SPE 153307

Analytical Estimation of CO2 Storage Capacity in Depleted Oil and Gas

Reservoirs Based on Thermodynamic State Functions
Ernesto Valbuena, Maria Barrufet, and Gioia Falcone, Texas A&M University

Copyright 2012, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Latin American and Caribbean Petroleum Engineering Conference held in Mexico City, Mexico, 1618 April 2012.
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 necessar ily 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 prohi bited. 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
Numerical simulation has been used, as common practice, to estimate the CO2 storage capacity in depleted reservoirs.
However, this method is time consuming, expensive, and requires detailed input data. This investigation proposes an
analytical method to estimate the ultimate CO2 storage in depleted oil and gas reservoirs by implementing a volumeconstrained thermodynamic equation of state (EOS) given average reservoir pressure and fluid composition.
This method was implemented in an algorithm which allows fast and accurate estimations of final storage, which can be used
to select target storage reservoirs and design the injection scheme and surface facilities. Impurities such as nitrogen and
carbon monoxide, usually contained in power plant flue gases, are considered in the injection stream and can be handled
correctly in the proposed algorithm by using their thermodynamic properties in the EOS.
Results from analytical method presented excellent agreement with those from reservoir simulation. Ultimate CO 2 storage
capacity was predicted with an average difference of 1.26 wt% between analytical and numerical methods; average oil, gas,
and water saturations were also matched. Additionally, the analytical algorithm performed several orders of magnitude faster
than numerical simulation, with an average of 5 seconds per run.
Introduction
Greenhouse gas (GHG) emissions have been continuously increasing in the past 3 decades. More than 70% of these
emissions are composed of CO2, which reached 30 Gt in 2009 (EIA 2011). Several environmental agencies and governments
have shown concern about this statistic and its potential relation with global warming.
Coal consumption accounted for the release of nearly 14 Gt of CO2 during 2009, almost 45% of worldwide carbon dioxide
emissions (EIA 2011). Given that coal is mainly used in power plants to generate electricity, these locations are large sources
of CO2 and become the most important target for carbon capture and sequestration (CCS) processes. A large coal-fired power
plant, generating 500 MW, emits approximately 2.9 Mt-CO2 per year or 55.2 BSCF of CO2 per year.
Metz et al. (2005), Dooley et al. (2006), and EPA (2011) defined carbon capture as a process consisting first of removing the
impurities from a CO2-based stream to increase the CO2 concentration and improve the efficiency of the final storage
process; and secondly, compressing the gas stream to transport it to a storage location, which can be geologic formations
such as aquifers and depleted oil and gas reservoirs, to achieve long-term isolation from the atmosphere.
Geological storage of CO2 in aquifers and depleted oil and gas reservoirs represent an attractive option to reduce carbon
emissions to the atmosphere, as it has been studied in the oil and gas industry for several years. Particularly, interest now
exists in using depleted reservoirs taking advantage of the higher storage density in comparison with aquifers; additionally,
extensive knowledge of the reservoirs static and dynamic properties, acquired during the developing phase, are available to
optimize the efficiency of the project and increase the final storage capacity and profits.
Numerical simulation has been used as common practice to understand and estimate the ultimate CO 2 storage capacity
(UCSC) of aquifers and depleted reservoirs. However, this method is expensive, time consuming, and requires detailed input

