SPE Design of Carbon Dioxide Storage in Oilfields Ran Qi, Tara C. LaForce and Martin J. Blunt, SPE, Imperial College London

Transcription

1 SPE Design of Carbon Dioxide Storage in Oilfields Ran Qi, Tara C. LaForce and Martin J. Blunt, SPE, Imperial College London Copyright 2008, Society of Petroleum Engineers This paper was prepared for presentation at the 2008 SPE Annual Technical Conference and Exhibition held in Denver, Colorado, USA, September This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright. Abstract We extend our study of the design of carbon dioxide, CO 2, storage in aquifers (Qi et al., 2007) to oilfields. We demonstrate that pore-scale capillary trapping is an effective and rapid mechanism to render the CO 2 immobile in oil reservoirs. We construct analytical solutions to the transport equations, accounting for relative permeability hysteresis. We use this to design an injection strategy where CO 2 and brine are injected simultaneously followed by chase brine injection. We study field-scale oil production and CO 2 storage using a streamline-based simulator that captures dissolution, dispersion, gravity and ratelimited reactions in three dimensions. While injecting at the optimum WAG ratio gives the fastest oil recovery, this allows CO 2 to channel through the reservoir leading to rapid CO 2 breakthrough and extensive recycling of the gas. We propose to inject more water than optimum. This allows to CO 2 to remain in the reservoir, increases the field life and leads to improved storage of CO 2 as a trapped phase. A short period of chase brine injection at the end of the process traps most of the remaining CO 2. Introduction Carbon Capture and Storage (CCS), the collection of CO 2 from industrial sources and its injection underground, could contribute significantly to reductions in atmospheric emissions of greenhouse gases (IPCC, 2005). Possible sites for injection include coalbeds, deep saline aquifers, and depleted oil and gas reservoirs. Although aquifers have the greatest storage potential, injecting CO 2 into depleted oil and gas reservoirs can provide additional hydrocarbon production and improved storage security. The principal concern with CO 2 storage is its long-term fate: can it be guaranteed that the CO 2 will remain underground for hundreds to thousands of years? Since oil and gas has been trapped for geological time in hydrocarbon reservoirs, they should all contain impermeable seals preventing escape. However, the top seal could leak, have gaps or be penetrated by wells through which CO 2 could migrate to the surface (Bruant et al., 2002). Dissolution in water and reaction with rock could also contribute to safe CO 2 storage, but both are slow processes taking thousands to billions of years (Ennis-King and Paterson, 2005; Hesse et al., 2006; Xu et al., 2003). Capillary trapping (residual non-wetting phase trapping) has been recognized as the most rapid method to immobilize CO 2 with time scales in the order of years to decades (Ennis-King and Paterson, 2002; Kumar et al., 2005; Obi and Blunt, 2006; Juanes et al., 2006; Taku Ide et al., 2007; Ghomian et al., 2008). Our previous study on the design of CO 2 storage in aquifers (Qi et al., 2007) has recommended an injection strategy in which CO 2 and brine are injected simultaneously followed by chase brine injection to render more than 85% of injected CO 2 immobile in a very short period. In the oil industry, CO 2 flooding has been used worldwide as a tertiary Enhanced Oil Recovery (EOR) mechanism for more than 30 years, particularly for reservoirs with pressures above the minimum miscibility pressure (MMP) where miscible displacement would occur. The ideal reservoirs for miscible CO 2 flooding usually have oil densities ranging from 29º to 48º API ( kg/m 3 ) and reservoir depths from 760m to 3700m below the surface (Taber et al., 1997). CO 2 flooding has the disadvantage that the unfavorable mobility ratio between the oil and CO 2 can result in early CO 2 breakthrough because of channeling of CO 2 through the reservoir fluids. Water alternate gas (WAG) injection can be successfully applied to improve the sweep efficiency and delay early CO 2 breakthrough (Lake, 1989). However, CO 2 flooding EOR projects have been designed to minimize the amount of CO 2 injected to recover the oil, since the CO 2 costs money to transport and inject, while for CCS, injection needs to maximize both CO 2 storage and oil recovery. It has been recommended by other researchers that the fluids be injected at the optimal WAG ratio, the injection gas composition be adjusted to reach the MMP, while the well type and completions are designed to maximize both oil recovery and CO 2 storage (Jessen, et al., 2004; Kovscek and Cakici, 2004; Malik and Islam, 2000).

