Phenomenological Modeling of Critical
Condensate Saturation and
Relative Permeabilities in Gas/
Condensate Systems
Kewen Li, SPE, and Abbas Firoozabadi, SPE, Reservoir Engineering Research Inst.
Summary
The effects of gravity, viscous forces, interfacial tension, and wet-
tability on the critical condensate saturation and relative perme-
ability of gas condensate systems are studied using a phenomeno-
logical simple network model. The results from the simple model
show that wettability significantly affects both critical condensate
saturation and relative permeability. Relative permeability at some
saturations may increase significantly as the contact angle is al-
tered from 0° 共strongly liquid-wet兲 to 85° 共intermediately gas-
wet兲. The results suggest that gas well deliverability in condensate
reservoirs can be enhanced by wettability alteration near the well-
bore.
Introduction
Much attention has been paid to the study of in-situ liquid forma-
tion and fluid flow mechanisms in gas condensate systems in re-
cent years. When gas condensate reservoirs are developed by
pressure depletion, gas well deliverability is affected by the
amount and the distribution of the condensate formed around the
wellbore. Understanding of the parameters that affect the distribu-
tion and the amount of condensate saturation (S c ) in the wellbore
and the effect of these parameters on gas and liquid flow are
important in the development of the methods for increased gas
deliverability.
There are numerous field examples of gas condensate reser-
voirs that experience a sharp drop in gas well deliverability at
high pressures due to condensation near the wellbore.1-3 There are
examples that show that the condensation may even kill gas
production.2 The sharp reduction in gas deliverability may be due
to the shape of the gas phase relative permeability (k rg ); the
mechanisms of gas productivity impairment are not yet clear. Two
main parameters affect condensate recovery and gas well deliver-
ability. These two parameters are critical condensate saturation
共S cc , which affects liquid recovery兲 and gas phase relative per-
meability 共k rg , which affects well deliverability兲.
Various authors4-9 have measured S cc and a wide range of ob-
servations has been reported. The measured values of S cc are gen-
erally in the range of 10 to 50%. Gravier et al.4 measured S cc in
low-permeability carbonate cores with permeability in the range
of 0.5 to 40 md and interfacial tension 共␴兲 variations from 0.5 to
1.5 dyn/cm. S cc varied from 24.5 to 50.5% with an average of
35.0%. These measurements were conducted in the presence of
connate water. Morel et al.9 reported a critical condensate satura-
tion of less than 1% for ␴ ⫽0.05 dyn/cm and connate water satu-
ration of 20%. The critical condensate saturation was measured in
a vertical dolomite core. The measured S cc is apparently related to
various effects including gravity and interfacial tension.
Some authors have accounted for the effect of gravity by con-
ducting flow experiments both in horizontal and vertical direc-
tions. Danesh et al.10 studied the effect of gravity on S cc in water-
wet micromodels and sandstone cores. They found that S cc in a
vertical core was lower than that in a horizontal core; their tests
Copyright © 2000 Society of Petroleum Engineers
Original SPE manuscript received for review 11 February 1999. Revised manuscript re-
ceived 4 January 2000. Paper (SPE 56014) peer approved 9 February 2000.
138 SPE Journal 5 共2兲, June 2000
showed that gravity reduced the critical condensate saturation.
These authors also examined the dependence of critical conden-
sate saturation on ␴. As expected, S cc increases with an increase
in ␴. Ali et al.11 observed an improved condensate recovery of
17.2% in a vertical gas injection compared with that in a horizon-
tal injection. Their experiments were conducted with a high inter-
facial tension for gas condensates ( ␴ ⫽0.92 dyn/cm). Sandstone
cores and a synthetic six-component fluid were used by these
authors. When ␴ was decreased to 0.04 dyn/cm, condensate re-
covery improved to 27.5% with vertical injection. Henderson
et al.12 also reported a significant reduction of the critical conden-
sate saturation in long vertical core samples in comparison with
those in horizontal cores. Munkerud,6 however, did not find a
significant difference in condensate recovery and saturation be-
tween horizontal and vertical pressure depletion. He used Berea
cores and a synthetic six-component fluid.
A number of authors have studied gas phase relative permeabil-
ity (k rg ) of gas condensate systems. Gravier et al.4 found an
abrupt decrease in k rg at saturation close to S cc . Chen et al.5
measured k rg at reservoir pressures with reservoir rock and reser-
voir fluids from two North Sea gas condensate reservoirs 共reser-
voirs A and B兲. The interfacial tension ␴ varied from 0.01 to 0.42
dyn/cm. The results showed that k rg reduced about 10-fold due to
condensation when S c reached 20% for the core and fluid from
reservoir A. Reduction of k rg with the increase of S c for the core
and fluid from reservoir B is less pronounced but an abrupt de-
crease of k rg at the saturation close to S cc was also observed.
Chen et al.5 showed that k rg is rate dependent; it increases with
the increase of the flow rate. Henderson et al.13 measured gas
condensate relative permeabilities using a long Berea core and a
five-component synthetic gas condensate fluid; ␴ varied from 0.05
to 0.40 dyn/cm. They found that the relative permeabilities of both
gas and liquid phases decreased with the increase of ␴ and in-
creased with the increase of the flow rate. Gas relative permeabil-
ity was more sensitive to ␴ and flow rate than liquid relative
permeability. The results from the work of Munkerud6 and of
Haniff and Ali14 differ from the other authors. They did not ob-
serve the abrupt decrease of k rg .
The brief review above points out that there are many unre-
solved issues in the study of critical condensate saturation and gas
condensate relative permeability. The sharp decrease in gas rela-
tive permeability at saturation close to S cc that was observed ex-
perimentally has not yet been studied theoretically. The depen-
dence of S cc , k rg , and k rc on interfacial tension, viscous force,
and gravity are not clear from the experimental data. The question
remains how to measure S cc , k rg , and k rc in the laboratory and
how to scale them to field conditions. The conclusions drawn
from the experimental results are sometimes not consistent. A
theoretical understanding of the effect of ␴, gravity, viscous force,
and wettability on S cc , k rg , and k rc provides insight into the
recovery of liquid condensate from rich gas condensate reservoirs,
and gas well deliverability from both rich and lean gas condensate
reservoirs.
Theoretical work on the critical condensate saturation and rela-
tive permeabilities of gas condensate systems is very limited. Mo-
hammadi et al.15 found that the relative permeabilities in gas con-
densate systems follow the conventional displacement process.
1086-055X/2000/5共2兲/138/10/$5.00⫹0.50
 Fig. 1–Size and distribution of the 20Ã20 network.
These authors used percolation theory in their work for horizontal
porous media. The effect of interfacial tension, viscous force,
gravity, and wettability on critical condensate saturation and gas
condensate relative permeability was not included in the work of
Mohammadi et al.15 Fang et al.16 developed a phenomenological
network model for critical condensate saturation. The main con-
clusion from this model is that critical condensate saturation is a
function of interfacial tension and contact angle in gas-oil-rock
systems. The model of Fang et al.16 is based on the assumption of
negligible viscous forces and large tube length.
