CO2 -hydrate formation in depleted gas reservoirs–A methodology for CO2 storage

Available online at www.sciencedirect.com

   

Energy Procedia 4 (2011) 3949–3956

www.elsevier.com/locate/procedia

Energy Procedia 00 (2010) 000–000

www.elsevier.com/locate/XXX

GHGT-10

CO2-hydrate formation in depleted gas reservoirs – A methodology for CO2 storage Olga Ye. Zatsepinaa and Mehran Pooladi-Darvish a,b* a

b

University of Calgary, 2500 University Drive NW, Calgary, AB, T2N 1N4, Canada Fekete Associates Inc., #200, 540-5th Avenue SW, Calgary, Alberta, Canada T2P 0M2 Elsevier use only: Received date here; revised date here; accepted date here

Abstract With the growing concern about climate change, interest towards reducing CO2 emissions has increased. Geological storage of CO2 is perceived to be one of the most promising methods that could provide significant reduction in CO2 emissions over the short and medium term. Since a major concern regarding geological storage is the possibility of leakage, trapping CO2 in the solid form is quite attractive. Unlike mineral trapping, the kinetics of CO2-hydrate formation is quite fast, providing the opportunity for long-term storage of CO2. In this paper, we study storage of CO2 at conditions similar to those at depleted gas pools of Northern Alberta. Thermodynamic calculations suggest that CO2 hydrate is stable at temperatures that occur in a number of formations in Northern Alberta, in an area where significant CO2 emissions are associated with production of oil sands and bitumen. Numerical simulation results presented in this paper suggest that, upon CO2 injection into such depleted gas reservoirs, pressure would initially rise until conditions are appropriate for hydrate formation, enabling storage of large volumes of CO2 in solid form. These results show that, because of tight packing of CO2 molecules in the solid (hydrate), the CO2 storage capacity of these pools is many times greater than their original gas-in-place. This provides a local option for storage of a portion of the CO2 emissions there. c 2010 ⃝ 2011 Elsevier Published by All Elsevier © Ltd. rightsLtd. reserve

Keywords: CO2; storage; hydrate; depleted gas reservoir

* Corresponding author: Tel: 403-210-8662; fax: 403-284-4852 E-mail address: [email protected]

doi:10.1016/j.egypro.2011.02.334

39502

O.Y. Zatsepina, M. Pooladi-Darvish / Energy Procedia 4 (2011) 3949–3956 Author name / Energy Procedia 00 (2010) 000–000

1. Introduction Sequestration of CO2 in the form of hydrate is quite attractive since its solid state provides a low probability of leakage and an effective trapping mechanism. Hydrates are solids composed of the framework of water molecules, with gas molecules captured in the lattice vacancies. This solid structure provides a high density for the CO2 storage, while kinetics of the hydrate formation is relatively fast (Zatsepina and Buffet, 2002). This paper focuses on storing CO2 through hydrate formation in depleted reservoirs of natural gas (Koide et al., 1995). Since natural gas has been kept in the reservoir beneath impermeable rocks for millions of years, these gas pools could represent safe CO2-storage facilities. We present a modeling study of hydrate formation by injecting CO2 into a depleted gas reservoir. The CMG STARS [2008] reservoir simulator employed in the study has previously been used to model formation/dissociation of hydrate at reservoir conditions (Uddin et al., 2008). We start with a discussion of the phase diagram of mixed hydrates of CH4 and CO2.

2. Mixed hydrate phase equilibrium Phase diagram for CH4-CO2 hydrate is shown in Figure 1 (Adisasmito et al., 1991). It shows that formation conditions of mixed hydrates depend on the composition of the vapor phase (the phase boundaries are shown at 0, 20, 40, 60, 80, and 100% CO2 in the vapor phase). At a fixed temperature, the phase boundary of the mixed hydrate spreads to lower pressures when the fraction of CO2 in the vapor phase increases. Consider three-phase equilibrium at a temperature of 6° C. For CH4 hydrate, the equilibrium pressure at 6° C is 4.61 MPa, while for CO2 hydrate it is 2.54 MPa. When hydrate forms from a gas mixture of, for example, 70% CO2 and 30% CH4, the equilibrium pressure is 2.80 MPa. At a P-T condition denoted by “X”, gas mixtures with CO2 mole fraction of 60% and more would form a stable hydrate. As we shall see, injection of CO2 will lead to displacement methane. Therefore, hydrate formation in the displaced area will follow the three-phase equilibrium curve of pure CO2. 4500 HYDRATE - VAPOR CH4

