REE161933 DOI: 10.

2118/161933-PA Date: 30-April-14

Stage:

Page: 177

Total Pages: 13

A New Method To Predict the

Performance of Gas-Condensate
Reservoirs
Ali Al-Shawaf, SPE, Saudi Aramco; Mohan Kelkar, SPE, University of Tulsa; and
Mohammad Sharifi, SPE, Amirkabir University of Technology

Summary
Gas-condensate reservoirs differ from dry-gas reservoirs. The
understanding of phase and fluid flow-behavior relationships is
essential if we want to make accurate engineering computations
for gas-condensate systems. Condensate dropout occurs in the reservoir as the pressure falls below the dewpoint, resulting in significant gas-phase production decreases.
The goal of this study is to understand the multiphase-flow
behavior in gas-condensate reservoirs and, in particular, to focus on
estimating gas-condensate-well deliverability. Our new method
analytically generates the inflow-performance-relationship (IPR)
curves of gas-condensate wells by incorporating the effect of condensate banking as the pressure near the wellbore drops below the
dewpoint. The only information needed to generate the IPR is
the rock relative permeability data and a constant-compositionexpansion (CCE) experiment.
We have developed a concept of critical oil saturation near the
wellbore by simulating both lean and rich condensate reservoirs
and have observed that the loss in productivity caused by condensate accumulation can be closely tied to critical saturation. We are
able to reasonably estimate re-evaporation of liquid accumulation
by knowing the CCE data.
We validated our new method by comparing our analytical
results with fine-scale-radial-simulation-model results. We demonstrated that our analytical tool can predict the IPR curve as a
function of reservoir pressure. We also developed a method for
generating an IPR curve with field data and demonstrated its
application with field data. The method is easy to use and can be
implemented quickly. Another advantage of this method is that it
does not require the knowledge of accurate production data
including the varying condensate/gas ratio (CGR).
Introduction
Well productivity is an important issue in the development of most
low- and medium-permeability gas-condensate reservoirs. Liquid
buildup around the well can cause a significant reduction in productivity, even in lean gas-condensate reservoirs in which the maximal
liquid dropout in the CCE experiment is as low as 1% (Afidick
et al. 1994). Subsequently, accurate forecasts of productivity can be
difficult because of the need to understand and account for the complex processes that occur in the near-well region.
The production performance of a gas-condensate well is easy
to predict (similar to a dry-gas well) as long as the wells flowing
bottomhole pressure (FBHP) is above the fluid dewpoint pressure.
After the wells FBHP falls below the dewpoint, the well performance starts to deviate from that of a dry-gas well. Condensate
begins to drop out first near the wellbore. Immobile initially, liquid condensate accumulates until the critical condensate saturation (the minimal mobile condensate saturation) is reached. This
rich liquid zone grows outward deeper into the reservoir as depletion continues (Fevang and Whitson 1996).
C 2014 Society of Petroleum Engineers
Copyright V

This paper (SPE 161933) was accepted for presentation at the Abu Dhabi International
Petroleum Exhibition and Conference, Abu Dhabi, 1114 November 2012, and revised for
publication. Original manuscript received for review 7 January 2013. Revised manuscript
received for review 3 April 2013. Paper peer approved 30 January 2014.

The loss in productivity caused by liquid buildup is mostly

influenced by the value of gas relative permeability (Krg) near the
well compared with the value of Krg in the reservoir farther away.
The loss in productivity is more sensitive to the relative permeability curves than to fluid pressure/volume/temperature (PVT)
properties (Mott 1997).
The most-accurate way to calculate gas-condensate-well productivity is by fine-grid numerical simulation, either in singlewell models with a fine grid near the well or in full-field models
with use of local grid refinement. A large part of the pressure
drawdown occurs within 10 ft of the well, so that radial models
are needed with the inner grid cell having dimensions of approximately 1 ft (Mott 2003; Xiao and Al-Muraikhi 2004; Sharifi and
Ahmadi 2009).
Several investigators have estimated the productivity of gascondensate reservoirs, but not one of these methods is simple to use.
Some methods require the use of the modification of a finite-difference simulation process, whereas other methods use simplified simulation models (Guehria 2000; Xiao and Al-Muraikhi 2004; Ahmadi
et al. 2014). Our objective is to develop a simple, yet accurate, analytical procedure to estimate the productivity of gas-condensate reservoirs without running simulations. The only required data in our
method are the CCE data and the relative permeability curves. As
with other simplified methods, our new technique allows well-performance evaluation quickly without reservoir simulations.
In this study, we propose a new methodology for predicting
gas-condensate-reservoir performance with limited data (relative
permeability and PVT data). We know that the productivity index
(J) for single-phase gas is always higher than the productivity
index (J*) for two-phase flow; therefore, the productivity ratio
(J*/J) is always less than unity. By testing many cases, we found
that productivity ratio (J*/J) is strongly correlated to Krg (S*o) for
each relative permeability curve used. S*o is the oil saturation at
which oil becomes reasonably mobile. We define this critical saturation later. The main advantages of our method are being simple,
direct, and yet reasonably able to capture the behavior of the reservoir for most of the cases. It can predict the behavior that the
reservoir has before and after the average pressure is less than the
dewpoint pressure. The procedure does not need knowledge of the
producing CGR. The application of the proposed method is divided into two cases. In the first case, the average reservoir pressure is above the dewpoint pressure (Pd); in the second case, the
average reservoir pressure is below the Pd. For the first case, in
which initial reservoir pressure is above the dewpoint, we found
that the IPR curve can be explained by two straight linesone in
which bottomhole pressure (BHP) is above and one in which BHP
is below the dewpoint. Above the dewpoint, we can use a singlephase-flow equation; and below the dewpoint, we can modify the
slope with the relative permeability of oil at residual oil saturation
(ROS). We need only one set of observed production data to predict the entire IPR curve.
For the case in which average reservoir pressure (current reservoir pressure away from the wellbore) is below dewpoint, we
have only one curve below the dewpoint. We can predict the IPR
under these conditions if we have one set of production data available or if we have the relative permeability and CCE data available. We provide the details next.

May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:01 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

177

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Stage:

1.0

0.6

Rich Fluid

Near Wellbore
Region 1

Condensate
Buildup
Region 2

Single Phase
Gas
Region 3

Total Pages: 13

Lean Fluid

Intermediate Fluid

30%
25%

Two-Phase Gas-Oil
Flow

Liquid Dropout (So)

Liquid Saturation

0.8

Page: 178

Only Gas Flowing

0.4

20%
15%
10%
5%

0.2
rw = 0.35 ft

Swi = 0.25
0%

0.0

1
1

10

100

1000

10000

1000

2000

3000
4000
Pressure, psi

5000

6000

7000

Radius, ft

Fig. 2CCE data for synthetic gas-condensate compositions.

Fig. 1Three regions of flow behavior in gas-condensate well.

