CO2 Leakage Behaviors in Typical Caprock-Aquifer System during Geological Storage Process.

In this study, a 3D reactive flow simulation model is built to simulate the leakage processes though assumed leakage channels. The geochemical reactions are coupled with fluid flow simulation in this model with consideration of reservoir minerals calcite, kaolinite, and anorthite. As an essential trigger for geochemical reactions, changes in pH value are investigated during and after the CO2 injection process. By comparing CO2 migration with/without geochemical reactions, the influence of geochemical processes on CO2 leakage is illustrated. The leakage behaviors through leakage channels with different permeabilities are evaluated. Influence of reservoir temperature on CO2 leakage is also exhibited. Furthermore, the effects of the distance between the injection well and leakage zone on the leakage potential are studied. The results indicate that the geochemical reactions have impact on the leakage processes, which can decrease the leakage level with the presence of geochemical reactions. The region of low pH enlarges with continuous injection of CO2. Hence, monitoring changes in pH can reflect the migration of CO2, which can provide an alert for CO2 leakage. The occurrence of the leakage phenomenon is postponed with increasing the distance between the CO2 injection well and the leakage channel. However, the leakage level tends to be consistent with injecting more CO2. The CO2 leakage risk can be reduced through the leakage channels with lower permeability. With the presence of higher reservoir temperatures, the leakage risk can be improved. These results can provide references for the application of monitoring methods and prediction of CO2 front associated with geochemical processes.

Introduction

Global warming is becoming a serious threat to the environment, which are caused by the significant man-made emissions of greenhouse gases.[1,2] CO2 is definitely the representative of greenhouse gases due to human’s overreliance on traditional fossil fuels (coal, oil, and natural gas).[3] Greenhouse gas control has become a focal point of the 21st Century. In the 2015 COP21, also known as the 2015 Paris Climate Conference, the issue of greenhouse gas control was also emphasized and attracts more and more attention from many aspects of the society.[4] An important route in dealing with this issue is via a CO2 geological storage as part of carbon capture, utilization, and storage (CCUS) systems.[5,6] There is a strong case for the need of CCUS in the future energy mix if modern human society wishes to maintain economic development following its historic and current dependence on traditional fossil energy. There are different kinds of geological structures, such as deep saline aquifers, depleted oil and gas reservoirs, and deep unminable coal seams,[7−9] that can be storage candidates for the CO2 geological storage process. CO2 geological storage associated with deep saline aquifers has a promising future because of its wide distribution and enormous storage potentials. Many demonstration projects conducted in recent years have proved that the CO2 storage in deep saline aquifers can be an effective method in eliminating the negative effects of greenhouse gas emissions. However, the safety, with regard to CO2 leakage, has always been an eminent issue for public concern, with many potential pathways and unknown factors. The rock-fluid interactions may change the property of the caprock and weaken the seal layer. Hence, it is very important to monitor CO2 migration and investigate the leakage mechanisms during CO2 storage process.[10] Meanwhile, geochemical processes could influence the CO2 migration front, which may affect the timeliness of monitoring technologies, especially in a large scale. Geochemical reactions can change the permeabilities of storage reservoirs and leakage channels, which serve a key role in altering CO2 migration rules.

In the early stage, CO2 geological storage capacity and mechanisms are hot issues during the application of CCUS. The available storage structures mainly include deep saline aquifer, unminable coal seams, and depleted oil and gas reservoir etc.[7,11−13] Combination of CO2 storage and enhanced oil recovery (EOR) is also welcome, with benefits in both environment and economic efficiency.[14] At the same time, it is theoretically feasible to inject CO2 to replace CH4 in natural gas hydrate deposits, which can serve as an alternative way to develop unconventional energy and eliminate CO2 emissions.[15,16] Evaluation on CO2 geological storage potential indicates that different storage geostructures have different storage capacities with various storage mechanisms. However, the leakage risks using different storage structures have diverse levels. The technical feasibility of CCUS has been demonstrated by field applications, but the economic and safety issues restrict the further development of CO2 geological storage. Implementation of carbon tax can improve the economic efficiency. Nevertheless, the public always worries and doubts about the potential leakage risk associated with CO2 geological storage projects. CO2 leakage will eliminate the environmental benefits, which is provided by CO2 geological storage activities or damage the ecosystem.[17,18] Therefore, it is important to investigate the leakage behaviors and mechanisms with the presence of leakage channels.