2 2 [115663] In this paper, we use a streamline-based simulator, which has been extended to solve for CO 2 transport in oil reservoirs, to design a strategy to maximize both oil recovery and CO 2 storage as an immobile, trapped phase. Streamline-based simulation for CO 2 transport We extend a streamline-based simulator based on the work of Batycky et al. (1997) to simulate CO 2 transport in aquifers and oil fields (Qi, 2008). There are three phases (hydrocarbon, aqueous and solid) and four components (CO 2, oil, water and salt) in our simulation, as shown in Fig.1. We assume incompressible flow and ignore capillary pressure. During the simulation, grid-block properties are mapped along streamlines. Along each streamline we solve the governing transport equations on a one-dimensional (1D) grid. For each cell face in the 1D streamline grid, we record the grid block in the underlying 3D model which it resides in. Then to compute the flux across this face, we use the relative permeabilities associated with this block in this way we can account for relative permeability hysteresis and different trapping models (Land (1968) or Spiteri et al., (2005)). Finally, gravity segregation and phase exchange are considered by solving the relevant equations on the grid. Fig. 1. Schematic graph of the phases (shown on the left) and components (columns on the right) for streamline-based simulation of CO 2 transport. To capture the physics of CO 2 transport, we implemented a thermodynamic model of mutual dissolution between CO 2 and water and resulting salt precipitation (Spycher et al., 2003, 2005). A simple rate-limited chemical reaction is used to account for mineral precipitation of CO 2 (Obi and Blunt, 2006). The resultant changes of porosity and permeability due to chemical reaction and salt precipitation are also considered. In oilfield simulation, we assume first contact miscibility between CO 2 and oil to form a single hydrocarbon phase and the CO 2 concentration in the hydrocarbon phase is solved along streamlines. We use the Todd & Longstaff model (Todd and Longstaff, 1972) to represent sub-grid-block viscous fingering. We apply the trapping model introduced in Spiteri et al. (2005) with the same parameters as in Qi et al. (2007), representing a weakly water-wet system. Using the Carlson (1981) model, we assume that we can compute the relative permeability for the hydrocarbon phase from the primary drainage relative permeability as a function of the flowing saturation S hf (see also Blunt (2000)) during imbibition (waterflooding and chase brine injection), secondary drainage (CO 2 injection) and we do not consider hysteresis in the aqueous phase relative permeability: imbibition sec drainage primary drainage k ( S ) = k ondary ( S ) = k ( S ) (1) rh h rh h rh hf Reservoir description and fluid properties We use the SPE 10 model to represent a sector of a North Sea sandstone oil reservoir. This model has dimensions 366m 670m 52m with 1,122,000 grid cells given by in the x, y and z directions respectively. The top 35 layers represent the Tarbert formation, and the bottom 50 layers represent the Upper Ness. This reservoir is highly heterogeneous with a variation in permeability of over 5 orders of magnitude and has an average porosity of 0.2 (Christie and Blunt, 2001). There is one injector located in the center of the reservoir and four producers at the four corners, as shown in Fig. 2. The injector is controlled by injection rate and producers are controlled using a bottom hole pressure (BHP) constraint. Despite the simultaneous injection of two fluids, no problems with injectivity were encountered at the rates used. We use the primary drainage relative permeability curve measured for Berea sandstone (Oak, 1990), as shown in Fig.3. The initial oil saturation is 59, which is the maximum hydrocarbon saturation the system could reach during a secondary drainage process. We compute the hydrocarbon relative permeability during imbibition and secondary drainage based on the primary drainage curve in Fig. 3, the trapping model of Spiteri et al. (2005) and Eq. 1. We assume that the oil/water relative permeability and CO 2 /water relative permeabilities are the same. Other reservoir and fluid properties are listed in Table 1. Note that the CO 2 viscosity is 50 times lower than oil and more than ten times lower than water. The CO 2 will tend to be very mobile at high saturation; an effective storage design will depend on managing injection to keep the CO 2 at low saturation and low mobility so that it stays in the reservoir.