SPE 153307

data to develop representative models. Analytical methods provide a much faster approach to estimate UCSC in geological
formations; nonetheless, current analytical methods are mostly focused on applications for storage calculations in aquifers,
which only consider rock compressibility, brine and CO 2 compressibility, and in some cases, CO 2 solubility in brine. These
methods do not represent interactions between hydrocarbon components and injected fluids into depleted oil and gas
reservoirs. In addition, storage efficiency in aquifers is low, below 5%, given that aquifers have large volumes for storage but
are already pressurized (Gallo et al. 2002; Okwen et al. 2010).
This investigation proposes an analytical method to estimate maximum CO 2 storage in depleted oil and gas reservoirs by
implementing a volume-constrained thermodynamic EOS using the reservoirs average pressure and fluid composition. This
method is based on the partial molar volume (PMV) definition, which relates the change in volume of a mixture with the
change of the amount of one component holding constant pressure and temperature. Separation procedures to remove the
impurities from the CO2-based injection stream are expensive, consequently a considerable amount of impurities remain in
the gas composition; hence, the PMV concept in this study has been generalized, to allow the change in the amount of more
than one component.
Required input data for the proposed model are average initial pressure, pore volume, water saturation, reservoir temperature,
depletion and target pressures, average fluids composition at depletion, rock and water compressibilities, produced water
volume, and the volume of hydrocarbon and injection fluid mixture is estimated with the PMV approach. Final outputs are
maximum amount of CO2 injected at target pressure, final average fluid composition, and saturations.
This method allows fast and accurate estimations of final storage that can help to design and schedule injection scheme and
surface facilities. Impurities such as nitrogen and carbon monoxide, usually contained in power plant flue gases, are
considered in the injection stream and will be correctly handled in the proposed algorithm by using their thermodynamic
properties into the EOS, properly modeling the interactions of hydrocarbons in the reservoir and injected fluid.
Key Facts about CCS
CO2 dissolution in water/brine is negligible during injection stage (Sifuentes et al. 2009). Dissolution effects are very slow
and they become noticeable only after a few hundred years of storage. Likewise, mineralization rates in a reservoir volume
base are very slow and current level of understating on long term chemical reactions does not allow to characterize this
phenomenon. Depending on reservoir rock and water properties, mineralization may occur nearwellbore during the CO2
injection process reducing injectivity (Gallo et al. 2002); this potential problem is usually addressed through fluid
compatibility laboratory studies.
Considering previous stated conditions, the analytical algorithm developed in this study neglects interactions between CO 2
and brine given the slow reaction rates. Nonetheless, a simple validation will be presented to demonstrate the fact that the
amount of CO2 dissolved in brine in depleted oil and gas reservoirs subject to CCS processes is very small in comparison
with the quantity of CO2 stored in supercritical phase and mixed with hydrocarbons in place.
Van der Meer and Ferhat (2009) corroborated the importance of injectivity to estimate costs and plan the operations.
Conversely, they mentioned that ultimate storage capacity is independent from the number of wells, spacing or stimulation
techniques used. In addition, Barrufet et al. (2010) demonstrated the independency between CO2 storage amount and
injection rates under different well constraints. These facts suggest that the number of wells, injectivity, and injection rates
affect only the duration of the injection process; nevertheless, the maximum amount of CO 2 stored in the reservoir depends
on the pressure differential achieved, fluid compositions and characteristics, temperature, and saturations.
Impurities removal is the most expensive process in a CCS project. Estimated costs for CO 2 capture and compression range
from USD 12 to USD 60 per ton of CO 2 (Dooley et al. 2006). However, this is an important procedure to maximize the
ultimate storage capacity in the target formation. Often, certain level of impurities is allowed in the injection stream due to
excessively high cost for very pure outlet gas.
Storage capacity and displacement are reduced by the presence of impurities in the CO 2-based injection gas (Nogueira and
Mamora 2005). This observation suggests that there must be a balance between separation cost and desired efficiency of the
storage process. For this reason, understanding the effect of impurities in the UCSC represents a key factor in the design of
any CCS project.
Storage efficiency in aquifers is low, below 5%. Aquifers exhibit large volumes for storage but are already pressurized,
leading to low storage density processes (Gallo et al. 2002; Okwen et al. 2010). CCS projects in depleted oil and gas
reservoirs offer a large pressure differential between depleted and target conditions producing large CO 2 storage density. In
addition, reservoir data are available, often with a good characterization. Wells are already drilled, decreasing capital
expenditure needs and the possibility of developing enhanced oil recovery (EOR) methods is available.