In this work, we first modify the S cc model of Fang et al.16 for
predicting critical condensate saturation to include the effect of
viscous forces and to relax the assumption of large tube length.
We then develop a relative permeability model 共k r model兲 for gas
condensate systems from a renormalization technique. Using the
model, the effects of interfacial tension, flow rate, and contact
angle hysteresis on the critical condensate saturation and gas and
condensate relative permeabilities are investigated. The results
from the modified S cc model and the k r model show that S cc and
k r are significantly affected by interfacial tension, gravity, flow
rate, and wettability. Wettability may have the most pronounced
effect on both S cc and k r .
Network Model for Condensation in Porous Media
The porous medium used in this study is represented by a simple
two-dimensional 共2D兲 pore network with circular capillary tubes
of length L and cubic intersections of side width b; we assume b
to be equal to the maximum diameter of the capillary tube in the
network. There are four tubes connected to each intersection;
therefore, the coordination number is four. A log-normal size dis-
tribution is assumed with a standard deviation of 5.74 and an
average radius of 23 ␮m for the capillary tubes. The size distri-
bution of the network with 20⫻20 intersections is sketched in
Fig. 1. In this sketch, the tube diameter is proportional to the
thickness. The total pore volume of the network with a tube length
of 6000 ␮m is about 8.93 mm3; the total volume of the intersec-
tions with a side width of around 32.67 ␮m in the network is
about 0.03 mm3 and the estimated permeability of the network is
about 1 darcy. The value of threshold radius r p used in this study
is 20 ␮m 共from condensation, liquid will first form in tubes with
r⭐r p 兲. Note that the simple network described is selected for
some sensitivity studies.
K. Li and A. Firoozabadi: Phenomenological Modeling of Critical Condensate Saturation
Fig. 2–Liquid configuration in a single circular tube.
Critical Liquid Length and Critical Condensate
Saturation in a Single Capillary Tube
Fang et al.16 developed a phenomenological model for the critical
condensate saturation in a porous medium represented by a net-
work. Their model is based on two assumptions: viscous forces
are negligible and the critical height (h cc ) of liquid formed in-situ
in the vertical capillary tubes is less than the tube length. The
critical height of liquid in a vertical capillary tube is the maximum
length of the liquid column before it becomes mobile. In this
work, we extend the model of Fang et al.16 to include the effect of
viscous forces, and provide an option for h cc ⬎L. When a liquid is
first formed in a capillary tube, we assume that it appears in its
middle in the form of a liquid bridge. During growth, interfacial
tension, gravity, contact angle hysteresis, viscous force, and the
size of capillary tubes are the main factors that affect the stability
of the new liquid phase.
The critical height h cc of liquid formed in-situ in a vertical
circular tube of radius r is expressed as 共see Fig. 2a兲16
2␴ p R ⫺p A
h cc ⫽ 共 cos ␪ R ⫺cos ␪ A 兲 ⫺ , 共1a兲
⌬ ␳ gr ⌬␳g
where p R is the pressure at the top 共or the receding side兲 and p A
the pressure at the bottom 共or the advancing side兲 of the liquid
column; ␪ R is the receding contact angle and ␪ A the advancing
contact angle; ⌬ ␳ is the density difference between the gas and
liquid phase. The equation above is derived from the balance be-
tween capillary pressure, viscous force, and gravity in a vertical
capillary tube. In this study, for simplicity it is assumed that the
gas density ␳ g is equal to zero 共that is, ⌬ ␳ ⫽ ␳ c 兲 and ␴ does not
change with the pressure. The effect of pressure on ␴ can be
readily considered in our model. In a horizontal capillary tube, the
critical length of liquid is indirectly calculated from
共 p R ⫺p A 兲 r⫽2 ␴ 共 cos ␪ R ⫺cos ␪ A 兲 . 共1b兲
The above equation implies that without viscous forces 共that is,
when p R ⫺p A ⫽0兲 liquid cannot be moved, unless ␪ R ⫽ ␪ A . In a
vertical tube, liquid will flow downward to the bottom of the tube
when the height of the liquid column exceeds h cc . The advancing
contact angle will then change to 90°. At this time, the critical
height of the liquid column will increase and can be calculated
from Eq. 1a. Once h cc is determined, the critical condensate satu-
ration in a single vertical capillary tube is calculated as h cc /L.
SPE Journal, Vol. 5, No. 2, June 2000 139
The calculation of S cc in a horizontal capillary tube is relatively
simple. S cc is equal to zero when the viscous force r(p R ⫺p A ) is
greater than the interfacial force 2 ␴ (cos ␪R⫺cos ␪A). Otherwise,
S cc is equal to 1 or less, according to Eq. 1b. S cc in a single
horizontal capillary tube may be less than 1 because of limited
condensation.

Process of Condensate Formation, Growth, and Flow

for h cc Ì L
Liquid condensation may take place in capillary tubes of smaller
diameters based on the thermodynamics of curved interfaces.17
The condensation process in a network is represented by discrete
steps.16 When pressure decreases below the dewpoint pressure,
the gas in those tubes with radius r⭐r p is allowed to condense.
The amount of condensation for each step in a liquid bridge is
assumed to be 200 ␮ m3. There is no initial condensation in
pores with r⬎r p . Presented in Fig. 3 is the process of in-situ
condensate formation, growth, and flow when h cc ⬎L in vertical
tubes and the viscous force is greater than the interfacial force in
horizontal tubes. It is known from our modeling results that criti-
cal condensate saturation and relative permeabilities are affected
by the choice of r p –the larger the value of r p , the greater the
critical condensate saturation. But, the trend of the effects of grav-
ity, wettability, and viscous forces on critical condensate satura-
tion and relative permeabilities at various values of r p is similar.
Therefore, only the results for r p ⫽20 ␮ m are presented in this
paper. For the study of viscous effect, it is assumed that gas is
injected at the top of the network.
As the pressure in the system decreases below the dewpoint
pressure, gas is assumed to condense in the middle of tube 3 (r 3
⬍r p ) in the form of a liquid bridge, as shown in step 1. The top
receding contact angle ␪ R 共measured through the liquid phase兲
will be less than the bottom advancing contact angle ␪ A due to
gravity and viscous effects.
Continuous condensation increases the liquid height in tube 3.
Assuming the critical liquid height h 3cc ⬎L, the advancing con-
tact angle will increase and reach 90°. This is shown in step 2.
With further condensation, the advancing contact angle will ex-
ceed 90°. The receding contact angle at the top will also increase
and reach 90° 共see step 3兲.