3 [115663] 3 P1 P2 I P3 P4 Fig. 2. The permeability field and well placement for the three-dimensional simulations presented. 1 kr k rh Imbibition and secondary drainage k rh primary drainage k rw S w Fig. 3. Relative permeability during primary drainage and imbibition (Oak, 1990). Table 1. Reservoir and fluid properties. Property Value Temperature 353K (80 ºC) Reference pressure, BHP 27 MPa Reservoir pore volume m 3 Initial water saturation Original oil inplace m 3 Water density 1050 kg/m 3 Water viscosity Pa s Oil density 850 kg/m 3 Oil viscosity Pa s CO 2 density 710 kg/m 3 CO 2 viscosity Pa s Brine density saturated with CO kg/m 3

4 4 [115663] Results and discussion One-dimensional simulations and analytical solutions To understand the displacement process and to validate our numerical model, we first ran a series of 1D simulations. We used 8000 grid blocks, with a total system size of 24,000m 300m 50m with a Darcy velocity of ⅔ m/day. The fluid parameters used are shown in Table 1. We ignore dissolution, dispersion and reaction. The system is initially at the maximum initial oil saturation. We first waterflood the system until all the grid cells reach residual oil saturation of We then inject CO 2 and brine simultaneously (SWAG) followed by brine injection only. We assume that the injected brine is fully saturated with CO 2 and that does not affect the viscosity of the brine. Fig. 4 shows the water/oil fractional flow (blue) and water/co 2 fractional flow (red) curves based on the imbibition relative permeabilities in Fig. 3 and fluid properties in Table 1. The slope of the dashed lines indicates the dimensionless CO 2 front velocity when different WAG ratios are injected. The steep water/oil fractional flow curve at high water saturation leads to a mobile oil bank that moves rapidly ahead of the CO 2 front. The optimal WAG ratio in this tertiary process is defined as the injected water fractional flow when the water and CO 2 fronts move at the same speed. It is found by finding the line that is tangent to the oil/water curve (Lake, 1989) and has a value of 0.15 see Fig. 4. This is equivalent to an injection CO 2 fractional flow, f ci = Chase brine injection will follow the water/co 2 fractional flow curve (red) from S w = 1-S or =29 falling to the injected f w. Again the slope of this line will be high, indicating that the chase brine will move very rapidly and quickly trap all the injected CO fw S w Fig. 4. Water fractional flow, f w, as a function of water saturation, S w, during secondary imbibition (waterflooding). The blue curve is the water/oil fractional flow and red curve is the water/co 2 fractional flow. The slope of the dashed lines indicate the dimensionless velocity of the CO 2 front when different WAG ratios are injected (f w=0.15, 0.3, 0.4 and are shown). The optimal WAG ratio is found from the line that is tangent to the water/oil curve and has a value of 0.15 (15% volume fraction water injected indicated by the horizontal dashed line). Note that at relatively high water saturation, both water/oil and water/co 2 fractional flow curves are extremely steep, which indicate very high wavespeeds. Fig. 5 shows a comparison of numerical and analytical solutions for an injected fractional flow f ci of (WAG ratio = 0.3). After waterflooding to residual oil, brine and CO 2 are injected simultaneously for 1100 days followed by chase brine injection. In this example, chase brine was injected for 5 days. Fig. 5 shows the hydrocarbon saturation as a function of distance. The injector is at a distance 0 and producer is 24,000 m away. A long, fast-moving oil bank is formed ahead of the injected CO 2 (a constant state followed by a rarefaction); it travels more than ten times faster than the CO 2 front. The chase brine front, behind the CO 2, is also moving much faster than the leading CO 2. After only 12 days, chase brine will have trapped all the CO 2. Numerical and analytical solutions are in good agreement, validating the numerical solver (Lake, 1989; Qi et al., 2007). Fig. 6 shows the solution for f ci =. This profile is shown after 2 days of chase water injection subsequent to 1100 days of brine and CO 2 injection. Again, the mobile oil front is much faster than the mobile CO 2 front. Chase brine will trap all the CO 2 injected in 8 days.