A fine-grid radial compositional model was built in a commercial

flow simulator to validate the results of our analytical approach.
Literature Review
Fevang and Whitson (1996) proposed a method for modeling the
deliverability of gas-condensate wells. Well deliverability is calculated with a modified Evinger-Muskat pseudopressure approach
(Evinger and Muskat 1942). The gas/oil ratio (GOR) needs to be
known accurately to use the pseudopressure integral method for
each reservoir pressure. Fevang and Whitson (1996) have developed a method to calculate the pseudopressure integral in the
pseudosteady-state gas-rate equation:
!
Pr
Krg
Kro
Rs
dp: . . . . . . . . . . . . . . . . 1
qg C
Bg lg
Pwf Bo lo
To apply their method, we need to break the pseudopressure integral into three parts, corresponding to the three flow regions as
discussed next:
 Region 1: An inner near-wellbore region as shown in Fig. 1 in
which both condensate and gas are mobile. It is the most important
region for calculating gas-condensate-well productivity because
most of the pressure drop occurs in Region 1. The flowing composition (GOR) within Region 1 is constant throughout, and a semisteady-state regime exists. This means that the single-phase gas
entering Region 1 has the same composition as the produced wellstream mixture. The dewpoint of the producing wellstream mixture
equals the reservoir pressure at the outer edge of Region 1.
 Region 2: This is the region in which the condensate saturation is building up. The condensate is immobile, and only gas is
flowing. Condensate saturations in Region 2 are approximated by
the liquid-dropout curve from a constant-volume-depletion (CVD)
experiment, corrected for water saturation.
 Region 3: This is the region in which no condensate phase
exists (above the dewpoint). Region 3 exists only in a gas-condensate reservoir that is currently in a single phase. It contains a single-phase (original) reservoir gas.
One of the major findings in Fevang and Whitson (1996) is that
the primary cause of reduced well deliverability within Region 1 is
Krg as a function of the relative permeability ratio (Krg/Kro) Fevang
and Whitsons approach is applicable for running coarse simulation
studies, in which the producing GOR is available from the prior
timestep.
In addition to the existence of multiphase flow, high-velocity
phenomena that include capillary number effect and non-Darcy
flow also can affect the well performance. Usually, at high velocity, non-Darcy flow has a negative effect, and the capillary number has a positive effect on well productivity (Blom and Hagoort
1998; Mott et al. 2000).
Guehria (2000) presented an approach to generate IPR curves for
depleting gas-condensate reservoirs without resorting to the use of
simulation. The producing GOR (Rp) is calculated with an expres178

sion derived from the continuity equations of gas and condensate in

Region 1. But his approach requires an iterative scheme by assuming that Rp at a given reservoir pressure generates IPR curves for
rich gas-condensate. Whereas, for lean gas-condensates, GOR values from the reservoir material-balance (MB) model are adequate to
achieve good results.
Later, Mott (2003) presented a technique that can be implemented in an Excel spreadsheet model for forecasting the performance of gas-condensate wells. The calculation uses an MB
model for reservoir depletion and a two-phase pseudopressure integral for well inflow performance. Motts method generates a
wells production GOR by modeling the growth of condensate
banking without a reservoir simulator.
All these methods do not provide a simple way of generating
IPR curves for condensate wells. The IPR curve as a function of
reservoir pressure is needed to accurately predict and optimize the
performance of the well.
Behavior of the Productivity Index in
Gas-Condensate Reservoirs
The pseudosteady-state rate equation for a gas well is given by
Fevang and Whitson (1996) and Kelkar (2008):
2p a Tsc khmPr  mPwf 
  
 ; . . . . . . . . . . . . . . . 2
re
 0:75 S
TPsc ln
rw
where a1 for the SI units, a1=2p  142 for field units,
qsc gas flow, K permeability, h thickness, m (p) pseudopressure, T reservoir temperature, re drainage radius, and rw
wellbore radius. We can use this equation to estimate the gas-production rate when the BHP is above the dewpoint of reservoir fluids. This means that this equation is applicable only for singlephase gas flow. As soon as the BHP drops below the dewpoint
pressure of reservoir fluid, condensate begins to drop out, first
near the wellbore, and the well performance starts to deviate from
that of a dry-gas well. Liquid condensate accumulates until the
critical condensate saturation (the minimal mobile-condensate saturation) is reached. This liquid-rich bank/zone grows outward
deeper into the reservoir as depletion continues.
Liquid accumulation, or condensate banking, causes a reduction in the gas relative permeability, and acts as a partial blockage
to gas production, which leads to a potentially significant reduction in well productivity. To quantify the impact of gas-condensation phenomena, we have developed a method to generate the IPR
of gas-condensate reservoirs with analytical procedures. The main
idea of our research is to combine fluid properties (CCE or CVD
data) with rock properties (relative permeability curves) to arrive
at an analytical solution that is accurate enough to estimate the
IPR curves of gas-condensate reservoirs.
qsc

Model Preparation. Fluid Description. Two different synthetic

gas-condensate compositions were used to generate the rich, intermediate, and lean fluids represented in Fig. 2. The rich fluid is
May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:01 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Stage:

Page: 179

Total Pages: 13

Distance X, ft

TABLE 1FLUID PROPERTIES

3000 2000 1000

Rich
Gas

Intermediate
Gas

Lean
Gas

Initial reservoir pressure (psia)

Dewpoint pressure (psia)
Reservoir temperature ( F)
Maximal liquid dropout (%)

7,000
5,400
200
26

5,500
3,250
260
20

5,000
2,715
340
8.5

1000 2000 3000

3000
2000

Distance Y, ft

Parameters

1000
0
1000
2000
3000

composed of three componentsmethane (C1) 89%, butane (C4)

1.55%, and decane (C10) 9.45%. The composition of lean and intermediate is the same, and we just use a 260 F temperature for
intermediate fluid and a 320 F for lean gas. The four components
are methane (C1) 60.5%, ethane (C2) 20%, propane (C3) 10%, and
decane (C10) 9.5%. The characteristics of the condensate mixtures
are outlined in Table 1. The Peng-Robinson three-parameter
equation of state (EOS) (PR3) was used to simulate phase behavior and laboratory experiments, such as CCE and CVD.
Reservoir Description. The Eclipse 300 compositional simulator was used for the simulation. A 1D radial compositional model
with a single vertical layer and 36 grid cells in the radial direction
is used as a test case, as shown in Fig. 3. Homogeneous properties
are used in the fine-scale model, as described in Table 2.
A single producer well exists at the center of the reservoir. The
model has been refined near the wellbore to accurately observe
the condensate-dropout effect on the production. For that purpose,
the size of the radial cells has been logarithmically distributed
with the innermost grid size at 0.25 ft, according to the following
equation:
 1=N
ri1
re

; . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
ri
rw
where N is the number of radial cells in the model. Besides, with
very small gridblocks around the well, the timesteps have been
refined at initial times, which led to a very smooth saturation profile around the well. The fully implicit method was chosen for all
the runs.
Relative Permeability Curves. It already has been discussed
in the literature that relative permeability curves affect the gas
flow significantly in a gas-condensate reservoir after the pressure
falls below the dewpoint pressure. Accurate knowledge about relative permeability curves in a gas-condensate reservoir is essential. Unfortunately, the relative permeability curves are rarely
known accurately. It would be worthwhile if we could investigate
the effect of different relative permeability curves and study the
uncertainty they bring to the saturation buildup in gas-condensate
reservoirs.
Different sets of relative permeability curves were used in the
study. For two-phase relative permeability (oil and gas), these
curves were generated on the basis of Corey equations, as illustrated here:

OilSat
0.00000

0.11686

0.23372

0.35057

0.46743

Fig. 3The 36-cell fine radial model.

where n is the gas relative permeability exponent, m is the oil relative permeability exponent, Sor is the ROS, a is endpoint gas relative permeability, and b is endpoint oil relative permeability.
Fractures (X-Curves) and intermediate and tight relative permeability curves were generated by changing n and m exponents
from 1 through 5 and changing Sor from 0 to 0.60. Fig. 4 shows
three sets of relative permeability curves. Corey-1 (X-curve) is
generated on the basis of n 1, m 1, and Sor 0. Corey-14 is
generated on the basis of n 3, m 4, and Sor 0.20. The third
curve, Corey-24 is generated on the basis of n 5, m 4, and
Sor 0.60.
Sensitivity Study. In this research, we have examined a large
number (more than 20) of relative permeability curves. The sensitivity study also examined the effects of fluid richness on gas productivity with two fluid compositions (lean and rich fluids).
The results of the sensitivity study were checked with simulation results. The simulation runs were performed under a constant-rate mode of production with the fine compositional radial
model. After testing this wide range of relative permeability
curves, we found that a very strong correlation exists between the
wells productivity-index (PI) ratio and Krg (S*o). Figs. 5 and 6
show that for all 24 different cases of relative permeability and for
both rich and lean fluids, respectively, the relationship between
the PI ratio and Krg (S*o) is linear with a strong correlation coefficient. The value of S*o is defined in a later section. The well PI ratio
is defined as
PI ratio

minimum well PI
: . . . . . . . . . . . . . . . . . . . 6
maximum well PI

Fig. 7 shows a typical example of the well PI as a function of

time for a producing well. For this case, we used a single-well

Krg aSng . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4


1  Sg  Sor m
; . . . . . . . . . . . . . . . . . . . . 5
Kro b
1  Sor

Krg(Corey-1)