If both the advancing and the receding contact angles reach 90°
at the same time, the liquid column becomes unstable and will
flow downward. The advancing contact angle will then exceed
90° The receding and advancing contact angles will continue to
increase with further condensation in tube 3 共see step 4兲. Given
r 2 ⬍Min(r 4 ,r 5 ,r 6 ), the condensate formed in-situ will contact
corner point ‘‘A’’ shown in step 5. The top interface has the
maximum radius. At this stage, the condensate becomes unstable
and forms two new interfaces, one in tube 2 and another in the
intersection as shown in step 6. Because the radius of the interface
in tube 2 is smaller than that of the intersection, the liquid pres-
sure in the intersection is higher than that in tube 2. The liquid
will be pulled into tube 2 as a result. The new configuration is
shown in step 7.
Now, the liquid interface in tube 2 serves as a new condensa-
tion site. After tube 2 is filled by the condensate, the contact angle
at the left end of tube 2 will increase until it reaches 90° 共i.e., flat
interface兲; then there will be no further condensation. However,
condensation will continue at the right end of tube 2; as a result, a
new liquid interface in the intersection will develop. The liquid
interface will touch corner point ‘‘D’’ with further condensation
共see step 8兲.
The liquid interface will then break into two new liquid inter- Fig. 3–Condensation in a network, h cc Ì L .
faces, one in tube 1 and another in tube 4, as shown in step 9.
Given that r 1 is less than r 4 , the liquid in tube 4 will be pulled
into tube 1. The condensation will continue in tube 1 until it
reaches the critical height (h 13cc ). The liquid formed later in tube
1 will flow downward into tube 4. The liquid height in tube 1
remains the same. The critical height h 13cc is calculated from After tube 4 is filled, the contact angle at the right end of tube

冉 冊
4 will increase until it reaches 90° 共dashed line shown in step 9兲.
2 ␴ cos ␪ R cos ␪ A p R ⫺p A Then, the contact angle at the left end of tube 2 will start to
h 13cc ⫽ ⫺ ⫺ . 共2兲
⌬␳g r1 r3 ⌬␳g increase. Because r 4 is larger than r 2 , the condensate will flow

140 K. Li and A. Firoozabadi: Phenomenological Modeling of Critical Condensate Saturation SPE Journal, Vol. 5, No. 2, June 2000
into the intersection connecting tubes 4, 8, 9, and 10 共see step 10兲.
The contact angle at the left end of tube 2 will exceed 90° with
further condensation.
When in-situ condensation continues 共see step 10兲, if r 8 is less
than r 10 , the liquid interface in the intersection will touch corner
point ‘‘F.’’ The liquid interface will also break into two new
interfaces, one in tube 8 and another still in the intersection after
the interface contacts point ‘‘F.’’ The configuration of the liquid
interface in the intersection near point ‘‘F’’ will change and ad-
just itself as shown in step 11. At the same time, the contact angle
at the left end of tube 2 will decrease. Because the radius of the
liquid interface in the intersection is larger than that in tube 8 and
the liquid pressure in the intersection is higher than that in tube 8,
the liquid will be pulled into it. After a force balance is estab-
lished between the capillary force and the gravity in the liquid
column in tube 8, the liquid in this tube will reach the critical
height. With further condensation, the liquid formed in tubes 1
and 8 will flow downward and the liquid in the intersection will
touch corner point ‘‘G.’’
The liquid interface in the intersection will break after it con-
tacts corner point ‘‘G.’’ If r 9 ⬍r 10 , the liquid formed in tubes 1
and 8 will be pulled into tube 9 until it is filled 共see step 12兲. With
further condensation, the liquid formed in tubes 1 and 8 will flow
downward into tube 10. Assuming the critical liquid height h 10cc
in tube 10 is less than the tube length, the liquid will flow down-
ward when the liquid height ⬎h 10cc . At the same time, the liquid
in tube 8 will also flow downward into tube 10. The configuration
developed is shown in step 13. With further condensation, the
liquid formed in tube 1 will flow downward through tube 4 to tube
10 and then to the next element. When the liquid interface in the
intersection touches point ‘‘G,’’ the liquid will first enter tube 10
instead of tube 9, if r 9 ⬎r 10 . When the liquid height in tube 10 is
equal to or larger than h 10cc , the liquid in tube 10 flows down-
ward 共see step 14兲.
If r 8 ⬎r 10 , the liquid interface will form two new interfaces
after contacting corner point ‘‘G,’’ one in tube 10 and another in
the intersection. Because the radius of the liquid interface in the
intersection is larger than in tube 10, the liquid will be pulled into
tube 10. The liquid formed later will also enter tube 10 instead of
tube 8. Assuming h 10cc ⬍L, the configuration shown in step 15 is
established after the liquid height in tube 10 reaches critical
height. With further condensation in this element, liquid will flow
downward through tube 4 to tube 10 and then to the next element.
As stated previously, the process of in-situ condensate forma-
tion, growth, and flow when h cc ⬍L is described in Ref. 16.
Calculation of S cc
The first step for the estimation of S cc is to determine the differ-
ential pressure in each tube. The pressure distribution in the entire
network is calculated as follows. At every intersection one can
write
4
兺 q ⫽0,
i⫽1
i 共3兲
where q i is the flow rate of fluid in tube i. The flow rate in a
circular capillary tube can be calculated from
冉 冊
␲ r i4 ⌬p i
q i⫽ ⫹␳g , 共4兲
8␮ L
provided that the fluid is incompressible. In the above equation, ␮
is fluid viscosity, ␳ is fluid density, and ⌬p i and r i are the differ-
ential pressure 共inlet to outlet兲 and radius of tube i, respectively.
Once pressures at inlet and outlet ends of the network are pro-
vided, Eqs. 3 and 4 can be used to obtain the pressure distribution
in the network. The Gauss iteration method can be used to solve
the set of linear equations. Note that for every tube i with h c
⬍h cc , q i ⫽0, as we will discuss later. S cc in each tube can be
calculated with the known pressure distribution according to Eqs.
1a, 1b, and 2. Then the value of the total S cc in the network is
readily calculated.
K. Li and A. Firoozabadi: Phenomenological Modeling of Critical Condensate Saturation
Calculation of Fluid Conductance
Consider single-phase flow in a capillary tube of radius r i and
length L. According to Darcy’s law, the flow rate across the tube
is q i ⫽(k i / ␮ L) ␲ r i2 (⌬p i ⫹ ␳ gL), where k i is the permeability of
tube i. An alternative form of the Darcy equation is q i ⫽G i (⌬p i
⫹ ␳ gL), where G i is the conductance of tube i. Comparison of
the Darcy equation with Eq. 4 provides the expression for the
absolute fluid conductance of tube i,
␲ r i4
Gi j⫽ 共 S j ⫽1 兲 , 共5兲
8␮ jL
where G i j and S j are the absolute conductance and the saturation
of fluid j in tube i, respectively; ␮ j is the viscosity of fluid j.