5 [115663] Chase water front Mobile CO 2 front Advancing oil front Sh Mobile Oil Simulation Analytical 0.42 Trapped CO 2 Mobile CO Distance (m) Fig. 5 One-dimensional hydrocarbon saturation profile where CO 2 and brine injection into residual oil at an injection CO 2 fractional flow f ci of for 1100 days is followed by 5 days of chase brine injection. The mobile oil bank moves very rapidly. The CO 2 front is followed by the chase brine, which also moves extremely fast, trapping all the CO 2 after only 12 days Mobile CO 2 front Advancing oil front Chase water front Sh 0.47 Mobile Oil Analytical Simulation Trapped CO 2 Mobile CO Distance (m) Fig. 6. Saturation profile corresponding to Fig. 5, but with an injection fractional flow of. The advancing mobile oil front moves even faster in this case. The chase water front is again moving much faster than the mobile CO 2, indicating that the CO 2 will be quickly trapped. Our 1D analysis has shown that simultaneous CO 2 and water injection (SWAG) is an effective tertiary recovery mechanism. The injection of CO 2 very rapidly leads to incremental production of oil, albeit with a low fractional flow. In these cases we inject more water than the optimum WAG ratio so that the injected brine moves ahead of the CO 2 front. As we show later, this impedes the movement of CO 2 to the production wells and leads to greater CO 2 storage. Chase brine injection after SWAG can trap all the CO 2 injected very rapidly in a residual phase. This conclusion is similar to that reached for chase brine injection after CO 2 injection in an aquifer (Qi et al., 2007) in both cases the advancing brine front moves very rapidly. However, in a 3D system, the water and CO 2 phases may not migrate in the same direction due to the effects of reservoir heterogeneity and buoyancy. Therefore, the process may be less efficient.

6 6 [115663] Three-dimensional simulations We performed a series of fine-grid 3D simulations of CO 2 storage and trapping. As in the 1D simulations, we first waterflood the reservoir until the average watercut of the four producers reaches 70%, which gives a recovery factor of 23.1%. The oil saturation distribution after waterflooding is shown in Fig m Z X y 366m Fig. 7. Oil saturation distribution after waterflooding. Left: three-dimensional profile. Right: two-dimensional cross section across the center of the reservoir, where the injection well is located. We then inject CO 2 and water together at different CO 2 reservoir-volume fractional flows, f ci (1.0, 0.9, 0.85,,,, 0.4, and 0.3). For different f ci, we inject CO 2 at the same rate ( kg/day) with the volume of water necessary to make the appropriate fractional flow of CO 2. CO 2 injection ceases when the total mass of CO 2 produced is equal to 20% of the total mass of CO 2 injected. We assume that there is a cost and energy penalty associated with recycling the CO 2 and so we do not allow significant production of the gas. Then we inject chase brine at a rate of 1000 m 3 /day. We account for CO 2 dissolution in brine, but do not allow water to enter the CO 2 phase. We assume that injected water will be fully saturated with CO 2 when they mix together in the wellbore. All the wells are vertical and fully completed from the top to the bottom of the reservoir, which maximizes the miscible contact area between CO 2 and oil. We compute the mass of CO 2 stored and oil production. We wish to maximize both CO 2 storage and oil production. Fig. 8 shows the cumulative oil production for different f ci. For the same pore volume (PV) of fluids injected, the optimal WAG ratio, f ci =0.85, has the maximum oil production. Lower f ci (more water than the optimal WAG ratio) can result in a higher final recovery because injection continues for longer before there is significant production of CO 2. In the lower f ci cases, injected water moves ahead of the injected CO 2, leaving CO 2 behind and minimizing CO 2 production. Therefore, injection at lower f ci (more water) reduces CO 2 cycling and leads to better ultimate oil recovery by extending the SWAG injection time. Fig. 9 plots the mass of CO 2 stored for different f ci when CO 2 injection ceases. The optimal WAG ratio, f ci =0.85, cannot store the maximum amount of CO 2. More CO 2 is stored as f ci decreases, although below there is very little difference in the storage capacity. Fig. 10 shows the ratio between the mass of incremental oil production (this is the oil produced by SWAG minus the oil recovery by continuous waterflooding) and the mass of CO 2 stored. Again, the lowest f ci (0.3) has the largest ratio, which is simply because it has the longest injection time. However, for the same PV injected, f ci =0.3 produces much less oil than higher f ci see Fig. 8. We suggest injecting more water than the optimum ratio: this lowers the mobility of the CO 2 in the reservoir and leads to greater storage. However, a very low f ci results in excessive water injection and production: it is best to inject at about f ci = as this gives good CO 2 storage and oil production, while keeping the project life and water handling manageable. Figs 11 and 12 show the oil and CO 2 volumetric fractions respectively at the end of CO 2 injection. The CO 2 volumetric fraction is the hydrocarbon saturation times the volume fraction of CO 2 in the hydrocarbon phase; the oil volumetric fraction is the hydrocarbon saturation times the oil volume fraction. Injecting more water than the optimal WAG ratio retains the CO 2 in the reservoir: the figures show efficient displacement of oil in high permeability regions; the CO 2 is then stored in these same portions of the field. The moderate hydrocarbon saturation leads to a low fractional flow and so the CO 2 is relatively immobile, despite its low viscosity.

