SPE 88717
Simulation Study of Miscible Gas Injection for Enhanced Oil Recovery in Low
Permeable Carbonate Reservoirs in Abu Dhabi
W. van Vark*, S. K. Masalmeh, and J. van Dorp** – Shell Abu Dhabi,
M. Abu Al Nasr and S. Al-Khanbashi – Abu Dhabi National Oil Company
Presently with: * Shell International Exploration and Production, ** Petroleum Development Oman LLC,
Copyright 2004, Society of Petroleum Engineers Inc.
This paper was prepared for presentation at the 11th Abu Dhabi International Petroleum
Exhibition and Conference held in Abu Dhabi, U.A.E., 10–13 October 2004.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in a proposal submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300
words; illustrations may not be copied. The proposal must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.
Abstract
In this paper we will present the outcome of a study
conducted to evaluate the feasibility of large scale injection of
sour and/or acid gas into a low permeable carbonate reservoir
to enhance oil recovery.
Other than for disposal, H2S containing mixtures have
rarely been injected as a miscible agent in oil recovery
projects. Moreover, the very few projects that have actually
been executed are relatively small (generally less than 10
MMscf/d).
In this study different recovery processes were evaluated
such as water flooding, lean gas injection, sour gas (natural
gas with a large H2S content) injection, acid gas injection,
acid gas (mixture of H2S and CO2) after a slug of sour gas and
CO2 injection. To evaluate these different (EOR) recovery
processes, a detailed reservoir description is essential and for
this purpose element-models were used.
The critical importance of a thorough understanding of
reservoir geology and rock properties for miscible gas
injection schemes has been confirmed by the experiences of
water breakthrough and over-ride in a number of reservoirs in
Abu Dhabi and the poor performance of some miscible gas
injection projects in the industry.
The simulation study shows that miscible acid gas
injection is the preferred recovery mechanism for part of the
reservoir under study. This is a result of several key factors,
including the favorable miscibility with the native oil (lower
miscibility pressure with reservoir crude), better solvent for
asphaltene, a more favorable mobility ratio due to high acid
gas viscosity and density and availability of large quantities of
acid gas from the underlying formation. Acid gas is therefore
an attractive, low cost, miscible, enhanced oil recovery agent,
provided fully adequate corrosion mitigation procedures and
HS&E management systems are implemented in the
development.
Introduction
The class of reservoirs that are the subject of this investigation
into the merits of alternative gas injection are large, high N/G,
about 50 m thick carbonate reservoirs. Water injection has
been the prime recovery process in these reservoirs that do not
have a substantially natural water drive. In these large
reservoirs properties vary and flank areas often show a severe
reduction in permeability. Vertically averaged permeability is
often more than 100 mD in the crest but deteriorates to less
than 10 mD at the periphery. With such low permeability,
waterflooding quickly looses its attractiveness, as it will
require a very dense well spacing or result in a protracted
field life. The use of gas as an injectant alleviates these
drawbacks to a certain extent, but due to the lower viscosity
the gas injection process is less efficient than water injection.
Moreover, injection of lean (high methane content) gas carries
significant cost, as there is a market for gas in the region.
Large volumes of sour gas are encountered in the region
and are often found directly below the oil reservoirs. Mainly
because of the vast HSE complexities associated with
producing large volumes of H2S, these sizeable volumes are
so far untapped. Safe handling of H2S remains a challenge, in
particular from a sustainable development perspective.
However, in-depth studies have demonstrated that the safety
aspect can be managed within a sound economical framework
[1].
This paper reports about a simulation study that was
undertaken with the objective of identifying the impact of
alternative (gas) injectants on the recovery factor for the low
permeable areas of these large carbonate oil reservoirs.
Injectants
A typical field in this region consists of a stack of reservoirs
where the type of hydrocarbon progresses with depth from
relatively heavy oils in the shallow reservoirs, through light
oils, rich gases to dry gases in the deeper closures. These deep
gas reservoirs contain very sour gases. For all practical
(subsurface modeling) purposes the gas from the deeper
reservoirs can be considered to contain CH4, H2S and CO2 in
the ratio 60/30/10. In this paper this mixture will be referred
to as “sour gas”. CO2 is often an important ingredient of
injection mixtures where a certain level of miscibility between
2 SPE 88717
injectant and reservoir oil is desirable. For the purpose of
achieving miscible conditions, H2S is a more suitable
component than CO2. The toxic aspects though have
prohibited large-scale use of H2S in oil recovery processes so
far [2-5]. However, this source of sour gas is still 60 %
methane, which has a substantial value. An alternative for