Prior to S cc , when there is some liquid saturation in the tube, gas
may not flow, and therefore, q i ⫽0. The effective conductance of
fluid j in tube i, which is related to the effective permeability of
phase j, can be calculated using the following equation when S c
⭐S cc :
G i j ⫽0 共 S c ⭐S cc 兲 . 共6兲
Now consider a tube with some liquid saturation S c 共see Fig. 2b兲
which is greater than S cc . In this case, liquid is located at the
bottom of the tube and both liquid and gas may flow. The equa-
tion for calculating the effective gas conductance G ig of tube i in
two-phase flow is
冉 冊
⫺1
␲ r i4 ␮c
G ig ⫽ 1⫺S ic ⫹ S 共 S c ⬎S cc 兲 , 共7兲
8 ␮ gL ␮ g ic
where ␮ g and ␮ c are the gas and liquid condensate viscosity,
respectively. S ic is the liquid condensate saturation in tube i. The
derivation of Eq. 7 is provided in Appendix A. Alternatively, Eq.
7 can be obtained from Eq. 5 using the weighted average viscosity
␮
¯ ⫽S g ␮ g ⫹S c ␮ c . Similarly, the effective liquid conductance of
tube i in two-phase flow can be calculated as
G ic ⫽
␲ r i4
冋
S ⫹
␮g
8 ␮ c L ic ␮ c
共 1⫺S ic 兲 册 ⫺1
共 S c ⬎S cc 兲 , 共8兲
where G ic is the conductance of the liquid phase for case S c
⬎S cc . The reason for presenting Eqs. 7 and 8 separately is to
show that the effective conductance of each phase in the two-
phase state can be a function of its absolute conductance in the
single-phase state. On the other hand, effective gas and liquid
conductance may not be identical for tube shapes other than cir-
cular.
The next step is to calculate the effective conductance of both
gas and liquid phases in the entire network based on conductances
of individual tubes. Once the effective conductance of both gas
and liquid phases in the network is calculated, gas and condensate
relative permeabilities can be readily calculated. In the following,
we will use the renormalization method18 to calculate the effective
conductance of the network from conductances of individual
tubes.
Renormalization Method and Relative Permeabilities
A useful technique to study transport in disordered systems is the
real-space renormalization group method, which has arisen in
condensed-matter physics.18 King19,20 has applied this technique
to single- and two-phase fluid flows. Filippi and Toledo21 also
used the method for heterogeneous and highly anisotropic sys-
tems. The renormalization method is accurate and computation-
ally cost effective compared to direct numerical simulation.
For a given phase and a known saturation distribution, the ef-
fective phase conductance is calculated in small regions, the so-
called renormalization cells, and then on larger scales. Fig. 4
shows grid units with different renormalization cell numbers n b
for a 2D network. In Fig. 4, the renormalization cell number var-
ies from 2 to 5. One unit cell with a renormalization cell number
SPE Journal, Vol. 5, No. 2, June 2000 141
of n b is a unit consisting of n 2b nodes and 2n 2b tubes. The effective
conductance of one phase in the whole system can be obtained
from repeated application of the renormalization procedure.
Fig. 5 shows the renormalization procedure for n b ⫽2. The pri-
mary network is comprised of four unit cells of n b ⫽2. After the
first step of renormalization, the network reduces to one unit cell
of n b ⫽2. In the last step, the primary network is converted into
G 8 G 4 G 6 ⫹G 8 G 4 G 7 ⫹G 8 G 6 G 7 ⫹G 8 G 3 G 7 ⫹G 6 G 7 G 3 ⫹G 4 G 6 G 3 ⫹G 4 G 3 G 7 ⫹G 6 G 8 G 3
G z⫽
G 6 G 7 ⫹G 4 G 6 ⫹G 6 G 8 ⫹G 4 G 7 ⫹G 4 G 8 ⫹G 3 G 7 ⫹G 3 G 4 ⫹G 3 G 8
and
G 4 G 3 G 2 ⫹G 4 G 3 G 1 ⫹G 4 G 2 G 1 ⫹G 4 G 5 G 1 ⫹G 2 G 1 G 5 ⫹G 3 G 2 G 5 ⫹G 3 G 5 G 1 ⫹G 2 G 4 G 5
G x⫽
G 2 G 1 ⫹G 3 G 2 ⫹G 2 G 4 ⫹G 3 G 1 ⫹G 3 G 4 ⫹G 5 G 1 ⫹G 5 G 3 ⫹G 5 G 4
respectively, where G 1 , G 2 , . . . , G 8 represent the phase conduc-
tance of each tube 共see Fig. 6兲. The derivation of Eq. 9 is provided
in Appendix B. Relative permeabilities for the network presented
in Fig. 1 are calculated according to the following procedure: 共1兲
calculate the absolute gas conductance in the vertical direction
G abs for the whole network first saturated completely with gas.
This is realized by repeating the renormalization steps at different
scales, shown in Fig. 5; 共2兲 calculate the effective conductance of
gas G zg or liquid phase G zc at a given condensate saturation S c in
the vertical direction when liquid is formed in the network due to
the pressure decrease below the dewpoint pressure; 共3兲 calculate
gas relative permeability k rg from G zg /G abs and liquid relative
permeability k rc from ( ␮ c G zc )/( ␮ g G abs ). Calculations at differ-
ent values of S c provide relative permeabilities.
Results
The model described above is used to calculate S cc , k rg , and k rc .
The results and the sensitivity of the calculation to various param-
eters are described in the following.
Critical Condensate Saturation. Critical condensate saturation
is calculated using the modified S cc model in which viscous forces
are considered and the critical height (h cc ) of liquid formed in-
situ in vertical capillary tubes can be less than the tube length. The
results show that critical condensate saturation is a function of
interfacial tension, flow rate, gravity, and contact angle hysteresis.
Effect of Interfacial Tension and Gravity. The effect of ␴ and
gravity on S cc is shown in Fig. 7 for two different values of the
viscous force ⌬p across the network. The results are based on a
tube length of 6000 ␮m, ␪ R ⫽0° and 0°⬍ ␪ A ⭐90°. Fig. 7 shows
that S cc increases as ␴ increases (0⬍ ␴ ⬍4.0 dyn/cm). Fang
et al.16 made a similar observation for ␴ in the range of 0 to 0.45
dyn/cm. Fig. 7 shows that S cc decreases with an increase of ⌬ ␳ .
When ␴ is large enough, S cc reaches the same maximum value for
Fig. 4–Schematic of renormalization for different cell numbers.
142 K. Li and A. Firoozabadi: Phenomenological Modeling of Critical Condensate Saturation
two conductors, one in the horizontal and the other in the vertical
direction. Let us consider a unit cell with eight tubes 共that is, n b
⫽2兲, sketched in Fig. 6a. The effective networks in the vertical
and horizontal directions are shown in Figs. 6b and 6c, respec-
tively. The effective conductance, G z , in the vertical direction
and the effective conductance in the horizontal direction, G z , for
n b ⫽2 are calculated from
共9兲
, 共10兲
different ⌬ ␳ values, which implies that the effect of gravity on
S cc becomes negligible at high values of ␴. The results in Fig. 7
are consistent with the experimental data by Danesh et al.,10 Chen
et al.,5 Ali et al.,11 and Henderson et al.12 These authors observed
that S cc increases as ␴ increases and the increase in gravity lowers
S cc . Various calculations from the model show that, as long as ␪ A
is not zero ( ␪ A ⬎0), the critical condensate saturation is not af-
fected by the magnitude of ␪ A .
