Anda di halaman 1dari 15

SPE 115786

Improved Recovery in Gas-Condensate Reservoirs Considering


Compositional Variations
Chunmei Shi and Roland N. Horne, Stanford University

Copyright 2008, Society of Petroleum Engineers

This paper was prepared for presentation at the 2008 SPE Annual Technical Conference and Exhibition held in Denver, Colorado, USA, 2124 September 2008.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
The objective of this work was to develop a methodology to increase the productivity of gas/condensate from gas-condensate
reservoirs. Presently, gas-condensate reservoirs experience reductions in productivity by as much as a factor of 10 due to the
dropout of liquid close to the wellbore. The liquid dropout blocks the flow of gas to the well and lowers the overall energy
output by a very substantial degree. The combination of condensate phase behavior and rock relative permeability results in a
composition change of the reservoir fluid, as heavier components separate into the dropped-out liquid while the flowing gas
phase becomes lighter in composition. This effect has been sparsely recognized in the literature, although there is clear
evidence of it in field observations. This work quantified the effect, developed a scientific understanding of the phenomena,
and used the results to investigate ways to enhance the productivity by controlling the liquid composition that drops out close
to the well. By optimizing the producing pressure strategy, it should be possible to cause a lighter liquid to be condensed in the
reservoir, after which the productivity loss would be more easily remedied.
The research made use of experimental measurements of gas-condensate flow, as well as compositional numerical
simulations. Different strategies have been compared, and the optimum producing sequences are suggested for maximum
condensate recovery. Results show that composition varies significantly as a function of fluid phase behavior and producing
sequence; condensate recovery can be improved with proper producing strategy, and productivity loss can be reduced by
changing the producing sequence.
This study can be used to determine the optimum producing strategy when the well is brought into production and reduce
the productivity loss caused by the condensate banking effect.

Introduction
Gas-condensate reservoirs represent an important source of hydrocarbon reserves and have long been recognized as a reservoir
type, possessing the most intricate flow and complex thermodynamic behaviors. The gas-condensate reservoir is initially gas at
the reservoir condition. Liquid forms in the reservoir when the bottom-hole pressure drops below the dew-point pressure. The
accumulated condensate in the vicinity of the well bore causes a blocking effect and reduces the effective permeability
appreciably, and also causes the losses of well productivity and of the recovery of heavy components at the surface.
The effect of condensate blocking on well productivity is a broad and active research area that has attracted many
researchers, including Fussell (1973), Hinchman and Barree (1985), Aziz (1985), Clark (1985) and Vo et al. (1989). The
productivity loss caused by condensate buildup is striking. According to Whitson (2005), in some cases, the decline can be as
high as a factor of 30. Several examples of severe productivity decline are available in the literature (Engineer, 1985, Duggan,
1972, Allen and Roe, 1950, Abel et al., 1970 and Afidick et al., 1994). Even in very lean gas-condensate reservoirs with a
maximum liquid dropout of only 1%, the productivity may be reduced by a factor of about two as the pressure drops below the
dewpoint pressure (Fevang and Whitson, 1996). Barnum et al. (1995) reviewed data from 17 fields, conducted a survey on
field examples from published industrial cases, and concluded that a severe drop in gas recovery occurs primarily in low
productivity reservoirs with a permeability-thickness below 1000 md-ft.
Condensate blocking is not the only factor that influences the degree of productivity loss. It has been recognized in the
literature that the relative permeability does impact the degree of productivity loss below the dewpoint. Hinchman and Barree
(1985) showed how the choice between the imbibition and the drainage relative permeability curves used in the numerical
reservoir simulations could dramatically alter the productivity forecast below the saturation pressure for gas-condensate
reservoirs.
2 SPE 115786

