CO2 Utilization and Geological Storage in Unconventional Reservoirs After Fracturing

Jinzhou Zhao; Lele Wang; Bing Wei; Valeriy Kadet

doi:10.1016/j.eng.2025.01.005

Engineering ›› 2025, Vol. 48 ›› Issue (5) :98 -113. DOI: 10.1016/j.eng.2025.01.005

Research Carbon Capture, Utilization, and Storage—Article

research-article

CO₂ Utilization and Geological Storage in Unconventional Reservoirs After Fracturing

Author information +

History +

PDF (3645KB)

Abstract

Cyclic injection holds great potential for CO₂ emission reduction coupled with enhanced unconventional oil recovery. There is, however, a lack of a thorough understanding of carbon distribution, migration, and transformation underground over time at the reservoir scale. To address this issue, we conducted a rigorous numerical simulation integrating microseismic events, multi-geomechanics, and multi-geochemistry to represent the complex fracture geometry, rock stress sensitivity, and CO₂–oil–brine–rock interactions. The fluid model, reservoir model, and geochemical reaction kinetics were carefully validated and calibrated using experimental data. The performance of CO₂ utilization and geological storage was comprehensively investigated in terms of changes in oil production, CO₂ storage, carbon distribution, and petrophysical properties. The results indicate that 48.3% of the injected CO₂ was stored stably underground after ten cycles (ten years), with a 3.4% increase in oil recovery. The presence of multiple CO₂ storage forms, such as dissolved in water and mineralized carbonate, impeded CO₂–oil interaction, leading to a 25.9% reduction in the volume of the CO₂–oil mixing zone and a 2.2% decrease in cumulative oil production, albeit with a 7.7% increase in the storage rate. The cyclic injection mode had a significant impact on the migration and transformation of CO₂ in the reservoir. While dissolved CO₂ in oil accounted for over half of the total storage, it had the possibility of being released during production. After ten cycles, 20% of the injected CO₂ (approximately 12 000 t) reached long-term storage in four forms: mineralized carbonate (6%), water-dissolved CO₂ (6%), aqueous ions (4%), and trapped gas (4%). Notably, the non-fracture zone within the stimulated reservoir volume (SRV) served as the primary trapping area for residual gas. This work provides valuable insights into dynamic CO₂ transport and transformation processes under cyclic injection and presents a more comprehensive and precise framework for assessing CO₂ capture, utilization, and storage with enhanced oil recovery (CCUS-EOR) performance in unconventional reservoirs after fracturing.

Graphical abstract

Keywords

CO₂ utilization and geological storage / Cyclic injection / Unconventional reservoir / Carbon distribution, migration, and transformation / Emission reduction

Cite this article

Download citation ▾

Jinzhou Zhao, Lele Wang, Bing Wei, Valeriy Kadet. CO₂ Utilization and Geological Storage in Unconventional Reservoirs After Fracturing. Engineering, 2025, 48(5): 98-113 DOI:10.1016/j.eng.2025.01.005

登录浏览全文

4963

注册一个新账户忘记密码

1. Introduction

Global warming is intensifying and poses a significant risk to the sustainable development of humankind. According to the International Energy Agency (IEA), global energy-related carbon dioxide (CO₂) emissions grew by 1.1% in 2023 to a record high of 3.74 × 10¹⁰ t [1]. Reducing greenhouse gas emissions to mitigate climate change has thus emerged as a critical focus for the international community. CO₂ capture, utilization, and storage (CCUS) is an emerging strategy for large-scale CO₂ reduction. This technology is anticipated to facilitate the low-carbon utilization of fossil fuels and address global climate change [2]. Among the ten CO₂ utilization pathways enumerated by Hepburn et al. [3], CO₂-enhanced oil recovery (CO₂-EOR) demonstrates substantial potential for CO₂ utilization and geological storage. CCUS with enhanced oil recovery (CCUS-EOR) involves injecting CO₂ into hydrocarbon reservoirs to alter the physicochemical properties of the reservoir and fluids, thereby boosting crude oil production and achieving net-zero CO₂ emissions [4]. This technology serves as a pivotal technical support for China in achieving its goals of carbon peak and carbon neutrality [5]. It is estimated that approximately 5.1 × 10⁹ t of CO₂ could be sequestered in China using CO₂-EOR technology [6].