7 [115663] E+05 Cumulative oil production (m 3 ) 8.86E E E E E E E E E+05 Waterflooding E PVI (Total) (a) 5.36E+05 Cumulative oil production (m 3 ) 5.26E E E E E E E E E+05 Waterflooding E PVI (Total) (b) Fig. 8. Cumulative oil produced as a function of total injected pore volume (PVI) (water and CO 2) after initial waterflooding at different injection fractional flows f ci indicated on the legend. All recovery curves stop when the total production of CO 2 is 20% of the injected mass of CO 2. Figure 8(b) is an enlarged section of Figure 8(a) showing the early-time behavior.

8 8 [115663] 3.50E E+08 CO2 mass (kg) 2.50E E E E E E f ci Fig. 9. The mass of CO 2 stored in the reservoir at the end of CO 2 injection as a function of injected fractional flow f ci Ratio f ci Fig. 10. Ratio between the mass of incremental oil production and the mass of stored CO 2 at the end of CO 2 injection as a function of the injected fractional flow f ci. During chase brine injection, we define two trapping efficiencies: the fraction of the injected mass of CO 2 that is either trapped or dissolved (trapping efficiency I); and the fraction of the stored mass of CO 2 (CO 2 in the reservoir) that is either trapped or dissolved (trapping efficiency II). The 1D analysis indicates that chase brine is a very effective way to render the CO 2 immobile in a residual phase in the absence of dissolution. In the 3D simulations we start to inject chase brine when 20% of the total injected CO 2 has been produced. The simulation stops when trapping efficiency I no longer increases. During chase brine injection the producers are still producing mobile CO 2 and trapped CO 2 continues to dissolve into the injected brine. Hence, the stored mass can decrease, leading to a decrease in trapping efficiency I. Fig. 13 plots the trapping efficiency as a function of chase brine injection time. Using f ci = can reach the highest trapping efficiencies in only 50 days: over 95% of the CO 2 underground is either dissolved or trapped and is very unlikely ever to escape. The chase brine injection time for different cases corresponds to the chase brine velocity in Fig. 4. The higher chase brine velocity, the faster CO 2 is trapped. The storage efficiency can be defined as follows (Obi and Blunt, 2006; Qi et al., 2007): it is the mass of CO 2 stored divided by the mass of CO 2 assuming that it fills the entire pore space. This is calculated by taking the stored mass from Fig. 9 and dividing by the reservoir pore volume multiplied by the CO 2 density (710 kg/m 3 ). For f ci around, the storage efficiency is approximately 17% at the end of chase brine injection. This is significantly better than that achieved for aquifer injection using the same reservoir model (Qi et al., 2007). The reason for this is that we have a closed system in a reservoir that allows recycling of produced CO 2.

9 [115663] 9 Z 52m X y CO 2 volumetric fraction 366 m Fig. 11. CO 2 volumetric fraction profile for f ci = at the end of CO 2 injection in three dimensions (left) and two-dimensional cross section (right). 52m Z y 366 m X Oil volumetric fraction Fig. 12. Oil volumetric fraction profile for f ci = at the end of CO 2 injection in three dimensions (left) and two-dimensional cross section (right) Trapping efficiency I Trapping efficiency II Chase water injection time (days) Chase water injection time (days) Fig. 13. Trapping efficiency I (left) and II (right) as a function of chase brine injection time for f ci = 0.85,, and.