Fevang and Whitson (1996) addressed the well deliverability problem in their gas-condensate modeling, in which they
observed that the impairment of the well deliverability resulting from the near wellbore condensate blockage effect depends on
the phase behavior, absolute and relative permeabilities, and how the well is being produced. According to Fevang and
Whitson (1996), the well deliverability impairment resulting from the near wellbore condensate blockage depends on the
relative permeability, especially for gas and oil relative permeability ratios (krg/kro) ranging from 0.05 to 0.3. In their well
deliverability calculations, they approximated the condensate saturation in region 2 (where the liquid below the immobile
liquid saturation) with the liquid dropout curve from a CVD experiment. This approximation, however, did not account for the
condensate accumulation and the variations of the overall compositions in the reservoir caused by the liquid buildup; hence it
can not accurately estimate the well deliverability for the condensate blockage effect.
Variations of the fluid flow properties at the time of discovery have been observed and discussed for many reservoirs
around the world (examples include Riemens and de Jong (1985) for Middle Eastern reservoirs and Schulte (1980) for North
Sea reservoirs). Lee (1989) also presented an example to show the variation of the composition and the saturation of a gas-
condensate system due to the influences of the capillary and gravitational forces. The composition change has also been
observed in the field (Yuan et al., 2003). The fluid samples from two wells of the Kekeya gas field in China show that as the
reservoir pressure drops, the produced fluid become leaner and leaner. Furthermore, because the flowing fluid becomes lighter,
the liquid trapped in the reservoir ends up being richer in the heavy components and therefore is more difficult to remedy at
the surface.
Roussennac (2001) illustrated the compositional change during the depletion in his numerical simulation. According to
Roussennac (2001), during the drawdown period, the overall mixture close to the well becomes richer in heavy components as
the liquid builds up in the well grid cell, and the fluid behavior changes from the initial gas-condensate reservoir to that of a
volatile/black oil reservoir.
To characterize the condensate banking dynamics, Wheaton and Zhang (2000) presented a general theoretical model to
show how the compositions of the heavy components in a gas-condensate system change with time around the production
wells during depletion. According to Wheaton and Zhangs model, the rate of change in heavy component composition is
higher for a rich gas-condensate system than for a lean gas-condensate system for the same reservoir, and the condensate
banking problem is particularly acute for low permeability high yield condensate systems.
Bengherbia and Tiab (2002) also demonstrated in their study that both the production history and the simulation prediction
show an increase in lighter components in the flowing phase since the pressure drops below the dewpoint, but it is not clear
how the composition change affects the gas production and the condensate recovery.
The well-producing scheme may impose significant impacts on the phase behavior, hence the composition configuration of
the reservoir fluid. However, the manner by which the producing scheme influences the phase behavior, especially the
composition configuratioin, has not yet been sufficiently addressed. Other parameters, such as the relative permeability and the
initial fluid property (lean or rich gas-condensate), are also important to the production strategy and have not been investigated
so far. In spite of the numerous methods proposed for measuring the relative permeability, investigating the phase behavior
and optimizing the condensate recovery for gas-condensate systems, there is still no completely general approach for phase
behavior analysis, especially for the effect of compositional variations. Since the phase behavior directly influences the
relative permeability and defines the condensate recovery schemes of gas-condensate systems, this work has focused on this
area.

Experimental Design, Apparatus and Procedures


Experimental Design
To investigate the composition change resulting from the pressure variation and the condensate build-up, we need to select a
proper gas-condensate system to conduct the core flooding experiment. In this study, we chose a binary gas-condensate system
based on the following principles:
The system should be easy to handle in the laboratory, thus two to four components are preferred;
The critical temperature of the system should be slightly below 20 C, which makes the experiment easy to perform at
room temperature, and the critical pressure should be such that it can be conducted within a safe pressure range;
A broad condensate region is desirable in order to achieve considerable condensate dropout during the experiment;
Gas and liquid should show large discrepancies in density and viscosity so as to be easily distinguished by the phase.
Figure 1 shows the phase envelope for a binary gas-condensate system which satisfies these four principles. This system is
composed of a mix of 85% methane and 15% butane. At a temperature of 20C and a pressure from 125 atm to 75 atm, this
phase diagram has a good retrograde region.
SPE 115786 3

Figure 1: Phase diagram for a two-component methane-butane gas-condensate system (PVTi, 2003a). For this system, the critical
temperature and the critical pressure are Tc = 6.3C and Pc = 128.5 atm respectively. At room temperature, the system produces a
moderate retrograde region.

Experimental Apparatus
As shown in Figure 2, the gas-condensate experiment consists of four major parts: gas supply and exhaust, the core flow
system, gas sampling, and the data acquisition.

HC Detector Gas In

Water In

C1
&
C4

High
pressure
N2 Water pump LabView

Computer

Figure 2 : Schematic diagram of the gas-condensate flow system.