CO₂-EOR has been utilized as a tertiary oil recovery technique in conventional reservoirs for many years. Nevertheless, with the depletion of conventional resources, significant attention has shifted to unconventional resources such as tight and shale oil. Unlike conventional reservoirs, tight reservoirs possess extremely low in situ permeability (≤ 0.1 md; 1 md = 0.987 × 10⁻³ μm²), with crude oil residing within tiny micro- and nano-pores [7]. This makes it challenging for conventional methods such as CO₂ flooding to overcome the difficulty of hydrocarbon flow. The emergence of horizontal drilling with multi-stage hydraulic fracturing technology has boosted tight oil production. Hydraulic fracturing establishes numerous channels for hydrocarbon flow from the tight matrix to the wellbore. High oil production is attainable in the initial stage, but the oil rate declines rapidly, resulting in a primary oil recovery less than 8% of the original oil in place (OOIP), while more than 90% of the OOIP remains untapped. Recent experimental studies [8], numerical simulations [9], and field projects [10] have confirmed the success of cyclic CO₂ injection in promoting oil recovery from tight formations. Cyclic CO₂ injection, also called CO₂ “huff-n-puff” (HnP), is a one-well process that includes the three stages of “huff,” “soak,” and “puff.” This process is typically repeated several times until the economic feasibility limits are reached.

Numerous studies regarding CO₂-EOR have focused on the interactions between CO₂ and crude oil, including miscibility, re-pressurization, oil viscosity reduction, oil swelling (OS), evaporation, and molecular diffusion [9], [11]. The reservoir characteristics and development methods of unconventional reservoirs after fracturing have made the mechanisms of CO₂-EOR and geological storage distinct from the traditional understanding [12], [13]. For example, CO₂ flow in tight reservoirs is dominated by fracture flow, with the CO₂ penetrating unfractured rocks, driven by pressure and concentration gradients [12]. The minimum miscibility pressure (MMP) holds little significance in diffusion-dominated ultratight reservoirs, since the MMP relies on advection-dominated conditions [14], and the oil recovery increases with an injection pressure even beyond the MMP [15]. Analogous phenomena have been reported by Tovar et al. [16] and Kumar et al. [17]. The formation of a solvent–oil mixed zone in the injection stage and the expansion of oil along with the evaporation of oil components in the production stage are considered to be the primary mechanisms behind CO₂ HnP [15]. These distinctions require studies that focus more on the material exchange process between the matrix and the fracture. Aside from CO₂–crude oil interactions, CO₂-EOR in unconventional reservoirs relies on reservoir characterization, rock mechanics, and injection strategies. Cyclic-injection-induced effective stress fluctuations significantly affect the physical properties of a reservoir, especially the permeability of artificial fractures. This has a profound impact on production history matching [18] and oil production prediction [19]. In addition, artificial fracture conductivity exhibits hysteresis under multiple loading/unloading cycles, further influencing the fluid transport behavior in fracture-matrix systems [20]. Hence, for a given tight reservoir with artificial fractures, full coupling of the multiphase fluid characteristics and petrophysical properties is particularly essential for an accurate assessment of CO₂-EOR performance.

Early CO₂-EOR processes mainly focused on improving oil recovery, with less attention on CO₂ storage in porous formation [21]. Recent studies have demonstrated that cyclic injection is a viable technology for achieving CCUS-EOR in unconventional reservoirs. Lashgari et al. [22] proposed that 25%–50% of the total CO₂ injected could be trapped in the Middle Bakken through cyclic CO₂ injection. In a life-cycle assessment of Eagle Ford Shale, Bao et al. [23] found that cyclic CO₂ injection resulted in a 20% increase in oil recovery and a 31.3% effective CO₂ sequestration rate. Ning and Tura [24] demonstrated that the efficiency of CO₂ storage in unconventional reservoirs was highly dependent on reservoir conditions, operational parameters, and production strategies.