SPE 153307

Description of Solution
The main purpose of this study is to develop an analytical model to estimate the UCSC of depleted oil and gas reservoirs at
certain target reservoir pressures. An essential step to achieve this goal is to demonstrate that the UCSC can be estimated
using average reservoir pressure and fluid compositions at depletion stage, setting the final target pressure and applying
thermodynamic principles. A brief delineation of the life cycle of depleted oil and gas reservoirs, candidates for CCS
processes, will set a base of understanding for the presentation of the solution approach based on thermodynamic state
functions.
Field Operations
Tracking key reservoir and fluid properties during depletion/production and CO 2 injection phases in an oil and gas reservoirs
is crucial to accurately model and predict a successful process. During production, the reservoir is characterized through
seismic, electrical logs, core and fluid samples, well tests, measurements of produced fluids, etc. Once the reservoir cannot be
economically produced, the wells are abandoned and operations finish, it is at this stage that the CCS process may begin. The
average reservoir pressure at the time production ends is defined in this work as depleted or depletion pressure.
Production phase allows the estimation of original fluids in place (oil, gas, and water), reservoir size and temperature,
porosity, rock and water compressibility, initial and depleted reservoir pressure, and produced fluids during depletion.
Laboratory fluid analysis and production measurements can be used to determine average reservoir fluid composition at
depletion.
CO2 injection starts at the average depleted reservoir pressure and as the fluid is injected (without production), the average
pressure increases until reaching a designed target pressure. This target pressure can be limited by the initial reservoir
pressure or by operational constraints. Composition of the CO 2-based injection stream is known in advance; gas composition
from the source, e.g. power plant, is characterized and a separation process can be designed to provide the desired
composition of the fluid. Separation costs highly increase as the purity of the outlet stream increases (concentrated CO 2); the
economic success of the project may be subject to the balance between fluid purity and the ultimate amount of CO 2 that can
be stored. For this reason, performing fast sensitivities to understand the storage capacity for different injection fluid
compositions becomes a strategic activity.
Fig. 1 shows the typical behavior of average reservoir pressure, hydrocarbon and injected fluid moles/mass in place, pore
volume, and water saturation for a reservoir with no water production subject to CO 2 storage after depletion. Initial reservoir
pressure depletes as the reservoir is produced to economical limit at depletion pressure; injection starts and pressure is
restored to a target pressure, smaller than initial reservoir pressure for seal assurance purposes. Moles or mass in place is
reduced as hydrocarbons are produced; when injection starts, it increases as injected fluid mixes with the hydrocarbons. Pore
volume (PV) diminishes during depletion and increases during injection due to pressure drop and rock compressibility. Water
saturation presents the opposite behavior to PV, given that water expands with decreasing pressure and compresses with
increasing pressure. PV and water saturation, and consequently hydrocarbon pore volume (HCPV), the space filled by the
hydrocarbons and CO2-based injected fluid, can be calculated using compressibility principles. The relationship between
pressure and moles/mass in place is performed through thermodynamic principles using and EOS.

Fig. 1Tracking key properties during depletion and injection processes in oil and gas reservoirs subjected to CO2 storage
projects. Average pressure, moles or mass in place, pore volume, and water saturation depend upon each other. Relationships can
be calculated using compressibility and thermodynamic principles.

SPE 153307

Solution Approach
Previous section described typical life-cycle for depleted oil and gas candidate reservoirs for CCS projects. Properties shown
on Fig. 1 can be either designed or calculated using compressibility principles and thermodynamic relationships. The
thermodynamic process followed by the proposed analytical method, displayed on Fig. 2, starts at depletion conditions when
oil and gas production has ceased and injection begins. HCPVdep is the volume occupied by the hydrocarbon fluid, pdep is the
average reservoir pressure, and ndep is the total amount of reservoir fluid expressed in moles. Following a path from depletion
to target conditions involves the change of all three variables (pressure, volume, and fluid moles) at the same time, which is
inconveniently complex to solve.
Subsequently, the concept of State Functions, which states that fluid properties at the end of the process are independent from
the path followed to reach final target conditions, is used to perform the thermodynamic analysis of the reservoir fluid
volume change (with respect to pressure increase and mixing with the injected fluid) dividing it in two paths. The first path
starts at depletion conditions and ends at a fictitious condition by increasing the pressure to ptarget at constant fluid moles; the
pressure increase causes fluid compression to the fictitious volume Vndep. The second path starts at the fictitious condition and
ends at target conditions; this path is performed at constant pressure by adding n moles of fluid until the hydrocarbon and
injected gas mixture volume, Vndep+n , equals HCPVtarget.

Fig. 2Thermodynamic relationships between fluid volume, pressure, and moles during injection. Applying the State Function
allows an easier estimation of added moles by creating a fictitious condition in the path between depletion and target states.

Oil, gas, and water average saturations at depletion can be estimated through compressibility analysis and performing vaporliquid equilibrium (VLE) calculations using an EOS, e.g. Peng-Robinson (PR). Hydrocarbon fluid volume at the fictitious
condition is calculated also performing a VLE compressing the fluid from depletion pressure pdep to the designed target
pressure ptarget. Finally, the amount of stored CO2-based fluid n (the unknown), is estimated by solving Eq. 1, derived from
fundamental thermodynamic relationships (Valbuena 2011):