straight injection of sour gas is applying a sweetening process,
where the CH4 is separated out, and the remaining waste
product, which is then roughly a 75/25 mixture of H2S and
CO2, is injected. This latter mixture, which will be referred to
as “acid gas”, has excellent miscibility characteristics.
The reservoir oil was characterized by tuning the
properties of pseudo components in a phase equilibrium
package. An accurate match with experimental laboratory test
data provides confidence that the thermodynamic description
of the reservoir oil is sufficient to the extent that it can be
used to explore interactions with potential gaseous injectants.
Because of the reputation of CO2 as a suitable ingredient of
injection gas mixtures and the possibility of a future CO2
source, injection of pure carbon dioxide was also included in
this evaluation. The difference in the way CO2 and acid gas
interact with the reservoir oil is illustrated in Figure 1. This
figure depicts the saturation pressure vs. the concentration of
the injection gas in the mixture. In case of CO2 injection the
bubble point pressure increases gradually when CO2 is added
to the oil before it reaches a maximum (it is a dew point by
then). On the other hand, with acid gas as the injectant the
bubble point decreases monotonically with increasing
injectant concentration. For the relevant reservoir pressures
and temperatures the mixing of acid gas and reservoir oil will
always result in a single phase fluid, i.e. acid gas is first
contact miscible with the oil. Injection of CO2 will result in a
miscible condition only after continued mass exchange
between the two phases have caused a certain convergence of
the respective compositions. That process is called developed
or dynamic miscibility.
Injection of lean gas (>94 % CH4) is currently taking place
in dedicated areas of the subject reservoirs. This has the
advantage of increasing the total system mobility, but mass
exchange between the injection gas and the native oil is
minor. The global properties of the four subject injectants
with respect to interaction with the reservoir oil are
summarized:
Miscibility ρoil - ρgas µoil / µgas
(lb/cuft)
Lean gas None 27 6.0
Sour gas Some swelling 20 4.4
CO2 Dynamic ~ zero 2.6
Acid gas First-contact -5 1.4
Not surprising, this indicates that the richer the injectant
the more easy to achieve an efficient microscopic
displacement. Therefore, on the basis of injected pore
volumes, the recovery factor might be expected to vary
accordingly. However, there is also the factor of time (or well
spacing) that is required for a certain recovery to materialize,
which would favor injection of a low viscosity gas.
Furthermore, excellent mixing results observed in a PVT cell
will only turn into real value if the reservoir heterogeneity
allows for mixing to take place at a sufficient scale. For this
reason a numerical model with a great deal of geological
detail is required to realistically assess the relative merits of
the various injection processes.
One of the common problems with gas injection is
asphaltene precipitation. This problem has been reported
when injecting CO2 [e.g. 6-7, and references therein]. A
dedicated laboratory study was performed using both CO2 and
acid gas as injectants. It was found that addition of CO2
renders asphaltene in the crude oil unstable. Both experiments
and modeling work indicated that asphaltene onset pressure
decreases upon mixing with the acid gas while it increases
upon mixing with CO2.
Reservoir Model
The basis of all subsequent simulation work is a detailed
reservoir geological model. This static model has a vertical
resolution of one foot, and areal block dimensions measure
some 100 m. The strongly layered reservoirs do allow for an
adequate representation in a model with this resolution. All
log and core data were honored and stochastic methods were
applied in populating the space in between wells. Limited
vertical upscaling was applied in the construction of the
subsequent dynamic model. Only blocks with the same
properties were merged, in that way no geological detail was
lost. Areally, the dynamic model grid is denser than that of
the static model. The PVT description in the simulator is an
eleven component Equation-of-State model. The proximity of
the free water and the high capillary rise in the low permeable
bottom layers imply that a large portion of the reservoir
segment is in the transition zone (see Figure 2). Therefore, it
was critical to ensure that proper modeling of the transition
zone that takes into account the impact of initial oil saturation
on saturation functions was implemented [8].
The reservoirs are traditionally developed by inverted five
spot patterns and as there are no obvious alternatives to this
approach, the dynamic models were squares cut out of the
static model. Injector producer spacing is just over one km,
and in order to appreciate the time factor in comparing the
recovery processes the simulations were constrained by a
fixed pressure difference between injector and producer.
Effectively a pressure difference of 1500 psi is imposed, in
line with current practice, where injection pressures are
limited by rock competence and at the producing end the
pressures are not allowed to drop (too) much below bubble
point.
The base dynamic model is therefore a square measuring
750 * 750 m, with injector and producer at diagonal corners