Pressure (kPa)

3500 20%

x

40%

2500

60% 80%

CO2

1500

VAPOR - AQ. SOLUTION 500 2

4

6

8

10

o

Temperature ( C)

Figure 1. Methane-CO2 mixed hydrate phase diagram. The cross represents a shallow gas pool in Northern Alberta at 6° C that may be pressurized to 3 MPa. Thick lines represent phase boundaries for CH4 and CO2 hydrate, while thin lines represent phase boundaries for mixed hydrate (mole fraction of CO2 in the vapor phase is written above the lines).

3. Assumptions

33951

O.Y. Zatsepina, M.Author Pooladi-Darvish (2011) 3949–3956 name / Energy/ Energy ProcediaProcedia 00 (2010) 4 000–000

Hydrate Composition. The vapor phase composition affects not only P/T conditions of hydrate stability, but also fractionation of species in hydrate (Uchida et al., 2005). We assume that the coefficient of proportionality between compositions of the vapor and hydrate phases is close to 1 (Buffett and Zatsepina, 2004). In our modeling, the mole fraction of the vapor phase changes with injection of CO2. If the newly formed hydrate reflects a new gas composition, the previously formed hydrate could either remain unchanged or consume part of the CO2 and adjust its CO2-hydrate fraction. In this work, we assume that the previously formed hydrate changes to reflect new gas composition. As a result, the vapor and hydrate phases in the modeling have the same compositions. Consequently, the CO2 fraction in the hydrate phase would increase with injection time as CO2 molecules substitute methane in the structure. Rate-limiting Mechanisms, Enthalpy. There are three major rate-controlling mechanisms in the process of hydrate formation in reservoirs that include intrinsic kinetics, fluid flow, and heat transfer (Hong and Pooladi-Darvish, 2005). In this study, we assume fast intrinsic kinetics, so that the pressure and temperature of hydrate formation follow equilibrium conditions. Since heat transfer controls the rate of hydrate formation, latent heat of the reaction plays an important role. Enthalpy of mixed hydrates is calculated using values for the pure components in the proportion they have in the hydrates. For methane hydrate, a value of 54.44 kJ/mol is used (Gupta et al., 2008). For CO2 hydrate, an averaged value of 60 kJ/mol from available data is used (Anderson, 2003). 4. Results We consider a depleted gas reservoir with a thickness and radius of 5 m and 300 m, respectively, with porosity of 30% and permeability of 500 mD. The (initial) reservoir pressure is 500 kPa, initial temperature is 5° C, and saturations of gas and water are 75% and 25%, respectively. The initial volume of gas-in-place (methane) is 1.66×106 sm3. There is a layer of shale on the top and bottom of the reservoir (the shale thickness is chosen so that the effect of temperature changes in the reservoir is not felt at the top or the bottom boundaries). A vertical well is fully open, with CO2 gas injected at a constant rate of 0.1×106 sm3/day at a temperature of 10° C. (The injection temperature is chosen so that no hydrate forms in the vicinity of the wellbore: the three-phase equilibrium pressure at 10° C is 4.4 MPa.) The reservoir pressure is restricted to no greater than 4 MPa in order to avoid appearance of liquid CO2. The reservoir is discretized in radial direction: 300 cells × 1 meter; in vertical direction: 5 cells × 1 meter in the reservoir and 5 cells × 5 meters in the shale. Thermal properties of the reservoir and shale are given in Table 1.  Table 1: Thermal properties of the reservoir and shale  Reservoir volumetric heat capacity (J/m3 - °C)

2.6×106

Rock thermal conductivity (J/m - day - °C)

2.47×10

Hydrate heat capacity (J/gmole - °C)