Kro(Corey-1)

Krg(Corey-14)

Kro(Corey-14)

Krg(Corey-24)

Kro(Corey-24)

1
0.9
0.8
0.7
0.6
Kr

TABLE 2RESERVOIR PROPERTIES USED IN THE FINE

RADIAL MODEL

0.5
0.4
0.3
0.2

Porosity (%)
Absolute permeability (md)
Reservoir height (ft)
Irreducible water saturation (%)
Rock compressibility (psi1)

20
10
100
0
4.0  106

0.1
0

0.1

0.2

0.3

0.4

0.5
Sg

0.6

0.7

0.8

0.9

Fig. 4Different sets of Corey relative permeability curves.

May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:01 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

179

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Stage:

Page: 180

Rich-Krg(So*) vs PI Ratio

Total Pages: 13

lean-Krg(So*) vs PI Ratio
1.00

1.00
y = 1.1827x + 0.0337

0.80

0.80

y = 1.4857x + 0.0153
R2 = 0.9445

0.60

Krg(So*)

Krg(So*)

0.60
R2 = 0.9222

0.40

0.40

0.20

0.20

0.00
0.00

0.20

0.40
0.60
PI Ratio

0.80

0.00
0.00

1.00

Fig. 5Krg (S*o) vs. PI ratio for rich fluid.

0.40
0.60
PI Ratio

0.80

1.00

Fig. 6Krg (S*o) vs. PI ratio for lean fluid.

Productivity Index (Fine Radial Model)

600

Radial Cell-1

Radial Cell-5

Radial Cell-10

Radial Cell-20

Radial Cell-30

Radial Cell-36

0.45

500
Min Well PI
PI Ratio =
Max Well PI

400

0.4
0.35
0.3

300

0.25

So

Well PI, Mscfd/psi

0.20

200

0.2
0.15

100

0.1
0

10
15
Time, Years

20

25

Fig. 7Well PI as a function of time.

model that has rich gas-condensate. Initial reservoir pressure was

7,000 psi, and the well was producing with a constant-rate constraint of 20 MMscf/D. In this example, initial reservoir pressure
was above the dewpoint. We found that the PI ratio value is very
close to that obtained from the corresponding relative permeability curve Krg (S*o). This shows that, by using relative permeability
data, we are able to capture the behavior of the PI of a reservoir
without performing any simulation. The abrupt fall of PI happens
shortly after the FBHP falls below the Pd. After analyzing several
cases, we found that the PI ratio is very close to Krg (S*o).
Rich Fluid

Lean Fluid

500
450
Well PI, Mscfd/psi

400
350
300
250
200
150
100
50
0

10

15

20

25

30

35

40

45

50

Time, Years

Fig. 9Well PI vs. time for the rich and lean fluids (same relative permeability curves).
180

0.05
0

10
15
Time, Years

20

25

Fig. 8Oil-saturation profiles around the well as a function of

time.

Fig. 7 shows that the well PI drops quickly and eventually

increases slightly before stabilizing. The slight increase in PI is a
result of the revaporization of oil and changing composition
(leaner fluid). This type of PI behavior is observed in the field
data in which, after the BHP drops below the dewpoint pressure,
PI decreases suddenly before stabilizing. Similar behavior also
was observed by Kamath (2007).
As we can see from Fig. 8, after oil saturation reaches a maximal value, So drops gradually after a period of production, which
will enhance the gas relative permeability and, therefore, the gas
productivity. Figs. 5 and 6 show clearly that for both rich and lean
fluids, respectively, the relationship between the PI ratio and krg
(S*o) is linear with a strong correlation coefficient.
Another important outcome of this sensitivity analysis is that
the loss in productivity is more sensitive to the relative permeability curves than to fluid PVT properties. Fig. 9 shows the well PI
vs. time for the rich and lean fluids with the same relative permeability set. Fig. 9 demonstrates that the loss in productivity is
much more dependent on relative permeability curves than on the
fluid composition.
Fig. 10 summarizes the results of the sensitivity study performed on the rich and lean fluids with the wide range of relative
permeability curves. It appears that the loss is similar for both
rich and lean gases when the same relative permeability curves
are used. The slope is very close to a value of unity.
Generation of IPR Curves. IPR curves are very important to
predict the performance of gas or oil wells; however, generating
IPR curves with a simulator is not straightforward because the
May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:01 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Stage:

Page: 181

Rich vs Lean PI Ratio

1.00
1.E+09

Pr (5400 psi)

Pr (5000 psi)

Pr (4000 psi)

Pr (3000 psi)

Pr (2000 psi)

Pr (1000 psi)

1.E+09
R2 = 0.9621

0.60

m(P), psi2/cp

Rich PI Ratio

0.80

Total Pages: 13

0.40

8.E+08
6.E+08
4.E+08

0.20
2.E+08
0.00
0.00

0.20

0.40
0.60
Lean PI Ratio

0.80

0.E+00

1.00

10

20

30

40

50

60

70

80

qg, MMscfd

Fig. 10Rich vs. lean PI ratio.

Fig. 11Pseudopressure vs. gas-rate plot.

IPR represents an instantaneous response of the reservoir at a

given reservoir pressure for a given BHP. This cannot be generated in a single run, because if we change the BHP in a simulation
run, depending on how much oil or gas is produced, the average
pressure will change.
To generate IPR curves, we used a composite method. For different initial reservoir pressures, we ran a simulator at a fixed
BHP. We varied the BHP from high to low values. After we had
the results for different BHPs, we combined the rates for different
BHPs for a fixed average pressure and constructed an IPR curve.
We repeated this procedure for different average pressures so that
we can construct the IPRs for different average pressures.
Important Observations From the Simulation Study. We
wanted to develop simple relationships for generating the IPR
curves under the field conditions. We used some important observations from our simulation study to develop our procedure:
 As previously discussed, we observed that the relative permeability curves have a bigger impact on the loss of productivity
than the type of fluid that is produced.
 The loss of productivity is significant and almost instantaneous after the pressure drops below the dewpoint. After that point,
the PI recovers because of the revaporization of the liquid and the
changing composition (with leaner fluid).
 The oil-saturation profile near the wellbore is a strong function of the average reservoir pressure (Dyung et al. 1987). Consequently, for a given reservoir pressure, the oil-saturation profile
does not change significantly for different BHPs, which is shown
in the next section.