1. Gas supply and exhaust. The upstream gas mixture is stored in a piston cylinder (HaiAn, China, capacity 4,000 ml,
pressure range 0-32 MPa), and the cylinder pressure is controlled by a nitrogen cylinder with a maximum pressure of 6000 psi.
The downstream gas exhaust is discharged to fumehood directly since the total volume of the exhaust is small and safe to
release into the atmosphere.
2. Core flow system. The core flow system consists of a titanium core-holder (Shiyi Science and Technology, model J300-
01), which can support a system pressure as high as 40 MPa, and a Berea sandstone core-plug with a length of 27.04 cm and a
diameter of 5.06 cm. This core is homogeneous with an average porosity around 17% and an average absolute permeability
about 5 md.
3. Gas sampling. One of the unusual aspects of the experiment is the ability to measure the in-place composition, as well as
the usual pressure and temperature data. The in-place composition samples are collected with Tedlar gas sampling bags
(SKCwest, model 232-02) attached to six sampling ports along the core-holder and one at the upstream end. In order to protect
the gas samples from being polluted by other gases, the gas sampling bags are connected to the system so that the bags can be
vacuumed before the experiment. The sampling pressure is regulated by the 25 psi relief valves to ensure the sampling
4 SPE 115786

pressure is lower than the pressure limits of the sampling bags. The collected gas samples are sent to an Agilent 6890N series
gas chromatograph for compound composition analysis. Since the experiment gas is a combination of light methane and
butane, a special 60m 0.32mm ID capillary column (Agilent, model 113-4362) is configurated for a TCD detector to detect
the light components.
4. The data acquisition part. The in-place pressures are measured by pressure transducers with pressure range of 2000psi.
The saturation of core is measured using a CT scanner (GE HiSpeed CT/i).

Experimental Procedures
Two flow experiments were performed to verify composition variation during production and to test the impact of different
producing strategies on the composition configuration at different core locations. Before the flow tests, the core was saturated
with the mixture of methane and butane at a pressure of 1,956 psi for both runs, then the first batch of composition samples
was taken. The pressure on the core was maintained by the pressure regulator to ensure that seven samples were taken at
roughly the same pressure.
The second batch samples were taken during flow. When the flow reached steady state flow, and the final upstream and
downstream pressures reached 1,954.1 psi and 1,000 psi respectively for the first experiment and 1,974 psi and 500 psi
respectively for the second experiment.The second batch of samples was taken while the flow was taking place.
All gas samples were collected in the sampling bags and sent to the gas chromatograph for composition analysis.
After the second flow experiment, the inlet of the upstream gas supply and the outlet of the downstream gas discharge were
both closed. Pressure in the isolated core equilibrated, and one hour later the still isolated coreholder was detached from the
system and taken to the X-ray computed tomography (CT) scanner for saturation analysis. According to Akin and Kovscek
(2003), the CT number of the core lies on the straight line connecting complete saturation by phase 1 (say liquid) to complete
saturation by phase 2 (say gas). Thus, a single energy scan is sufficient to measure two-phase saturations as shown in Eqn.1:
CTglr = (1 ) r + Sl l + S g g (1)

1 = Sl + S g (2)

Where the subscript glr refers to rock containing both gas and liquid phases. Here, r, g and l are the attenuation
coefficients for the rock matrix, core fully saturated with gas and liquid respectively and Sg and Sl are gas and liquid
saturations respectively. The CT number is defined by normalizing with the linear attenuation coefficients of water, w, as
shown in Eqn. 3:
w
CT = 1000 (3)
w
The porosity, , is defined as:
CTlr CTar
= (4)
CTl CTa
Where the subscripts l and a represent liquid-phase and air-phase CT numbers, whereas lr and ar refer to liquid- and air-
saturation rock respectively.
Thus, the saturation of gas in each voxel is defined as Eqn 5:
CTlr CTglr
Sg = (5)
CTlr CTgr

Simulation models
Binary-component simulation model
The objective of the simulation was to understand the impact of producing scheme on the condensate banking and
compositional variations, and to understand the effects of relative permeability on composition variations. A hypothetic
cylindrical reservoir model, with radius of 9699 ft and near the well region permeability-thickness of 162.5 md-ft was chosen.
In the simulations, small radii grid blocks around the wellbore were chosen to allow for accurate pressure drop calculation in
the near wellbore region. A simulator E300 (2005a, Eclipse) with fully implicit (FULLIMP) method was used to simulate the
performance under different producing strategies.
The reservoir fluid is a binary (C1/C4 = 85%/15%) synthetic gas-condensate system, characterized by Figure 3. This is the
same composition as the mixture used in the lab experiments. The simulation was performed under reservoir temperature 60
F. The producer in this simulation was controlled by gas rate and minimum bottom-hole pressure (BHP). The well initially
produced at the designated gas rate and switched to bottomhole flowing pressure BHP control if the BHP was below the BHP
minimum limit.
SPE 115786 5

Figure 3: A lean binary gas-condensate system with charateristics of (a) PT digram and (b) Liquid dropout curve at 60F.