A comparison of Figs. 7a and 7b reveals that S cc decreases
when ⌬p increases. The effect of gravity on S cc becomes negli-
gible at high ⌬p when ␴ varies in the range of 0 to 2 dyn/cm 共see
Fig. 7b兲. The results presented in Fig. 7 demonstrate that gravity
may or may not affect S cc . When ⌬p is small 共that is, at low flow
rates兲, S cc is affected by gravity, whereas for high ⌬p, the effect
of gravity is less pronounced. The results in Fig. 7 explain the
effect of gravity on S cc and reveal that there may be no inconsis-
tency in the experimental data reported by various authors. The
effect of viscous forces is discussed in more detail next.
Effect of Flow Rate. The effect of viscous force 共that is, flow
rate兲 on S cc is shown in Fig. 8 for ⌬ ␳ ⫽0.1 and 0.5 g/mL. Other
parameters are the same as those used previously. ⌬p varies in the
Fig. 5–Schematic of the renormalization procedure; n b Ä2.
SPE Journal, Vol. 5, No. 2, June 2000
Fig. 6–Effective conductance in the horizontal and vertical di-
rections; n b Ä2.
range of 0.05 to 0.4 psi in Fig. 8. Fig. 8 shows that the critical
condensate saturation decreases with the increase of the viscous
force 共that is, increase in pressure drop兲. A comparison of Figs. 8a
and 8b reveals that the effect of ⌬p on S cc is more pronounced at
lower gravity when ␴ varies in the range of 0 to 2 dyn/cm 共see
Fig. 8a兲. The effect of ⌬p on S cc reduces at high ␴ ( ␴
⬎3 dyn/cm). The calculated results shown in Fig. 8 are consistent
with the experimental data of Chen et al.5 who found that S cc
decreases as the flow rate increases.
Fig. 7–Effect of ␴ and gravity on S cc ; ␪ R Ä0°, 0°Ë ␪ A Ë90°.
K. Li and A. Firoozabadi: Phenomenological Modeling of Critical Condensate Saturation
Fig. 8–Effect of ␴ and viscous force on S cc ; ␪ R Ä0°, 0°Ë ␪ A
Ë90°.
Effect of Contact Angle Hysteresis. In a gas-oil-rock system,
oil phase is considered as the wetting phase and gas is the non-
wetting phase. It is often assumed that the contact angle through
the liquid phase is zero. The assumption of liquid wetting in a
gas-liquid system is valid but the assumption of ␪ ⫽0° may be
invalid. If the wettability of the fluid-rock system in a gas con-
densate reservoir near wellbore could be altered to intermediate
gas wetting by some chemicals, S cc and k rg may also change. Fig.
9a presents S cc vs. ␴ for ␪ R ⫽85° and 85°⬍ ␪ A ⭐90°. The other
parameters are the same as those used in Fig. 7. This figure shows
that S cc becomes very small and variations of ␴ or ⌬ ␳ have very
little effect on S cc when the receding contact angle increases to
85°. Fig. 9b shows the effect of receding contact angle; S cc de-
creases as the receding contact angle increases from 0 to 85° at a
fixed ␴. Fig. 9 demonstrates that wettability alteration to prefer-
ential gas wetting has the most pronounced effect on S cc .
The observation that there may be no effect of ␴ on S cc for
intermediate gas wettability 共see Fig. 9兲 is interesting. The critical
condensate saturation is sensitive to ␴ if the fluid-rock system is
strongly liquid wet as was shown in Fig. 7. However, for an
intermediately gas-wet system, the critical condensate saturation
becomes small and does not increase much even at very high ␴, as
shown in Fig. 9.
Effect of Network Size. We calculated S cc with different net-
work sizes to study the sensitivity of our S cc model to the network
size. The results showed that for ␴ ⫽0.01 dyn/cm with or without
viscous force, S cc becomes approximately constant when the net-
work size is larger than 15⫻15. This demonstrates that the net-
work size chosen for this study, 20⫻20, is appropriate.
Relative Permeabilities. In the following we will present the re-
sults of the study of relative permeability from the model.
Features of Gas Relative Permeability. Fig. 10a depicts a
sketch of the measured gas phase relative permeability curve for
low-permeability carbonate rocks in gas condensate systems.4 Fig.
10b presents the calculated gas phase relative permeability curve
from our model and shows that the model reproduces the features
of the experimental data. Both the model and experimental data
SPE Journal, Vol. 5, No. 2, June 2000 143
 Fig. 11–Effect of ␴ on the relative permeability; ⌬ p Ä0.3 psi,
⌬ ␳ Ä0.5 gÕmL, ␪ R Ä0°, 0°Ë ␪ A Ë90°.

show an abrupt drop in gas phase relative permeability at conden-

sate saturation close to S cc . When gas phase relative permeability
decreases abruptly, the gas well deliverability reduces sharply.
Effect of Interfacial Tension. The effect of ␴ on gas conden-
sate relative permeabilities is plotted in Fig. 11. The saturation
interval is selected to be for S c ⬎S cc to compare model results
with experimental data. ␴ varies in the range of 0.01 to 0.50
dyn/cm. Fig. 11 shows that both gas and liquid phase relative
permeabilities decrease with the increase of ␴. Liquid phase rela-
tive permeability is less sensitive to ␴ than gas phase relative
permeability. The model’s results are in agreement with the ex-
perimental data of Henderson et al.13 These authors found that the
relative permeabilities of both phases decrease with the increase
of ␴. The gas relative permeability reduction was, however, more
pronounced.
Effect of Flow Rate. The effect of flow rate on gas and con-
Fig. 9–Effect of ␴, gravity, and viscous force on S cc .
densate relative permeabilities is graphed in Fig. 12. Fig. 12
shows that both k rg and k rc increase with the increase of ⌬p. Our
results are in agreement with the experimental work of Chen
et al.5 who found that the gas relative permeability increased as
the flow rate increased. The liquid phase relative permeabilities
shown in Fig. 12 are less sensitive to viscous forces than the gas
phase relative permeability. The model’s results of relative per-
meabilities are also consistent with the experimental observation
by Henderson et al.9 These authors found that the relative perme-
ability of both phases increased with the increase of flow rate but
that the liquid relative permeability increased less than the gas
relative permeability.
Effect of Contact Angle Hysteresis. If the wettability of a
fluid-rock system in a gas condensate reservoir near wellbore can
be altered from strongly liquid-wet to intermediately gas-wet, the
gas and condensate relative permeabilities behavior may also
change. Fig. 13 depicts model results for k rg and k rc at different
receding contact angles. Fig. 13 shows that both k rg and k rc in-
crease as the wettability to liquid decreases 共␪ R ⫽0, 40°, and 80°
and ␪ R ⬍ ␪ A ⫽90°, ␴ ⫽0.4 dyn/cm兲. There is a significant increase

Fig. 10–Gas phase relative permeability in gas condensate sys- Fig. 12–Effect of viscous force on the relative permeability; ␴
tems. Ä0.5 dynÕcm, ⌬ ␳ Ä0.5 gÕmL, ␪ R Ä0°, and 0°Ë ␪ A Ë90°.

