PETROLEUM SOCIETY PAPER 2006-101

CANADIAN INSTITUTE OF MINING, METALLURGY & PETROLEUM

Gas Condensate Reservoir Performance:

Part II Fluid Characterization
F.B. THOMAS, Hycal Energy Research Laboratories Ltd.
G. ANDERSEN, Chevron/Texaco
D.B. BENNION, Hycal Energy Research Laboratories Ltd.

This paper is to be presented at the Petroleum Societys 7th Canadian International Petroleum Conference (57th Annual Technical
Meeting), Calgary, Alberta, Canada, June 13 15, 2006. Discussion of this paper is invited and may be presented at the meeting if
filed in writing with the technical program chairman prior to the conclusion of the meeting. This paper and any discussion filed will
be considered for publication in Petroleum Society journals. Publication rights are reserved. This is a pre-print and subject to
correction.

Abstract
The salient findings of this work indicate that:
Gas condensate reservoirs may exhibit very poor
performance due to serious phase interference effects.
Engineer1 describes the Cal Canal field which exhibited
exceptional productivity decline incident to liquid drop
out. The operators recommendation for field
development was to abandon the reservoir after the
reservoir pressure reached the dew point; the forecasted
productivity was so poor that most of the hydrocarbon
was going to be left unrecovered.
Times and gas price have changed but gas condensate
production can still be very challenging. Phase
behavior, interfacial tension, velocity and pore size
distribution all affect how a condensate reservoir will
produce. Much can be done in the laboratory to gain
insight as to how serious phase interference effects are
going to be in the field, long before field problems are
encountered. This paper describes experimentation that
was done to quantify the impact of pressure depletion on
well productivity. The fluid preparation for this project
was described in part I of this two part series2.
1. Retrograde condensate resulted in very rapid
decrease in gas permeability 60 and 84%
reduction in Krg by the attainment of critical
condensate saturation.
2. End-point saturations such as trapped gas and
residual condensate saturation are sensitive to
the level of interfacial tension (therefore
pressure). Critical condensate saturation was
much less sensitive to level of IFT (pressure).
3. Two-phase testing as a function of capillary
number indicated an effective gas permeability
very close to that measured by the explicit
relative permeability measurements at
saturations just higher than the critical
condensate saturation.
4. The condensate and gas permeabilities
measured during the gas-phase hysteresis
injection did not agree with those measured
during the two-phase injection.
5. The two-phase injection approach and the
protocol whereby explicit measurement of

