See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/254528075

Underground Gas Storage in a Partially Depleted Gas Reservoir

Article  in  Oil & Gas Science and Technology · November 2008

DOI: 10.2118/113588-MS

CITATIONS READS

20 502

4 authors, including:

Reza Azin
Persian Gulf University
125 PUBLICATIONS   586 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Pre-Darcy Flow and Klinkenberg effect in Dense, Consolidated Carbonate Formations View project

Montan University Ecoindustrial tour in Iran View project

All content following this page was uploaded by Reza Azin on 17 October 2014.

The user has requested enhancement of the downloaded file.

Iran. J. Chem. Chem. Eng. Research Note Vol. 29, No. 1, 2010

Investigation of Underground Gas Storage

in a Partially Depleted Naturally Fractured Gas Reservoir

Jodeyri Entezari, Ali

Tehran Energy Consultants, Khoramshahr St., Tehran 1554614313, I.R. IRAN

Azin, Reza*+
Department of Chemical Engineering, Engineering Faculty, Persian Gulf University, Bushehr, I.R. IRAN

Nasiri, Amir
Tehran Energy Consultants, Khoramshahr St., Tehran 1554614313, I.R. IRAN

Bahrami, Hassan●
Faculty of Chemical and Petroleum Engineering, Sharif University of Technology, Tehran, I.R. IRAN

ABSTRACT: In this work, studies of underground gas storage (UGS) were performed on
a partially depleted, naturally fractured gas reservoir through compositional simulation. Reservoir
dynamic model was calibrated by history matching of about 20 years of researvoir production.
Effects of fracture parameters, i.e. fracture shape factor, fracture permeability and porosity were
studied. Results showed that distribution of fracture density affects flow and production of water,
but not that of gas, through porous medium. However, due to high mobility of gas, the gas
production and reservoir average pressure are insensitive to fracture shape factor. Also, it was
found that uniform fracture permeability distribution enhances communication within reservoir and
consequently more pressure support is obtained by water bearing of aquifer. Effect of aquifer on the
reservoir performance was studied, and it was found that an active aquifer can reduce condensate
drop out around the well bore. On the other hand, water invasion is an important issue which may
kill the well. Results showed that use of horizontal wells is superior to vertical wells in order to
avoid detrimental effects of active aquifer.

KEYWORDS: Underground gas storage, Naturally fractured reservoir, Fracture shape factor,
Aquifer, Horizontal well.

INTRODUCTION
The demand for natural gas depends heavily on means of balancing demand and supply of natural gas
weather. Underground gas storage (UGS) is an economical during year.
* To whom correspondence should be addressed.
+ E-mail: [email protected]
●
Current address: Faculty of Petroleum Engineering, Sharif University of Technology, Tehran,, I.R. IRAN
1021-9986/10/1/103 8/$/2.80