Pr (5400 psi)

Pr (5000 psi)

Pr (4000 psi)

Pr (3000 psi)

Pr (2000 psi)

Pr (1000 psi)

 Relative change in well performance before and after the

dewpoint depends only on relative permeability, but, overall, well
performance depends on other reservoir properties (i.e., permeability). This is incorporated in PI calculations automatically
(Eqs. 2 and 16).
Development of New Approach
New Analytical Approach for Estimating Gas-CondensateWell Productivity. When we plotted IPR curves (i.e., for intermediate fluid with Pd 3,250 psi), as a function of both pressure
and pseudopressure, we noticed that plotting the pseudopressure
vs. the gas rate resulted in two clear straight lines for every reservoir pressure (Fig. 11). This is consistent with our observation
that below the dewpoint, the oil saturation does not change significantly for a given reservoir pressure. Fig. 12 represents the normal depiction of IPR curves. Notice a peculiar behavior of IPR
curves when plotted as a function of pseudopressure. The lines
are parallel above the dewpoint as expected because the productivity does not change. Below the dewpoint, for different reservoir
pressures, the lines are parallel for a certain pressure range; however, as the reservoir pressure depletes, the slope becomes gentler.
This is an indication of improved productivity. This is a result of
the re-evaporation of the liquid phase and changing composition
(with leaner fluid) as the pressure declines. This type of trend is
difficult to capture with the pressure data.
As soon as reservoir pressure drops below the dewpoint (Pd),
which is 3,250 psi in this example, there will be a productivity loss
that is characterized by the straight line below Pd in the pseudopressure plot, shown in Fig. 11. To illustrate our method, we will take
Pr 5,400 psi as an example for illustration (Fig. 13).

m(P) vs qg, Pr = 5400 psi

6000
1.2E+09

5000

1.0E+09
Slope = (1/j)

m(P), psi2/cp

BHP, psi

4000
3000
2000
1000
0

8.0E+08
6.0E+08

Pd = 3250 psi

4.0E+08
Slope = (1/j*)

2.0E+08

20

40

60

qg, MMscfd

Fig. 12BHP-vs.-gas-rate IPR plot.

80

0.0E+08

10

20

30

40

50

60

70

80

qg, MMscfd

Fig. 13IPR for Pr 5 5,400 psi.

May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:01 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

181

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

The pseudopressure plot in Fig. 13 clearly shows that there are

two distinct PIs. The first PI (J) is constant for a single-phase gas
flow (in which FBHP is above Pd), and the second PI (J*) is for a
two-phase flow (in which FBHP is below Pd). Referring to the
preceding pseudosteady-state gas-rate equation (Eq. 2), one can
see that
qsc

2p aTsc Khmpr  mpwf 

  
:
re
TPsc ln
 0:75 S
rw

The PI in terms of pseudopressure is given by

J

qsc
: . . . . . . . . . . . . . . . . . . . . . . 7
m Pr  mPwf 

Looking at Fig. 13, we can define the slopes as

slope of the line above Pd being 1=J . . . . . . . . . . 8
and
slope of the line below Pd being 1=J  : . . . . . . . . . .9
After analyzing several cases, we found that the PI ratio can be
determined by dividing the slope above Pd by the slope below Pd as
 
1

slope of the line above Pd
J
J
  PI ratio:

1
slope of the line below Pd
J
 
J
10
Because the PI (J) for a single-phase gas is always higher than the
PI (J*) for a two-phase flow, the PI ratio (J*/J) is always less than
unity. By conducting many numerical experiments, we found that
the PI ratio (J*/J) is strongly correlated to Krg (S*o) for each relative permeability curve used (i.e., Figs. 5 and 6). The application
of the proposed method is divided into two cases. The first case is
when the average reservoir pressure is above the Pd, and the second case is when the average reservoir pressure is below the Pd.
General ProcedureAverage Reservoir Pressure Is Above
the Pd. In this case in which the average reservoir pressure is
above the Pd, the pseudosteady-state gas-rate equation (Eq. 2) is
used to estimate the gas rate when the BHP > Pd. When the BHP
drops below the Pd, we need to estimate (J*) first to be able to calculate the gas rate analytically. Because, in this case, initial reservoir pressure is above the Pd, we can estimate the PI (J), which will
be constant for all BHPs above the Pd. The procedure is as follows:
1. Estimate the PI above dewpoint with Eq. 7.
2. Calculate J* with our knowledge of Krg (S*o) and J from the
previous section as
J
PI ratio  Krg So : . . . . . . . . . . . . . . . . . . . . 11
J
3. Generate the IPR curve below Pd and calculate the flow rate.
After estimating J*, which has a constant but higher slope than
J, as shown before on the pseudopressure plot, we can use J* to
estimate the gas rate for all BHPs below the Pd. With the following equation, we assume
y mx b . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
and


1
mPwf   q b: . . . . . . . . . . . . . . . . . . . . 13
J
We can use our knowledge of the rate and FBHP at the Pd
with the pseudosteady-state gas-rate equation (Eq. 2) above dewpoint. Then, the intercept b can be calculated as
182

Stage:

b mPd

Page: 182

Total Pages: 13

qd
; . . . . . . . . . . . . . . . . . . . . . . . . . . 14
J

where b in field units is in (psi2/cp).

Now, our straight-line pseudopressure equation is complete to
estimate the gas rate for any FBHP less than the Pd as
q b  mPwf J  ; . . . . . . . . . . . . . . . . . . . . . . . . 15
where q is in (Mscf/D), b and mPwf are in (psi2/cp), and J* is in
[(Mscf/D)/psia2/cp].
General ProcedureAverage Reservoir Pressure Is Less
Than the Pd. In this case, the pseudopressure vs. rate plot, IPR,
will have only one straight line. Fig. 11 shows three examples of
IPR lines in which initial reservoir pressure is below the Pd. To be
able to generate the IPR curves for cases in which initial reservoir
pressure is below the Pd, the following procedure should be
followed:
1. Estimate the PI (J).
If an IPR curve for the case in which reservoir pressure is
above the Pd is available, the PI (J) of this case could be used to
estimate J* as a function of pressure with CCE data, as will be
explained in the next step. For cases in which IPR curves above
the Pd are not available, the PI (J) could be estimated with a pseudosteady-state gas-rate equation,
J

qsc

mPr  mPwf 

2p aT Kh
  sc
:
re
TPsc In
 0:75 S
rw
16

2. Estimate the PI (J*).

As we stated earlier, the PI ratio (J*/J) is correlated to krg (S*o),
but in these cases in which initial reservoir pressure is below Pd,
liquid revaporization plays a very important role in determining
the productivity of gas-condensate reservoirs. By examining the
CCE data, as shown in Fig. 2, we can see that, as soon as the pressure drops below the Pd, liquid saturation reaches a maximal
value (Max_SoCCE) around the Pd quickly. Although the rate of
accumulating condensate bank to its maximal saturation could be
affected by many variables (e.g., rate, pressure drop, and fluid
composition), in general, the pressure drop near the wellbore is
large, and the volume near the wellbore is small. The combination
of a higher pressure gradient and a smaller volume results in a
rapid accumulation of condensate as soon as the pressure drops
below the dewpoint. As the pressure declines further, the liquid
saturation gradually declines. Our idea of using CCE data in generating the IPR curves is really to account for this phenomenon of
liquid vaporization and changing composition (with leaner fluid)
as pressure drops below the Pd (Dyung et al.1987). We found that
the use of a fixed value of Krg (S*o) or Krg (Max_SoCCE) will underestimate the gas productivity for cases in which initial reservoir
pressure is below the Pd.
Therefore, for any reservoir pressure below Pd, the Krg needs
to be estimated at the corresponding pressure and oil saturation
from the CCE data according to the following equation:
J
Pr PI ratio  Krg SoCCE . . . . . . . . . . . . . . . . 17
J
3. Estimate the gas rate.
Finally, the gas rate can be estimated directly from the following equation:
q mPr  mPwf J  . . . . . . . . . . . . . . . . . . . . . 18
The preceding outlined procedure for generating IPR curves
assumes that Sor Max_SoCCE, but this is not always the case in
real field applications. Because Sor is a rock property whereas
Max_SoCCE is a fluid property, we expect them to be different in
most of the cases in field applications.
May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:02 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Stage:

Page: 183

Total Pages: 13

Krg(Corey-14)

TABLE 3GENERAL PROCEDURE FOR ESTIMATING

PI RATIO

Kro(Corey-14)