1
 relative permeability is performed provide a
very thorough gas-condensate reservoir data
set, which are amenable for use in simulation
and reservoir production forecasting.
Background
When retrograde condensation occurs during pressure
decrease the liquid that accumulates fills some of the
pores of the porous media. As the liquid increases in
saturation the cross-sectional area available for flow may
reduce the permeability to gas. The condensate may be
mobilized if the draw-down pressure is large enough to
overcome capillary pressure; this case results in the
liquid having less severe impact on gas production. In
some cases, once liquids accumulate, the gas
permeability drops quickly and is difficult to restore.
Laboratory testing helps the operator to identify which
of the two scenarios will be germane to his specific
reservoir. If appropriate testing is done some of the
uncertainty can be removed from field development
(references 3 5).
To gain an idea of whether a reservoir is going to
experience severe liquid phase interference effects one
can saturate the core stack, in the laboratory, with
equilibrium fluid and then measure regain gas
permeability (defined as the percentage of the single-
phase permeability that is achieved by flowing
equilibrium gas through the core sample) as a function
of drawdown pressure at difference absolute pressures.
Figure 1 shows the type of testing that can be performed.
The change in absolute pressure (back pressure
decreases from3000 to 2000 psi) had a drastic
deleterious impact on gas permeability. Although overly
pessimistic due to the initial 100% liquid-saturated
condition this behavior teaches a valuable lesson: in the
field the absolute pressure and the differential pressure
are closely coupled and the capillary pressure influences
are linked to both. As figure 2 shows however, the IFT
and therefore the capillary pressure (assuming the Pcap
is proportional to IFT) increases exponentially with
decreasing pressure whereas the drawdown is only
linear. Consequently, as the bottom-hole flowing
pressure is reduced it becomes more difficult to
overcome capillary pressure effects; the corollary is that
the harder you pull some wells the worse they perform.
The quantification of these effects was desired for a rich
HPHT gas condensate reservoir. Was liquid drop out
going to severely impact gas production?. The next
section describes work done to answer this.
Gas Condensate Relative Permeability
Testing
The experimental testing was done using two separate
core stacks, each comprised of four core plugs of 3.81
2
cm diameter. The properties of the stacks are shown in
Table 1.
The stacks were prepared with the relevant water
saturations and then a base-line permeability to
humidified methane was measured. The methane
permeabilities for the two stacks at Swi were 122.1 and
28.4 mD respectively at a pore pressure of 9000 psi at
246 F. Compared to the single-phase gas permeabilities
of 136.1 and 36.0 mD this decrease in gas permeability
in the presence of water would be expected since it is
thought that most of the water would be associated with
the smaller pores which contribute little to the overall
permeability of the samples. The core stack was then
mounted in an oven at reservoir temperature and with
two pressure-control valves (PCV): one on the injection
end and one on the production end. With the PCV in
place the pressure of the injection fluid could be
maintained above dew point pressure so that single
phase fluid would enter the valve, flash into two phases
and then liquid and gas flow through the stack. With the
liquid drop-out characteristics previously measured (See
part I of this series) 1 hydrocarbon pore volume of dew
point fluid could be injected to the stack at a pressure
corresponding to 5.9% liquid drop out (3000 psi); by
injecting a pore volume of fluid 5.9% HCPV liquid
could be accumulated (or any fraction thereof).
Although there would be an initial distribution of
condensate along the stack eventually the condensate
would reach a critical saturation above which it would
start to flow. It was considered that due to the relatively
homogeneous stack the average critical condensate
saturation (CCS) measured would be representative of
the overall stack. Once the specific volume of dew point
fluid was injected, with its corresponding volume of
condensate, equilibrium gas was injected and the end-
point permeability to equilibrium gas measured. Figure
3 shows the relative permeability relationship for the
second stack. It showed a critical condensate saturation
of about 19% at which point the gas permeability had
decreased to 17% of its original value. This Scc was
considered high but some porous media has been
reported with Scc values well in excess of this (1).
Once the condensate was observed to be mobile, by
recording accumulations of condensate in the effluent
visual cell maintained at 3000 psi and 246 F, the stack
was flooded with equilibrium liquid. Part I of this paper
series suggests that it is required to follow a path for
equilibrium liquid preparation. At 3000 psi sufficient
liquid drop-out occurs to provide volumes suitable for
flooding (12.8 % or 91 cc for every 700 cc of dew point
fluid). Moreover, 3000 psi was thought to be
conservative from a bottom-hole flowing pressure over
the first years of the project and that phase interference
effects would not be expected to be this serious for many
years (reservoir pressure of 12,025 psi). For the higher
pressure regime however, with only 2.2% liquid drop
out, insufficient separator liquid was available and so a
synthetic liquid was prepared. It had properties that
were not representative (GOR too high, IFT too low). It
resulted, when flooded into the core stack, in a high-
mass-transfer displacement and was not deemed as
representative of the process.
Once the Scc had been attained, the stack was then
flooded with equilibrium condensate at 3000 psi and the
displacement history-matched with a two-phase
compressible relative permeability model. The trapped
gas saturation was 12.6% of the pore volume or 16.6%
of the initial hydrocarbon in place. Following the
condensate saturation the equilibrium gas was then
injected at increasing rates up to a rate corresponding to
field rates of 45 MMscfD (well-bore of 8.5 inches and a
completion interval of 200 feet 1013 scf/ft2-day). The
residual condensate saturation to this high rate was
43.3% of the trapped gas end-point condensate
saturation. The high-rate gas permeability was
Forchheimer corrected to provide a value of 24.8 mD or
a regain permeability of 87.5% of the initial Kg value.
Figure 4 shows the same relationship but at 9000 psi.
Very little change was observed from the perspective of
gas permeability reduction. The condensate injection
cycle was very different however and is due to the
inadequacy of the synthetic liquid phase at representing
the equilibrium condensate phase at 9000 psi (see part I
of this series).
Similar testing was done for the higher-permeability
stack. The results from this testing are summarized as
follows:
1. The high- and low-permeability stacks
exhibited Scc values of 14 and 19%
respectively. Critical condensate saturation
changed very little as a function of IFT change
from 3.4 to 0.25 dyne/cm (3000 and 9000 psi).
This is not unexpected since the larger porous
features are thought to govern the first