................................................................................................................................ (1)

where Vndep+n is the volume occupied by the hydrocarbon and injected gas mixture at the end of the injection process and it
coincides with the HCPVtarget, Vndep is the fictional volume representing the volume that would be occupied by ndep moles of
hydrocarbons at target pressure and temperature, ndep is the number of moles of the reservoir fluid before injection, and n is
the cumulative number of injected moles at the end of the process.
The argument of the integral, , is the change of fluid volume (hydrocarbon and injected gas mixture) with respect to the
amount of injected gas, a generalized expression based on the partial molar volume (PMV) definition. The PMV represents
the change of volume of a fluid mixture with respect to the change of one of its components (Michelsen and Mollerup 2004);
given the multi-compositional nature of the CO2-based injection stream, a generalization of the PMV was developed for this
study, see Appendix A.
Solution of Eq. 1 provides the amount of injected CO 2-based fluid, therefore final average reservoir fluid composition can be
estimated. Oil, gas, and water average saturations at target conditions (end of injection) are computed performing VLE
calculations.

SPE 153307

Analytical Method
The rigorous analytical method was implemented in an algorithm to estimate the UCSC and fluid saturations in depleted oil
and gas reservoirs under CO2 storage processes. Required input data are reservoir temperature, average initial reservoir
pressure, depletion reservoir pressure at the end of the production stage, average reservoir fluid composition at the end of the
production stage, initial pore volume, initial water saturation, produced water volume, and rock and water compressibility.
Controls of the process are: target pressure at the end of the injection, desired EOS, and the composition of the injection gas.
Solution is completed in 4 steps, as presented in Fig. 3. First step is to calculate PV, water saturation, and HCPV at depletion
and target conditions using the isothermal compressibility principle (Dake 1978). Recall that, for previous presented reasons,
water is assumed to be inert during production and injection, meaning that there is no mutual solubility between water and
other fluids, and no chemical reactions take place. Second step involves estimating fluid (hydrocarbon) properties at
depletion; molar volume, vapor fraction, among other properties are calculated through VLE, allowing to estimate oil and gas
saturations and number of moles of hydrocarbon in place ndep. Third step is to compute the fictional volume Vndep occupied by
ndep moles of hydrocarbons at target pressure and temperature by using VLE calculations.

Fig. 3Analytical method flow chart to estimate CO2 storage capacity. This method is based on the generalization of the partial
molar volume definition considering a mole fraction change in more than one component (Valbuena 2011).

Finally, the fourth step comprises the rigorous solution of the integral of the PMV presented in Eq. 1, evaluated at final
average reservoir pressure and temperature. All the terms in this expression are known, except n which is the objective of
this study. Given the complexity of the generalized PMV expression, the integral in Eq. 1 is solved numerically. Once the
amount of injected fluid is known, final average fluid composition can be estimated; oil, gas and water saturations at target
conditions (end of injection) are computed using VLE calculations.
Validation
Validation of the analytical method results was performed through comparison with reservoir numerical simulation. A total
of 24 reservoir models were created representing the wide variety of scenarios varying reservoir pressures, temperature,
initial water saturation, cases with and without water production, presence of single- and two-phase fluids at depletion and
target conditions, and reservoir and injected fluid compositions. Fluid models were constructed by calibration of the PR-EOS
with experimental PVT data.
Numerical Simulation Models
Reservoir static properties are representative of the Cupiagua field in Colombia. The selected section of the stratigraphic
column is formed by layered sandstone units, which display sheet-like packages of shoreface bodies (Ramon and Fajardo
2006). Reservoir model areal extension and thickness are 1,740 acres and 270 ft respectively. Porosity ranges from 1.54% to
18.22%, with an average value of 13%, and 4.2% standard deviation. Horizontal permeability in the reservoir ranges from
0.21 to 85.4 mD, with an average value of 28.5 mD, and 28.7 mD of standard deviation. Dykstra-Parsons coefficient for this
model is 0.79, indicating high heterogeneity. Permeability anisotropy is given by vertical to horizontal ratio of 0.1.