1
0.9

Cases in
Which Pr Below Pd

0.8

Case

Cases in
Which Pr Above Pd

Sor Max_SoCCE
Sor < Max_SoCCE
Sor > Max_SoCCE

J*/J Krg (Sor)

J*/J Krg (Max_SoCCE)
J*/J Krg (Sor)

J
Pr Krg SoCCE
J

0.6
Kr

0.7

0.5
0.4
0.3
0.2

For that purpose, we have analyzed several cases in which Sor

could be equal to, less than, or greater than Max_SoCCE. On the basis of our evaluation, we believe that the maximum of the two values should be used to correctly capture the fluid behavior around
the wellbore and thereby accurately estimate the gas productivity.
The procedure to estimate PI (J*) here is exactly the same as
the preceding procedure outlined for the case in which Sor
Max_SoCCE, but with some modifications, as given by Table 3.
This procedure is used for flowing pressure less than the dewpoint. In effect, what we are saying in this table is that, if we have
reservoir pressure above the dewpoint, then to calculate the IPR
curve for the BHP below the dewpoint, we can use a constant
slope (J*) on the basis of the Krg estimate, as stated in Table 3.
Subsequently, after the reservoir pressure drops below dewpoint,
we will need to use Krg as a function of average reservoir pressure.
Importance of Threshold Oil Saturation (S*o). We found that an
accurate estimation of gas productivity depends not only on Sor but
also on the threshold oil saturation (S*o) for reservoirs with tight oil
relative-permeability curves. Fig. 14 shows an oil relative-permeability curve that was generated on the basis of Sor 0.20 and a
high value of oil exponent (m 4). This higher value of the oil relative permeability exponent makes the oil relative permeability very
low and eventually makes oil practically immobile until its saturation exceeds the threshold value (S*o), which is, in this case, 0.48, as
shown in Fig. 14. After testing several tight relative permeability
curves, we found that, for practical applications, we can determine
that the threshold (S*o) corresponds to Kro 1% of the maximal
value of oil relative permeability.
Therefore, in generating IPR curves, it is more important to
know S*o than Sor. We define S*o as a minimal saturation needed to
make oil mobile (i.e., Kro is at least 1% of the endpoint value). It
is a strong function of the curvature of the relative permeability
curve. Therefore, all we need to do is use Table 3 as it is, but
replace Sor with S*o, as shown in Table 4.
Cases in Which So* > Max_SoCCE. This is probably the most
common case in the field. Even for rich condensates, it is not unusual to find that the minimal mobile oil saturation is greater than
maximal oil saturation in the CCE experiment. The rich condensate fluid with maximal liquid dropout (26%) is being used for
this case in which it is less than S*o 0.48, as shown previously in
Fig. 14. Referring to Table 4, we can see that in this case the PI
ratio is determined by Krg (S*o).
Initial Reservoir Pressure Is Above the Pd. Fig. 15 shows
the So distribution around the wellbore as a function of the BHP.
We would like to examine this figure carefully. As soon as the

0.1
0

So* = 0.48

0.1

0.2

0.3

0.4

0.5

0.6

Sor = 0.20

0.7

0.8

0.9

Sg

Fig. 14Illustration of threshold (S*o) in tight relative permeability curves.

BHP drops below Pd, which is 5,400 psi for this fluid, liquid starts
dropping out close to the wellbore first. The radius of oil banking
expands inside the reservoir, and oil saturation away from the
wellbore increases as BHP decreases. We can see that, at a BHP
of 3,000 psi, So reaches a maximal value of approximately 0.62,
and this value almost stays constant even though the BHP drops
to 1,000 psi and then to atmospheric conditions.
What we want to highlight here is that when reservoir pressure
is above the Pd, the oil saturation near the wellbore remains reasonably constant, irrespective of the BHP. Because, in this case,
the threshold (S*o) is higher than Max_SoCCE, the value of S*o
should be used to obtain the corresponding krg and therefore estimate the well productivity for the cases in which reservoir pressure is above Pd. By following the general procedure outlined
previously for the case in which the initial reservoir pressure is
above the Pd, we can generate the IPR curve, as shown in Figs.
16 and 17. It can be seen that, as long as the BHP is above the
dewpoint (Pd 5,400 psi) that corresponds to a gas rate below 40
MMscf/D, our method can exactly predict the behavior of the simulation model. Below the dewpoint pressure, although there is a
small difference, our method can reproduce the simulation behavior with reasonable accuracy. Just keep in mind that the only
change you would make for this case in which the threshold
(S*o) > Max_SoCCE is to use the larger value of the two, which, in
this case, is the S*o,
J
PI ratio Krg So : . . . . . . . . . . . . . . . . . . . . . 19
J

BHP(5000 psi)

BHP(3000 psi)

BHP(1000 psi)

BHP(15 psi)

0.7
0.6
0.5
So

0.4
0.3

TABLE 4GENERAL PROCEDURE FOR ESTIMATING PI

RATIO FOR TIGHT ROCKS

0.2
0.1

Case

Cases in
Which Pr Above Pd

Cases in
Which Pr Below Pd

S*o Max_SoCCE
S*o < Max_SoCCE
S*o > Max_SoCCE

J*/J Krg (S*o)

J*/J Krg (Max_SoCCE)
J*/J Krg (S*o)

J
Pr Krg SoCCE
J

0
0.1

10
Radius, ft

100

1000

Fig. 15Oil-saturation distribution as a function of BHP for

Pr 5 6,900 psi.

May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:02 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

183

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Fine Radial Model

Stage:

Page: 184

Total Pages: 13

Analytical Solution

Analytical Solution

Fine Radial Model

1.6E+09

8000

1.4E+09
m(P), (psia2/cp)

7000

BHP, psi

6000
5000
4000

1.2E+09
1.0E+09
8.0E+08
6.0E+08
4.0E+08

3000

2.0E+08
2000

0.0E+00
0.0

1000
0

10

20

30
40
qg, MMscfd

50

60

70

10.0

20.0

30.0
40.0
qg, MMscfd

50.0

60.0

70.0

Fig. 17Pseudopressure vs. gas-rate plot for Pr 5 6,900 psi.

Fig. 16IPR curve for Pr 5 6,900 psi.

BHP(1,000 psi)

BHP(15 psi)

0.6
0.5
BHP(3,000 psi)

BHP(1,000 psi)

BHP(15 psi)

0.7

0.4

0.6

So 0.3

0.5
0.2
So

0.4
0.1

0.3
0.2

0
0.1

0.1
0
0.1

10
Radius, ft

100

1,000

10
Radius, ft

100

1,000

Fig. 19Oil-saturation distribution as a function of BHP for

Pr 5 3,000 psi.

Fig. 18Oil-saturation distribution as a function of BHP for

Pr 5 5,000 psi.

Initial Reservoir Pressure Is Below the Pd. Now, we will

illustrate an example for the case in which initial reservoir pressure is below Pd. To correctly generate the slope of the IPR curve
on the pseudopressure plot, we need to account for revaporization.
Again, we need to use the fine grid model to capture the condensate behavior near the wellbore.
By examining three figures, Figs. 18 through 20, that show
the So distribution for saturated reservoirs, one will notice that So
decreases gradually as a function of reservoir pressure from

BHP(15 psi)