Some research works have been carried out to study the leakage mechanisms and behaviors. Nordbotten et al.[19] applied a novel framework for predicting the leakage channels connecting multiple subsurface permeable formations. The capabilities of the model on typical field data were demonstrated. Their results showed the flexibility and utility of the solution methods. Song and Zhang[20] gave a comprehensive review of caprock-sealing mechanisms for a CO2 geological storage project. The review exhibited that CO2 leakage can be rapid and catastrophic through faults or fracture networks, whereas diffusive loss is usually low. Jordan et al.[21] built a reduced-order model to predict CO2 and brine leakages along the wellbores into the surface or the overlying aquifer. The influences of wellbore properties and the state of the CO2 plume on leakage profiles are also investigated. The results indicate that minima in flow rates exhibited in the response surface are induced by a complex nonlinear phase behavior along the wellbore. CO2 leakage can be increased with the presence of a shallow aquifer by comparing with the cases that CO2 directly leak to the land surface. Class et al.[22] carried out a benchmark study on the problem related to CO2 storage in geologic formations. They applied different simulators to study the CO2 leakage through an assumed leakage channel, methane recovery enhanced by CO2 injection and a field-scale injection scenario into a heterogeneous formation. A description of the benchmark problems and a brief introduction on participating codes are provided in their study. The results of the benchmark study are also presented and discussed. Zhang et al.[10] illustrated and analyzed the monitoring process of CO2 storage safety and leakage in the CCS demonstration project of the Jilin oilfield in China. They demonstrated that the monitoring of targets should be focused on the reservoir, near-surface, and injection and production systems, as shown by field experience. It can not only prevent CO2 leakage, but also avoid the blind expansion of monitoring program scopes by applying monitoring methods to ensure the integrity of wellbores. Their works can provide valuable guidance for the further enlarged Jilin project and other CO2 EOR and geological storage operations.

In this paper, a geological scale 3D reactive flow simulation model is built to simulate the leakage processes though the assumed leakage channels in typical caprock–aquifer geological storage systems. The dissolution/precipitation reactions are coupled with fluid flow simulations in this model with considerations for reservoir minerals calcite, kaolinite, and anorthite. As an essential trigger for geochemical reactions, the changes in pH value are investigated during and after the CO2 injection process. The influence of the reservoir temperature on CO2 leakage is illustrated. CO2 leakage potential through leakage channels with different permeabilities is evaluated. The effects of the distance between the CO2 injection well and the leakage channel on the leakage potential are investigated.

Reactive Flow Simulation Model

Literature researches indicate that the existence of a leakage channel is key risk point for CO2 storage safety and the leakage mechanisms can be influenced by environmental and anthropogenic factors. It is meaningful to study the dynamic leakage behaviors and their mechanisms with the presence of a leakage channel in typical CO2 geological storage systems. Meanwhile, the study on CO2 leakage can provide a reference for the implementation of CO2 monitoring methods and projects, which is beneficial for increasing the CO2 geological storage safety and enhancing the public confidence in CO2 geological storage. The background data of this study are according to topic 1, which is given at the Workshop on Numerical Models for CO2 Storage in Geological Formations that took place from 2–4 April, 2008, in Stuttgart, Germany. Compositional simulator is adopted to set up the 3D reactive flow simulation model of typical caprock–aquifer system for investigating CO2 leakage behaviors and mechanisms during the CO2 geological storage process.[22] The existence of leakage channels with high permeability inside the caprock or seal layer can induce an upward migration of CO2 and increase the leakage risk potential. Hence, a 3D reactive flow simulation model is built for studying CO2 leakage behaviors and mechanisms in typical caprock–aquifer system as shown in Figure .

Figure 1

3D model to simulate CO2 leakage during CO2 geological storage in a typical caprock–aquifer system.