Although the oil production and storage performance of cyclic gas injection in unconventional reservoirs have been evaluated, yielding valuable insights, these studies significantly simplified the reservoir characterization, fluid–solid interactions, and transport phenomena in order to save computation time. This simplification has restricted the comprehension of the synergistic effects of the CCUS-EOR process in unconventional reservoirs and the transformation of CO₂ storage, along with their contributions to storage capacity. Accurate prediction of the fate of injected CO₂ is essential for designing injection and storage strategies aimed at optimizing oil production and carbon storage efficiency. Consequently, a thorough assessment of CCUS-EOR-related responses is required on a fully coupled physical model that integrates multiphase flow, multicomponent transport, fluid phase behavior, geomechanics, and chemical reactions [25], [26].

The success of conventional geological CO₂ sequestration projects has led to the identification of four main types of subsurface CO₂ storage: structural trapping (hindering CO₂ upward movement), residual trapping (rendering CO₂ immobile due to capillary forces), solubility trapping (dissolving CO₂ in fluid), and mineral trapping (precipitating CO₂ as carbonate) [27], [28]. These four trapping mechanisms correspond to six forms of subsurface CO₂: supercritical mobile gas, CO₂ physically dissolved in oil and water, aqueous ions, trapped gas, and mineralized CO₂. The state of the CO₂ is significantly influenced by multiple factors, including temperature, pressure, salinity, CO₂ injection scenario, reservoir heterogeneity, and geomechanics [29], [30], [31], resulting in transformations among different CO₂ storage forms. For example, pressure-sensitive storage mechanisms, such as solubility trapping, are affected by pressure fluctuations from the intermittent mode, which complicates the migration and transformation of CO₂ in the subsurface. Moreover, mineral precipitation/dissolution, especially in reservoirs containing high feldspar and chlorite contents, can alter the rock’s physical properties (e.g., porosity, permeability, and wettability), thereby affecting fluid flow behavior and CO₂ distribution [32]. Hu and Rui [33] employed the TOUGHREACT simulation program to investigate the carbon migration and phase distribution patterns in CO₂ geological utilization and storage. They found that the forms of CO₂ storage were dynamically transforming, emphasizing that accurately quantifying the transformation and contribution of various CO₂ storage mechanisms under in situ conditions is crucial for comprehending the CO₂ transport process and assessing storage potential. However, to date, few studies have systematically investigated the dynamic CO₂ transport behaviors and the CCUS-EOR response during cyclic injection in unconventional reservoirs after fracturing.

In short, although recent works have demonstrated the potential of cyclic CO₂ injection for CCUS-EOR in unconventional reservoirs, few models have fully integrated the multi-geomechanical and multi-geochemical impacts, resulting in poor elucidation of the dynamic carbon distribution, migration, and transformation underground. To bridge these knowledge gaps, we have developed a robust component model of a specific tight reservoir considering these important impacts. The primary objective of this work was to assess the CCUS-EOR response in this reservoir by capturing detailed changes with respect to oil production, CO₂ storage, and petrophysical properties. The reservoir model and fluid model were well-calibrated using experimental data. Rock stress sensitivity and fracture permeability hysteresis were used to account for the influence of geomechanics, and a complex geochemical reaction kinetics model related to the mineral composition was appropriately set in the modeling to capture tiny changes in rock volume. Our model reveals the dynamic migration and transformation of CO₂ and its storage contribution over multiple cycles, shedding light on the interaction mechanisms between CO₂-EOR and geological storage and clarifying the process of CO₂ transport underground. This comprehensive understanding offers novel insights for the application of CCUS-EOR in unconventional reservoirs.