0.35
0.3
0.25
So

0.2
0.15
0.1
0.05
0
0.1

10
Radius, ft

100

1,000

Fig. 20Oil-saturation distribution as a function of BHP for

Pr 5 1,000 psi.
184

approximately 0.62, when Pr 6,900 psi (Fig. 15) to almost 0.30

when Pr 1,000 psi (Fig. 20). Looking at Fig. 20, we can see
that, for low pressure (BHP 1,000 psi), condensate saturation
around wellbore is negligible and is smaller compared with the
saturation far away from the wellbore. We conclude from these
figures that oil revaporization and changing fluid composition
close to the wellbore are strong functions of decreasing reservoir
pressure that results in lower saturation.
For each saturated reservoir pressure, we can see that So builds
up to a uniform value close to the wellbore. This uniform So
remains almost constant as BHP decreases when reservoir pressure remains constant. Therefore, a valid assumption for the application of our method is to assume a uniform So for every saturated
(average) pressure under consideration. Fig. 20 shows an example
of an extreme case in which all the oil evaporates at a very low
flowing pressure.
After we have understood gas-condensate behavior around the
wellbore, we can see that we really need to use a method that can
mimic condensate revaporization as reservoir pressure depletes.
The best tool to use is the CCE or CVD data for each condensate
fluid being used (Dyung et al. 1987). The CCE data of the rich
fluid are shown previously in Fig. 2. Following Fevang and Whitson (1996), we will assume that CCE provides the best description
of fluid near the wellbore (i.e., we can use CCE data to predict the
revaporization of oil). The reason we choose CCE is because the
accumulation near the wellbore is minimal, which means that
near the wellbore region, whatever flows from the reservoir will
flow into the wellbore without any change in the composition.
Because in this case the threshold (S*o) is greater than Max_
SoCCE, our approach is to develop a linear relationship between
the S*o and the CCE data, as shown in Fig. 21. In other words, we
assume that relative gas permeability should be evaluated at a
May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:02 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Stage:

Page: 185

Rich Fluid

Total Pages: 13

Fine Radical Model

100%

4,500

90%

4,000

80%

3,500

60%

Krg

So 50%
40%

BHP, psi

70%
So* = 0.48

3,000
2,500
2,000
1,500
1,000

CCE

500

30%

0
20%

10
qg, MMscfd

10%
0%

Analytical Solution

1,000

2,000

3,000
4,000
P, psi

5,000

Fig. 21Developing linear relation between

S*o

6,000

value between S*o and Max_SoCCE. So, for any pressure, we can go
to the figure and choose corresponding saturation and relative permeability. A careful examination of Fig. 15 and Figs. 18 through
20 tells us that an actual liquid dropout around the wellbore is
much greater than Max_SoCCE and is closer to the threshold S*o.
We have seen after testing several cases under this category that
the use of Krg (Max_SoCCE) will overestimate the gas rate because
it will not account for the relative permeability of the oil phase;
however, we use CCE data to account for changes in oil saturation
as the reservoir pressure declines.
It is very clear that condensate banking (accumulation) is tied
up with two factors. The first factor is fluid properties (maximal So
from CCE), and the second factor is rock properties (immobile So).
Accordingly, we have asked ourselves this question: Although
we know that the actual liquid dropout around the wellbore is
much greater than Max_SoCCE, how can we still use the CCE
data along with relative permeability curves to arrive at a robust analytical procedure that is accurate enough to estimate the well
productivity?
Other researchers have shown that relative permeability has a
first-order effect on condensate banking rather than the PVT properties (Mott 1997). As we have concluded from the results of the
sensitivity study, different fluids will have a similar productivity
loss for the same relative permeability curve used, confirming to
us that it is the relative permeability that is the most important in
determining the productivity loss.
We used engineering approximation to model the behavior
below the dewpoint pressure. As we have stated before, we are
going to assume that the area around the wellbore behaves similarly to the CCE data for every designated saturated pressure. Following the general procedure outlined previously for the case in
which (the Initial Reservoir Pressure Is Below the Pd subsection),
we are going to explain our approach at Pr 4,000 psi. After estimating the PI (J), as shown in Step 1 of the procedure, we can
estimate PI (J*) as the following:

J
Pr Krg SoCCE : . . . . . . . . . . . . . . . . . . . . . . . 20
J
At Pr 4,000 psi, we should estimate So from the linear relation between the S*o and CCE data, as shown in Fig. 21. The next
step is to go back to relative permeability curves to estimate Krg
at the corresponding So from this linear relation. After that, J* can
be calculated directly from Eq. 20. The last step before generating
the IPR curve is to estimate the gas rate directly from Eq. 18. Just
keep in mind that in this case the pseudopressure vs. rate plot will
have only one straight line because this is what we expect to see
in saturated reservoirs. The IPR curve is shown in Fig. 22 along
with the pseudopressure plot in Fig. 23. The complete IPR curves
of this case are shown in Fig. 24.
Fig. 25 shows the well PI producing at a constant-rate control.
We used rich fluid, and relative permeability parameters are
n 3, m 4, Sor 0.2, a 0.9, and oil endpoint relative
permeability 0.58.
We found that the PI ratio from flow-simulation data is
minimum well PI
 0:09: . . . . . . . . . . . . . . . . . . . . 21
maximum well PI
On the basis of the relative permeability data, we have
Krg(S*o) 0.08. As was expected, by use of our new method, we
are able to capture the change in PI below the dewpoint.
On the basis of the PI ratio, we can define the productivity loss
as
productivity loss 1 

minimum well PI
: . . . . . . . . . 22
maximum well PI

Therefore, in this example, the productivity loss is 0.91. This

means that this well will experience a 91% productivity loss as
soon as BHP reaches the Pd. Looking at Fig. 25, we can see that
the well restores some of its productivity after approximately 5
years of production. Fig. 26 shows the saturation profiles as a
function of time, which shows the re-evaporation process.
Fine Radial Model

Analytical Solution

8,000

Analytical Solution

9.0E+08
8.0E+08
7.0E+08
6.0E+08
5.0E+08
4.0E+08
3.0E+08
2.0E+08
1.0E+08
0.0E+00

7,000
6,000
BHP, psi

m(P), (psia2/cp)

Fine Radial Model

20

Fig. 22IPR curve for Pr 5 4,000 psi.

7,000

and CCE data.

15

5,000
4,000
3,000
2,000
1,000

10
qg, MMscfd

15

20

Fig. 23Pseudopressure vs. gas-rate plot for Pr 5 4,000 psi.

10

20

30
40
qg, MMscfd

50

60

70

Fig. 24IPR curves for rich gas (Corey-14).

May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:02 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

185

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Stage:

Productivity Index (Fine Radial Model)

500
450

Total Pages: 13

Radial Cell 1

Radial Cell 5

Radial Cell 10

Radial Cell 20

Radial Cell 30

Radial Cell 35

0.7

350

0.6

300
250

0.5

200

0.4
So

PI, Mscfd/psi

400

Page: 186

150

0.3

100
0.2

50
0

10

20
30
Time, Years

40

0.1

50

10

20
30
Time, Years

Fig. 25Well PI as a function of time.

Liquid Dropout Curve (CCE)

40

50

Fig. 26Oil-saturation profiles around the well as a function of

time.

4%
3%

Liquid Dropout (So)

TABLE 5FLUID PROPERTIES FOR THE FIELD CASE

3%

Initial reservoir pressure (psia)

Dewpoint pressure (psia)
Reservoir temperature ( F)
Maximal liquid dropout (%)

2%
2%

9,000
8,424
305
3

1%
1%
0%

1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000

Pressure, psi

Fig. 27CCE data for field-case fluid.

In summary, in the proposed method, with accurate relative

permeability and CCE data, one can predict the change in performance of the reservoir above and below the dewpoint without
conducting any flow simulation. We should explain that we have
not included non-Darcy and capillary effects in our analysis. On
the basis of a very limited number of runs (which are not included
here), we believe that our method should reasonably work when
both the effects are present.
Validation by Field Applications
In this section, we will show the application of our method for a
field case. Both compositional-model data and relative permeability curves were provided for a field case. A nine-component compositional model is used with the PR3 to simulate phase behavior
and laboratory experiments (CCE tests), as shown in Fig. 27.
Tables 5 and 6 show fluid properties and composition for the field
case, respectively.