144 K. Li and A. Firoozabadi: Phenomenological Modeling of Critical Condensate Saturation SPE Journal, Vol. 5, No. 2, June 2000
Fig. 13–Effect of ␪ R on the relative permeability; ␴
Ä0.4 dynÕcm, ⌬ p Ä0.2 psi, ⌬ ␳ Ä0.5 gÕmL, and ␪ R Ë ␪ A Ë90°.

in the gas phase relative permeability over the whole saturation

range when receding contact angle increases from 0° 共strongly
liquid-wet兲 to 80° 共intermediately gas-wet兲. These results suggest
that if the rock wettability around the wellbore can be altered from
liquid-wet to intermediately or preferentially gas-wet, the gas well
deliverability may increase significantly. Gas phase relative per-
meability, k rg , reduces very rapidly for the strongly liquid-wet
condition ( ␪ R ⫽0°) and approaches zero when S c increases to
22%. Fig. 13 gives a clear indication of the significance of the
wettability and contact angle hysteresis on the relative permeabil-
ity.
We also studied the effect of network size on relative perme-
abilities and found that it was very small for a network size up to
35⫻35. Although the network size used in this study is small, the
relative permeabilities from the model could indeed represent
some critical features of experimental results, as indicated in Fig.
10b.
To our knowledge, the effect of wettability on relative perme-
abilities in gas condensate systems has not yet been experimen-
tally studied. We have embarked on an experimental research pro-
gram to measure the relative permeabilities of gas condensate
systems under varying wettability conditions. Some preliminary
data on relative permeabilities are presented in Ref. 22.

Blocking Effect Due to In-Situ Condensation. In some gas res-

ervoirs, a well may be blocked completely and the gas production
may stop at high pressures.4 This phenomenon is a result of in-situ
liquid formation near the wellbore as the reservoir pressure drops
below the dewpoint. The liquid newly formed in the porous me-
dium blocks both small and large pores and reduces the effective
permeability of the gas phase significantly.
The distribution of the liquid condensate formed in-situ at dif-
ferent stages of condensation in the pore network is presented in
Fig. 14. Fig. 14a shows the condensate distribution at S c
⫽5.0%. The black represents the liquid condensate and the white
the gas phase. The condensate is first formed in the middle of
smaller tubes in the form of liquid bridges. The capillary tubes
that contain liquid bridges may be blocked for gas phase flow. As
a result, the number of tubes for gas flow is reduced and the
effective gas permeability through the whole network decreases.
Gas production will, therefore, be reduced when the pressure
drops below the dewpoint. With a further pressure decrease, more
liquid drops out and condensate saturation S c increases. Fig. 14b
shows the liquid distribution at S c ⫽15.0%. The liquid occupies Fig. 14–„a… Saturation distribution at S c Ä5%; ␴ Ä0.5 dynÕcm,
some large capillary tubes. The number of capillary tubes avail- ⌬ p Ä0.02 psi, ⌬ ␳ Ä0.1 gÕmL, ␪ R Ä0°, and 0°Ë ␪ A Ë90°. „b… Satu-
ration distribution at S c Ä15%; ␴ Ä0.5 dynÕcm, ⌬ p Ä0.02 psi,
able for gas flow is further reduced. Fig. 14c shows the liquid ⌬ ␳ Ä0.1 gÕmL, ␪ R Ä0°, and 0°Ë ␪ A Ë90°. „c… Saturation distribu-
distribution at the critical condensate saturation, S cc ⫽58.0%; tion at S c Ä58%; ␴ Ä0.5 dynÕcm, ⌬ p Ä0.02 psi, ⌬ ␳ Ä0.1 gÕmL,
there are almost no continuous paths available for gas flow. The ␪ R Ä0°, and 0°Ë ␪ A Ë90°.
effective gas permeability is zero; when this occurs, gas produc-
tion may stop completely.
relative permeabilities in gas condensate systems. These include
Discussion 共1兲 keeping ⌬ ␳ and ␴ constant with pressure decline, 共2兲 neglect-
There are a number of simplifying assumptions that we have made ing the effect of water to film and connate water, 共3兲 simultaneous
in the simple network model for critical condensate saturation and formation of liquid bridges in all the tubes with r⭐r p , and 共4兲 a

K. Li and A. Firoozabadi: Phenomenological Modeling of Critical Condensate Saturation SPE Journal, Vol. 5, No. 2, June 2000 145
two-dimensional model representation of porous media. Assump-
tions 1 and 3 could be relaxed without much difficulty, but the
extension of the model from two to three dimensions complicates
the bookkeeping. Extension of the model to include the effect of
initial water may not be straightforward. The model may also be
extended to include tube shapes other than circular. The main goal
of the study was to find the most sensitive parameter that can
influence S cc and k r . The results presented above have clearly
shown that wettability alteration from strongly liquid condensate
wetting to weak liquid condensate wetting is the key parameter.
The experimental work presented in Ref. 22 and work in progress
are in line with the model’s results. Interestingly, both with and
without initial water saturation, the alteration of wettability to
preferential gas wetting affects S cc and k r the most.
Conclusions
The major conclusions drawn from the work are the following.
䊉 The predictions from our simple model show that gravity gen-
erally reduces the critical condensate saturation; but as interfacial
tension increases, the gravity effect becomes less pronounced. The
effect of gravity is pronounced when viscous forces are small.
These results are in agreement with experimental data.
䊉 Viscous forces may have a pronounced effect on S
cc , espe-
cially at low interfacial tension. However, the most important pa-
rameter is wettability expressed in terms of the receding contact
angle; wettability alteration to intermediate gas wetting reduces
critical condensate saturation dramatically, irrespective of gravity
and interfacial tension.
䊉 The model predicts a significant effect of interfacial tension
and viscous forces on gas phase relative permeability. These pre-
dictions are in line with experimental data.
䊉 The model’s results show that there is a very significant effect
of gas wetting on relative permeability. The results imply that the
most effective method for increasing gas well deliverability may
be the alteration of wettability around the wellbore.