10 10 [115663] Fig. 14 shows the 3D saturation profile after 30 days of chase brine injection for f ci =. CO 2 accumulates at the top of the reservoir and channels along high permeability streaks. The chase brine tends to follow the CO 2 to trap the vast majority of the injected solvent very quickly, leaving only a relatively small amount of mobile CO 2. After another 20 days chase brine injection (50 days in total), 76.5% of injected CO 2 will be trapped, which is 96.5% of the total mass of CO 2 stored. Z 52m X y 366m 52m Z X y 366m CO 2 volumetric fraction Fig. 14. CO 2 volumetric fraction profiles for a three-dimensional simulation with an injected CO 2 fractional flow of : trapped CO 2 (top) and mobile CO 2 (bottom) showing both the three-dimensional profile and two-dimensional cross section. 548 days of CO 2 and brine injection together is followed by 30 days of chase brine injection. Note that the case brine effectively traps the majority of the CO 2 near the wellbore and in high permeability regions. Figs. 15 and 16 show the cumulative oil and CO 2 production during SWAG and chase water injection for f ci = and respectively. Dashed lines separate the production during SWAG injection and chase water injection. During chase water injection, the producers are still producing oil. During water injection the water cut increases rapidly and so CO 2 production is significantly reduced because CO 2 is trapped as an immobile phase by the fast-moving injected water. 9.00E+05 Cumulative Oil/CO 2 production ( m 3 ) 8.00E E E E E E E E+05 Oil production CO 2 production SWAG Chase water 0.00E Time (days) Fig. 15 Cumulative oil and CO 2 production during SWAG and chase water injection when f ci =.

11 [115663] E+06 Cumulative oil/co 2 production (m 3 ) 9.00E E E E E E E E E+05 Oil production CO 2 production SWAG Chase water 0.00E Time (days) Fig. 16. Cumulative oil and CO 2 production during SWAG and chase water injection when f ci =. Our 3D simulations have shown that the traditional optimal WAG ratio is not ideal when both CO 2 storage and enhanced oil recovery are important to an injection scheme. The results indicate that injecting more water during WAG injection will significantly reduce CO 2 cycling and store more CO 2 in the reservoir. The optimal ratio is found for an injected CO 2 fractional flow around 50% in this case considering the combination of CO 2 stored, oil production and cost (related to the amount of water injected and produced). Our proposed design criteria are to inject CO 2 simultaneously with brine with a SWAG ratio more than optimal followed by chase brine injection. Our simulation studies, in a very heterogeneous system, indicate that more than 75% of the injected CO 2 is trapped, or more than 90% of the CO 2 underground, with an overall storage efficiency of 17%, while the cumulative oil production is almost twice the production that could be achieved through waterflooding alone. Conclusions We have extended streamline-based simulation to study CO 2 transport in oil reservoirs. We can simulate displacement in million-cell models using a standard PC. The simulator incorporates a state-of-the-art trapping model and relative permeability hysteresis based on pore-scale modeling verified by experimental data. We studied CO 2 storage and tertiary recovery in a heterogeneous oilfield. In this example, extensive channeling led to early breakthrough of water and CO 2. To retain the CO 2 in the reservoir, we propose an injection strategy where CO 2 and water are injected simultaneously at a higher WAG ratio (more water) than the traditional optimum value. Water moves ahead of the CO 2, keeping the CO 2 at low saturation and relatively immobile in the reservoir. Where there are concerns over long-term storage security, a brief period of chase brine injection is sufficient to render more than 90% of the CO 2 underground trapped or dissolved with an overall storage efficiency of approximately 17%. Further work is required to validate the trapping model used here and to test our design criteria for other reservoir models. Acknowledgments We would like to acknowledge funding from the following agencies: Shell under the Grand Challenge on Clean Fossil Fuels; Qatar Petroleum, Shell and the Qatar Science and Technology Park under the Qatar Carbonates and Carbon Storage Research Centre; and NERC (grant number NE/C516401/1). Ran Qi is grateful for the Schlumberger Faculty of the Future Scholarship. References 1. Batycky, R.P., Blunt, M.J. and Thiele, M.R.: A 3D Field-scale Streamline-based Reservoir Simulator, SPE Reservoir Engineering, 12, (1997) 2. Bedrikovetsky, P., Marchesin, D., and Ballin, P.R.: Mathematical Model for Immiscible Displacement Honouring Hysteresis, SPE 36132, Proceedings of the SPE 4 th Latin American and Caribbean Petroleum Engineering Conference, Port-of-Spain, Trinidad and Tobago, April (1996). 3. Blunt, M.J.: An empirical model for three-phase relative permeability, SPEJ, 5(4), , December (2000). 4. Bruant, R.G., Jr., Celia, M.A., Guswa, A.J. and Peters, C.A.: Safe storage of CO2 in deep saline aquifers, Environmental Science and Technology, 36(11), 240A 245A (2002). 5. Carlson, F.M.: Simulation of relative permeability hysteresis to the nonwetting phase, SPE 10157, Proceedings of the SPE Annual Technical Conference and Exhibition, San Antonio, TX, October 5 7 (1981). 6. Chiquet, P., Broseta, D., and Thibeau, S.: Wettability alteration of caprock minerals by carbon dioxide, Geofluids, 7, doi: /j x (2007).