The relative permeability curves are shown in Fig. 28. As is

common in field applications, what matters here is the threshold
S*o. Although Sor 0.20, the threshold S*o 0.32, which corresponds approximately to Kro 1% as a practical value. As we
have stated before, an accurate estimation of gas productivity in
this case depends on the value of Krg estimated at the threshold
S*o, which equals 0.32 in this case.
In this section, we illustrate the application of our method
from the production-operations view in the field. We are going to
show that our method can be used to generate IPR curves on the
basis of production-test data in the same way as the Vogel, Fetkovich, and other IPR correlations.
Assume that we are working in a production environment in a
field in which we have no knowledge about relative permeability
curves. The only thing that we have is some production data.
Because initial reservoir pressure is above the Pd, we know that the
pseudopressure vs. gas-rate plot will have two straight lines, as
explained earlier. Therefore, to generate the IPR curve for a given
Krg-Field Case

Kro-Field Case

1
0.9
0.8
0.7
0.6

Component

Composition (Fraction)

Kr

TABLE 6FLUID COMPOSITION FOR THE FIELD CASE

0.5
0.4

H2S
CO2
N2
C1
C2C3
C4C6
C7C9
C10C19
C20

186

0
0.0279
0.0345
0.7798
0.1172
0.0215
0.0132
0.00445
0.00145

Krg(So*) = 0.31

0.3
0.2
0.1

So* = 0.32 Sor = 0.2

0
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Sg

Fig. 28Relative permeability set of the field case.

May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:02 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Stage:

Production Test Data, (Pr = 8,900 psi)

2.38E+09

9,000

2.36E+09

8,000

2.34E+09
m(P), (psia2/cp)

10,000

7,000
BHP, psi

6,000
5,000
4,000
3,000

Total Pages: 13

Test Data Above Pd

2.32E+09
2.30E+09
2.28E+09
2.26E+09
2.24E+09

2,000

2.22E+09

1,000
0

Page: 187

2.20E+09
0

10

20

30
40
qg, MMscfd

50

60

70

8
10
qg, MMscfd

12

14

16

18

Fig. 30Pseudopressure vs. gas-rate plot.

Fig. 29Production data of two tests.

reservoir pressure, all we need is two test points. One point should
be above the Pd, and the other point should be below the Pd.
Fig. 29 shows an example of data from two production tests. To
explain how our method will work, we have chosen one of the test
data to be at the Pd. Otherwise, any available test data above the Pd
will suffice. We will illustrate our method by the following procedure.
Summary of New PI-Generation Procedure. Estimate the PI
(J) by using Pr and the test data at the Pd with the previously cited
Eq. 7:
qsc
:
J
mPr  mPwf 
Or another way to estimate J is to plot the test points above the
Pd on the pseudopressure plot, as shown in Fig. 30, and then J can
be calculated from the previously cited Eq. 8:
1
slope  :
J
2. Using J, you should be able to generate the first portion of
the IPR curve with Eq. 7.
3. In the same way as conducted previously, we need to plot
the test points below the Pd on the pseudopressure plot, as shown
in Fig. 31. Then, J* can be calculated from the slope as in the previously cited Eq. 9:
slope 

1
:
J

4. The intercept of the straight line (b) should be estimated as

with the previously cited Eq. 14:
qd
b mPd  :
J

5. Finally, we can estimate the gas rate for any BHP less than
the Pd with the previously cited Eq. 15:
q b  mPwf J  :
Results. The generated IPR curve and the pseudopressure plot
are shown in Figs. 32 and 33, respectively. We have shown that
one can use our method to generate the IPR curves on the basis of
available test data without any basic knowledge about the relative
permeability curves. Before we proceed to an example in which
initial reservoir pressure is below the Pd, we would like to highlight one important observation. We found that the PI ratio equals
0.30 with the preceding Eq. 10 as noted in Eq. 23:
 
1

slope of the line above Pd
J
J


1
slope of the line below Pd
J
 
J
 productivity ratio 0:30: 23
We would have expected this value to be equal to Krg (S*o).
From Fig. 28, we can see that S*o 0.32 and Krg (S*o) 0.31. We
can easily quantify the productivity loss after we have an idea
about the PI ratio or Krg (S*o) with Eq. 22 as the following:
productivity loss 1  productivity ratio:
Therefore, in this, after the FBHP drops below the Pd, the well
will lose 70% of its productivity. We would like to highlight two
major points here:
 Having two tests of production data clearly helped us to
characterize one value of relative permeability, which is Krg at
(S*o). By knowing Krg (S*o), one can easily use relative permeability
Fine Radical Model

Test Data Below Pd

Analytical Solution

10,000

3.E+09

9,000
8,000
7,000
2.E+09

BHP, psi

m(P), (psia2/cp)

2.E+09

1.E+09

6,000
5,000
4,000
3,000
2,000

5.E+08

1,000
0.E+00

10

20

30
40
qg, MMscfd

50

Fig. 31Pseudopressure vs. gas-rate plot.

60

70

10

20

30

40
50
60
qg, MMscfd

70

80

90

100

Fig. 32IPR curve for Pr 5 8,900 psi.

May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:02 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

187

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Fine Radial Model

Stage:

Analytical Solution

Page: 188

Total Pages: 13

Fine Radical Model

Analytical Solution

2.50E+09
10,000
9,000

m(P), (psia2/cp)

2.00E+09

8,000
1.50E+09

7,000
BHP, psi

1.00E+09
5.00E+08
0.00E+00

6,000
5,000
4,000
3,000

20

40
60
qg, MMscfd

80

100

2,000
1,000

Fig. 33Pseudopressure vs. gas-rate plot for Pr 5 8,900 psi.

curves (if available) to estimate (S*o) or maximal oil saturation

around the well.
 Although this field case has very lean gas with a maximal
liquid dropout of only 3%, as shown in Fig. 27, the loss in productivity is significant (70%). This clearly tells us that the most-important parameter in determining the productivity loss is the gas/
oil relative permeability curves, expressed in terms of Krg estimated at residual (or threshold) oil saturation.
We can easily generate the IPR curves for reservoir pressure
above the dewpoint pressure. The slope of IPR curves above and
below the dewpoint will remain the same, except that the curve
will shift downward as the reservoir pressure declines. If the reservoir pressure is below the dewpoint, to generate the entire IPR
curve, we will need just one observation.
The PI (J*) can be calculated directly on the basis of the test
data and reservoir pressure with the previously cited Eq. 16:
J

qsc
:
mpr  mpwf 

After that, the gas rate can be directly estimated from the previously cited Eq. 18:
q mPr  mPwf J  :
If we want to predict the future IPR on the basis of the current
IPR curve, we will need to have information about relative permeability and CCE data. The complete IPR curves of this case are
shown in Fig. 34 in which the generated IPR curves are validated
with the results of the fine-radial-compositional model.
With our methodology, we examined 10 years of the production history of a gas-condensate well with known CCE and rela-

Analytical Solution

10,000
9,000
8,000
Field Observed Data

BHP, psi

7,000
6,000
5,000
4,000
3,000
2,000
1,000
0

20

40
60
qg, MMscfd

80

100

Fig. 35Validation of the new method with field data.