103
Iran. J. Chem. Chem. Eng. Jodeyri Entezari, A., et al. Vol. 29, No. 1, 2010

Depleted gas and/or oil reservoirs are the best candidates The fracture parameters, namely fracture permeability (Kf),
for UGS. These reservoirs contain a cap rock with fracture storativity (ω) and fracture conductivity (λ), can
ensured prevention of gas/oil migration over geological be obtained from welltest analysis. Average reservoir
time. Also, it takes advantages of existing wells, surface permeability can be estimated from Eq. (3);
facilities, and pipeline systems which reduce investment
K Average = [162.6 q.µ o .Bo ] /[m.h] (3)
costs. These reservoirs are cheapest to develop, operate
and maintain compared to other candidates for gas storage. The fracture permeability can be estimated from
A NFR is the one in which fractures have direct effect equation (4) [4]:
on fluid flow, reservoir anisotropy, hydrocarbon recovery
and storage [1]. The most common model normally used k2
kf = (4)
for fracture characterization is dual-porosity and km
dual-permeability model introduced by Warren & Root [2],
Fracture conductivity characterizes the ability of
where the reservoir is considered as the rock matrix and
the matrix blocks to flow into fracture system[2]:
fractures. Flow takes place through fractures and matrix
act as fluid source. Fractures may have a positive or Km 2
λ=δ rw (5)
negative effect on oil or gas production. Wells in Kf
a fractured reservoir have higher deliverability,
The shape factor (δ) is related to fracture density.
an important issue in UGS where high rate of injection
From Warren & Root theory [2]:
and especially withdrawal is essential.
In NFR, horizontal wells intersect fractures and drain 1 1 1
δ=4 ( + + ) (6)
fractures and reservoir effectively. Horizontal wells can L2x L2y L2z
be employed effectively to reduce water conning, near
well bore turbulence effect and number of wells, and A large shape factor implies smaller block size or
increase rate of gas or oil production in tight reservoirs. higher density of fractures. By above equations, shape
In UGS cases, it is necessary to have a high injection and factor and average matrix block size are related to
withdrawal rates during a relatively short time, use of each other:
horizontal wells can be very useful [3]. δ = [λ.K f ] /[K m .rw 2 ] (7)
In this work, a partially depleted, naturally fractured
gas reservoir was used to study for UGS. A compositional L = rw 12 K m / (λ.K f ) (8)
and dual- porosity model of the reservoir was constructed.
After history matching, new wells were Shape factor and matrix block size can be obtained
defined to speed up the depletion phase. Then, using fracture conductivity from welltest, matrix
injection/withdrawal (I/W) of gas was defined in the permeability from core analysis, and knowing fracture
model and effects of fracture parameters, aquifer and permeability.
horizontal wells were studied. Fracture storativity is the fraction of fluid stored in the
fracture system [2]:

(ϕ f C f ) (9)
NATURALLY FRACTURED RESERVOIRS ω=
(ϕ f C f + ϕ m C m )
NFR contain two porosity systems, that of rock matrix
( ϕ m ) and of fractures and vugs ( ϕ f ) [4]: Using definition of fracture storativity, the fracture
porosity can be calculated as follows [4]:
Matrix void volume Cm ω
ϕm = (1) ϕf = ϕm (10)
Matrix bulk volume C f 1− ω

Fracture void volume Fracture compressibility might be different from

ϕf = (2)
Total bulk volume matrix compressibility by an order of magnitude.

104
Iran. J. Chem. Chem. Eng. Investigation of Underground Gas Storage ... Vol. 29, No. 1, 2010

Table 1: Initial reservoir fluid com.

Component N2 CO2 C1 C2 C3 IC4 NC4 IC5 NC5 C6 C7+

Composition, mole % 2.655 1.56 89.159 3.804 1.166 0.234 0.334 0.16 0.117 0.22 0.591

100000 350.0 300 0.8

Gas Production Rate, MSCF/D

90000 0.7

Cumulative Oil Production,

Oil Production Rate, STB/D
300.0

Cumulative Gas Production,

250
80000
0.6
70000 250.0
200
60000 0.5

MMSTB
MMMSCF
200.0
50000 150 0.4
40000 150.0
0.3
100
30000 100.0 0.2
20000 50
50.0 0.1
10000
0 0.0 0 0.0
0 2 4 6 8 10 12 14 16 18 0 2 4 6 8 10 12 14 16 18

Years (a) (b) Years

Fig. 1: Daily and cumulative production history from reservoir (a: gas production; b: condensate production).

METHODOLOGY the initial and lumped reservoir fluid and injection gas
In this work, simulation study was conducted on compositions, respectively. Fig. 2 illustrates phase
an Iranian gas reservoir using compositional module of envelope of reservoir fluid. Three wells (one existing and
GeoQuest software, Eclipse 300, Version 2004 [5]. 2 new wells) were used for I/W. Each cycle took 6
The reservoir was initially at 3130 psia and 171°F, months for injection and 5 months for withdrawal.
and contained about 1 TCF original gas in place. The injection period in each year was from April 15th to
It has produced for about 16 years with a single well. October 15th, and production period was from November
Fig. 1 illustrates the daily and cumulative production 1st to March 31st of next year.
history from this reservoir.
A comprehensive fracture study revealed that RESULTS & DISCUSSION
the reservoir contains a network of fractures which contribute A total of 6 different scenarios were run to investigate
to production. The fracture density on top of structure effects of fracture parameters, well direction (horizontal
where dip is high is higher than flanks. A dual porosity vs. vertical), and aquifer on the performance of UGS.
compositional model was then employed to simulate fluid Table 3 summarizes these scenarios
flow in this reservoir. Average permeability of matrix and
fractures is 0.72 and 93 md, respectively. Also, average FRACTURE NETWORK AND RESERVOIR
porosity of matrix and fractures is 0.051 and 0.00085, HETEROGENEITY
respectively. The reservoir was discretized into A. Shape Factor
111×41×10 cells which first half of grids were allocated Fig. 3 shows distribution of sigma in layer 13. The top
to matrix blocks and second half of grids to of reservoir has a higher shape factor. This seems
corresponding fractures. to be reasonable, as the higher dip of layers increases
Steps in the simulation study of UGS were described the likelihood of fracturing in the reservoir rock.
elsewhere [10]. The IRAP RMS 7.5.1 software [6] Based on the real model, distribution of fracture shape factor
was used to construct a geological model. Also, the was taken non-uniform throughout the reservoir. However,
Peng-Robinson EOS [7] was used and tuned to predict as we are dealing with a gas reservoir, using a uniform
the phase behavior of reservoir fluid. Tables 1 and 2 show shape factor throughout the model might be reasonable.