(a) Effect of different BHP strategies


For each set of relative permeability input, different bottomhole flowing pressure BHP schemes (Figure 4) were performed
to explore the effect of different producing schemes on compositional variations.

decreasing BHP

Figure 4: BHP schemes for binary gas-condensate system simulation.

(b) Relative permeability effect


Two sets of relative permeability curves, as shown in Figure 5, were used as inputs to two different simulations to
investigate the effect of different values of critical gas saturation Scc on condensate blockage and composition variation
phenomena.

Figure 5: Two set of relative permeability curves for simulation input.


6 SPE 115786

(c) Rate schemes with Genetic Algorithm


To find an optimal producing rate scheme for gas production and condensate recovery, the Genetic Algorithm was used as
a search technique in this study. A typical Genetic Algorithm is processed as the procedure shown in Figure 6 and in the
following pseudocode.

mutation

crossover

F(Xj ) reproduction repeat for


objective function N generations
initial population
evaluation
Figure 6: Schematic diagram of the Genetic Algorithm.

Pseudocode for the Genetic Algorithm:


1. Choose initial population.
2. Evaluate the fitness of each individual in the population.
3. Repeat:
Select any two individuals to reproduce.
Breed new generation through crossover and mutation and give birth to offspring.
Evaluate the individual fitness of the offspring.
Replace worst ranked part of population with offspring.
4. Until termination.

In this study, a flow rate is defined as a combination of 10 random rates from 6,200 to 62,000 scf/day, and the individual
population fitness is evaluated by the total gas production (WGPT) and the fraction of the heavier component in the liquid
phase (BXMF). Total 15 generations were performed during the optimization processs.

Multicomponent simulation model


The multicomponent fluid properties are shown in Table 1. Relative permeability scenario 1 was used in the simulation.
Unlike the binary cases studies in the previous section, different gas rate controls but the same minimum bottomhole flowing
pressure BHP = 500 psi constraints (Figure 7) were explored in this simulation.

Table 1: Fluid characterization for a multicomponent gas-condensate system.

Component Composition PC(atm) TC(K) Acentric factor Mol. Weight


N2 0.0121 72.8 304.2 0.22500 44.01
CO2 0.0194 33.5 126.2 0.04000 28.01
C1 0.6599 45.4 190.6 0.00800 16.04
C2 0.0869 48.2 305.4 0.09800 30.07
C3 0.0591 41.9 369.8 0.15200 44.10
C46 0.0967 35.0 448.1 0.22500 66.86
C7+1 0.0475 27.9 465.6 0.31230 107.78
C7+2 0.0152 16.8 587.8 0.55670 198.56
C7+3 0.0033 10.9 717.7 0.91692 335.20
SPE 115786 7

decreasing BHP decreasing BHP

Figure 7: BHP and gas rate schemes for multicomponent gas-condensate system simulation.

Results and Discussion


Experiment results
(a) Pressure profiles
Figure 8 shows the pressure profile for the first experiment run. In this run, the upstream and downstream pressures were
regulated to 1,954.1 psi and 1,000 psi respectively, so the flow was under constant pressure control and reached steady state
flow conditions. The core is fairly homogenous, so the continuously increasing pressure gradients from port 1 to port 5
indicate the decrease of flow mobility, which is caused by the condensate drop-out in the core. As the flow passed port 5, the
pressure gradient starts to increase, which reflects that some part of the liquid was revaporized as the pressure further
decreased.
2000.0

Port 1
1800.0
Port 2

Port 3
1600.0

Port 4
Pressure (psi)

1400.0

Port 5
Port 6
1200.0

Flow direction

1000.0

800.0

core length

600.0
0 50 100 150 200 250 300

Core location (mm)

Figure 8: Pressure profile along the core for the first experiment.