203.7

5

Shale volumetric heat capacity (J/m3 - °C)

3.76×106

Shale thermal conductivity (J/m - day - °C)

1.5×105

Hydrate thermal conductivity (J/m - day - °C)

3.40×104

Figure 2 (left) shows changes in average values of pressure, temperature, and hydrate saturation with time. During the first 90 days, pressure increases with injection while no hydrate forms. With continued injection, the pressure increases and hydrate forms. However, as the hydrate forms, gas is consumed and the rate of pressure increase lessens. Formation of hydrate is also signified by an increase in temperature, which continues to rise with increasing hydrate saturation. After 270 days and injection of more than 15 times the original gas-in-place, the pressure and temperature have increased to 3.95 MPa and 9.2° C,

39524

O.Y. Zatsepina, M. Pooladi-Darvish / Energy Procedia 4 (2011) 3949–3956 Author name / Energy Procedia 00 (2010) 000–000

respectively. Figure 2 (right) shows the CO2-hydrate phase boundary together with changes in P/T conditions and saturation of hydrate at a radius of 100 m half-way between the top and bottom of the reservoir (the CO2-hydrate diagram is shown because methane is displaced in this region). Initially, pressure increases at a nearly constant temperature. Thereafter, it follows the equilibrium curve as saturation of hydrate increases.

T

6

0.6

4

0.4

2

0.2

P

Sh

0

0 0

50

100

150

200

0.8 3500 0.6 2500 Peq(T)

0.4

1500 P

Hydrate Saturation

0.8

Eq. Pressure, Pressure (kPa)

8

1

4500

1

Average Hydrate Saturation

Av. Pressure (MPa), Temperature (ο C)

10

0.2 Sh

500

250

0 4

5

Time (days)

6

7

8

9

10

Temperature (C)

Figure 2. Average values for pressure, temperature, and hydrate saturation as functions of time (left); equilibrium pressure, pressure, and hydrate saturation at a distance of 100 m from the wellbore in the middle depth of the hydrate reservoir as functions of temperature (right)

Figures 3 (left) and (right) show radial distributions of temperature and hydrate saturation, respectively, at the middle depth prior to hydrate formation at 60 days and afterward at 120, 180, and 240 days at various stages of hydrate formation. Radial distributions of temperature in Figure 3 (left) are affected by the temperature of injected CO2, initial temperature of the reservoir, and latent heat of hydrate formation. Each distribution could be divided into three regions: the vicinity of the wellbore, the outskirts, and in-between. Before hydrate formation, temperature changes from 10° C of injected CO2 to the initial reservoir temperature of 5° C. Once hydrate formation occurs, an “in-between” region develops. Temperature there rises from 5.2° C at 60 days prior to hydrate formation to 9° C at 240 days. From the “in-between” region to the outskirts, temperature decreases with radius. 10

0.18 240 days

0.15

Hydrate saturation

Temperature (ο C)

9 180 days

8

7 120 days

6

0.12 0.09 240 days

0.06 180 days

0.03

120 days

60 days

5

60 days

0 0

50

100

150

200

Radial distance (m)

250

300

0

50

100

150

200

250

300

Radial distance (m)

Figure 3. Radial distributions of temperature (left) and hydrate saturation (right) in the middle of hydrate reservoir at 60, 120, 180, and 240 days.

O.Y. Zatsepina, M.Author Pooladi-Darvish (2011) 3949–3956 name / Energy/ Energy ProcediaProcedia 00 (2010) 4 000–000

53953

Radial distributions of hydrate saturation in Figure 3 (right) reflect changes in temperature. It shows the absence of hydrate around the wellbore followed by constant saturation in the “in-between” region with a hump prior to hydrate disappearing on the outskirts. Methane is displaced by CO2 so that pure CO2 hydrate forms in most of the reservoir. Mixed hydrates appear in a narrow zone at the CO2 front, with the highest methane fraction of 30%. Testing of the results has revealed that the hump in hydrate saturations is associated with a decreased temperature at the CO2 front. Presence of the saturation hump is very sensitive to the distribution of the CO2 and local temperatures. Numerical experimentations have shown that the hump may totally disappear when a large diffusion coefficient is considered.