Nomenclature
G abs ⫽ absolute gas conductance of the network in single-
phase
G ic ⫽ effective conductance of liquid condensate in tube i
in two-phase
G i j ⫽ absolute conductance of the fluid j in tube i in
single-phase
G ig ⫽ effective conductance of gas in tube i two-phase
G zg ⫽ effective gas conductance of the network in two-
phase
G zc ⫽ effective liquid conductance of the network in two-
phase
h c ⫽ height of liquid column
h cc ⫽ critical height of liquid column
I ⫽ total electrical current through the network in the
vertical direction
I i ⫽ electric current through the conductance G i 共i⫽3, 4,
6, 7, 8兲
k rg ⫽ gas-phase relative permeability
k rc ⫽ liquid-phase relative permeability
L ⫽ capillary tube length
p A ⫽ pressure at the bottom
p R ⫽ pressure at the top
P c ⫽ capillary pressure between the gas and condensate
⌬p ⫽ total pressure differential
⌬p c ⫽ pressure differential across the condensate phase
⌬p g ⫽ pressure differential across the gas phase
⌬p i ⫽ pressure differential between two ends of capillary
tube i
q g ⫽ flow rate of gas
q i ⫽ flow rate of fluid through capillary tube i
q c ⫽ flow rate of the condensate
r ⫽ radius of a capillary tube
r i ⫽ radius of capillary tube i
r p ⫽ threshold radius
S c ⫽ condensate saturation
146 K. Li and A. Firoozabadi: Phenomenological Modeling of Critical Condensate Saturation
S cc ⫽ critical condensate saturation
V0 ⫽ voltage at the top of the effective network
V1 ⫽ voltage at point A
V2 ⫽ voltage at point B
Vt ⫽ voltage at the bottom of the effective network
␪A ⫽ advancing contact angle
␪R ⫽ receding contact angle
␳c ⫽ condensate density
␳g ⫽ gas density
⌬␳ ⫽ density difference between the condensate and gas
␴ ⫽ interfacial tension
␮ ⫽ fluid viscosity
␮g ⫽ gas viscosity
␮c ⫽ condensate viscosity
Acknowledgments
P.G. Toledo initiated work on renormalization calculation of gas
phase relative permeability at the Reservoir Engineering Research
Institute 共RERI兲. We have benefited from his work considerably.
This work was supported by the Technology Research Center of
the Japan Natl. Oil Corp., by U.S. DOE Grant No. DE-FG22-
96BC14850, and by member companies of RERI. Their support is
greatly appreciated.
References
1. Allen, F.H. and Roe, R.P.: ‘‘Performance Characteristics of a Volu-
metric Condensate Reservoir,’’ Trans., AIME 共1950兲 189, 83.
2. Engineer, R.: ‘‘Cal Canal Field, California: Case History of a Tight
and Abnormally Pressure Gas Condensate Reservoir,’’ paper SPE
13650 presented at the 1985 SPE California Regional Meeting,
Bakersfield, California, 27–29 March.
3. Barnum, R.S. et al.: ‘‘Gas Condensate Reservoir Behavior: Produc-
tivity and Recovery Reduction Due to Condensation,’’ paper SPE
30767 presented at the 1995 SPE Annual Technical Conference and
Exhibition, Dallas, 22–25 October.
4. Gravier, J.F. et al.: ‘‘Determination of Gas-Condensate Relative Per-
meability on Whole Cores Under Reservoir Conditions,’’ SPEFE
共February 1986兲 9; Trans., AIME, 281.
5. Chen, H.L., Wilson, S.D., and Monger-McClure, T.G.: ‘‘Determina-
tion of Relative Permeability and Recovery for North Sea Gas Con-
densate Reservoirs’’ paper SPE 30769 presented at the 1995 SPE
Annual Technical Conference and Exhibition, Dallas, 22–25 October.
6. Munkerud, P.K. ‘‘Measurement of Relative Permeability and Flow
Properties of a Gas Condensate System During Pressure Depletion
and Pressure Maintenance,’’ paper SPE 19071 presented at the 1989
SPE Gas Technology Symposium, Dallas, 7–9 June.
7. Saeidi, A. and Handy, L.L.: ‘‘Flow and Phase Behavior of Gas Con-
densate and Volatile Oils in Porous Media,’’ paper SPE 4891 pre-
sented at the 1974 SPE California Regional Meeting, San Francisco,
4–5 April.
8. Munkerud, P.K. and Torsaeter, O.: ‘‘The Effects of Interfacial Ten-
sion and Spreading on Relative Permeability in Gas Condensate Sys-
tems,’’ presented at the 1995 European Symposium on Improved Oil
Recovery, Vienna, Austria, 15–17 May.
9. Morel, D.C. et al.: ‘‘Mobility of Hydrocarbon Liquids in Gas Con-
densate Reservoirs: Interpretation of Depletion Laboratory Experi-
ments,’’ paper SPE 24939 presented at the 1992 SPE Annual Tech-
nical Conference and Exhibition, Washington, DC, 4–7 October.
10. Danesh, A., Henderson, G.D., and Peden, J.M.: ‘‘Experimental Inves-
tigation of Critical Condensate Saturation and Its Dependence on In-
terstitial Water Saturation in Water-Wet Rocks,’’ SPERE 共August
1991兲 336; Trans., AIME, 291.
11. Ali, J.K. et al.: ‘‘The Influence of Interfacial Tension on Liquid Mo-
bility in Gas Condensate Systems,’’ paper SPE 26783 presented at the
1993 Offshore European Conference, Aberdeen, 7–10 September.
12. Henderson, G.D. et al.: ‘‘An Investigation Into the Processes Govern-
ing Flow and Recovery in Different Flow Regimes Present in Gas
Condensate Reservoirs,’’ paper SPE 26661 presented at the 1993 SPE
Annual Technical Conference and Exhibition, Houston, 3–6 October.
13. Henderson, G.D. et al.: ‘‘Measurement and Correlation of Gas Con-
densate Relative Permeability by the Steady-State Method,’’ paper
SPE 30770 presented at the 1995 SPE Annual Technical Conference
and Exhibition, Dallas, 22–25 October.
14. Haniff, M.S. and Ali, J.K.: ‘‘Relative Permeability and Low Tension
Fluid Flow in Gas Condensate Systems,’’ paper SPE 20917 presented
at the 1990 European Petroleum Conference, The Hague, The Neth-
erlands, 22–24 October.
SPE Journal, Vol. 5, No. 2, June 2000
15. Mohammadi, S. et al.: ‘‘Pore-Level Modeling of Gas-Condensate Appendix B–Effective Conductance in the Vertical
Flow Through Horizontal Porous Media,’’ paper SPE 20479 pre- Direction for n b Ä2
sented at the 1990 SPE Annual Technical Conference and Exhibition, The effective network in the vertical direction for n b ⫽2 is shown
New Orleans, 23–26 September. in Fig. 6c. One can write the following equations based on the
16. Fang, F. et al.: ‘‘A Phenomenological Modeling of Critical Conden-
sate Saturation,’’ paper SPE 36716 presented at the 1996 SPE Annual
theory of electric circuits:
Technical Conference and Exhibition, Denver, Colorado, 6–9 Octo- I 3 ⫽G 3 共 V 2 ⫺V t 兲 , 共B-1兲
ber.