Figure 9 shows the pressure profile for the second experiment run. In this run, the upstream and downstream pressures
were regulated to 1,974 psi and 500 psi respectively. The flow in the second experiment experienced greater pressure drop
than that of the first experiment. Compared with the first experiment, the second flow experienced greater flow deliveriblity
loss as the flow had greater pressure gradient increase from port 1 to port 2.
8 SPE 115786

2000.000

1800.000 Port 1
Port 2

1600.000
Port 3
Pressure (psi)

1400.000
Port 4

1200.000
Flow direction

1000.000 Port 5

Port 6
800.000

core length

600.000
0 50 100 150 200 250 300

Core location (mm)

Figure 9: Pressure profile along the core for the second experiment.
(b) CT images
Before the flow experiment, the core was first fully saturated with liquid-phase butane, and seven cross-sectional images
along the core were taken in the CT scanner. The core was then vacuumed and resaturated with gaseous phase methane, and
another seven cross-sectional images along the core were taken. The core was put back in the flow system and vacuumed again
for the flow experiment. One hour after the second flow experiment, the coreholder was isolated, still under pressure, and
detached from the flow system, and sent back to the CT scanner for saturation measurement. Table 2 shows the average CT
numbers for the seven cross-sectional slices during the three CT scans. The gas saturation (as shown in Figure 10) in the core
was then calculated with Eqn. 5, and the liquid saturation by Eqn.2. From Figure 10, we can see that after the pressure in the
core had built up to 1871psi, the liquid did not totally vaporize as would have been suggested by the original fluid phase
diagram and there is still about 20% liquid left in the core. This is due to two reasons: the accumulation of the liquid in the
core, and the composition variation in the core. Due to the liquid build-up, the fluid composition inside the core was changed.
Hence the fluid phase diagram was different than the original. Even though the core pressure went back to a pressure above or
close to original dew-point pressure, it seemed to be still below the the new dew-point pressure, hence liquid still existed in the
core. In a more dramatic scenario, the phase diagram of the fluid in the core may shift to the left side of the critical point at the
same reservoir temperature. Hence the fluid changes from a gas-condensate to a volatile oil at reservoir temperature. In this
case, repressurizing is not a good strategy to remove the liquid accumulation.

Figure 10: CT images of the core saturated with (a) liquid butane (b) gas methane (c) the mixture of methane and butane and (d) the
difference between liquid butane and gas methane and (e) the difference between liquid butane and the mixture of methane and
butane at l = 74 mm.
SPE 115786 9

Table 2: Average CT numbers for slides at different core locations.

Slice #1 Slice #2 Slice #3 Slice #4 Slice #5 Slice #6 Slice #7


Core with liquid C4 2439.7 2441.5 2571.6 2480.7 2483.1 2465.3 2512.5
Core with gas C1 2280.2 2291.2 2475.1 2343.9 2381.4 2334.5 2403.4
Core with the mixture 2324.4 2320.5 2495.4 2369.4 2406.6 2362.5 2426.7
of C1 and C4
Gas saturation (Sg) 0.723 0.805 0.790 0.814 0.752 0.786 0.786
Liquid saturation (So) 0.277 0.195 0.210 0.186 0.248 0.214 0.214

0.9
Gas saturation, Sg (fraction)

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0
0 50 100 150 200 250 300

CT slice location (mm)

Figure 11: Gas saturation profile after the pressure build up following the second flow.