condensate mobilized and is not expected to be
a strong function of IFT. Residual condensate
saturation is considered to be a strong function
of IFT although the data measured herein were
insufficient to judge conclusively (lack of
available low-IFT condensate phase).
2. The gas-phase permeability at Scc decreased
by 60 and 84% for the high- and low-
permeability stacks respectively.
3. Trapped gas saturations were 13.5% PV for the
high-permeability stack at 3000 psi and 12.6%
PV for the low-permeability stack. The higher
pressure trapped gas saturations were not
measured since the synthetic liquid phase had
excessive mass transfer and trapped gas values
were unrealistic (apparent value of 3% was
determined).
3
4. The importance of using representative
equilibrium fluids became paramount since
subtle changes in IFT could affect the
displacement efficiency significantly; analogue
fluids are very difficult to create that possess
the viscosity, density, IFT characteristics of the
true equilibrium phases.
Two-Phase Steady-State Injection
Some authors6 have suggested that the high rates
experienced near production wells clean up any
condensate that might drop out contrary to what has
been measured here and elsewhere. To simulate more
directly what might happen in the field the experimental
protocol was modified so that multiple volumes of dew
point fluid could be injected until steady state was
observed. Once stable differential pressure was attained
the permeability of each phase was calculated based on
the volume fraction condensed, as measured in the
constant composition expansion. Figure 5 shows the
liquid drop-out as a function of pressure below the dew
point and figure 6 shows an example of the differential
pressure history from fluid injection into the second core
stack at different rates.
A number of rates were examined corresponding to a
range of capillary number from 6.99x10-7 to 4.44x10-5.
Appendix A shows the way the authors calculated the
capillary number.
Tables 2 and 3 show the results of the two-phase
injection and include the viscosity of the phases, the
phase fraction, the calculated end-point permeabilities,
the back pressure, fluid type, differential pressure,
superficial and pore-scale capillary numbers. The
superficial capillary numbers were corrected for porosity
and water saturation to reflect pore-scale velocities: this
corresponds to the columns of pore-scale capillary
numbers.
Contrasting the permeabilities for the gas to those
measured as part of the critical condensate
experimentation provides corroboration of the different
methods. Figure 3 shows the gas-phase permeability to
be approximately 5 mD at the critical condensate
saturation whereas table 2 reports gas-phase
permeabilities of 1.5 to 2.4 mD. The value of 2.4 mD
would be expected to occur at about 25% condensate
saturation from figure 3 which may be the equilibrium
saturation at which the accumulation of condensate
ceased. This would be expected since the residual
condensate saturation is always higher than critical
condensate saturation and the residual condensate
saturation will largely dictate the end-point gas
permeability. The two techniques agree very well with
each other. On the basis of flux calculations, the highest
capillary number that was possible to run in the
laboratory corresponded to rates of 11.25 MMscfD
compared to 45 MMscfD which was contemplated as the
expected production rate at which the operator would
produced the well. Figure 7 shows the plot of
permeability as a function of capillary number. This
relationship suggests that although rate may be
increased, the permeability to gas is already approaching
asymptotic and therefore the rate used in the
experimentation is thought to yield gas permeabilities
that would be representative.
Figure 8 shows the condensate permeability as a
function of capillary number. The values again appear
to be approaching asymptotic behavior at the rates
reached in the laboratory. Figure 9 shows all of the data
on the same figure with an R2 value of 0.54.
Discussion
The values that derive from this plot place the
permeabilities in the range of gas-phase permeabilities
that were observed just above the critical condensate
saturation in the critical condensate testing. This seems
intuitively consistent and corroborates both experimental
procedures.
Ultimately flow in porous media should yield to
quantification by use of an expression such as equation
1. However, the function K(r) is not known and appears
to be relatively specific to each fluid and rock system.
Some simulators use explicit relative permeability
curves such as those in figures 3 and 4 whereas others
use a permeability as a function of capillary number.
The authors have insufficient experience with both to
judge the superiority of either.
RMax
P
K (r ) f (r )2rdr
dQ = Equation 1
L O
The relatively close agreement between the two
techniques used in this work is pleasing from a
forecasting perspective. The explicit relative-
permeability measurements allow for a more complete
set of data to be measured including trapped gas,
residual condensate and critical condensate saturations,
providing that fluids can be created which represent
those experienced in situ, but the two-phase continuous
flow approach provides empirical evidence of what
permeability reduction may be experienced. The two
data-sets when taken together are conclusive as to what
productivity challenges will experienced during pressure
depletion.
Summary
The results of this work can be summarized as follows:
4
1. Retrograde condensate resulted in very rapid
decrease in gas permeability 60 and 84%
reduction in Krg by the attainment of critical
condensate saturation.
2. End-point saturations such as trapped gas and
residual condensate saturation are sensitive to
the level of interfacial tension (therefore
pressure). Critical condensate saturation was
much less sensitive to level of IFT (pressure).
3. Two-phase testing as a function of capillary
number indicated an effective gas permeability
very close to that measured by the explicit
relative permeability measurements at
saturations just higher than the critical
condensate saturation.
4. The condensate and gas permeabilities
measured during the gas-phase hysteresis
injection did not agree with those measured
during the two-phase injection.
5. The two-phase injection approach and the
protocol whereby explicit measurement of
relative permeability is performed provide a
very thorough gas-condensate reservoir data
set, which are amenable for use in simulation
and reservoir production forecasting.
REFERENCES
1. Engineer, R., SPE paper 13650, 1985.
2. Thomas, F.B., G. Andersen and D. B. Bennion,
Gas-Condensate Reservoir Performance:
Part I Fluid Characterization, CIPC 2006,
paper 100.
3. Chen, H.L., S.D. Wilson and T.G. Monger-
McClure, Determination of Relative
Permeability and Recovery for North Sea
Gas-Condensate Reservoirs, August 1999,
Volume 2, Number 4, pp 393.
4. Fevang, O., C.H. Whitson, SPE 30714,
Modeling Gas Condensate Well
Deliverability.
5. Saevareid, A., C.H. Whitson and O. Fevang,
An Engineering Approach to Measuring
and Modeling Gas Condensate Relative
Permeabilities, personal communication.
6. M.R. Carlson and J.W.G Myer, Reduced
Productivity Impairment for Fracture
Stimulated Gas Condensate Wells, JCPT,
1999, Vol. 38, No. 13.
 Table 1
Routine Permeability and Porosity Results

