Scholarly article on topic 'Remediation of possible leakage from geologic CO2 storage reservoirs into groundwater aquifers'

Remediation of possible leakage from geologic CO2 storage reservoirs into groundwater aquifers Academic research paper on "Earth and related environmental sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Energy Procedia
Keywords
{Groundwater / Remediation / Storage}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — Ariel Esposito, Sally M. Benson

Abstract Maintaining the long term storage of CO2 is an important requirement for a large scale geologic CO2 storage project. Nevertheless, the possibility remains that the CO2 will leak out of the formation into overlying groundwater aquifers. A site specific remediation plan is also important during the site selection process and necessary before storage begins. Due to the importance of protecting drinking water resources, this study analyzes the optimal remediation scenario for various leakage conditions. The three objectives for remediation considered here are removing any mobile CO2, reducing the quantity of CO2 in the reservoir, and reducing the aqueous phase concentration of CO2. The first part of our research was to determine the processes that control the size and shape of the leakage plume in the groundwater aquifer. We used the multiphase flow simulator TOUGH2 with CO2 leakage from a point source to analyze the plume at various leakage rates. We next determined that during remediation the important physical processes include capillary trapping as a result of hysteresis in the relative permeability curves, dissolution, and buoyancy induced flow. We compared the effectiveness of using vertical and horizontal extraction wells to remove the CO2. We next examined injecting water to dissolve the gaseous CO2 and reduce the overall concentration and increase capillary trapping. Finally, we analyzed the combination of water injection and extraction with multiple wells to determine the optimal spacing and flow rate. Based on the simulations analyzed for this study, multiple conclusions can be made on the effectiveness of various remediation scenarios. With one vertical extraction well the optimal scenario for the larger leakage cases is a multistep extraction process that removes mobile CO2 from the areas with high gas saturation first. A horizontal extraction well in the middle of the aquifer is much more efficient than vertical wells at removing CO2. Water injection is effective at quickly reducing the mobile phase CO2 with tradeoffs between injection rate and increases in pressure. The most effective scenario over a longer time period includes injection for a short time followed by extraction from four vertical wells. To reduce the CO2 most rapidly, four injector wells with high flow rates and one extraction well is the most effective for the large leakage cases.

Academic research paper on topic "Remediation of possible leakage from geologic CO2 storage reservoirs into groundwater aquifers"

Available online at www.sciencedirect.com

m%' ScienceDirect

Energy Procedía 4 (2011) 3216-3223

Energy Procedía

www.elsevier.com/locate/procedia

GHGT-10

Remediation of possible leakage from geologic CO2 storage reservoirs into groundwater aquifers

Ariel Espositoa and Sally M. Bensona

aStanford University, 367 Panama St., Stanford, CA, 94305, USA

Abstract

Maintaining the long term storage of CO2 is an important requirement for a large scale geologic CO2 storage project. Nevertheless, the possibility remains that the CO2 will leak out of the formation into overlying groundwater aquifers. A site specific remediation plan is also important during the site selection process and necessary before storage begins. Due to the importance of protecting drinking water resources, this study analyzes the optimal remediation scenario for various leakage conditions. The three objectives for remediation considered here are removing any mobile CO2, reducing the quantity of CO2 in the reservoir, and reducing the aqueous phase concentration of CO2.

The first part of our research was to determine the processes that control the size and shape of the leakage plume in the groundwater aquifer. We used the multiphase flow simulator TOUGH2 with CO2 leakage from a point source to analyze the plume at various leakage rates. We next determined that during remediation the important physical processes include capillary trapping as a result of hysteresis in the relative permeability curves, dissolution, and buoyancy induced flow. We compared the effectiveness of using vertical and horizontal extraction wells to remove the CO2. We next examined injecting water to dissolve the gaseous CO2 and reduce the overall concentration and increase capillary trapping. Finally, we analyzed the combination of water injection and extraction with multiple wells to determine the optimal spacing and flow rate.