c) Composition profiles
Figure 12 shows the composition results for the first experiment run. The blue diamand dots in Figure 12(a) are the
compositions for the samples collected before the flow test. At this stage, the core was saturated with the mixture of methane
and butane at a pressure of 1,956 psi. Due to the instantaneous pressure drop during the sampling process, the actual sampling
pressures are depicted by the blue diamand dots in Figure 12(b). From Figure 12(b), we can see that samples collected at port
1, port 2, port3 and port 4 are roughly at the same pressure of 1865psi, which is above the dew-point pressure (1837psi). The
samples collect at these four ports have 15% butane (Figure 12(a)) in the mixture. The sampling pressures at port 5 and port 6
are lower than the dew-point pressure, and hence the heavy component butane is below 15% due to the liquid drop-out in the
core.
The flow test was performed under constant pressure drop since the upstream and downstream pressures were regulated to
1,954.1 psi and 1,000 psi respectively. The pressure profile for the flow tests is shown in Figure 8. Again, due to the
instantaneous pressure drop during sampling process, the actual sampling pressures are depicted by the pink square dots shown
in Figure 12(b). Notice that both port 1 and port 5 underwent bigger pressure drop during the sampling because of long
sampling time. The composition results in Figure 12(a) show that as the pressure drops, the fraction of the heavy component
C4 also decreases. This is due to the fact that as the pressure dreases, liquid drops out into the reservoir, and the accumulated
liquid remains immobile until the liquid saturation exceeds Scc. Since the liquid is mainly made up of the heavy components,
the liquid becomes richer as the pressure decreases, and on the contrary, the flowing phase becomes leaner as the pressure
drops. In the PVT cell, the liquid is supposed to revaporize when the pressure is below 1500psi according to the liquid drop-
out curve in Figure 3(b). However, in the porous medium, due to the constraint of relative permeability and the interfacial
tension between the rock and the fluid, the immobile liquid accumulates locally. The accumulated liquid alters the local fluid
composition, and hence the fluid might not be able to revaporize at the pressure where the original fluid can. This is confirmed
by the composition results of port 1, port 4, port5, and port 6, where the sampling pressures are below 1500psi. Both port 2 and
port 3 have butane percentage higher than the original composition. This might due to the fact that at the sampling pressures of
port 2 and port 3, the fluid has the maximum liquid drop-out, and the accumulated liquid saturation exceeds Scc, so some part
of the liquid becomes mobile. This brings more heavy component to the flowing phase.
10 SPE 115786

30
Butane (C4) in the flowing phase (%)

original composition
25 composition during flow

20

15

port 1 port 2 port 3


10 port 4
port 5
5 port 6

0
0 50 100 150 200 250 300

Core location (mm)


(a)
2000

1800

1600
Sampling pressure (psi)

1400
port 2
1200
port 3
1000 port 4
port 1
800
port 6
600 port 5

400
original pressure
200 pressure during flow
0
0 50 100 150 200 250 300

Core location (mm)


(b)
Figure 12: Composition results showing (a) Compositin variation during the flow experiment under (b) Sampling pressures for the
first experiement.

Simulation results
Two-component simulation results
(a) Effect of different BHP strategies
For the binary gas-condensate simulation, we tested three scenarios, different bottomhole pressure BHP scheme, different
relative permeability input and different rate control.
In a PVT cell, the liquid drop-out from the gas-condensate system can be revaporized if we either lower the BHP or
increase the BHP. However, in a porous medium, the liquid drop-out is immobile unless the liquid accumulation reaches the
critical condensate saturation value on the relative permeability curve. The accumulated condensate is generally made up of
heavier components and hence changes the local phase composition. Whether the condensate build-up can be revaporized is
mainly determined by the local fluid composition and its consequent phase diagram. From the CT saturation measurements,
we already saw that pressure increase is not a good strategy to remove the liquid accumulation in the core. Figure 13 shows the
simulation results for a field scale binary gas-condensate system. From Figure 13 (a), we can see that the total gas production
increases as the BHP decreases. In this case, we can decrease the BHP to achieve greater pressure difference; hence
temporarily to produce more gas from the reservoir. However, lowering the BHP will cause the expansion of the two-phase
region, and the accumulation of more heavy-component in the reservoir, and less heavy-component recovered at the surface
(Figure 13 (b) and (c)). Hence, lowering the BHP may be a good strategy for tatal gas production, but in the long run, not be a
good strategy for maximizing total fluid recovery because the heavy component is generally difficult to remove from the
reservoir.
SPE 115786 11

decreasing BHP

(a)

decreasing BHP decreasing BHP

(b) (c)
Figure 13: Effects of BHP schemes on (a) total gas production, (b) yC4 and (c) yC4 in the well for a binary gas-condensate system.

(b) Relative permeability effect


Once the liquid drops out to the reservoir, it is immobile unless the accumulated liquid saturation exceeds the critical
condensate saturation value (Scc). Hence different rocks with different Scc will impact the amount of liquid saturation in the
reservoir differently. Figure 14 shows the results of two simulations with different relative permeability inputs (Figure 5). As
expected, more liquid accumulated in the reservoir with higher critical condensate saturation. To improve the condensate
recovery, one way would be to reduce the interfacial tesion (IFT) factor of the rock and hence lower the critical condensate
saturation.

(a) Kr scenario 1 with Scc = 0.25 (b) Kr scenario 2 with Scc = 0.45

Figure 14: Effects of relative permeability on liquid buildup.