2. Reservoir description

The targeted Triassic Baikouquan Formation in Mahu Sag is located in Junggar Basin, northwestern Xinjiang, China. In terms of its lithological characteristics, this formation is divided into three layers. The upper layer is characterized by the presence of a set of mudstone caprocks with a certain thickness, while the lithology of the middle and lower layers is mainly composed of sandstone and conglomerate [34]. The target area is a typical tight reservoir with an average porosity and permeability of 9.2% and 0.11 md, respectively. The oil saturation is 52%–64%, the average gas–oil ratio (GOR) is 108 m³·m⁻³, and the average dead oil density is 826 kg·m⁻³.

3. Methodology

A reservoir model was established by means of CMG–GEM software to simulate the interactions among CO₂, crude oil, brine, and rock [35]. Further evaluation was conducted on CO₂-EOR and CO₂ storage during the CO₂ cyclic injection process. CO₂ solubility, molecular diffusion, gas relative permeability hysteresis, geochemical reactions, and porosity/permeability change caused by multi-geomechanical and multi-geochemical impacts were all integrated into the modeling, as described below.

3.1. Reservoir model

The reservoir model was developed based on a typical horizontal well drilled in the Baikouquan Formation in the Mahu Sag. The microseismic data monitored during horizontal well fracturing was used to create hydraulic fractures, as shown in Fig. 1(a). The model domain of 905.1 m × 230.6 m × 52.6 m was meshed using Cartesian grids of 60 × 45 × 3 in the I, J, and K directions, respectively (Fig. 1(b)). The geometry and direction of the fractures were described using microseismic events. The distribution of the stimulated reservoir volume (SRV) is shown in Fig. 1(c). The logarithmically spaced-locally refined (LS-LR) method was adopted to refine the grids where fractures are located into 5 × 5 × 1 along the I, J, and K directions. The SRV was divided into a fracture zone and a non-fracture zone. Table 1 lists the parameters used for the simulation.

3.2. Fluid composition and model

The ion composition of the formation water was determined by means of an IC761 ion chromatograph with a total salinity of 11 776.33 mg·L⁻¹. Detailed information is provided in Table 2. The composition of the crude oil was identified through gas-chromatography analysis. More than 35 components were lumped into seven pseudo-components to improve computational efficiency, as listed in Table 3. The Peng–Robinson equation of state (PR-EOS) in CMG-WinProp was employed to calculate the fluid properties and phase equilibrium [36]. In order to accurately describe the phase behaviors and fluid interactions between oil and gas and between brine and gas, relevant tests were conducted, including constant composition expansion (CCE), OS, CO₂ solubility, and MMP. State parameters such as critical pressure, critical temperature, omega A, omega B, and binary interaction coefficients were carefully tuned until a satisfactory agreement between the measured and simulated values was attained, as depicted in Figs. 2(a) and (b). The Jossi–Stiel–Thodos viscosity correlation was used to match the viscosity of the crude oil [36]. The measured and calculated values of the fluid properties are given in Table S1 in Appendix A. It was assumed that the gaseous and aqueous phases were in thermodynamic equilibrium, and the fugacities of CO₂ in the gas and aqueous phase were equal (Eq. (1)). The PR-EOS and Henry’s law (Eq. (2)) were respectively applied to calculate the fugacities of CO₂ in the gas and aqueous phases. The reference Henry’s constant and the partial molar volume of CO₂ in water at infinite dilution were essential parameters used to fit the CO₂ solubility experimental data, as shown in Fig. 2(c).

(1)

f_{i g} - f_{i a} = 0

where f_i_g is the fugacity of component i in the gas phase; and f_i_a is the fugacity of component i in the aqueous phase.