105
Iran. J. Chem. Chem. Eng. Jodeyri Entezari, A., et al. Vol. 29, No. 1, 2010

Table 2:Lumped reservoir fluid composition and Composition of injection gas.

Lumped Component C1-N2 C2-CO2 C3-NC4 IC5-NC5 FC6 C7-C11 C13+

Reservoir Fluid Composition, mole % 91.814 5.364 1.734 0.277 0.22 0.525 0.066

Injection Fluid Composition, mole% 97.5 2.46 0.04 0 0 0 0

P-T Diagram
2500

2000 Layer 13
Pressure (psia)

1500

1000

500 Temperature (deg F)

Sigma
0
-200 -100 0 100 200 300 0.000 0.015 0.030 0.045 0.060
Temperature (deg F)
Fig. 2: Phase envelope of reservoir fluid Fig. 3: Distribution of sigma in Layer 13

This has the advantage of decreasing simulation run time. B. Fracture Permeability
To compare the effect of shape factor, 2 models were Fig. 6 shows distribution of fracture permeability
defined. In the first model (Run 015), sigma distribution in layer 13. To test the effect of fracture permeability on
map based on fracture study was used; in second model the reservoir performance for UGS, a case (Run 018) was
(Run 012), a uniform sigma value (0.01, which run with uniform fracture permeability of 90 md, which is
corresponds to a block dimension of 3.5 meters in all the mean value of fracture permeability in the model, and
directions) was used throughout reservoir. Figs. 4 and 5 the results were compared with Run 015. Figs. 7-10 show
compare gas and water production rate in these two the reservoir gas production, average pressure, and GOR
models, respectively. According to Fig. 4, the gas and water production rate, respectively.
production is insensitive to sigma. The same result According to Fig. 7, there is a minor difference in gas
was obtained for and reservoir average pressure. production rate and both models can produce with
The high mobility of gas makes it easy to flow through the anticipated target rate. But, Fig. 8 shows a considerable
fracture network to well bore. Generally, shape factor is difference in reservoir average pressure (FPR), especially
more important in oil reservoirs than gas reservoir [8]. in injection-withdrawal cycles.
On the other hand, due to lower mobility, water In uniform fracture permeability case (018), the FPR
production from reservoir may be affected by the is higher. The idealistic uniform fracture permeability
distribution of fracture density. Higher water production distribution enhances communication within reservoir
in model 015, shown in Fig. 5 confirms this idea. Higher and consequently more pressure support is obtained by
density of fractures around wells in case 015 causes water bearing cells.
breakthrough of water, while the uniform case 012 As can be seen in Fig. 6, many water bearing cells
produced no water. In other words, shape factor affects below gas-water contact have low fracture permeability
movement of water through porous medium as a fluid that makes them less effective in pressure maintenance
with lower mobility more than gas. As there is no aquifer of reservoir. In the case with a uniform permeability
defined in these 2 models, the produced water origin is of fracture for these cells, they can effectively contribute
from connate water. in pressure maintenance. In addition, higher average pressure

106
Iran. J. Chem. Chem. Eng. Investigation of Underground Gas Storage ... Vol. 29, No. 1, 2010

Table 3: Summary of different scenarios in this work.

