , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-2)
DM Ⲑ 共M−1兲
Appendix B A.k1tx
tDx = , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-3)
Published Front-Radius Calculation Methods. Different meth- 共␾ ⭈ ␮ ⭈ ct兲1rw2
ods have been proposed over the years to estimate the condensate-
bank radius from well-test data, but they are difficult to apply in 共k Ⲑ ␾ ⭈ ␮ ⭈ ct兲1
D= , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-4)
practice when based on the identification of straight lines on a 共k Ⲑ ␾ ⭈ ␮ ⭈ ct兲2
Horner plot. They all require the independent evaluation of the
total compressibility in the condensate bank, which is the subject 共k Ⲑ ␮兲1
M= , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-5)
of the present paper. 共k Ⲑ ␮兲2
The deviation time method (van Poollen 1964, 1965) calculates
where RD is the dimensionless bank radius, rw is the well radius,
a front radius from the time tend at the end of the first semilog
M is the mobility ratio, and D is the diffusivity ratio. The other
Horner straight line, which represents the two-phase inner-region
parameters are the same as in Eq. B-1. The intersection-time
mobility:
method depends on the ability to observe the two semilog straight

冑
lines. In practice, the first semilog straight line may be masked by
A.k1 tend wellbore storage, whereas the second straight line may not have
R= . , . . . . . . . . . . . . . . . . . . . . . . . . (B-1)
共 ␾ ⭈ ␮ ⭈ ct 兲1 共 De 兲end
t been reached or may be superseded by boundary effects, thus
making the intersection-time method often inapplicable.
where k, ␾, ␮, and ct are the Region 1 permeability, porosity, Another method uses type-curve matching of well-test pressure
viscosity, and total compressibility, respectively. A is a conversion and pressure-derivative data (Olarewaju and Lee 1989; Olarewaju
factor, and (tDe)end is a theoretical dimensionless deviation time et al. 1991). The pressure match can be used to calculate the
based on the front radius; (tDe)end values of 0.25 (van Poollen inner-region mobility:

冉冊 冉 冊
1964), 0.389 (Merrill et al. 1974), or 0.18 (Ambastha 1988) have
k qB dpwD Ⲑ d logtDe
been used, depending on how (tDe)end is calculated. Obtaining an = . , . . . . . . . . . . . . . . . . . . . (B-6)
accurate deviation time for small mobility contrasts, however, may ␮ 1 h d⌬pw Ⲑ d log⌬t match
be difficult. ␣k2h共pi − pwf兲
The intersection-time method (Merrill et al. 1974; Odeh 1969; pWD = , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-7)
Ramey 1970) uses the intersection time tx of the two semilog q␮B
Horner straight lines corresponding to the mobilities of the inner ␤k1t
(gas and condensate) and outer (gas) regions, respectively, and a tDe = , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (B-8)
theoretical dimensionless intersection time (tDX): ␾1␮1ct1R21
where pwD is the dimensionless wellbore pressure drop, pwf is the
wellbore flowing pressure, tDe is the dimensionless deviation time,
⌬t is the elapsed time, ⌬pw is the wellbore pressure drop, h is the
formation thickness, ␣ and ␤ are unit conversion constants, and 1
and 2 represent the inner and outer regions, respectively. The time

Fig. 36—Condensate-bank radius comparison (analytical and

numerical results with capillary number and non-Darcy effects).

610 October 2006 SPE Reservoir Evaluation & Engineering

match yields an estimate of front radius, R, if the inner-region
properties are known:

R= 冑 A.k1
共␾ ⭈ ␮ ⭈ ct兲1
. 冋 册⌬t
共tDe兲
. . . . . . . . . . . . . . . . . . . . . . (B-9)
match
Fig. A-1—Gas and oil relative permeabilities at different capil-
lary numbers.

a BS degree in petrochemical engineering from Amir-Kabir U.

SI Metric Conversion Factors of Technology (Tehran Polytechnic) and MS and PhD degrees
bbl × 1.589 873 E−01 ⳱ m3 in petroleum engineering, from Imperial College London. Alain
ft × 3.048* E−01 ⳱ m C. Gringarten is a professor of petroleum engineering and Di-
rector of the Centre for Petroleum Studies at Imperial College
ft3 × 2.831 685 E−02 ⳱ m3 London. e-mail: a.gringarten@imperial.ac.uk. Before joining
psi × 6.894 757 E+00 ⳱ kPa Imperial in 1997, he held a variety of senior technical and man-
agement positions with Scientific Software-Intercomp, Schlum-
*Conversion factor is exact.
berger, and the French Geological Survey in Orléans, France.
Gringarten holds MS and PhD degrees in petroleum engineer-
ing from Stanford U. and an engineering degree from Ecole
Manijeh Bozorgzadeh is a Research Fellow at Imperial College Centrale Paris. He is a recognized expert in well-test analysis
London. e-mail: Manijeh.Bozorgzadeh@imperial.ac.uk. She and was the recipient of the 2005 Cedric K. Ferguson Certifi-
has 14 years of experience in the field of petroleum and pet- cate, the 2003 SPE John Franklin Carll Award, and the 2001 SPE
rochemical engineering, having worked for the Natl. Iranian Formation Evaluation Award. He was elected a Distinguished
Oil Co. Research Inst. of Petroleum Industry. Bozorgzadeh holds Member of SPE in 2002.

October 2006 SPE Reservoir Evaluation & Engineering 611