The blank area between two color layers represents a sealing layer. The assumed aquifer overlaying the caprock is 1000 m in length, 500 m in width, and 30 m in thickness, respectively. The length, width, and thickness of the reservoir for CO2 storage are the same as the aquifer. The CO2 storage reservoir is covered with a caprock of 100 m in thickness. Initial porosity and permeability of the reservoir and aquifer are set to be 0.15 and 20 × 10–3 μm2, respectively. Moreover, the reference depth is set to be 800 m with a reference pressure of 8.5 MPa. All the layers in the reservoir are perforated. Injection of CO2 in a liquid phase continues over a period of 20 years with a constant injection rate of 20,5000 m3/day. Mover, aqueous, and mineral reactions during the simulation process are taken into consideration with the related geochemical reaction rate models as follows.[23]

Equilibrium condition of aqueous reactions in the compositional simulator can be exhibited aswhere is the activity product for aqueous reaction α, and it can be expressed as eq where ak is the activity coefficient for component k; Raq is the amount of aqueous reactions; andis the chemical equilibrium constant of aqueous reaction α, which can be expressed by eq where R is the molar gas constant; T is the temperature; and ΔG0 is the Gibbs energy under standard condition.

The chemical equilibrium constants of geochemical reactions including aqueous reactions and mineral reactions vary with temperature. Therefore, a fourth-order polynomial with varying temperatures is used in this study as shown in eq .

Given that relative slow reaction rate of mineral reactions, the following reactive rate formula is applied in this modelwhere Âβ is the reaction surface area of the mineral reaction β per volume of porous medium rock; kβ is the reaction rate constant of the mineral reaction β; Qβ is the activity product of the mineral reaction; Keq,β is the chemical equilibrium constant of mineral reactions.

Mineral reactions may induce changes on porosity, which will influence the permeability of CO2 geological storage reservoirs and leakage zones. In the compositional simulator, porosity change can be depicted as eq where ϕ is the porosity; r̂k is the mean grain size of mineral particles; N̂k is the amount of mineral particles per rock volume; and rk is the dissolution reaction rate of mineral k.

In order to reflect the relationship between porosity and permeability, the Kozeny–Carman equation is inserted into the reactive simulation model and its expression is as follows[24]where k is the real-time permeability; k0 is the initial permeability; ϕ is the real-time porosity; ϕ0 is the initial porosity; and n is the correlation coefficient. In this study, the value of n is assumed to be 1. With the presence of a stronger relationship between porosity and permeability, the influence of mineral reactions on reservoir properties can be enhanced with the value of n larger than 1. In addition, the equations and correlation coefficients of the aqueous and mineral reactions in this study are shown in Tables –3.

Table 1

Aqueous Species and their Initial Concentrations

ion	H+	Ca2+	SiO2	Al3+	OH–	HCO3–	CO3–
C, mol·L–1	1.00 × 10–7	9.12 × 10–6	2.35 × 10–8	2.32 × 10–11	5.46 × 10–7	2.49 × 10–2	1.17 × 10–5

Table 3

Geochemical Reaction Coefficients

reaction formula	a0	a1	a2	a3	a4
H2O = H+ +OH	–14.92816	4.187619 × 10–2	–1.973673 × 10–4	5.549507 × 10–7	–7.581087 × 10–10
CO2 + H2O = H+ +HCO3–	–6.549243	9.001740 × 10–3	–1.02115 × 10–4	2.761879 × 10–7	–3.561421 × 10–10
CO2 + H2O = 2H+ +CO32–	–17.1572	2.17693 × 10–2	–2.22373 × 10–4	5.77919 × 10–7	–6.25514 × 10–10
calcite +H+ = Ca2+ + HCO3–	2.068889	–1.42667 × 10–2	–6.060961 × 10–6	1.459215 × 10–7	–4.189284 × 10–10
kaolinite +6H+ = 5H2O + 2SiO2 + 2Al3+	9.729544	–9.889756 × 10–2	2.915576 × 10–4	–3.270281 × 10–7	–3.311012 × 10–10
anorthite +8H+ = 4H2O + Ca2+ + 2SiO2 + 2Al3+	31.74573	–2.012538 × 10–1	5.958903 × 10–4	–9.041158 × 10–7	9.153776 × 10–11

Results and Discussion

Influence of Geochemical Reactions on CO2 Leakage

By comparing with the scenario without considering geochemical reactions, the leakage behaviors can be analyzed with the presence of geochemical reactions. At the same time, the mineral trapping mechanism during CO2 geological storage process can also be exhibited at some extent. As shown in Figure , grid (15, 13, 1/5, 1, 1), and grid (15, 13, 1/5, 5, 1) are selected as the no.1 and no.2 reference points for monitoring the change in CO2 saturation, respectively. The distance between the leakage channel and the CO2 injection well is set to 100 m. The permeability of leakage channel is assumed to be 1000 × 10–3 μm2.