SPE 153307

Sensitivities to reservoir temperatures were performed between 100 and 285 F. Initial pressures were calculated for each
case to maintain thermodynamic equilibrium, ranging between 5,600 and 9,000 psia. Depletion and target pressures were
designed to evaluate cases with small and large pressure differential, ranging from 500 to 6,000 psia between depletion and
target conditions. Three compositional fluid descriptions were also used to validate the analytical method for different fluid
types: black oil, volatile oil, and gas condensate. Initial reservoir water saturation sensitivity was performed by changing the
water oil contact depth, generating cases with and without water production. Six injection gas compositions were used with
mixtures CO2, nitrogen (N2), and carbon monoxide (CO); CO2 molar composition ranged between 100% and 60%, N 2 from
0% to 40%, and CO from 0 to 10%. Table 1 presents the 24 cases developed for comparison and validation.
Table 1SENSITIVITY CASES. 24 SCENARIOS COVER DIFFERENT RESERVOIR AND INJECTED FLUIDS, TEMPERATURES,
SINGLE- AND TWO-PHASE CONDITIONS, AND WATER PRODUCTION

Case
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

Res.
Fluid
V-O
V-O
V-O
V-O
V-O
V-O
V-O
V-O
G-C
G-C
G-C
G-C
G-C
G-C
G-C
G-C
B-O
B-O
B-O
B-O
B-O
B-O
B-O
B-O

Inj.
Fluid
1
2
3
4
5
6
1
2
3
4
5
6
1
2
3
4
1
2
3
4
5
6
1
2

Sensitivity Cases
Inj. Fluid
T,
Depletion
Comp.
F Condition**
CO2
100
S
CO2-N2
100
T
CO2-N2
150
T
CO2-N2
150
S
CO2-N2-CO 200
T
CO2-N2-CO 200
T
CO2
250
S
CO2-N2
250
T
CO2-N2
270
T
CO2-N2
270
S
CO2-N2-CO 270
T
CO2-N2-CO 285
T
CO2
285
S
CO2-N2
285
T
CO2-N2
285
T
CO2-N2
285
S
CO2
250
S
CO2-N2
200
T
CO2-N2
200
T
CO2-N2
250
S
CO2-N2-CO 250
T
CO2-N2-CO 250
T
CO2
285
S
CO2-N2
285
T

Target
Condition**
S
S
S
S
S
T
S
S
T
S
S
T
S
S
T
S
S
S
T
S
S
T
S
S

Water
Prod.
No
No
No
No
Yes
No
No
Yes
No
No
Yes
No
No
Yes
No
Yes
No
Yes
Yes
No
No
Yes
No
Yes

** (S) Single-phase hydrocarbon fluid. (T) Two-phase hydrocarbon fluid.

Discussion of Results
Case 18, from Table 1, was selected as a complex representative case to analyze in this section. Overall analysis of results
from the 24 cases is summarized later.
Case 18 is represented by a black oil reservoir fluid at 200 F temperature. Injected fluid was 90% CO2 and 10% N2. Initial
reservoir pressure was 6,094 psia. The reservoir was produced until average pressure reached 978 psia and injection was
performed to repressurize the reservoir to 6,038 psia average pressure. Water production during depletion stage was 80.9
MMSTB. Cumulative oil production during depletion was 8 MMSTB, representing 7.8% recovery factor. Low recovery
factor is caused by high initial water saturation and water production during depletion, nearly 81 MMSTB of water. Reservoir
fluid falls in two-phase during depletion and after injection the fluid returns to single-phase.
The CO2-based stream was injected to repressurize the reservoir from 978 psia to 6,038 psia average pressure, a 5,059
pressure differential, ensuring not to exceed initial reservoir pressure. Final amount of injected fluid calculated from reservoir
numerical simulation was 809.96 MMlb-mol of fluid, corresponding to 728.96 MMlb-mol of CO2 and 81.00 MMlb-mol of
N2. Fig. 4 illustrates the CO2 molar density (moles per reservoir volume) distribution at the end of the injection stage. The
CO2 reaches great extension into the reservoir. High vertical and horizontal compositional gradients are imbedded in the
analytical calculations by using average reservoir pressures and compositions.

SPE 153307

MLSC1

0.002

1.133

2.265

4.530

3.400

Fig. 4Case 18CO2 molar density distribution at the end of injection, from numerical simulation model. The CO2 reaches great
extent of the reservoir with heterogeneous plume growth.