(quarter wells). The geological model of the reservoir is fairly
layer cake and the layer properties vary only gently over well
distances. However, a few wells encountered thin high
permeable zones. These high permeable streaks cannot be
correlated, but are expected to have a lateral extent of
hundreds of meters. It is unlikely that they provide a complete
short-cut between injector and producer, but where these high
permeable events occur, they impact the flow profile and the
(vertical) sweep efficiency. Therefore, as a sensitivity an
alternative simulation model was constructed that contained
such a high permeable streak.
SPE 88717
The simulations primarily address lean gas, sour gas and
acid gas injection. The first process is currently practiced in
selected areas and the latter two are potential resources from
the deeper sour reservoirs. For completeness forecasts with
CO2 and water injection were also carried out.
Modeling results
Injection of lean gas results in a plateau off-take rate of about
2.5 % of STOIIP per annum. This rate can be sustained for
some ten years, but after gas break through, the GOR
increases rapidly (Figure 3).
Sour gas injection yields qualitatively a similar production
profile. The latter displacement process is somewhat more
stable resulting in a 25 % longer plateau, but after gas
breakthrough the oil rate declines in a similar way, if
anything, the decline is even a little faster.
The production profile that results from acid gas injection
shows a different character. The more viscous injectant causes
a lower plateau rate, but lasts longer (60 % more than with
lean gas) and importantly, the GOR increases much more
gradually after breakthrough, which shows as a less steep
reduction in oil rate. Figure 4 compares these predictions on
the basis of injected pore volumes. The acid gas injection
(first contact miscible) process results in a very high recovery
factor in the areally fairly homogeneous reservoirs.
CO2 has properties that fall in between of those of sour
and acid gas, and the production response is similarly in
between. Although not first contact miscible, in case of CO2
injection miscibility develops relatively quickly and the
response is closer to the acid gas case than to that of sour gas.
Simulations based on a static model that contains a high
permeable streak (see Figure 2) resulted in a higher (about
twice as high) plateau rate, but the length of the plateau period
is more than halved, resulting in a poorer ultimate recovery.
Figure 5 illustrates the differences in the calculated oil rates.
Figure 6 shows the recovery on the basis of subsurface
injection volumes. Interestingly, the difference in
dimensionless recovery between the respective models is of
the same order as the difference between the farthest apart
recovery processes (lean gas vs. acid gas injection) for the
same reservoir model. This illustrates that in this particular
case the reservoir geological aspects are as important as the
choice of injectant.
Finally, a simulation with water injection was carried out.
Results are shown in Figure 7. It will take 100 years to
achieve a Recovery Factor that can be obtained with gas
injection in 25 years. Moreover, the ultimate waterflood
recovery will be limited because a finite fraction of the oil
cannot be moved by water, whereas miscible displacements
do not suffer from such a limitation.
Discussion
The observation that the recovery factors achieved with
alternative gas injection processes correlate strongly with the
degree of miscibility between injectant and reservoir oil when
compared on the basis of injected pore volumes, is fully in
line with expectation. This study shows that this conclusion
also holds if the comparison is made on a time basis (for the
particular reservoir conditions, under same pressure
constraints and well spacing). The plateau rate correlates
3
inversely with the viscosity of the injectant, and therefore
would favor the leanest injectant, but that advantage is short-
term and small (difference in initial rate between the lean gas
and acid gas injection cases is some ten percent). Moreover,
and probably more important is that if the Recovery Factors
are compared using the realistic basis of a limiting GOR cut-
off, injection of acid gas is the superior process.
Miscibility primarily influences the microscopic sweep
efficiency, but frequently of equal or even more importance is
the macroscopic sweep. This study showed the potentially
large impact high permeable streaks, known to occur in these
reservoirs, can have. The effect is that a relatively
homogeneous sector of the reservoir could yield an amount of
oil that is twice that is achievable in an area where high
permeable layers cause low vertical sweep efficiencies.
The choice of injection process is eventually subject to
economic screening, which is not the subject of this paper. It
is noted that switching from lean gas to sour or acid gas, will
result in progressively more oil for a certain amount of
injected reservoir volumes, but the operations become
similarly more complex and significantly more costly.
However, on the proceeds side there is not just the extra oil
recovery. In case of sour gas injection, the lean gas that is
currently injected will be available for sale. Injection of acid
gas has a further merit in that it generates vast quantities of