12 SPE 115786

(c) Rate schemes with Genetic Algorithm


Instead of fixing the BHP or producing rate at a predefined value, combinations of random rates were used in the Genetic
algorithm in an attempt to define an optimal strategy. The final top three optimized rates are shown in Figure 15. From Figure
15 (b) and (c), we can see although we gained total gas production by optimizing the producing rate, we left more heavy-
component C4 in the reservoir. The final composition configuration is determined by the final rate control, the lower the
producing rate, the lower the amount of heavy-component left in the reservoir.
4
x 10 Three highest WGPT cases
12
first
second
10 third

8
WGPR(mscf/day)

0
0 1 2 3 4 5 6 7
Time (day)

(a)
5
x 10 Three highest WGPT cases Three highest WGPT cases
3.5 0.75

3 0.7

0.65
2.5
0.6
WGPT (mscf)

BXMF of C4

2
0.55

1.5 0.5

0.45
1
first
0.4 first
second
0.5 second
third 0.35
third
0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Time (day) Time (day)

(b) (c)
Figure 15: Effects of (a) rate schemes on (b) total gas production and (b) the heavier componet C4 in the well for a binary gas-
condensate system.

Multicomponent simulation results


Figure 16 shows the simulation results for a multicomponent gas-condensate system as defined earlier in Table 1. The general
conclusions for the BHP strategy are the same for both multicomponent and binary-component systems. That is, the total gas
production increases as a result of greater pressure difference between the reservoir and the well, but at the same time, lower
BHP also recovers less heavy-component at the surface(Figure 16 (b) and (c)), hence deopsits more heavy-component into the
reservoir and generates a larger two-phase region in the local phase diagram.
SPE 115786 13

decreasing BHP

(a)

decreasing BHP
Well yC4-6

decreasing BHP

(b) (c)
Figure 16: Effects of BHP schemes on (a) total gas production, (b) yC4-6 and (c) yC7+2 in the well for a multicomponent gas-condensate
system.

In summary from the simulation results, we can conclude that there is no standard way to optimize the producing strategy.
Using low BHP, we can achieve high total gas production temporarily, however, to minimize the condensate banking blockage
and hence to enhance the ultimate gas and liquid recovery, higher BHP may be a better strategy. The optimal approach is
likely to be dependent on the original composition and phase diagram of the fluid as well as the relative permability curves for
the rock.

Conclusions
1. In gas-condensate flow, local composition changes due to the influence of relative permeability and interfacial
tension effects.
2. The reservoir flow might not revaporize as suggested by the CVD experiment in the PVT cell due to the local
composition variation.
3. Repressurizing might not be a good strategy to remove the liquid accumulation in the reservoir.
4. Composition and condensate saturation change significantly as a function of producing sequence. The higher the
BHP, the less the condensate banking and a smaller amount of heavy-component is trapped in the reservoir. The
lower the producing rate, the lower the amount of heavy-component left in the reservoir.
5. Gas productivity can be maximized with proper producing strategy. The total gas production can be achieved by
lowering the BHP or optimizing the producing rate.
6. Productivity loss can be reduced by optimizing the producing sequence.
7. The condensate drop-out will hinder the flow capability, due to relative permeability effects.
14 SPE 115786

Nomenclature

BHP bottom-hole pressure


CO2 carbon dioxide
C1 methane
C2 ethane
C3 propane
C4-6 Butane to hexane
C7+1 The mix1 of Heptane and heavier components
C7+2 The mix2 of Heptane and heavier components
C7+3 The mix3 of Heptane and heavier components
CTa Air-phase CT number
CTl Liquid-phase CT number
CTar CT number for air-saturated rock
CTlr CT number for liquid-saturated rock
CTgr CT number for gas-saturated rock
CTglr CT number for the rock saturated with the mix of liquid and gas
CVD constant volume depletion
N2 nitrogen
MW molecular weight
pc critical pressure (psi or atm)
Scc critical condensate saturation
Sg Gas saturation
Sl liquid saturation
Tc critical temperature (K or C)
WGPT well total gas production (mscf or scc)
WBHP Well bottom hole pressure(psi or atm)
WGPR Gas production rate for a single well (mscf/day)
porosity
g the attenuation coefficients for the core fully saturated with gas
l the attenuation coefficients for the core fully saturated with liquid
r the attenuation coefficients for the rock matrix

Acknowledgements
We would like to express our appreciation to the members of SUPRI-D (Research Consortium on Innovation in Well Testing)
for financial support and useful discussions.