(2)

f_{C O_{2} (a)} = y_{C O_{2} (a)} \cdot H_{C O_{2}}

Henry’s law constant is calculated as follows [37]:

(3)

ln(H_{C O_{2}}) = ln(H_{{CO}_{2}}^{*}) + \frac{{\bar{v}}_{C O_{2}} \cdot (p - p^{*})}{RT}

where

f_{C O_{2} (a)}

is the fugacity of CO₂ in the aqueous phase;

y_{C O_{2} (a)}

is the mole fraction of CO₂ in the aqueous phase;

\ln (H_{C O_{2}})

is the natural logarithm of the Henry’s constant for CO₂ (

H_{C O_{2}}

);

\ln (H_{{CO}_{2}}^{*})

is the natural logarithm of the reference Henry’s constant for CO₂ at the reference condition;

{\bar{v}}_{C O_{2}}

is the partial molar volume of the components in water at infinite dilution; p is the pressure; p* is the reference pressure; R is the universal gas constant; and T is the temperature.

3.3. Rock stress sensitivity

Fluid injection or production modifies the pore pressure and effective stress, causing rock dilation or compaction, further affecting the formation porosity and permeability. This phenomenon becomes more frequent and pronounced with cyclic injection. The changes in matrix and fracture permeability follow different paths under loading and unloading cycles. The main reason is that hydraulic fractures undergo plastic deformation due to the inelastic responses of the proppant, and the permeability evolution displays obvious hysteresis. Describing this complex evolutionary behavior using geomechanics at the field scale poses a challenge, as it demands the development of a constitutive model for the inelastic response of the proppant. Instead, the multiplier correlations of the pressure-dependent porosity and permeability offer a satisfactory description of the hysteresis behavior of the fracture permeability [20], [35], as depicted in Fig. S1 in Appendix A. The figure includes one main depletion path and two hysteresis paths at the upper and lower pressure boundaries; the paths for the intermediate pressures are determined through linear interpolation. The propagation of artificial fractures falls outside the scope of this work.

3.4. Geochemical reactions and models

Geochemical reactions are used to describe and predict the physical and chemical evolution in porous media. Aqueous reactions are governed by thermodynamic principles and are typically described through equilibrium constants obtained from experimental data or thermodynamic databases. The reactive transport model, like mineral reactions, usually requires the introduction of kinetic models to simulate the rates of mineral dissolution/precipitation reactions over time. Geochemical reactions coupled with convective-diffusive transport processes can capture the spatial and temporal evolution of minerals and the variations in porosity and permeability.

3.4.1. Aqueous reactions

A portion of dissolved CO₂ in the formation water participates in reactions, as shown in Eqs. (4), (5), (6), thus generating a substantial quantity of bicarbonate ions and a minor quantity of carbonate ions. The ion reactions occurring within the aqueous phase are notably rapid processes. It is reasonable to assume that these reactions are controlled by chemical equilibrium, as given in Eqs. (7), (8), (9), (10), (11). The aqueous reactions of brine components are presented as follows:

(4)

C {O_{2}}_{(aq)} + H_{2} O ⇌ H^{+} + HC O_{3}^{-}

(5)

C O_{3}^{2 -} + H^{+} ⇌ HC O_{3}^{-}

(6)

H^{+} + O H^{-} ⇌ H_{2} O

The equations for the chemical equilibrium are as follows [38]:

(7)

Q_{α} - K_{eq, α} = 0, α =1, ..., R_{aq}

(8)

Q_{α} = \prod_{i =1}^{n_{aq}} a_{i}^{v_{i α}}

(9)

a_{i} = γ_{i} m_{i}, i = 1, ..., n_{aq}

(10)

\log γ_{i} = - \frac{A_{γ} z_{i}^{2} \sqrt{I}}{1+ {\dot{a}}_{i} B_{γ} \sqrt{I}} + \dot{B} I

(11)

I = \frac{1}{2} \underset{i = 1}{\sum^{n_{aq}}} m_{i} z_{i}^{2}

(12)

\log (K_{eq}) = a_{0} + a_{1} T + a_{2} T^{2} + a_{3} T^{3} + a_{4} T^{4}

where

Q_{α}

is the activity product of the aqueous reaction α; K_eq,_α is the chemical equilibrium constant of the aqueous reaction α; R_aq is the number of intra-aqueous chemical equilibrium reactions; n_aq is the number of aqueous components; a_i is the activity of component i; v_iα represents the stoichiometry coefficients of component i of the aqueous reaction α; γ_i is the activity coefficient of component i; m_i is the molality of component i;

A_{γ}

B_{γ}

, and

\dot{B}

are temperature-dependent parameters; z_i is the ion charge of component i;

{\dot{a}}_{i}

is the ion size parameter; I is the ionic strength; a₀–a₄ are the chemical equilibrium coefficients.