Figure 2

The location of reference point.

CO2 saturation decreases from 0.69047 to 0.68997 at the no.1 reference point with considering geochemical reactions after 20 years of CO2 injection. This phenomenon indicates that the beneficial mineral reactions can trap CO2 during CO2 geological storage processes to some extent. Given that the selected CO2 injection period is relatively short and the relationship between porosity and permeability is a weak correlation, the mineral trapping mechanism is not very significant with relatively small changes in CO2 saturation. However, the total storage capacity of CO2 based on mineral trapping mechanisms can be considerable in view of actual injection volume and storage period. Therefore, it is beneficial to enlarge the positive effects of mineral trapping mechanisms on the CO2 permanent storage by mastering the mineral distribution and total amount when selecting storage sites. Meanwhile, the accuracy of the leakage monitoring method can be improved with further insights on corresponding mineral reactions, which causes the delayed leakage with the presence of related geochemical reactions.

Changes in pH

A change in pH is induced after CO2 dissolves in water. pH is an important parameter to evaluate the tendency of different geochemical reactions. At the same time, the change in pH can reflect the CO2 migration front at some extent in CO2 storage projects. Hence, monitoring the change in pH can provide an alert for CO2 leakage and predict the potential zones for special reactions, which need an environment of low pH. The distance between the CO2 injection well and the leakage channel is set to be 100 m. Permeability of the leakage channel is assumed to be 1000 × 10–3 μm2. Reservoir temperature of 60 °C is adopted in this reactive flow simulation model. As can be seen in Figure , a low pH region becomes larger with the continuous injection of CO2 during the storage process and can spread to the aquifer overlying the caprock with the presence of a leakage channel.

Figure 3

Change in pH during the CO2 injection process at reservoir temperature of 60 °C (note: red color of the ruler in the figure represents the pH range of 6.9–7.7. With the hue becoming cold, the pH decreases and the dark blue color indicates a pH range of 0.0 to 0.8.) (a) pH before CO2 injection, (b) pH after 50 days of CO2 injection, (c) pH after 500 days of CO2 injection, and (d) pH after 20 years of CO2 injection

Therefore, monitoring changes in low pH areas can reflect the degree of CO2 escape and can also provide an early warning of some harmful reactions that require a low pH environment. Based on modeling and experimental studies, a decrease in aqueous pH associated with CO2 leakage into the overlaying aquifer will induce increased aqueous concentrations of a wide range of metals (such as Pb, Cd, Cu, Fe, Mn, Zn, Cr, V, and U).[25,26] CO2-induced decrease in pH can also weaken the integrity of the injection and production wells by damaging their cement ring and wellbore.[10] After 20 years of CO2 injection, the pH can be as low as 4.7 in the aquifer overlaying the caprock.

Influence of Leakage Channel Permeability

Leakage channels with different permeabilities can induce CO2 leakage at different scales. It is of great significance to study the influence of permeability of the leakage channel on the CO2 leakage performance during the CO2 storage process. Four kinds of leakage channels with the permeability of 1, 10, 100, and 1000 × 10–3 μm2, respectively, are selected to study the influence of leakage channel permeability on CO2 leakage behaviors. The distance between the CO2 injection well and the leakage channel is set to be 100 m and the reservoir temperature is assumed to be 60 °C.

As can be achieved in Figure , the CO2 leakage risk increases with increasing permeability of the leakage channel, which can be expressed by an enlarged distribution range of CO2 in the overlying aquifer. It is notable that the leakage risk can be reduced significantly with the presence of a relatively lower permeability (10 × 10–3 μm2), which indicates that CO2 leakage can be controlled with blocking the leakage channel to some extent, and it is unnecessary to plug the leakage channels thoroughly. In addition, the leakage control costs can also be reduced with applying a relatively small amount of blocking materials.

Figure 4

CO2 saturation with the presence of leakage channels with different permeabilities after 20 years of CO2 injection

Influence of Distance between Leakage Channel and CO2 Injection Well

For some storage cases, there may exist some potential leakage channels, which can be triggered by the injection activity. These kinds of potential leakage channels can keep a close state without CO2 injection, which provide barriers for the subsurface fluid migration. However, CO2 injection activity may disturb the relatively stable state of leakage channels that leads the leakage channels to an open state. Therefore, the influence of distance between the leakage channel and CO2 injection well on the leakage performance is estimated.