12 12 [115663] 7. Christie, M.A. and Blunt, M.J.: Tenth SPE Comparative Solution Project: A Comparison of Upscaling Techniques, SPE Reservoir Evaluation and Engineering, 4(4), , August (2001). 8. Ennis-King, J. and Paterson L.: Role of convective mixing in the long-term storage of carbon dioxide in deep saline formations, SPEJ, 10(3), (2005). 9. Ennis-King, J. and Paterson, L.: Engineering Aspects of Geological Sequestration of Carbon Dioxide, SPE 77809, Proceedings of the Asia Pacific Oil and Gas Conference and Exhibition, Melbourne, Australia, 8-10 October (2002). 10. Furati, K.M.: Effects of Relative Permeability History Dependence on Two-Phase Flow in Porous Media, Transport in Porous Media, 28, (1997). 11. Ghomian, Y., Pope, G.A. and Sepehmoori, K.: Hysteresis and Field-Scale Optimization of WAG Injection for Coupled CO 2 -EOR and Sequestration, SPE , Proceedings of the SPE/DOE Improved Oil Recovery Symposium held in Tulsa, Oklahoma, April (2008). 12. Hesse, M.A., Tchelepi, H.A. and Orr, F.M.: Scaling Analysis of the Migration of CO2 in Saline Aquifers, SPE , Proceedings of the SPE Annual Technical Conference and Exhibition, September 24-27th, San Antonio, Texas, USA (2006). 13. Intergovernmental Panel on Climate Change, Special Report on Carbon Dioxide Capture and Storage, B Metz et al., editors, (2005) Jerauld, G.R.: Prudhoe Bay Gas/Oil Relative Permeability, SPE Reservoir Engineering, 12, 66-73, (1997). 15. Jessen, K., Kovscek, A.R. and Orr, F.M.: Increasing CO2 storage in oil recovery, Energy Conversion Management, 46(2), (2005). 16. Juanes, R., Spiteri, E.J., Orr, F.M. and Blunt, M.J.: Impact of Relative Permeability Hysteresis on Geological CO2 Storage. Water Resources Research, 42, W12418, doi: /2005WR (2006). 17. Killough, J.E.: Reservoir Simulation with History-Dependent Saturation Functions, SPEJ, 16, 37-48, (1976). 18. Korbøl, R., and Kaddour, A.: Sleipner Vest CO2 disposal Injection of removed CO2 into the Utsira formation, Energy Conversion Management, 36(6 9), (1995). 19. Kovscek, A.R. and Wang, Y.: Geologic Storage of Carbon Dioxide and Enhavced Oil Recovery. Ι. Uncertainty Quantification Employing a Streamline Based Proxy for Reservoir Flow Simulation, Energy Conversion & Management, 46, , November (2005). 20. Kovscek, A.R. and Cakici, M.D.: Geologic Storage of Carbon Dioxide and Enhanced Oil Recovery. Π. Cooptimization of Storage and Recovery, Energy Conversion & Management, 46, , November (2005). 21. Kumar, A., Ozah, R., Noh, M., Pope, G.A., Bryant, S., Sepehrnoori, K. and Lake, L.W.: Reservoir simulation of CO2 storage in deep saline aquifers, SPEJ, 10(3), (2005). 22. Lake, L.W.: Enhanced Oil Recovery. Prentice-Hall, Inc. (1989). 23. Land, C.S.: Calculation of Imbibition Relative Permeability for Two- and Three-Phase Flow From Rock Properties, SPEJ, 8, (1968). 24. Malik, Q.M. and Islam, M.R.: CO2 Injection in the Weyburn Field of Canada: Optimization of Enhanced Oil recovery and Greenhouse Gas Storage with Horizontal Wells, SPE 59327, Proceedings of the SPE/DOE Improved Oil Recovery Symposium, Tulsa, Oklahoma, 3-5 April (2000). 