Based on the simulations analyzed for this study, multiple conclusions can be made on the effectiveness of various remediation scenarios. With one vertical extraction well the optimal scenario for the larger leakage cases is a multistep extraction process that removes mobile CO2 from the areas with high gas saturation first. A horizontal extraction well in the middle of the aquifer is much more efficient than vertical wells at removing CO2. Water injection is effective at quickly reducing the mobile phase CO2 with tradeoffs between injection rate and increases in pressure. The most effective scenario over a longer time period includes injection for a short time followed by extraction from four vertical wells. To reduce the CO2 most rapidly, four injector wells with high flow rates and one extraction well is the most effective for the large leakage cases. (©5 2010 Published by Elsevier Ltd. Keywords: Groundwater; remediation; storage

1. Introduction

Carbon capture and storage (CCS) is considered as a viable option for significantly reducing anthropogenic CO2 emissions to the atmosphere from large point sources such as coal-fired or natural gas power plants. Potential storage sites include geologic formations with cap rocks of low permeability to trap CO2 and prevent migration back the atmosphere. However, leakage through this caprock via wells or faults and fractures is possible. Leakage degrades the benefit of reducing emitted greenhouse gases, could result in forfeiture of carbon credits under a carbon trading system and, if not controlled, could ultimately result in the closure of the carbon storage project. Also, if the

doi:10.1016/j.egypro.2011.02.238

CO2 leaks into a groundwater aquifer utilized for drinking water or agricultural purposes it may pose human health risks or damage crops.

Both physical and chemical effects from leakage are possible. Physically, a large plume of gaseous CO2 in a shallow groundwater aquifer could interfere with groundwater extraction and conveyance systems because these systems are usually designed to only pump liquids. Chemical impacts could occur either from CO2 directly, or from displaced brine. When CO2 dissolves in the formation water it forms carbonic acid which reduces the pH and could lead to increased levels of trace metals such as arsenic and lead [1]; [2]; [3]. There are already many areas in the US where the groundwater has arsenic levels above the maximum contaminant level specified by the EPA at 5 ^g/L [4]. For this reason, in areas susceptible to trace metal contamination it is also important to reduce the aqueous CO2 concentration to minimize the reduction in pH and limit the possible increases in trace metal concentrations.

The goal of this study is to identify and compare options for groundwater remediation through the use of numerical simulation. Three objectives were chosen as the most important for reducing the effects of leaked CO2 on the groundwater: reduce the mobile separate phase CO2 and limit continued growth of the leakage plume; remove CO2 from the aquifer in both gas and aqueous phase; reduce the aqueous phase concentration of CO2 and consequently minimize the decrease in pH from the formation of carbonic acid. Three remediation techniques were chosen to meet these objectives: CO2 extraction in both aqueous and gas phase, water injection, and a combination of water injection and extraction.

2. Methods

The multiphase flow simulator TOUGH2 with the ECO2N fluid property module was used to simulate the leakage cases and remediation scenarios [5]; [6]. The TOUGH2 algorithm solves the mass balance equations for each phase over a finite volume for the H2O, NaCl, and CO2 systems. A simple model system is chosen to evaluate the leakage and remediation scenarios. The aquifer is 100 m thick and confined at the top with an impermeable layer. The top of the groundwater aquifer is 100 m below the ground surface and has an initial hydrostatic head of 80 m. The reservoir is homogeneous and anisotropic with a horizontal permeability of 100 md and a vertical permeability of 10 md. The porosity is 15% and the NaCl mass fraction is 0.01. The groundwater aquifer is modeled with both a 2D radial axisymmetric grid and a 3D Cartesian grid as shown in Figure 1.

3000 m

--—-- R=10,000 m —___Deptl:

9=15% Z=100m

XNaCl-0.01

A —s Rleak=5 m —v

k,. = 100 md k. = 10 md

Leak- =25 m2

XNaCl=0.01

^.=100 md k, =10 md

Figure 1. 2D Radial Axisymmetric Grid and 3D Cartesian Grid

Initially there is a hydrostatic pressure gradient with the pressure at the top of the aquifer equal to 0.884 MPa and 1.864 MPa at the base of the aquifer. There is also a temperature gradient of 0.03°C/m which leads to a range of temperature from 23°C at the top to 26°C at the base of the aquifer. The van Genuchten capillary pressure curve was used for the simulation [7]. In Eq. 1 we set A = 0.457, S;r = 0, 5;s = 0.999, Pmax = 1x107 Pa, and P0 = 1.96 x104.