References
Abel, W., Jackson, R., and Wattenbarger, R. (1970). Simulation of a Partial Pressure Maintenance Gas Cycling Project with a
Compositional Model, Carson Creek Field, Alberta. JPT January, pp 3846.
Afidick, D., Kaczorowski, N., and Bette, S. (1994). Production Performance of a Retrograde Gas: A Case Study of the Arun Field. In SPE
Asia Pacific Oil and Gas Conference, Melbourne, Australia. SPE 28749.
Akin, S., and Kovscek, A.R. (2003). Computed Tomography in Petroleum Engineering Research. Geological Society, London, Special
Publications 2003; v.215; pp 23-38.
Allen, F. and Roe, R. (1950). Performance Characteristics of a Volumetric Condensate Reservoir. Petroleum Transactions, AIME, pp 83
90.
Aziz, R. (1985). Deliverability Projection Model for Overpressured Gas-Condensate Reservoirs. In SPE Middle East Technical
Conference and Exhibition, Bahrain. SPE 13706.
Barnum, R., Brinkman, F., Richardson, T., and Spillette, A. (1995). Gas Condensate Reservoir Behavior: Productivity and Recovery
Reduction Due to Condensation. In SPE Annual Technical Conference and Exhibition, Dallas, TX. SPE 30767.
Bengherbia, M. and Tiab, D. (2002). Gas-Condensate Well Performance Using Compositional Simulator: A Case Study. In SPE Gas
Technology Symposium, Calgary, Canada. SPE 75531.
Clark, T. (1985). The Application of a 2-D Compositional, Radial Model to Predict Single-Well Performance in a Rich Gas Condensate
Reservoir. In SPE Annual Technical Conference and Exhibition, Las Vegas, NV. SPE 14413
Duggan, J. (1972). The Anderson L-an Abnormally Pressured Gas Reservoir in South Texas. JPT February, pp 132138.
Engineer, R. (1985). Cal Canal Field, California: Case History of a Tight and Abnormally Pressured Gas Condensate Reservoir. In SPE
1985 California Regional Meeting, Bakersfield, CA. SPE 13650.
Fevang, O. and Whitson, C. (1996). Modeling Gas-Condensate Well Deliverability. SPE Reservoir Engineering, pp 221230.
Fussell, D. (1973). Single-Well Performance Predictions for Gas Condensate Reservoirs. JPT July, pp 860870.
Hinchman, S. and Barree, R. (1985). Productivity Loss in Gas condensate Reservoirs. In SPE Annual Technical Conference and
Exhibition, Las Vegas, NV. SPE 14203.
SPE 115786 15

Lee, S. (1989). Capillary-Gravity Equilibria for Hydrocarbon Fluids in Porous Media. In SPE Annual Technical Conference and
Exhibition, San Antonio, TX. SPE 19650-MS.
Riemens, W. and de Jong, L. (1985). Birba Field PVT Variations Along the Hydrocarbon Column and Confirmatory Field Tests. In SPE
Middle East Oil Technical Conference, Bahrain. SPE 13719.
Roussennac, B. (2001). Gas Condensate Well Test Analysis. Masters thesis, Stanford University, Stanford, CA.
Schulte, A. (1980). Compositional Variation within a Hydrocarbon Column Due to Gravity. In SPE Annual Technical Conference and
Exhibition, Dallas, TX. SPE 9235.
Vo, D., Jones, J., and R.Raghavan (1989). Performance Prediction for Gas Condensate Reservoirs. SPE Formation Evaluation December,
pp 576584.
Wheaton, R. and Zhang, H. (2000). Condensate Banking Dynamics in Gas Condensate Fields: Compositional Changes and Condensate
Accumulation Around Production Wells. In SPE Annual Technical Conference and Exhibition, Dallas, TX. SPE 62930.
Whitson, C. (2005). Khuff Gas Condensate Development. In International Petroleum Technology Conference, Doha, Qatar. IPTC 10692.
Yuan, S., Ye, J., and Sun, Z. (2003). Theory and Practices in Gas-Condensate Reservoir Development. China Petroleum Industry Press,
Beijing, China.

SI metric conversion Factors

atm 1.013250 * E+05 = Pa


ft3 1.589873 E-01 = m3
(F-32)/1.8 = C
in3 1.638706 E+01 = cm3
psi 6.894757 E+00 = kPa

* Conversion factor is exact.