17. Firoozabadi, A.: Thermodynamics of Hydrocarbon Reservoirs, I 4 ⫽G 4 共 V 2 ⫺V 1 兲 , 共B-2兲
McGraw-Hill Book Co. Inc., New York City 共1999兲 Chap. 2. I 6 ⫽G 6 共 V 0 ⫺V 2 兲 , 共B-3兲
18. Stinchcombe, B. and Watson, B.P.: ‘‘Renormalization Group Ap-
proach for Percolation Conductivity,’’ Phys. Rev. C 共1976兲 9, 3221. I 7 ⫽G 7 共 V 0 ⫺V 1 兲 , 共B-4兲
19. King, P.R.: ‘‘The Use of Renormalization for Calculating Effective
I 8 ⫽G 8 共 V 1 ⫺V t 兲 . 共B-5兲
Permeability,’’ Transp. Porous Media 共1989兲 4, 237.
20. King, P.R.: ‘‘Renormalization Calculations of Immiscible Flow,’’ At points A and B
Transp. Porous Media 共1993兲 12, 237.
21. Filippi, G. and Toledo, P.G.: ‘‘Prediction of Effective Permeability
from Local Distributions of Conductances,’’ Reservoirs, PDVSA, Ca-
兺 I ⫽0 共 at point
i
i A:i⫽4, 7, 8 兲 , 共B-6兲
racas 共1993兲.
22. Li, K. and Firoozabadi, A.: ‘‘Experimental Study of Wettability Al-
teration to Preferential Gas-Wetness in Porous Media and Its Effect,’’ 兺 I ⫽0 共 at point
i
i B:i⫽3, 4, 6 兲 . 共B-7兲
SPEREE 共April 2000兲 139.
The total electric current through the system, I⫽I 6 ⫹I 7 , is given
by
I⫽G z 共 V 0 ⫺V t 兲 . 共B-8兲
Appendix A–Gas Conductance in a Single Circular
In the above equations, I i 共i⫽3, 4, 6, 7, 8兲 is the electric current
Tube
through conductance G i 共i⫽3, 4, 6, 7, 8兲; V 0 and V t are the
When both gas and liquid phases flow in a circular tube 共see Fig. voltages at the top and the bottom of the effective network, re-
2b兲, the gas flow rate can be calculated from the Poiseuille equa- spectively; V 1 and V 2 are the voltages at points A and B, respec-
tion, tively. G z is the effective conductance of the network in the ver-
␲ r 4 ⌬p g tical direction.
q g⫽ , 共A-1兲 Combining Eqs. B-1 to B-7 gives
8 ␮ g 共 L⫺h c 兲
G 4 共 G 6 V 0 ⫹G 3 V t 兲 ⫹ 共 G 3 ⫹G 4 ⫹G 6 兲共 G 7 V 0 ⫹G 8 V t 兲
where ⌬p g is the differential pressure across the gas phase, and h c V 1⫽ ,
is the liquid height 共or length兲. The effect of gravity and com- 共 G 4 ⫹G 7 ⫹G 8 兲共 G 3 ⫹G 4 ⫹G 6 兲 ⫺G 24
pressibility are both neglected. The expression for the liquid con- 共B-9兲
densate flow rate is G 4 共 G 7 V 0 ⫹G 8 V t 兲 ⫹ 共 G 4 ⫹G 7 ⫹G 8 兲共 G 6 V 0 ⫹G 3 V t 兲
V 2⫽
冉 冊
.
␲ r 4 ⌬p c 共 G 4 ⫹G 7 ⫹G 8 兲共 G 3 ⫹G 4 ⫹G 6 兲 ⫺G 24
q c⫽ ⫹ ␳ cg , 共A-2兲 共B-10兲
8␮c hc
Substituting Eqs. B-3 and B-4 into Eq. B-9 gives
where ␮ c is the condensate viscosity and ⌬p c the differential
pressure across the condensate phase. The differential pressure 共 G 6 ⫹G 7 兲 V 0 ⫺ 共 G 6 V 2 ⫹G 7 V 1 兲
G z⫽ . 共B-11兲
⌬p between the two ends of a tube is given by V 0 ⫺V t
⌬p⫽⌬p g ⫹⌬p c ⫹ P c , 共A-3兲 Eq. 9 in the text can be obtained by substituting Eqs. B-9 and
B-10 into Eq. B-11.
where P c is the capillary pressure between the gas and condensate
liquid phase. The gas flow rate is equal to the liquid flow rate Kewen Li is a research associate in the Dept. of Petroleum
when the fluid compressibility is neglected. That is, Engineering of Stanford U., Stanford, California. e-mail:
kewenli@pangea.stanford.edu. Previously, he was a research
q g ⫽q c . 共A-4兲 engineer at the Reservoir Engineering Research Inst.; prior to
Combining the above equations 共Eqs. A-1 to A-4兲, the gas phase that, he was a visiting scholar at IFP in France, and he served
differential pressure can be obtained as as director and senior engineer of the Core Analysis Labora-
tory at RIPED in China from 1992-1995. Li’s research interests
1 include the theoretical and experimental study of multi-phase
⌬p g ⫽ 共 ⌬p⫺ P c ⫹ ␳ c gh c 兲 . 共A-5兲 flow in porous media, numerical study of multi-phase flow in
␮c hc porous media, numerical reservoir simulation, and improved
1⫹
␮ g L⫺h c oil recovery. He holds a BS degree from Jianghan Petroleum
Inst., China, and MS and PhD degrees from Jianghan Petro-
Substituting Eq. A-5 into Eq. A-1 gives leum Inst., China, and MS and PhD degrees from RIPED, China,
all in petroleum engineering. Abbas Firoozabadi is senior sci-
␲r4 1 entist and director of RERI in Palo Alto and teaches at Imperial
q g⫽ 共 ⌬p⫺ P c ⫹ ␳ c gh c 兲 . 共A-6兲
8 ␮ g 共 L⫺h c 兲 ␮c hc College in London. e-mail: af@rerinst.org. He previously taught
1⫹ at the U. of Texas, Austin, and Standford U. His research inter-
␮ g L⫺h c
ests include equilibrium, nonequilibrium, and irreversible ther-
The gas conductance is then determined as modynamics; multiphase flow in fractured and layered per-
meable media; and tidal-force-influenced pressure transients
␲r4 1 for estimation of key parameters of fractured reservoirs.
G gas⫽ . 共A-7兲
8 ␮ g 共 L⫺h c 兲 ␮c hc Firoozabadi holds a BS degree from Abadan Inst. of Technol-
1⫹ ogy, Iran, and MS and PhD degrees from lllinois Inst. of Tech-
␮ g L⫺h c nology, all in gas engineering. He has served on the Editorial
Review Committee since 1986 and was a 1988–89 member
For tube i in the network, the condensate saturation is given by
and 1992–93 chairman of the Forum Series in North America
S ic ⫽h c /L. 共A-8兲 Steering Committee, a 1991–92 member of the Western Re-
gional Meeting Program Committee, a 1987–89 member of
Eq. 7 in the text can be obtained by substituting Eq. A-8 into Eq. the Reservoir Simulation Symposium Program Committee, and
A-7. a 1992–96 Short Course Instructor.

K. Li and A. Firoozabadi: Phenomenological Modeling of Critical Condensate Saturation SPE Journal, Vol. 5, No. 2, June 2000 147