3.4.2. Mineral reactions

The dissolved CO₂ will react with formation rocks under acidic conditions to generate stable carbonate minerals, thereby achieving permanent carbon sequestration. This is a long-term and slow process. Transition state theory (TST) is employed to describe the reaction kinetics of mineral dissolution and precipitation. The reaction rate of minerals is determined by the reaction surface area (RSA), reaction rate constant, activity product, and chemical equilibrium (Eq. (13)) [27], [38]. The direction of the mineral reaction is governed by the ratio Q_β/K_eq,_β, which is referred to as the saturation index (SI). Mineral precipitation occurs when the SI is greater than 1, and mineral dissolution occurs when the SI is less than 1 (Eq. (14)). The temperature-dependent rate constant can be expressed as Eq. (15). The RSA of minerals is another important parameter that significantly affects the reaction rate; it varies with the dissolution and precipitation of minerals, as shown in Eq. (16). The determination of the reaction kinetics parameters is elaborated in Section 3.4.4.

(13)

r_{β} = {\hat{A}}_{β} k_{β} (1 - \frac{Q_{β}}{K_{eq, β}}), β = 1, ..., R_{mn}

(14)

SI = \frac{Q_{β}}{K_{eq, β}} \{\begin{matrix} > 1, m i n e r a l p r e c i p i t a t i o n \\ < 1, m i n e r a l d i s s o l u t i o n \end{matrix}

(15)

k_{β} = K_{0 β} \exp [- \frac{E_{a β}}{R} (\frac{1}{T} - \frac{1}{T^{*}})]

(16)

{\hat{A}}_{β} = {\hat{A}}_{β}^{0} \cdot \frac{N_{β}}{N_{β}^{0}}

where r_β is the reaction rate of the mineral reaction β;

{\hat{A}}_{β}

is the RSA of the mineral reaction β; k_β is the rate constant of the mineral reaction β; Q_β is the activity product of the mineral reaction β; K_eq,_β is the chemical equilibrium constant of the mineral reaction β; R_mn is the number of mineral reactions; E_a_β is the activation energy of the mineral reaction β; T* is the reference temperature (usually 25 ^oC); k₀_β is the reaction rate constant of the mineral reaction β at T*;

{\hat{A}}_{β}^{0}

is the initial RSA of the mineral reaction β; N_β is the mineral mole number of the mineral reaction β per unit grid block volume at the current time; and

N_{β}^{0}

is the initial mineral mole number of the mineral reaction β per unit grid block bulk volume.

3.4.3. Porosity and permeability evolution due to geochemical reactions

Mineral dissolution and precipitation can also alter the void volume of the porous medium, leading to variations in porosity. They can be calculated through the following equations [39]:

(17)

{\hat{ϕ}}^{*} = ϕ^{*} - \underset{β =1}{\sum^{n_{m}}} (\frac{N_{β}}{ρ_{β}} - \frac{N_{β}^{0}}{ρ_{β}})

(18)

ϕ = {\hat{ϕ}}^{*} [1 + c_{ϕ} (p - p^{*})]

where

ϕ

is the actual porosity;

ϕ^{*}

is the reference porosity without mineral precipitation/dissolution;

{\hat{ϕ}}^{*}

is the reference porosity including mineral precipitation/dissolution; n_m is the number of minerals; ρ_β is the mineral molar density of the mineral reaction β; and

c_{ϕ}

is the rock compressibility.