Case Sigma Analytical Aquifer Horizontal Wells Fracture Permeability

011 constant 0.01 Yes No Map

012 constant 0.01 No No Map

015 Map No No Map

015a Map Yes No Map

017 Map Yes Yes Map

018 Map No No constant 90 md

180000 50
Case 015 Case 015
160000 Case 012 Case 012
40
FGRP MSCF (day)

140000

FGRP STB (day)

120000
30
10000
80000 20
60000
40000 10
20000
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Time (years) Time (years)

Fig. 4: Comparison of gas production rate for models 012 and 015. Fig. 5: Comparison of water production for models 012 and 015.

causes less GOR, which is reflected in Fig. 9. Also, process like reducing reservoir volume in successive I/W
higher water production shown in Fig. 10 indicates that cycles, increasing compressor power, and reducing
water flow from water bearing cells is faster. relative permeability to gas [10].
To study the effect of aquifer on UGS performance,
C. Fracture Porosity two models (case 011 and case 012) were run and
A model with constant fracture porosity (case 020) their results were compared. Case 011 has an analytical
was run and compared with case 015. The constant aquifer which its properties are given in Table 4, and
porosity value was selected so that the initial gas in place case 012 does not have an aquifer. The results show that
in fractures was kept unchanged. As just 4.5% of original reservoir average pressure, GOR and water production
gas in place was stored in the fractures, the main act are very different in two models. In case 011, pressure
of fractures is to enhance total permeability and increases continuously in successive I/W periods.
communication between matrix blocks rather than However, in case 012, the rate of pressure increase is
reservoir storativity. The results of these cases are not much less, as can be seen in Fig. 11. Maintaining
very different. reservoir at higher pressure during successive I/W cycles
by the act of an active aquifer leads to less condensate
ANALYTICAL AQUIFER EFFECT drop out in reservoir. As a result, the gas-oil ratio will be
The Carter-Tracy analytical aquifer model [9] was set lower. Fig. 12 compares condensate saturation near
in the dynamic model. Properties of analytical aquifer the well bore for cases 011 and 012. It is clear from this
are given in Table 4. Active aquifer can reduce ultimate figure that an active aquifer can reduce condensate drop
recovery of a gas reservoir, and impose problems in UGS out around the well bore. Such a high water cut kills

107
Iran. J. Chem. Chem. Eng. Jodeyri Entezari, A., et al. Vol. 29, No. 1, 2010

6000
Case 018
Case 015
5000

FGOR MSCF/STB
Layer 13

4000

3000

2000

1000
Fracture Permeability, md
0
0 100 200 300 400 0 10 20 30 40 50
Time (years)
Fig. 6: Distribution of fracture permeability in different Fig. 9: Comparison of GOR for models 015 and 018.
layer 13.

800
180000
Case 018 Case 018
160000 Case 015 700 Case 015
140000 600
FGPR MSCF (day)

FWPR STB/DAY

120000 500
100000
400
80000
300
60000
200
40000
20000 100
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Tim (years) Time (years)

Fig. 7: Comparison of reservoir gas production for models Fig. 10: Comparison of water production rate for models
0.15 and 0.18. 015 and 018.

3200 3200
Case 018 Case 012
3000 Case 015 300 Case 011

2800 2800
FPR (Psia)

2600 2600
FPR (Psia)

2400 2400
2200 2200
2000 2000

1800 1800

1600 1600
0 10 20 30 40 50 0 10 20 30 40 50
Tim (years) Time (years)

Fig. 8: Comparison of reservoir average pressure for models Fig. 11: Comparison of reservoir average pressure for models
0.15 and 0.18. 011 and 012.

108
Iran. J. Chem. Chem. Eng. Investigation of Underground Gas Storage ... Vol. 29, No. 1, 2010

Table 4: Properties of analytical aquifer.