25. Medeiros, H.B., Marchesin, D. and Paes Leme, P.J.: Hysteresis in Two-Phase Flow: A Simple Mathematical Model, Comp. Appl. Math., 17(1), (1998) 26. Oak, M.J.: Three-phase Relative Permeability of Water-wet Berea. SPE 20183, Proceedings of the SPE/DOE Seventh Symposium on Enhanced Oil Recovery, Tulsa, (1990). 27. Obi, E.I. and Blunt, M.J.: Streamline-based Simulation of Carbon Dioxide Storage in a North Sea Aquifer. Water Resources Research, 42, W03414, doi: /2004WR (2006). 28. Qi, R., Beraldo, V., LaForce, T.C. and Blunt, M.J.: Design of Carbon Dioxide Storage in a North Sea Aquifer Using Streamline- Based Simulation, SPE , Proceedings of the SPE Annual Technical Conference and Exhibition held in Anaheim, California, November (2007). 29. Qi, R.: Simulation of Geological Carbon Dioxide Storage, PhD thesis, Imperial College London (2008). 30. Qi, R., LaForce, T.C. and Blunt, M.J.: Design of Carbon Dioxide Storage in Aquifers, International Journal of Greenhouse Gas Control, in press, Spiteri, E. J., Juanes, R., Blunt, M.J. and Orr, F.M. Jr.: Relative Permeability Hysteresis: Trapping Models and Application to Geological CO2 Sequestration, SPE 96448, Proceedings of the SPE Annual Meeting, Dallas, Texas, 9 12 October (2005). 32. Spycher, N., Pruess, K. and Ennis-King, J.: CO 2 -H 2 O Mixtures in the Geological Sequestration of CO 2. Ι. Assessment and Calculation of Mutual Solubilities from 12 to 100 ºC and up to 600 bar, Geochimica et Cosmochimica Acta, 67(16), , (2003). 33. Spycher, N., Pruess, K. and Ennis-King, J.: CO 2 -H 2 O Mixtures in the Geological Sequestration of CO 2. Π. Partitioning in chloride brines at 12 to 100 ºC and up to 600 bar, Geochimica et Cosmochimica Acta, 69(13), , (2005). 34. Taku Ide, S., Jessen, K. and Orr, F.M. Jr.: Storage of CO 2 in saline aquifers: Effects of gravity, viscous, and capillary forces on amount and timing of trapping, International Journal of Greenhouse Gas Control, 4, (2007). 35. Taber, J.J, Martin, F.D. and Seright, R.S.: EOR Criteria Revisited. Part 1: Introduction to Screening Criteria and Enhanced Recovery Field Projects, SPE Reservoir Engineering, 12, , August (1997). 36. Taber, J.J, Martin, F.D. and Seright, R.S.: EOR Criteria Revisited. Part 2: Application and Impact of Oil Prices, SPE Reservoir Engineering, 12, , August (1997). 37. Todd, M. and Longstaff, W.: The development, Testing and Application of a Numerical Simulator for Predicting Miscible Flood Performance, SPE 3484, Journal of Petroleum Technology, July, Valvatne, P.H. and Blunt M.J.: Predictive pore-scale modeling of two-phase flow in mixed wet media, Water Resources Research, 40, W07406, doi: /2003WR (2004).

13 [115663] van Kats, F.M. and van Duijn, C.J.: A Mathematical Model for Hysteretic Two-Phase Flow in Porous Media, Transport in Porous Media, 42, (2001). 40. Xu, T., Apps, J. A. and Pruess, K.: Reactive geochemical transport simulation to study mineral trapping for CO2 disposal in deep saline arenaceous aquifers, J. Geophys. Res., 108(B2), 2071, doi: /2002jb (2003).