The proposed analytical model predicted a total amount of injected fluid of 808.91 MMlb-mol, corresponding to 728.02
MMlb-mol of CO2 and 80.89 MMlb-mol of N2, given the injection fluid composition. For this case, the amount of CO2
injected in the reservoir is equivalent to 14.51 Mt-CO2. This amount corresponds to 13 years of CO2 emissions from a 200
MW coal-fired power plant.
Table 2 shows the comparison of results between numerical simulation and analytical models for Case 18. The difference
between the analytical and numerical simulation results is 0.13%. Oil, gas, and water average saturations from the analytical
model match with those observed in numerical simulation. In addition, the analytical model performs 291 times faster than
numerical simulation.
Table 2CASE 18COMPARISON OF RESULTS FROM NUMERICAL SIMULATION AND ANALYTICAL MODELS. ANALYTICAL
MODEL RESULTS MA TCH THOSE FROM NUMERICAL SIMULATION AND PERFORMS NEARLY 291 TIMES FASTER
Case 18Comparison of Results
Numerical
Property
Analytical
Simulation
Depletion PV, ft3x109
3.00
3.00
Final PV, ft3x109
3.07
3.07

Difference,
%
0.00
0.00

Molesdep, lb-mol x 109

Mol Volumedep, ft3/lb-mol
So,dep, %
Sg,dep, %
Sw,dep, %

0.3782
3.4974
22.22
22.05
55.72

0.3778
3.5005
22.30
21.71
55.99

0.08
0.09
0.08
0.34
0.26

Molesinjected, lb-mol x 109

Mol Volumefinal, ft3/lb-mol
So,final, %
Sg,final, %
Sw,final, %

0.8099
1.1862
45.10
1.13
53.77

0.8089
1.1875
45.96
0.0
54.04

0.13
0.11
0.86
1.13
0.27

CPU time, sec

872

291x

Results from the analytical model matched those from reservoir simulation in the 24 comparison cases with small differences,
1.26 wt% in average and a maximum of 3.6 wt%, Fig. 5a. Average oil, gas, and water saturations were also matched with a
maximum difference of 2.4% at target conditions. In addition, the analytical model performed several orders of magnitude
faster than numerical simulation in all cases, with an average CPU time of 5 seconds versus 1600 seconds from numerical
simulation, see Fig. 5b. CPU time in the numerical method is highly dependent on reservoir size/number of cell in the model,
while the analytical method CPU time is independent from those parameters.

SPE 153307

a) CO2-Stream Stored Numerical and Analytical

Models

30

b) CPU Time in Numerical and Analytical Models

10000

Numerical Simulation

1.26 wt%
Avg. Difference

20

1000

CPU time, s

Simulation n, Mt

25

15

10

100

10

5
0
0

10

15
20
Analytical n, Mt

25

Analytical Model

30

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Case

Fig. 5Performance of proposed analytical method against compositional numerical simulation. a) Analytical model results match
reservoir simulation within 1.26 wt% average difference. b) Analytical model performs orders of magnitude faster than numerical
simulation in all 24 cases.

Cases with large pressure differential, between depletion and target conditions, achieved UCSCs between 12 and 30 Mt-CO2
(see Fig. 5a) for this relatively small reservoir, allowing to store between 10 and 25 years of CO2 from a 200 MW coal-fired
power plant.
CO2 Dissolution in Brine
Previous studies have suggested that during injection stage, CO 2 dissolution in water is negligible (Sifuentes et al. 2009).
This dissolution effect becomes noticeable only after a few hundred years of storage. Additionally, the rate of CO 2
dissolution in brine is usually appraised, from numerical simulation, using diffusion coefficients, which have to be estimated
from laboratory experiments and have high uncertainty.
An additional analysis on previously presented Case 18 was performed to estimate the amount of CO2 that would dissolved in
the brine assuming instantaneous dissolution, implementing an analytical model (Chang et al. 1998; Rowe and Chou 1970).
This analytical model uses pressure, temperature, and brine salinity as inputs.
CO2 distribution in the reservoir at the end of injection is presented in Fig. 6. Active cells show reservoir zones with CO2
composition higher than 8%, where the injected fluid is in contact with brine. Average water saturation in the zones where
CO2 is present is 27.26% and total volume of water in these zones is 31.99 MMSTB; total water volume in the CO2 zone was
calculated adding the volume of water contained in each cell with CO 2 composition equal or greater than 8%.

ZMF1

0.080

0.285

0.490

0.695

0.9

Fig. 6CO2 composition distribution at the end of injection, from numerical simulation. Active cells show zones where CO 2
composition is larger than 8%. The total water volume in this zone will be used to estimate maximum CO 2 solubility in brine for this
scenario.

Table 3 displays the maximum amount of CO2 that can instantaneously dissolve in brine with different salinities and the
percentage of the total CO2 injected that this represents; recall that total CO2 injected in Case 18 was 736.41 MMlb-mol.
Assuming completely fresh water yields a maximum of 2.25% of dissolved CO 2, which is equivalent to only 3 months of
CO2 emissions from a 200 MW coal-fired power plant. More realistic salinities yield further lower quantities of CO2
dissolution.