^cap = -P„([ST1/A - 1)1_A with 5* = (5, - Sir)/(Sis - 5(r) subject to - Pma, < Pcap < 0; (1)

The van Genuchten-Mualem model was used for the relative permeability curves (Figure 2) [8]; [7]. We set A = 0.457, 5;r = 0.3, 5;s = 1.0, and S„r = 0.05 (drainage), 0.15 (imbibition).

kri = toi - (1 - tsW} f/ 5, < 5 ; Kg = ^ 1-k;> _ f/ V = 0 (2)

1 if S(>Sis' ra l(l-S)(l-S2) i/ V>0

subject to 0 <krl,krg < 1 with 5= (S, - S,r)/(l - S,r - Sflr),S* = (S, - Slr)/(Sls - S,r) (3)

Figure 2. Relative Permeability Curves

3. CO2 Leakage Scenarios and Simulations

To better understand how leakage rates control the size and shape of the CO2 leakage plume, five leakage scenarios are simulated. Leakage is allowed to continue for five years before the hypothetical leak is detected and the poorly sealed well is plugged. Other leakage scenarios such as leakage along a fault or fracture are certainly possible. Table 1 lists the leakage rate, the total leakage, and the percent of the total geologically stored CO2 at the site that leaks into the groundwater. The leakage percentages are based on a total amount stored of 100 Mt of CO2 which corresponds to storage of 4 Mt of CO2 per year from a 500 MW coal plant for 25 years [9]. Leakage is simulated by imposing a constant mass flux into the base of the aquifer. The CO2 was water saturated at the base of the system. Movement of CO2 in the groundwater aquifer during leakage and remediation is complex and depends on the interplay of many factors. These factors include gravity effects, capillary forces, and viscous forces as well as the impacts from dissolution/exsolution of the CO2 with the water. The final gas saturation for each of the five CO2 leakage cases is shown in Figure 3. The leakage plume can be divided into the primary leakage plume and the secondary leakage plume or gravity tongue which is formed by the accumulation of CO2 at the base of the confining layer. The radius of the gravity tongue ranges from 50 m for the lowest leakage case to 425 m for the highest leakage case. The maximum gas saturation in the gravity tongue directly above the leakage zone ranges from 0.26 to 0.36. Importantly, for all of these scenarios, only about 20% of the CO2 is in gas phase, with the remaining CO2 dissolved in the aquifer.

Table 1. Five Leakage Case Parameters

Case Leakage Rate (kg/s) Leakage Rate (tons/year) Total Leakage Quantity (tons) % of Total

Stored

Case 1 0.006342 200 1,000 0.001%

Case 2 0.015855 500 2,500 0.0025%

Case 3 0.03171 1,000 5,000 0.005%

Case 4 0.06342 2,000 10,000 0.01%

Case 5 0.3171 10,000 50,000 0.05%

Figure 3. Five leakage cases, with the total amount of CO2 leaked from 1,000 to 50,000 tonnes.

4. Groundwater Remediation from Leakage of CO2

The first step of all the active remediation processes is to plug the leaking well and stop the flow of CO2 into the reservoir. During remediation, water flowing into the CO2 plume and changes in pressure are two processes that strongly affect the movement of CO2 and the fraction of CO2 in gas and aqueous phase. Under these conditions, the complex interaction of hysteretic multiphase flow, mass transfer between phases, gas compressibility and advective flow makes remediation challenging. When pressure decreases, the saturation of separate phase CO2 increases, due to a combination of decreasing the density of CO2 and exsolution of CO2 from the aqueous phase. Conversely, when pressure increases more CO2 dissolves in the groundwater, the density of CO2 increases, and the gas saturation decreases. When undersaturated water comes into contact with separate phase CO2, the CO2 plume dissolves. Depending on whether fluid is being injected or extracted, dissolution will either take place at the injection well or at the leading edge of the CO2 plume. Finally, it is not possible to completely extract all of the separate phase CO2 by advection alone because of residual trapping of CO2 [10].Some fraction of the gas is trapped due to snap-off and bypass of the non-wetting fluid during imbibition [11]. The only method to physically remove this CO2 is to dissolve it in water.