Stack 1
After
Provided Cleaning Target Swi (%) Length (cm) Order*
Plug # , % K, mD , % K, mD
203A 18.4 151 18.9 169 15.0 6.31 2
209A 17.9 101 18.6 116 15.0 6.33 1
210AH 18.5 120 19.4 118 15.0 6.50 4
212AH 18.5 166 19.2 157 15.0 6.44 3
Khavg
25.58 136.06

Stack 2
After
Provided Cleaning Target Swi (%) Length (cm) Order
Plug # , % K, mD , % K, mD
216A 17.3 32.2 18.0 39.9 25.0 6.49 3
224B 17.0 25.0 17.6 34.4 25.0 6.23 4
225 17.1 31.0 17.8 40.7 25.0 6.34 1
225AH 16.3 25.3 17.7 32.7 25.0 6.35 2
Khavg
25.41 36.63

* Huppler Ordering 1 at injection and 4 at production end.

5
 Table 2
Calculations for Two-Phase Injection at 3000 psi Back Pressure

RATE Kg Kc FLUID TYPE dP Nca Nca Pore Level Pore Level

(cc/hr) (mD) (mD) (psi) gas liquid Nca gas Nca liquid
150 1.0467099 GAS 85 6.996E-07 4.910E-06
150 1.51520152 CONDENSATE 85 1.013E-06 7.107E-06
2000 2.64792683 GAS 448 9.328E-06 6.546E-05
2000 3.83309909 CONDENSATE 448 1.350E-05 9.476E-05
150 1.48283903 GAS 60 6.996E-07 4.910E-06
150 2.14653549 CONDENSATE 60 1.013E-06 7.107E-06
2800 2.91364861 GAS 570 1.306E-05 9.165E-05
2800 4.21775394 CONDENSATE 570 1.891E-05 1.327E-04
1000 2.11834147 GAS 280 4.664E-06 3.273E-05
1000 3.06647927 CONDENSATE 280 6.752E-06 4.738E-05
500 1.57748833 GAS 188 2.332E-06 1.637E-05
500 2.28354839 CONDENSATE 188 3.376E-06 2.369E-05
2000 2.67178203 GAS 444 9.328E-06 6.546E-05
2000 3.86763151 CONDENSATE 444 1.350E-05 9.476E-05
2800 2.94987515 GAS 563 1.306E-05 9.165E-05
2800 4.27019493 CONDENSATE 563 1.891E-05 1.327E-04
1000 2.11834147 GAS 280 4.664E-06 3.273E-05
1000 3.06647927 CONDENSATE 280 6.752E-06 4.738E-05
500 1.57916829 GAS 187.8 2.332E-06 1.637E-05
500 2.28598029 CONDENSATE 187.8 3.376E-06 2.369E-05
150 1.03453886 GAS 86 6.996E-07 4.910E-06
150 1.4975829 CONDENSATE 86 1.013E-06 7.107E-06