SPE 153307

Table 3CO2 DISSOLVED IN BRINE. PERCENTAGE OF DISSOLVED CO2 IN BRINE IS NEGLIGIBLE.

Salinity EffectCO2 Dissolved in Brine

CO2 dissolved, Percentage of total
MMlb-mol
CO2 injected, %
0
16.59
2.25
5,000
13.99
1.90
10,000
11.79
1.60
25,000
7.07
0.96

Salinity, ppm

The quantity of dissolved CO2 in brine is negligible when compared to the amount stored in supercritical phase mixed with
the hydrocarbons. Additionally, CO2 dissolution in brine is not instantaneous and takes long time to occur. CO 2 dissolution in
brine does not play an important role in the UCSC of depleted oil and gas reservoirs during injection
Conclusions
The proposed analytical method allows fast and accurate estimations of ultimate CO2 storage capacity in depleted oil
and gas reservoirs. Difference between analytical and numerical simulation methods was in average 1.26 wt%, with
a maximum of 3.6 wt% in all studied cases.

The implementation of analytical method in an algorithm performed with an average CPU time of 5 seconds versus
1600 seconds from numerical simulation.

Average oil, gas, and water saturations at depletion and target conditions were accurately estimated.

Generalized partial molar volume allowed to complete the calculations required in cases with impure CO2-based
injection streams.

CO2 dissolution in brine is negligible during injection. Less than 1% of injected CO2 dissolves in brine.

Nomenclature
UCSC = ultimate CO2 storage capacity
VLE = vapor/liquid equilibrium
HCPV = hydrocarbon pore volume from rock-water compressibility, ft3 or bbl
( ) = parameter of EOS for mixtures (Quadratic Mixing Rule)
= parameter of EOS for mixtures (Linear Mixing Rule)
= mole fraction of component i in CO2-based injection stream
ndep = number of moles at depletion, lb-mol
n = cumulative injected moles, lb-mol
p = pressure, psia
PV = pore volume, ft3 or bbl
R = gas constant, 10.7316 ft3-psi-R-1/lb-mol
Sg = gas saturation
So = oil saturation
Sw = water saturation
T = temperature, F ([F] = 32 + [C] x 9/5)
= molar volume, ft3/lb-mol
= generalized partial molar volume of pseudo component x; where x could be a mixture of components with
constant composition, e.g. CO2-based injection stream (see Appendix A), ft3/lb-mol
= fictitious volume occupied by
moles of fluid at target p & T, ft3
= final volume of fluid at p and T from EOS, ft3
acres = area (1 acre = 43,560 ft2 = 4,046.86 m2)
ft = feet, length (1 ft = 0.3048 m)
MMlb-mol = quantity of substance (1 MMlb-mol = 106 lb-mol)
mD = milliDarcy, permeability (1000 mD = 1 D = 9.86923310 13 m)
MW = Megawatt, energy rate (1 MW = 106 W = 106 J/s = 106 kg-m2/s3)
ppm = parts per million
SCF = standard cubic feet, volume (5.615 SCF = 1 STB = 158.99 Litres)
BSCF = billion standard cubic feet, volume (1 BSCF = 10 9 SCF)
t or ton = tonne, mass (1 ton = 103 kg)
Mt = Megatonne, mass (1 Mt = 106 ton = 109 kg)
Gt = Gigatonne, mass (1 Gt = 109 ton = 1012 kg)

10

SPE 153307

Appendix A
This appendix presents the generalized partial molar volume (PMV) equation derived in this study. The PMV is defined as
the change of total volume of a mixture with respect to the change of moles of one component at constant pressure and
temperature; its original derivation can be found in the reference Michelsen and Mollerup (2004). Since the CO2-based
injection stream is multi-compositional, this study requires a generalization of the PMV concept for more than one
component changing in the mixture (Valbuena 2011).
Using the Peng-Robinson equation of state as example, one can derive pressure with respect to the change of moles and with
respect to volume. These derivations are somewhat tedious, however their manipulation yields the generalized PMV
expression for :
(

)
........................................................................................................ (A.1)

)
(

where
and
are the repulsion and attraction terms derivatives with respect to the change in moles of injected mixture
and
is the partial derivative of with respect to , expressed as follows:

(
)

(
(

..................................................................................................................................... (A.2)

(
(

) (
( )

)
(

...................................................................................... (A.3)

................................................................................................................. (A.4)

for:
.......................................................................................................................................... (A.5)
and:
[

)]

] ............................................................................................................... (A.6)

The variables in Eqs. A.2 to A.5 come from PR-EOS. is the molar volume of the fluid, R is the gas constant, parameters
and depend on the critical properties of the fluid (subscript m represents the fluid mixture, i represents the icomponent, and j the j-component), fi is the fraction of component i in the injection stream, and x is the liquid fraction if
the mixture is in liquid phase and can be substituted if the mixture is in single-vapor phase or even two-phase.