4.1 Remediation Using Extraction Wells

After the leak is stopped, the first step in an extraction scenario is to drill a vertical or horizontal well that penetrates the CO2 plume. After the well is drilled, extraction of fluid begins immediately. The well will operate until the amount of CO2 remaining in the reservoir is small or meets specific remediation requirements. The extraction wells operate with the deliverability well model found in the TOUGH2 Users Guide [5] with a prescribed flowing bottomhole pressure constraint and a productivity index [12].

The first scenario examined is for Case 1 (1,000 tons of CO2) where an extraction well is added that is fully screened. Over the five year extraction period 99% of the CO2 is removed as shown by the green line in Figure 4. First, the mobile phase CO2 flows rapidly into the well due to the pressure gradient formed by the pressure drop in the well. After producing the mobile gas, the remaining mobile CO2 is trapped at the residual gas saturation of 15%. Next, the leading edge of the plume is dissolved by water flowing towards the extraction well until all the CO2 is removed. Produced water is able to bypass the plume at the base of the aquifer because the rapid flow rate of CO2 gas from the high gas saturation pocket above the leakage plume leaves an area with relatively lower gas saturation. Partially screening the well over the top 90 m instead of the entire 100 m avoids water bypass of the plume and reduces the total amount of water produced, without decreasing the efficiency, shown by the purple line in Figure 4.

Vertical Extraction z=10 to z=100 m

eg s "S 'S EE

1000 800 600 400 200 0

Total Zw=90 Total Zw=100

Time (Years)

Figure 4. Case 1: Vertical Extraction with two well screening depths

Screening the well the entire depth of the aquifer was found to be ineffective for cases where a large gravity tongue had formed. For these cases when the well is screened over the entire depth, water bypass of the main plume becomes even more pronounced and after a short period the removal of CO2 decreases significantly. In an attempt to reduce water bypass, we developed a multi-step approach with extraction first from the areas with high gas saturation to reduce differences in the saturation along the depth of the aquifer. Once the plume is at a consistent CO2 saturation, then water will not bypass certain zones and will more uniformly remove CO2 with a well screened the entire depth. As an example, the following three step process was analyzed for Case 3 (5,000 tons). The extraction well was first screened from 40 to 90 m. After three years of extraction all the remaining mobile CO2 is at the top of the aquifer. To remove the mobile gas remaining fluid is extracted from the top 5 m of the aquifer for 325 days. Now the leakage plume is at consistent gas saturation close to the residual of 15%. The third stage is screening the well over the entire depth of the aquifer. However, even during this third phase the extraction of CO2 is very slow with approximately 240 tons removed per year. If the flow rate for the third step continues, this process will result in a total remediation time frame of approximately 21 years. Although this was determined to be a relatively efficient scenario by reducing the amount of water bypass, the remediation time frame is very long and consequently other options are considered.

To overcome the difficulties of extraction with vertical wells, we examine the use of horizontal wells for removing the plumes of CO2 using the 3D Cartesian grid. The optimal depth of the horizontal well was analyzed for Case 4(3D) (10,000 tons) by placing the horizontal well at multiple depths in the aquifer (Figure 5).The well at in the middle of the aquifer (z=50 m) leads to the highest removal of CO2. The extraction rate for a horizontal well closer to the bottom of the aquifer (z=30 m) is much faster than all the other well depths up until leveling off at 12 years. This is due to the fact that the pressure drawdown is the greatest. However, after 12 years continuing to extract fluids from the bottom of the aquifer is very ineffective because the remaining CO2 is at the top of the aquifer.