Table 3
Calculations for Two-Phase Injection at 6000 psi Back Pressure

RATE Kg Kc FLUID TYPE dP Nca Nca Pore Level Pore Level

(cc/hr) (mD) (mD) (psi) gas liquid Nca gas Nca liquid
160 6.07643998 GAS 14.8 2.542E-06 1.784E-05
160 3.49807601 CONDENSATE 14.8 1.464E-06 1.027E-05
300 2.89727164 GAS 58.2 4.767E-06 3.345E-05
300 1.66789707 CONDENSATE 58.2 2.744E-06 1.926E-05
2000 5.09814692 GAS 220.5 3.178E-05 2.230E-04
2000 2.93489371 CONDENSATE 220.5 1.83E-05 1.284E-04
2800 5.24599318 GAS 300 4.449E-05 3.122E-04
2800 3.02000562 CONDENSATE 300 2.561E-05 1.797E-04
160 2.36661346 GAS 38 2.542E-06 1.784E-05
160 1.36240855 CONDENSATE 38 1.464E-06 1.027E-05
300 2.98973775 GAS 56.4 4.767E-06 3.345E-05
300 1.72112782 CONDENSATE 56.4 2.744E-06 1.926E-05
2000 5.16846619 GAS 217.5 3.178E-05 2.230E-04
2000 2.975375 CONDENSATE 217.5 1.830E-05 1.284E-04
2800 5.377748 GAS 292.65 4.449E-05 3.122E-04
2800 3.09585405 CONDENSATE 292.65 2.561E-05 1.797E-04

6
 Figure 1
Effect of Absolute Pressure on Flow

90

80

70

60
Regain %

50

40

30

20

10

0
20 80 200 800 2000

Delta P (psi)

Pbhf = 3000 psia Pbhf = 2000 psia

Figure 2
Relative change in Capillary Pressure and Differential Pressure

3.5

3
Magnitude of Change

2.5

1.5

0.5

0
0 2000 4000 6000 8000 10000 12000
Pressure (psi)
Capillary Pressure Change Fractional Drawdown change

7
 Figure 3 - Stack #2
Temperature = 246 F and Pressure = 3000 psia
Results of Condensate Interference Testing
40

35

30
Scr
P erm e a b ility (m D )

25

20

15
Scc = 19.4%

10

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Gas Saturation
Gas Permeability 1-Swi Hysteresis Gas Liquid Perm

Figure 4 - Stack #2
Temperature = 246 F and Pressure = 9000 psia
Results of Condensate Interference Testing
40.00

35.00

30.00

25.00
Gas Permeability (mD)

20.00

15.00 Scc = 19.3%

10.00

5.00

0.00
0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 1.000

Gas Saturation
Gas Permeability 1-Swi Liq. Perm Hyst. Kg

8
 Figure 5
CCE - LIQUID FRACTION (Liquid / Total Volume)

0.1000

0.0900

0.0800

0.0700
LIQUID FRACTION

0.0600

0.0500

0.0400

0.0300

0.0200

0.0100

0.0000
0 2000 4000 6000 8000 10000 12000 14000
PRESSURE (psi)

Figure 6
Injection at 10,600 psi and BPR at 6000 psi
Differential Pressure History

250
2000 cc/hr at
10,600 psi

200
Delta P (psi)

150

300 cc/hr at
10,600 psi
100
160 cc/hr at
10,600 psi

50
refill pump

0
0 1000 2000 3000 4000 5000 6000
Time (s)

9
 Figure 7
Gas Permeability vs Capillary Number

6
Permeability (mD)

0
0.0E+00 1.0E-05 2.0E-05 3.0E-05 4.0E-05 5.0E-05
Capillary Number
Gas 3000 PSI Gas 6000 PSI

Figure 8
Condensate Permeability versus Capillary Number

4.5

3.5
Liquid Permeability

2.5

1.5

0.5

0
0.0E+00 5.0E-06 1.0E-05 1.5E-05 2.0E-05 2.5E-05 3.0E-05
Capillary Number
Condensate - 3000 PSI Condensate - 6000 PSI

10
 Figure 9
Phase Permeabilities vs Capillary Number

5
Capillary Number

3
y = 0.7541711Ln(x) + 12.0105453
2
R = 0.5342982
2

0
0.0E+00 1.0E-05 2.0E-05 3.0E-05 4.0E-05 5.0E-05
Permeability (mD)

11