methane as the gas resource contains 60 % CH4. Simulations
also show that preceding acid gas injection by immiscible
sour gas (one of the options under study) had only a marginal
negative effect on ultimate recovery compared to continuous
acid gas injection.
Well spacing and therefore off-take rates are also subject
to optimization. In this study well spacing has not been
addressed, but viscous forces dominate the displacement
process in this low permeable environment. Therefore the
outcome is basically directly proportional to the well spacing.
The base model is moderately homogenous, which results
in a fairly stable displacement and therefore (very) high
recovery factors. In case of a more heterogeneous reservoir,
the process will be progressively less stable. If that results in
truly premature gas breakthrough, a Water-Alternating-Gas
injection scheme where the water is used to stabilize the front
might become attractive. This opportunity has not been
addressed. Slug size and composition of the injectant are the
prime parameters that have to be optimized in a WAG
scheme. However, a classical miscible WAG scheme has as
disadvantage that it requires continuous disposal of the acid
gas. Moreover, different from most WAG operations, the cost
of the injection gas is low.
The PVT properties that are assumed in this work are the
results of calculations made with phase behavior programs.
Confidence in this approach is due to the fact that a number of
experimental PVT studies are available, and the 11-
component EOS model was calibrated on slim tube tests with
lean gas injection and swelling tests of H2S and CO2 with live
crude.
The processes that are discussed in this paper are of severe
complexity (surface probably even more so than subsurface)
and contain many novel aspects. Indisputably, large scale
application will have to be preceded by an extensive further
study effort, which will have to include field testing.
4 SPE 88717
Conclusions
Reservoir simulation techniques were used to evaluate the
relative merits of various gas injection processes that are
considered for application at the less permeable flanks of the
subject reservoirs. To this end high-resolution element models
were constructed. The PVT module is of the Equation-of-
State type and hence geological detail was combined with a
precise fluid description. The main observations and
conclusions are as follows.
• Injection of lean gas, sour gas, CO2, acid gas and
combinations thereof were found to result in similar
sustainable off-take rates.
• Therefore the ranking of the various gas injection
processes in terms of oil recovery is the same whether
compared on the basis of elapsed time or injected
(reservoir) volumes.
• The sweep efficiency improves with increased degree of
miscibility. This shows primarily in longer plateau
periods, as the sustainable rates are similar.
• In gas injection schemes it is usually the breakthrough of
the injection gas that terminates the process. Effectively,
when applying a realistic GOR limit, injection of acid gas
could easily recover twice as much oil as is attainable
with lean gas. The recoveries that may be expected from
sour gas and CO2 fall in between.
• It was found that the presence of local but lateral
extensive (hundreds of meters) high permeable streaks
would have a pronounced impact on the recovery. Such
phenomena make higher well rates possible but reduce
the ultimate recovery. Simulations indicate that the level
of heterogeneity has as much of an impact on the
recovery efficiency as the choice of injectant.
Acknowledgement
The authors would like to thank Abu Dhabi National Oil
Company and Shell Abu Dhabi management for permission
to publish this paper.
References
1- Kamath, A.R., van Dorp, J., Johnson, A., Alcock, A. and Keith,
I.: “Hazard Management Considerations for Sour Gas
Development During Concept Selection”, SPE 88724,
paper presented at the 11th ADIPEC Annual Technical
Conference and Exhibition, Abu Dhabi, Oct. 10-13, 2004.
2- Davison, R., Mayder, A., Hladiuk, D. and Jarrell, J.L.: “Zama
Acid Gas Disposal/Miscible Flood Implementation and
Results”, paper CIM 96-96.
3- Jamaluddin, A.K.M, Bennion, D.B, Thomas, F.B. and Clark
M.A.: “Acid/Sour Gas Management in the Petroleum Industry”,
SPE 49522, paper presented at the 8th ADIPEC Annual
Technical Conference and Exhibition, Abu Dhabi, Oct. 11-14,
1998.
4- Miwa, M. Shiozawa, Y. Saito, Y. and Taroom, I.O.: “Sour Gas
Injection Project”, SPE 78547, paper presented at the 10th
ADIPEC Annual Technical Conference and Exhibition, Abu
Dhabi, Oct. 13-16, 2002.
5- Thibeau, S. Barker, J.W. and Morel, D.: ”Simulation of Sour
Gas Injection in Low Permeability Oil Reservoirs”, SPE 84362,
paper presented at the SPE Annual Technical Conference and
Exhibition, Denver, Oct 5-8, 2003.
6- Leontaritis, K.J. and Mansoori, G.A.: ”Asphaltene Deposition:
A Survey of Field Experiences and Research Approaches”, J.
Pet. Sci. Eng. 1 (1988) 229.
7- Sarma, H.K.: ”Can We Ignore Asphaltene in a Gas Injection
Project for Light-Oils?”, SPE 84877, paper presented at the SPE
International Improved Oil Recovery Conference in Asia
Pacific, Kuala Lumpur, Oct. 20-21, 2003.
8- Masalmeh, S.K.: “High Oil Recoveries from Transition Zones”,
SPE 87291, paper presented at the 9th ADIPEC Annual
Technical Conference and Exhibition, Abu Dhabi, Oct. 15-18,
2000.
SPE 88717 5