Case 4(3D) Total CO2 Remaining

10,000 8,000 6,000 4,000 2,000 0

Time (years)

Figure 5. Case 4(3D) Total CO2 remaining with various horizontal well depths

4.2 Remediation Using Water Injection Wells

The second remediation technique evaluated is to inject water into the aquifer with the goal of immobilizing the separate phase CO2 and then dissolving it. If more water is injected after all the CO2 is dissolved in the water, the concentration of CO2 in the dissolved phase can be reduced due to advective mixing and diffusion. For the example presented here, 25 kg/s of water with a 0.01 mass fraction of NaCl is injected into the aquifer with a constant flow rate that is evenly distributed over the thickness of the aquifer. A lower flow rate of 5 kg/s was also investigated because the buildup in pressure with an injection rate of 25 kg/s was quite large. As expected, the injection rate has a significant impact on the time to dissolve the gas phase CO2 (Figure 6). If the primary objective of remediation is elimination of separate phase CO2, water injection is very effective.

Case 3: Injection Flow Rate Comparison

CS 30%

Ü g 25%

N 20%

u m SS 15%

— o JS P4 10%

c nj 5%

nj 0%

25 kg/s 5 kg/s

Time (years)

Figure 6. Mass Fraction in Gas Phase with water injection at 25 kg/s and 5 kg/s

4.3 Remediation Using a Combination of Injection and Extraction Wells

However, if the ultimate goal of remediation is to remove CO2 from the aquifer, dissolving all the CO2 to immobilize it could make subsequent extraction more difficult for two reasons. First, the CO2 is displaced away from the initial leakage site during injection, forming a donut shaped plume with an outer radius dependent on the flow rate and the duration of water injection. Second, part of the reason that the CO2 dissolves is because the pressure increases while water is injected into the aquifer. Once an extraction well is added the pressure drops again and the dissolved CO2 returns to mobile gas phase but farther away from the leakage site in the donut shaped plume.

The first technique analyzed that combines injection and extraction includes injection from one well followed by extraction from multiple wells. Case 4(3D) with 10,000 tons of leaked CO2 and a large gravity tongue is used to analyze the effectiveness of the multiple well scenarios. The injection well is placed in the center of the plume at and

is screened the entire depth of the aquifer. The injection rate is 15 kg/s of brine for 80 days. Four extraction wells are added at the interior edge of the donut shaped plume. The wells are screened the entire depth of the aquifer and operated for 15 years. The scenario with four extraction wells operating after water injection is very effective at reducing the total quantity of CO2 in the aquifer (Figure 7). After only 10 years of operation, only 2.28% of the total leaked of CO2 remains. This compares very favorably to the best horizontal extraction case, where only 80% of the CO2 is removed after 15 years.

Case 4(3D): Injection then Extraction from Four Wells

10,000

s © 6,000

u 2,000

Aqueous Total

6 8 Time (years)

Figure 7. Case 4-3D Injection then Extraction from Four Wells

The final combination method is simultaneous injection and extraction using multiple wells. For this case, the extraction well is at the center of the leakage plume and the injectors are around the outer boundaries and operate continuously. Multiple injection well flow rates of 5 kg/s, 2.5 kg/s, and 1 kg/s were analyzed for Case 4-3D (10,000 tons) to determine the optimal scenario. The wells are 300 m apart forming a diamond centered around the extraction well. The total amount removed over the 15 year remediation period for the three different injection flow rates is shown in Figure 8. The higher water injection rate of 5 kg/s leads to the most CO2 removed.

Five Spot: Total CO2 Remaining

10,000 8,000 6,000 4,000 2,000 0

1 kg/s 2.5 kg/s 5 kg/s

5.00 10.00

Time (years)

Figure 8. Five Spot: Total CO2 Remaining (Dw=300 m) 5. Conclusion

Remediating large accumulations of CO2 in groundwater aquifers was shown to be challenging due to complex multi-phase flow behavior. The bifurcated saturation gradient makes it difficult to extract CO2 efficiently from one vertical extraction well screened the entire depth of the aquifer if there is a large gravity tongue. The area with high gas saturation directly above the leakage zone is quickly depleted of gas phase CO2 when an extraction well is added. This allows for water to bypass the main portion of the plume and leads to inefficient removal of CO2. This study suggests that a multi-stage extraction process is more efficient for removing CO2 from the aquifer using

vertical extraction wells. Horizontal extraction wells are more effective than vertical extraction wells for larger leakage plumes at reducing the total mass of CO2 in the aquifer. A horizontal well placed at the middle of the aquifer is the most effective at removing CO2 over a longer time frame compared to other depths.