Permeability, md Water & Rock

Porosity, fraction Inner Radius,ft Thickness,ft Angle, Degree
Compressibility, 1/psi
5 0.05 0.0000035 50000 1000 360

Well Cell Condensate Saturation vs. Time

0.001 12000
Well Cell Condensate Saturation,

Aquifer Case 017

No aquifer Case 015a
0.00 10000

FWPR STB (day)

0.000 8000
Fraction

0.000 6000

0.000 4000

0.000 2000

0 0
0 5 10 15 20 25 30 35 40 45 50 0 10 20 30 40 50
Date Time (years)
Fig. 12: Condensate saturation near the well block for cases Fig. 13: Water production in reservoir with vertical wells (case
011 and 012. 015a) and horizontal wells (case 017)Table 1-Initial reservoir
fluid composition.

the well due to water invasion. One of the solutions is Abbreviations

to use horizontal well instead of vertical wells, discussed EOS Equation of State
in the next section. FPR Field Pressure
GOR Gas Oil Ratio
HORIZONTAL WELLS I/W Injection/ Withdrawal
As mentioned before, horizontal wells can intersect NFR Naturally Fractured Reservoir
more fractures and provide more area for gas UGS Underground Gas Storage
injection/withdrawal, as well as reducing the risk of high
water cut and conning problem during UGS. Two cases
were run to study and observe the role of horizontal wells Nomenclature
and compare reservoir performance with vertical wells Bo Oil formation volume factor, bbl/STB
(case 015a) and horizontal wells (case 017), completed Cm Compressibility of matrix, 1/psi
near the crest of reservoir. Water production in these Cf Compressibility of fracture, 1/psi
cases are plotted and compared in Fig. 13. According to K Average reservoir permeability from welltest
this figure, water production reduced considerably in the analysis, md
case 017 with horizontal wells. So, for this reservoir, Km Average matrix permeability from core
use of horizontal wells is superior to vertical wells. analysis, md
Kf Fracture permeability, md
CONCLUSIONS L Matrix block size, ft
Water production from reservoir is affected by the h Producing thickness, ft
distribution of fracture density; Fracture shape factor m Slope of pressure build-up data versus log of
affects movement of water through porous medium, time in Horner plot, dimensionless
but not for gas; Active aquifer can reduce condensate q Flow rate, STB/d
drop out around the well bore; Use of horizontal wells rw Well bore radius, ft
is superior to vertical wells in this case. λ Fracture conductivity, dimensionless

109
 Iran. J. Chem. Chem. Eng. Jodeyri Entezari, A., et al. Vol. 29, No. 1, 2010

µo Oil viscosity, cp
ϕf fracture porosity, dimensionless
ϕm Matrix porosity, dimensionless
δ Shape factor, 1/ft2
ω Fracture storativity, dimensionless

Acknowledgement
We are indebted to the Persian Gulf University
Deputy of Research for supporting this work under grant
no. ENG-3587

Received : May 1, 2008 ; Accepted : May 30, 2009

REFERENCES
[1] Nelson A.R., “Geologic Analysis of Naturally
Fractured Reservoirs”, 2nd ed., Gulf Professional
Publishing (2001).
[2] Van Golf-Racht T.D., “Fundamentals of Fractured
Reservoir Engineering”, 1st ed., Elsevier Scientific
Publishing Company (1982).
[3] Joshi S., “Horizontal Well Technology”, 1st ed.,
Pennwell Publishing Company (1991).
[4] Tiab D., Restrepo D.P., Igbokoyi A., “Fracture
Porosity in Naturally Fractured Reservoir”,
SPE 104056, Presented at First International Oil
Conference and Exhibition in Mexico, Cancun,
Mexico, 31 (2006).
[5] "Eclipse Reference Manual, version 2001",
Schlumberger (2001).
[6] "IRAP RMS Software Manual, Version 7.5.1",
ROXAR (2006).
[7] Peng D.Y., Robinson D.B., A New Two-Constant
Equation of State, Ind. Eng. Chem. Fundam., 15,
p. 59 (1976).
[8] Kazemi H., Numerical Simulation of Water-Oil Flow
in Naturally Fractured Reservoirs, SPEJ, p. 317
(1976).
[9] Carter R.D., Tracy G.W., An Improved Method for
Calculating Water Influx, Trans. AIME 219; J. Pet.
Tech, p. 58 (1960).
[10] Azin R., Nasiri A., Jodeyri Entezari A., Underground
Gas Storage in a Partially Depleted Gas Reservoir,
Oil & Gas Sci. Tech., 63 (6), p. 691 (2008).

110

View publication stats