References
Barrufet, M.A., Bacquet, A., and Falcone, G. 2010. Analysis of the Storage Capacity for CO2 Sequestration of a Depleted
Gas Condensate Reservoir and a Saline Aquifer. Journal of Canadian Petroleum Technology 49 (8). DOI:
10.2118/139771-pa
Chang, Y.-B., Coats, B.K., and Nolen, J.S. 1998. A Compositional Model for CO2 Floods Including CO2 Solubility in Water.
SPE Reservoir Evaluation & Engineering 1 (2). DOI: 10.2118/35164-pa
Dake, L.P. 1978. Fundamentals of Reservoir Engineering. Amsterdam, The Netherlands: Elsevier. Original edition. ISBN
978-0-444-41830-2.
Dooley, J., Dahowski, R., Davidson, C. et al. 2006. Carbon Dioxide Capture and Geologic Storage. A Core Element of a
Global Energy Technology Strategy to Address Climate Change. Columbus, Ohio, United States.
EIA, U.S., 2011. Total Carbon Dioxide Emissions from the Consumption of Energy. U.S. Energy Information
Administration. http://tonto.eia.doe.gov/cfapps/ipdbproject/IEDIndex3.cfm?tid=90&pid=44&aid=8. Accessed
January 18, 2011.

SPE 153307

11

EPA, U.S., 2011. Geologic Sequestration. U.S. Environmental Protection Agency.

http://www.epa.gov/climatechange/emissions/co2_geosequest.html. Accessed March 14, 2011.
Gallo, Y.L., Couillens, P., and Manai, T. 2002. CO2 Sequestration in Depleted Oil or Gas Reservoirs. SPE International
Conference on Health, Safety and Environment in Oil and Gas Exploration and Production, Kuala Lumpur,
Malaysia. SPE 74104-MS.
Metz, B., Davidson, O., Coninck, H.D. et al. 2005. IPCC Special Report on Carbon Dioxide Capture and Storage. New
York, NY, USA: Cambridge University Press. Original publication. ISBN 13 978-0-521-68551-1
Michelsen, M. and Mollerup, J. 2004. Thermodynamic Models: Fundamentals & Computational Aspects. Ed Stenby, E.
Denmark: Tie-Line Publications. Original edition. ISBN 87-989961-1-8.
Nogueira, M. and Mamora, D.D. 2005. Effect of Flue Gas Impurities on the Process of Injection and Storage of CO2 in
Depleted Gas Reservoirs. SPE/EPA/DOE Exploration and Production Environmental Conference. SPE 94906-STU.
Okwen, R.T., Stewart, M.T., and Cunningham, J.A. 2010. Analytical Solution for Estimating Storage Efficiency of Geologic
Sequestration of CO2. International Journal Of Greenhouse Gas Control 4 (1): 102-107.
Ramon, J.C. and Fajardo, A. 2006. Sedimentology, Sequence Stratigraphy, and Reservoir Architecture of the Eocene Mirador
Formation, Cupiagua Field, Llanos Foothills, Colombia. AAPG Memoir (88/SEPM Special Publication): p. 433469. DOI: 10.1306/1215884M883276
Rowe, A.M. and Chou, J.C.S. 1970. Pressure-Volume-Temperature-Concentration Relation of Aqueous Nacl Solutions.
Journal Of Chemical And Engineering Data 15 (1): 61-66.
Sifuentes, W.F., Giddins, M.A., and Blunt, M.J. 2009. Modeling CO2 Storage in Aquifers : Assessing the Key Contributors to
Uncertainty. Offshore Europe, Aberdeen, UK. SPE 123582-MS.
Valbuena, E. 2011. Analytical Estimation of CO2 Storage Capacity in Depleted Oil and Gas Reservoirs Based on
Thermodynamic State Functions. Master of Science, Texas A&M University.
Van Der Meer, L.G.H. and Ferhat, Y. 2009. CO2 Storage Capacity Calculations for the Dutch Subsurface. Energy Procedia 1
(1): 2615-2622.