Injecting water into the gaseous leakage plume is the most effective measure to quickly reduce the mobile phase CO2. Water injection can dissolve all the CO2 in months to less than a year with relatively high injection rates. The time to dissolve the CO2 increases with lower flow rate, but the pressure increase is much less.

The most effective scenarios to remove CO2 from leakage plumes with large gravity tongues combine injection and extraction from multiple wells. The scenario with the highest percentage removal rate begins with water injection into the center of leakage plume for a short period of time followed by CO2 removal with four extraction wells. Continuous water injection and extraction was also shown to be effective for large leakage cases. With a well spacing of 300 m higher flow rates performed better than the lower flow rate.

Remediation of CO2 leakage into a groundwater aquifer is a challenging activity that will vary for each leakage event based on the leakage rate of CO2, the time before the CO2 leak is stopped, and the permeability distribution in the aquifer. To reduce risks and decrease remediation difficulties, frequent monitoring of all aquifers overlying the leakage area is recommended. The sooner the leak is detected and stopped, the easier remediation becomes.

Reference List:

-[1] Wang, S. and P. R. Jaffe. 2004. Dissolution of a mineral phase in potable aquifers due to CO2 releases from deep formations; effect of dissolution kinetics. Energy Conversion and Management. 45(18-19): 2833-2848. -[2] Zheng, L. and J.A. Apps, Y. Zhang, T. Xu, J.T. Birkholzer. 2009. Reactive Transport Simulations to Study Groundwater Quality Changes in Response to CO2 leakage from Deep Geological Storage. Energy Procedia 1(1): 1887-1894.

-[3] Kharaka, Y.K. and J. Thordsen, E. Kakouros, G. Ambats, W. Herkelrath, S. Beers, J. 2010. Changes in the chemistry of shallow groundwater related to the 2008 injection of CO2 at the ZERT field site, Bozeman, Montana. Environ Earth Sci. 60:273-284

-[4] Reilly, T.E. and K.F. Dennehy, W.M. Alley, and W.L. Cunningham. 2008. Ground-Water Availability in the United States: U.S. Geological Survey Circular 1323, 70p.

-[5] Pruess, K. and C. Oldenburg, G. Moridis. 1999. TOUGH2 User's Guide Version 2.0. Berkeley, California: Lawrence Berkeley National Labs Report LBNL-43134, Earth Science Division, University of California, Berkeley -[6] Pruess, K. 2005. ECO2N: A TOUGH2 Fluid Property Module for Mixtures of Water, NaCl and CO2. Berkeley, California: Lawrence Berkeley National Labs Report LBNL-57952, Earth Science Division, UC, Berkeley -[7] van Genuchten, M.Th. 1980. A Closed-form Equation for Predicting the Hydraulic Conductivity of Unsaturated Soils. Soil Sci. Soc. Am. J., 44: 892-898.

-[8] Mualem, Y. 1976. A New Model for Predicting the Hydraulic Conductivity of Unsaturated Porous Media. Water Resour. Res., 12(3): 513 - 522.

-[9] Ansolabehere, S. and J. Beer, J. Deutch, A.D. Ellerman, S.J. Friedmann, H. Herzog et al. 2007. The Future of Coal: Options for a Carbon Constrained World. Massachusetts Institute of Technology, Cambridge MA. -[10] Benson, S.M. and R. Hepple. 2005. Prospects for Early Detection and Options for Remediation of Leakage from CO2 Storage Projects. Carbon Dioxide Capture for Storage in Deep Geologic Formations, Volume 2, D.C. Thomas and S.M. Benson (eds.), 2005 Elsevier Ltd.

-[11] Holtz, M.H. 2002. Residual Gas Saturation to Aquifer Influx: A Calculation Method for 3-D Computer Reservoir Model Construction. SPE Gas Technology Symposium, 30 April-2 May 2002, Calgary, Alberta, Canada -[12] Coats, K.H. 1977. Geothermal Reservoir Modeling. SPE-6892. 52nd Annual Fall Technical Conference and Exhibition of the SPE, Denver, Colorado.