5000 0

4500
20
Saturation Pressure (psia)

4000
CO2
3500 40

Depth (ft)
3000
60
2500 H2S/CO2 (3:1)
2000
80

1500
100
1000
120
500

0 140
0 10 20 30 40 50 60 70 80 90 100 0 1 10 100 1000
Concentration Injectant (mol %)
Permeability (mD)

Fig. 1. Saturation pressure of mixtures of reservoir Fig. 2. Typical permeability profiles at positions with
oil with carbondioxide and acid gas. and without high permeable streak.

0.04 20
1.0

0.9
Heavy lines : Acid Gas Injection
Intermediate : Sour Gas Injection 0.8
0.03 15
Oil rate (fr. STOIIP/y)

Thin lines : Lean Gas Injection

Fraction of STOIIP
GOR (MMscf/stb)

GOR 0.7

0.6
0.02 10 0.5
Lean Gas
0.4
Oil Sour Gas
0.3
0.01 5 Acid Gas
0.2

0.1

0.00 0 0.0
0 5 10 15 20 25 30 35 40 45 50 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
Year Injected Volume (Hydrocarbon Pore Volumes)

Fig. 3. Acid, sour & lean gas Injection - Oil rate Fig. 4. Cumulative oil vs. cumulative injected
and GOR vs. time. reservoir volume.
6 SPE 88717

0.09 18 0.7

0.08 16
0.6 Acid
0.07 14
Oil rate (fr STOIIP/y)

Heavy lines : Acid Gas Injection

Fraction of STOIIP
GOR (MMscf/stb)
Thin Lines : Lean Gas Injection 0.5
0.06 12
Lean Gas
0.05 10 0.4
Oil GOR
0.04 8 0.3

0.03 6
0.2
0.02 4 Heavy lines : With high K layer
Thin lines : Without High K
0.1
0.01 2

0.00 0 0.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5

Year Injected Volume (Hydrocarbon Pore Volumes)

Fig. 5. Acid & lean gas injection - Pattern includes Fig. 6. Impact of high permeable layer- Cumulative oil
high permeable layer. produced vs. cumulative injected

1.0

0.9

0.8

0.7
Fraction STOIIP

0.6

0.5

0.4
Lean Gas
0.3
Water
0.2 Sour Gas
0.1 Acid Gas
0.0
0 10 20 30 40 50 60 70 80 90 100
Year

Fig. 7. Cumulative oil vs. time for alternative

injectants.