10

0.25

8

0.2

6

0.15 CO2

4

0.1 CH4

2

0.05

0 0

50

100

150

200

Methane in Hydrate (106 sm3)

CO2 in Hydrate (106 sm3)

Figure 4 shows volumes of CO2 and methane stored in the form of hydrate as a function of time. The volume of CO2 considerably exceeds that of methane. A comparison between the original gas-in-place and the volume of CO2 stored in hydrate suggests that the latter is approximately five times larger than the initial gas-in-place.

0 250

Time (days)

Figure 4. Carbon dioxide and methane stored in hydrate form .

Two-dimensional distributions of temperature and hydrate saturation at 180 and 240 days are shown in Figures 5 (top) and (bottom), respectively. The two temperature distributions exhibit similar features. They show an increased temperature in the vicinity of the wellbore due to injection of warm CO2 and a reduced temperature on the hydrate top and bottom surfaces bordering shales, where heat diffuses into the cap- and base-rock. There is also a low temperature zone at the CO2 front close to the outer boundary. The front is associated with a vertical gradient in temperature. Testing of the results has revealed that vertical temperature profile at the CO2 front is associated with gravitational segregation of CO2 in the vapor phase. As mentioned earlier, diffusion could alter the composition gradients and associated equilibrium conditions. Two-dimensional maps of hydrate saturation in Figure 5 (bottom) reflect the same features discussed above. First, hydrate is absent in the vicinity of the wellbore because of the high temperature. There is no hydrate within 22 m and 10 m from the wellbore at 180 days and 240 days, respectively (hydrate forms closer to the wellbore at 240 days because pressure increases with injection time). Second, saturation of hydrate at the borders with shale is higher than in the middle of the reservoir; associated with the decreased temperature. Hydrate saturation at the CO2 front is also increased due to a reduced temperature

39546

O.Y. Zatsepina, M. Pooladi-Darvish / Energy Procedia 4 (2011) 3949–3956 Author name / Energy Procedia 00 (2010) 000–000

there. The maps indicate that radius of the hydrate zone at the bottom of the reservoir exceeds that at the top. This is caused by settling of CO2 which promotes formation of hydrate.

100

200

100

200

100

200

10.0

200.0

200.0

200.0

9.0

200.0

9.5

8.5 8.0 7.5 7.0 6.5 6.0 210.0

210.0

210.0

210.0

5.5 5.0

100

100

200

200

100

200

100

200

0.32 0.28 200.0

200.0

200.0

200.0

0.23 0.18 0.14 0.09

210.0

210.0

210.0

210.0

0.05 0.00

100

200

Figure 5. Two-dimensional distributions of temperature (top) and hydrate saturation (bottom) at 180 days (left) and 240 days (right).

5. Discussion In this study, we examined hydrate formation during a period of eight months. During this period, the sensible heat of the reservoir between the initial temperature and equilibrium temperature at maximum pressure controls the amount of hydrate formed. However, over longer periods of time (either slower injection or shut-in after the injection period), the increased temperature in the reservoir dissipates and the reservoir cooling would lead to formation of more hydrate and/or pressure reduction. In this case, the rate of hydrate formation would be controlled by conduction of heat. Diffusion and dispersion are not included in the modeling. So, the region with both CO2 and methane appears because of numerical errors. Diffusion of gas molecules would cause dispersion of the CO2 front promoting formation of the mixed gas hydrate. Such effects as tortuous flow paths and heterogeneity of

O.Y. Zatsepina, M.Author Pooladi-Darvish (2011) 3949–3956 name / Energy/ Energy ProcediaProcedia 00 (2010) 4 000–000