The permeability changes induced by mineral dissolution and precipitation are modeled through the Kozeny–Carman equation [39]:

(19)

\frac{k}{k^{*}} = {(\frac{ϕ}{ϕ^{*}})}^{3} \cdot {(\frac{1 - ϕ^{*}}{1 - ϕ})}^{2}

where k is the actual permeability; k* is the reference permeability without mineral precipitation/dissolution.

Fig. 3 presents an iterative procedure of coupling among the fluid flow, rock stress sensitivity, and geochemistry modules. In the coupling process, the reservoir pressure, temperature, and saturation distributions are initially computed based on the fluid flow equations in the reservoir simulator. Subsequently, the solutions from the fluid flow module are passed to the rock stress sensitivity and geochemical reaction modules to calculate the pressure-dependent porosity and permeability multipliers and the mineral molar change, respectively. The actual porosity and permeability values are updated by coupling these two variables, and the results are fed back into the reservoir simulator to acquire new pressure and temperature values. This iterative process is executed over time steps within a convergence criterion until the maximum time step is reached.

3.4.4. Reaction kinetics parameters

Previous studies have demonstrated that various factors affect geochemical reactions, including temperature, pressure, aqueous ion, rock mineral composition, RSA, activation energy, and reaction rate constant [40]. In our work, all uncertain parameters were carefully verified to improve the reliability of the simulation results. The mineral composition of the target area was derived from the average mineral content of similar layers in the Baikouquan Formation (T₁b) [41], as shown in Table 4. Quartz (43%), feldspar (42.5%), and clay (14.5%) are the dominant minerals in the studied reservoir. The RSA of the minerals is an important but uncertain parameter that depends on the heterogeneous distribution, particle size, shape, and complex contact interfaces of all minerals on the rock surface. Several studies have used specific RSA values to simulate mineral reactions, ignoring the heterogeneity of the accessible surface area among different minerals [42]. The geometric surface area (GSA) method based on ideal geometry, the specific surface area (SSA) measured by the Brunauer–Emmett–Teller (BET) approach, and the scanning electron microscopy (SEM) method for image analysis are universal methods used to characterize mineral RSAs [43]. However, the results measured by different methods may significantly vary in magnitude. It is challenging to acquire the RSA of each mineral in specific rocks because different types of minerals, particularly clay minerals, have large differences in particle size. Here, the RSA of each primary mineral was determined by combining the GSA and BET methods. The grain size range and BET limits for each pure mineral were derived by referring to the experimental BET data collected by Beckingham et al. [43], as presented in Table 4. The GSA was first calculated by assuming that each mineral is composed of spherical grains with different diameters. To eliminate the deviation caused by the assumption of ideal spherical grains, the roughness coefficient (rf) was introduced to ensure that the calculated SSA was within the BET range of pure minerals. Meanwhile, the total SSA of specific rocks (SSA*) was required to be consistent with the average BET result (4.0 m²·g⁻¹) of the Baikouquan Formation (T₁b) [41]. The final mineral RSAs were obtained by following Eqs. (20), (21), (22), (23). In our model, the initial RSA of the secondary minerals was set to 1 m²·m⁻³ to ensure that their precipitation reactions could be initiated.

(20)

GSA = \frac{A_{s}}{m} = \frac{3}{r ρ_{m}}

(21)

SSA = GSA \cdot rf

(22)

SS A^{*} = SSA \cdot V_{m}

(23)

RSA = SS A^{*} \cdot ρ_{m}

where A_s is the total surface area of the mineral grain; m is the grain mass; r is the grain radius; ρ_m is the mineral density; and V_m is the mineral volume fraction.

Other kinetic parameters, such as the activation energy and rate constant, were obtained from the work of Palandri and Kharaka [44], as shown in Table 5. The chemical equilibrium coefficients for aqueous and mineral reactions from the study of Kharaka et al. [45] are listed in Table S2 in Appendix A.