Changes of CO2 saturation in the no.2 reference point with the presence of different distances between the leakage channel and the CO2 injection is applied to investigate the distance influence on leakage behaviors in this study. The leakage channel is assumed to be open in this reactive flow simulation process. The reservoir temperature is set to be 60 °C. The permeability of leakage channels is assumed to be 1000 × 10–3 μm2. The distances of leakage channels and CO2 injection wells are set to be 20 m, 25 m, 30 m, 40 m, 60 m, 75 m, 90 m, and 100 m. As can be achieved in Figure , the leakage is postponed with increasing the distance. However, the leakage levels tend to be consistent with injecting more CO2.

Figure 5

Changes of CO2 saturation in the no.2 reference point with different distances between the leakage channel and the CO2 injection well.

Influence of Reservoir Temperature

It is of a certain guiding significance to investigate the influence of temperature on leakage behaviors for the selection of geological storage reservoirs and understanding the safety of geological storage reservoirs with different temperatures.

Changes of CO2 saturation in the no.2 reference point are applied here to study the influence of reservoir temperature on the leakage performance. The distance between the leakage channel and the CO2 injection well is assumed to be 100 m and the permeability of leakage channel is set to be 1000 × 10–3 μm2. As can be seen in Figure , the CO2 leakage rate accelerates with the increase in reservoir temperature, which can be exhibited by the steepening curve of CO2 saturation. Moreover, the leakage level enlarges with the presence of a higher reservoir temperature that can be achieved by the comparison of the CO2 saturation at higher reservoir temperatures and lower reservoir temperatures after 20 years of CO2 injection in the no.2 reference point. Hence, storage reservoirs with relatively lower reservoir temperatures are recommended to be candidates for CO2 geological storage when other conditions are similar, which may slow down the leakage rate or reduce the leakage risk with the presence of a potential leakage path. However, deeper storage can also decrease the leakage risk of CO2 to the atmosphere. Therefore, it is important to avoid the storage reservoirs with abnormally high temperatures.

Figure 6

Changes in CO2 saturation in the no.2 reference point with different reservoir temperatures

Conclusions

CO2 geological storage has become one of the research priorities worldwide in recent years, which can reduce greenhouse gas emissions. During the CO2 storage process, existing or newly generated leakage channels are inevitable because of natural and man-made geological activities. Hence, the storage safety in terms of CO2 leakage has always been an eminent issue for concern. In this study, a 3D reactive flow simulation model is built to investigate the influence of geochemical reactions, reservoir temperature, and the permeability of leakage channels on leakage behaviors. The influence of distance between the injection well and the leakage channel on leakage risk is also evaluated. At the same time, changes in pH during the leakage process are exhibited. The following conclusions can be drawn from the above study.

The leakage risk can be reduced with the presence of geochemical reactions. Storage reservoirs with suitable minerals are recommended to be the candidates for CO2 geological storage. A low pH region enlarges with the continuous injection of CO2, which can reflect the front of CO2 migration. Therefore, monitoring the change in pH can give an alert on CO2 leakage and predict the potential zone for harmful reactions, which need an environment of low pH.

The leakage level reduces with decreasing the permeability of the leakage channel. Leakage risk can be reduced significantly with the presence of a relatively lower permeability (10 × 10–3 μm2), which indicates that CO2 leakage can be controlled without plugging the leakage channels thoroughly.

The occurrence of a leakage phenomenon can be postponed with increasing the distance between the CO2 injection well and the leakage channel. However, the leakage level tends to be consistent with injecting more CO2.

CO2 leakage rate accelerates with increasing reservoir temperature. The leakage level enlarges with the presence of higher reservoir temperature. Hence, storage reservoirs with relatively lower reservoir temperatures are recommended to be candidates for CO2 geological storage when other conditions are similar.

Table 2

Values of Mineral Reactions Related Parameters

mineral	logK25°C, mol·m–2 s–1	activation energy, J·mol–1	initial reaction specific surface area, m2·m–3	mineral content (%)
calcite	–8.79588	41870.0	88.0	0.88
kaolinite	–13.00	62760.0	17600.0	1.76
anorthite	–12.0	67830.0	88.0	0.88