188

20

40
60
qg, MMscfd

80

100

Fig. 34IPR curves for the field case.

tive permeability data. We used the initial production data and

calculated the future IPR curves with our methodology. We show
the future IPR curves in Fig. 35. Superimposed on those curves,
we also show the actual production data as a function of time as
the reservoir pressure has declined. The match between predicted
rates and observed rates is good, validating our procedure.
Conclusions
In this paper, a new analytical procedure is proposed to estimate the
well deliverability of gas-condensate reservoirs. Our new method
generates IPR curves of gas-condensate wells by incorporating the
effect of condensate banking as the pressure near the wellbore drops
below the dewpoint. Other than basic reservoir properties, the only
information needed to generate the IPR is the rock relative permeability data and the CCE-experiment data. In addition to predicting
the IPR curve under current conditions, our method also can predict
future IPR curves if the CCE data are available.
We found that the most important parameter in determining
productivity loss is the gas relative permeability at immobile oil
saturation. We observed that at low reservoir pressures, some of
the accumulated liquid near the wellbore revaporizes. This revaporization can be captured with CCE data. In our method, we propose two ways of predicting the IPR curves:
Forward approach: With the basic reservoir properties, relative
permeability data, and CCE information, we can predict the
IPR curves for the entire pressure range. A comparison with
simulation results validates our approach.
Backward approach: With field data to predict the IPR curves
for the entire pressure range, this method does not require reservoir data; instead, similar to the Vogel method, it uses point information from the IPR curve and predicts the IPR curve for the
entire BHP range. Both synthetic and field data are used to validate our second approach.
Nomenclature
h formation thickness, ft
J PI, [(Mscf/D)/psia2/cp]
K absolute rock permeability, md
Krg gas relative permeability
Kro oil relative permeability
m ( p) pseudopressure function, (psi2/cp)
Pd dewpoint pressure, psia
Pr average reservoir pressure, psia
Psc pressure at standard condition, psia
Pwf flowing BHP, psia
qsc gas-flow rate at standard conditions, Mscf/D
re external reservoir radius, ft
Rp producing GOR, scf/STB
rw wellbore radius, ft
May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:02 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

REE161933 DOI: 10.2118/161933-PA Date: 30-April-14

Sg
So
S*o
Sor
Swi
T
Tsc

gas saturation, fraction

oil saturation, fraction
threshold oil saturation, fraction
ROS, fraction
connate water saturation
reservoir temperature,  R
temperature at standard condition,  R

Acknowledgments
The first author would like to acknowledge the support from the
Reservoir Description and Simulation Department at Saudi Aramco. All the authors thank the University of Tulsa for computational and other administrative support.

Stage:

Page: 189

Total Pages: 13

Xiao, J.J. and Al-Muraikhi, A. 2004. A New Method for the Determination
of Gas Condensate Well Production Performance. Paper SPE 90290 presented at the SPE Annual Technical Conference and Exhibition, Houston, Texas, 2629 September. http://dx.doi.org/10.2118/90290-MS.

SI Metric Conversion Factors

ft  3.048*
E01 m
ft3  2.831685
E02 m3
bbl  1.589 873
E01 m3
psi  6.894757
E00 kPa
ft3/D 2.831685
E02 m3/d
*
cp  1.0
E03 Pa S

R
( R460)
C
*Conversion factor is exact.

References
Afidick, D., Kaczorowski, N.J., and Bette, S. 1994. Production Performance of a Retrograde Gas Reservoir: A Case Study of the Arun Field.
Paper SPE 28749 presented at the SPE Asia Pacific Oil and Gas Conference, Melbourne, Australia, 710 November. http://dx.doi.org/
10.2118/28749-MS.
Ahmadi, M., Hashemi. A., and Sharifi. M. 2014. Comparison of Simulation
Methods in Gas Condensate Reservoirs. Journal of Petroleum Science
and Technology 4 (1): 761771. http://dx.doi.org/10.1080/10916466.
2011.604063.
Blom, S.M.P. and Hagoort, J. 1998. The Combined Effect of Near-Critical
Relative Permeability and Non-Darcy Flow on Well Impairment by
Condensate Drop Out. Res Eval & Eng 1 (5): 421429. http://
dx.doi.org/10.2118/51367-PA.
Dyung, T., Vo, J.J., and Raghavan, R. 1987. Performance Prediction for
Gas Condensate Reservoir. Paper SPE 16984 presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas, 2730 September. http://dx.doi.org/10.2118/16984-MS.
Evinger, H.H. and Muskat, M. 1942. Calculation of Theoretical Productivity Factor. Trans, AIME 146: 126139.
Fevang, O. and Whitson, C.H. 1996. Modeling Gas Condensate Well
Deliverability. Res Eval & Eng 11 (4): 221230. http://dx.doi.org/
10.2118/30714-PA.
Guehria, F.M. 2000. Inflow Performance Relationships for Gas Condensates. Paper SPE 63158 presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas, 14 October. http://dx.doi.org/
10.2118/63158-MS.
Kamath, J. 2007. Deliverability of Gas-condensate Reservoirs- Field
Experiences and Prediction Techniques. J. Pet Tech 59 (4): 9499.
http://dx.doi.org/10.2118/103433-PA.
Kelkar, M. 2008. Natural Gas Production Engineering. PennWell
Corporation.
Mott, R. 1997. Calculating Well Deliverability in Gas Condensate Reservoirs. Paper-104 presented at the IBC Technical Services Conference on
Optimization of Gas Condensate Fields, Aberdeen, UK, 2627 June.
Mott, R. 2003. Engineering Calculations of Gas-condensate Well Productivity. SPE Res Eval & Eng 6 (5): 298306. http://dx.doi.org/10.2118/
86298-PA.
Mott, R.E., Cable, A.S., and Spearing, M.C. 2000. Measurements of Relative
Permeabilities for Calculating Gas-condensate Well Deliverability. SPE
Res Eval & Eng 3 (6): 473479. http://dx.doi.org/10.2118/68050-PA.
Sharifi, M. and Ahmadi, M. 2009. Two-phase Flow in Volatile Oil Reservoir Using Two-Phase Pseudo-pressure Well Test Method. J. Cdn.
Pet. Tech. 48 (9): 611. http://dx.doi.org/10.2118/09-09-06-PA.

Ali M. Al-Shawaf is the lead simulation engineer for offshore

gas fields in Saudi Aramcos Reservoir Description and Simulation Department. He started his career as a production field
engineer with Saudi Aramco; then, he joined the Reservoir
Simulation Division, in which he conducted several full-field
simulation models in onshore and offshore development and
increment fields. In May 2012, Al-Shawaf completed his masters thesis program at the University of Tulsa through which he
developed a new method to predict the performance of gascondensate wells. He has authored or coauthored several technical papers and has participated in several technical conferences and symposiums regionally and internationally. Al-Shawaf
is an active member of the SPE, in which he serves on several
committees for SPE technical conferences and workshops. He
earned a BS degree from King Fahad University of Petroleum &
Minerals, Dhahran, Saudi Arabia, and an MS degree from the
University of Tulsa, both in petroleum engineering.
Mohan Kelkar is the director of the Tulsa University Center for
Reservoir Studies. He is currently working with several mediumand small-sized oil and gas companies in relation to reservoir
characterization and optimization of tight gas reservoirs. Kelkar has published more than 50 refereed publications and has
made more than 100 technical presentations. He is a coauthor of the book Applied Geostatistics for Reservoir Characterization, published by the SPE in 2002, and the book Gas
Production Engineering, published in 2008 by PennWell Books.
Kelkar earned a BS degree in chemical engineering from the
University of Bombay. He earned MS and PhD degrees in petroleum engineering and chemical engineering, respectively,
from the University of Pittsburgh, USA.
Mohammad Sharifi is assistant professor of the Petroleum Engineering Department at the Amirkabir University of Technology.
He has authored/coauthored several technical papers that
have been presented at international conferences or published in journals. Sharifis main research is in the area of bridging static to dynamic models through efficient upgridding and
upscaling techniques and the simulation of gas-condensate
and naturally fractured reservoirs. He is currently working on
developing a fast dynamic method for ranking multiple reservoir realization models. Sharifi earned BS, MEng, and MS
degrees from the Petroleum University of Technology (PUT),
the University of Calgary, and PUT, respectively, all in reservoir
engineering. He earned his PhD degree in petroleum engineering from the University of Tulsa (2012).

May 2014 SPE Reservoir Evaluation & Engineering

ID: jaganm Time: 16:02 I Path: S:/3B2/REE#/Vol00000/140011/APPFile/SA-REE#140011

189