73955

the porous medium also contribute to dispersion of the gas front. Therefore, mixing mechanisms in the reservoir (due to dispersion and/or heterogeneity) would lead to formation of mixed hydrate over a wider region of the reservoir compared to that of this work. These aspects as well as the reservoir cooling are the subject of on-going research. Composition of the in-situ and injected gas affects hydrate stability. Presence of heavier hydrocarbons such as C2 and C3 in the in-situ gas, and/or impurities in the injected gas such as SO2 and H2S, expand the stability zone of hydrate. This suggests expansion of the candidate reservoirs, with a possible inclusion of deeper and warmer reservoirs. The motivation for this work was the fast kinetics of hydrate formation that could trap CO2 in solid cages of CO2 -hydrate. It should be noted that, kinetics of hydrate dissociation is also fast. Therefore, if pressure and temperature conditions of the reservoir are altered, the hydrate could dissociate and the CO2 would be released once again. This may be of practical implications in Northern Alberta, where some of the depleted gas pools are underlain by bitumen reservoirs that may be developed by thermal methods. Work is underway to map those depleted gas pools, where risk of heating by thermal recovery from below is absent. 6. Summary and Conclusions We have studied CO2 injection into a depleted gas reservoir. It was shown that by controlling the temperature of the injected CO2, one could avoid hydrate formation in the vicinity of the wellbore, where formation plugging may occur. Formation of hydrate away from the wellbore and at saturations observed in this study is not expected to affect injectivity. Simulation results indicate that more hydrate forms at bottom and top, where heat of hydrate formation dissipates to the base- and cap-rock. The volume of CO2 that turned into hydrate was many times the original gas-in-place. A number of other factors, including effect of other hydrocarbons in the in-situ gas or impurities in the injected gas, and reservoir cooling after a period of CO2 injection, both of which could increase potential of CO2 storage in depleted gas pools are being studied. Acknowledgements The assistance of Dr. Dennis Coombe of Computer Modeling (CMG) in use of the STARS simulator for defining mixed-hydrates is gratefully acknowledged. Financial funding for this research was provided by the National Science and Engineering Council of Canada (NSERC). References [1] Zatsepina, O.Y. and Buffett, B.A., Nucleation of CO2-hydrate in a porous medium. Fluid Phase Equilibria. 2002; 200 (2): 263-275. [2] Koide, H., Takahashi, M., and Tsukamoto, H. Self-trapping mechanisms of carbon dioxide in the aquifer disposal. Energy Convers. Mgmt. 1995; 36 (6-9): 505-508. [3] CMG STARS. Advanced process and thermal reservoir simulator. Computer Modelling Group Ltd., Calgary, AB, Canada; 2008.

39568

O.Y. Zatsepina, M. Pooladi-Darvish / Energy Procedia 4 (2011) 3949–3956 Author name / Energy Procedia 00 (2010) 000–000

[4] Uddin, M., Coombe, D., and Wright, F. Modeling of CO2-hydrate formation in geological reservoirs by injection of CO2 gas. Journal of Energy Resources Technology, Transactions of the ASME. 2008; 130 (3): 0325021-03250211. [5] Adisasmito, S., Frank, R.J., and Sloan, E.D. Hydrates of carbon dioxide and methane mixtures. J. Chem. Eng. Data. 1991; 36: 68-71. [6] Uchida,T., Ikeda, I.Y., Takeya, S., Kamata, Y., Ohmura, R., Nagao, J., Zatsepina, O. Y., Buffett, B.A. Kinetics and Stability of CH4-CO2 mixed gas hydrates during formation and long-term storage. ChemPhysChem. 2005; 6 (4): 646-654. [7] Buffett, B., and Zatsepina, O. Research on mixed gas hydrate equilibrium. Report to Institute of Energy Utilization, AIST, Japan; 2004. [8] Hong, H. and Pooladi-Darvish, M. Simulation of Depressurization for Gas Production from Gas Hydrates Reservoirs. Journal of Canadian Petroleum Technology (JCPT), 2005; 44 (11): 39 – 46. [9] Gupta, A., Lachance, J., Sloan Jr., E.D., Koh, C.A., Measurements of methane hydrate heat of dissociation using high pressure differential scanning calorimetry. Chemical Engineering Science. 2008; 63: 5448-5853. [10] Anderson, G.K., Enthalpy of dissociation and hydration number of carbon dioxide hydrate from the Clapeyron equation. J. Chem. Thermodynamics. 2003; 33:1171-1183.