3.5. Molecular diffusion

Molecular diffusion plays a crucial role in mass transport for CCUS-EOR in tight reservoirs. The injected CO₂ penetrates the tight matrix from fractures through diffusion, thus displacing the oil while sequestering CO₂. Several empirical diffusion equations have been derived from extensive experimental data. The Wilke–Chang correlation (Eq. (24)) is used in our model to describe the molecular diffusion behavior of CO₂ in both oil and aqueous phases [46].

(24)

D_{i} = \frac{7.40 \times 10^{− 8} {(M_{i}^{'})}^{0.5} T}{μ v_{b i}^{0.6}}

where D_i is the diffusion coefficient of component i in the mixture;

M_{i}^{'}

is the molecular weight of the solvent; μ is the viscosity of the solution; and v_b_i is the partial molar volume of component i at boiling point.

3.6. Relative permeability

The relative permeability curves of the matrix and fracture were generated using the Corey-type power law equation [39]. The endpoints of the relative permeability curves (e.g., connate water saturation and irreducible oil saturation) were provided by the oilfield. The curvature exponents were determined by matching the historical production data of the horizontal well. The relative permeability curves for the matrix are shown in Figs. 4(a) and (b); straight lines were used for the fracture zone and miscible condition. The interfacial tension (IFT)-related relative permeability model proposed by Coats [47] was integrated into the model to perform an interpolation between the immiscible and miscible relative permeability curves. We primarily focused on the modeling of trapped gas during the cyclic injection process; the impact of geochemical reactions on relative permeability was not considered. The wettability tendency of rock leads to hysteresis in the relative permeability of the non-wetting phase (usually the gas phase) during the frequent transition from drainage to wetting phase imbibition, resulting in trapping of the gas phase. A two-phase gas hysteresis model developed using CMG–GEM was employed to elucidate the hysteresis behavior [39], and the land trapping model was used to predict the trapping saturation of the non-wetting phase [48], as presented in Eqs. (25), (26), (27), (28), (29). The gas relative permeability hysteresis model is shown below:

(25)

k_{rg} (S_{g}) = \{\begin{matrix} k_{rg}^{dr} (S_{g}), d u r i n g d r a i n a g e \\ k_{rg}^{dr} (S_{gf}), d u r i n g i m b i b i t i o n \end{matrix}

(26)

S_{gf} = S_{gcrit} + \frac{(S_{g} - S_{gr h}) (S_{g h} - S_{gcrit})}{(S_{g h} - S_{gr h})}

where k_rg is the gas relative permeability;

k_{rg}^{dr}

is the gas relative permeability on the drainage curve; S_g is the gas saturation; S_gf is the free gas saturation; S_gcrit is the critical gas saturation; S_grh is the trapped/residual gas saturation under the imbibition process; and S_gh is the value of gas saturation when the shift to imbibition occurs.

Land’s function defines the relationship between the reversal saturation S_gh and the maximum trapped gas saturation S_grh:

(27)

C = \frac{1}{S_{grmax} - S_{gcrit}} - \frac{1}{S_{gmax} - S_{gcrit}}

(28)

S_{gr h} = S_{gcrit} + \frac{S_{g h} - S_{gcrit}}{1+ C (S_{g h} - S_{gcrit})}

(29)

S_{gmax} = 1 - S_{wcon} - S_{oirg}

where C is land’s constant; S_grmax is the maximum residual gas saturation; S_gmax is the maximum gas saturation; S_wcon is the connate water saturation; and S_oirg is the irreducible oil saturation for the gas–liquid system.

In the model, S_grmax induced by wetting phase imbibition is assumed to be equivalent to the difference between the average water saturation of the reservoir and S_wcon subsequent to the completion of history-matching. Owing to the non-CO₂ wetting of the rock, the liquid phase relative permeability curves are considered to be the same for both the drainage and the imbibition processes. Fig. 4(c) shows the scanning curves of the gas relative permeability hysteresis.

4. Results and discussion

4.1. CO₂-EOR and geological storage induced by cyclic injection

Two schemes were devised to assess the dynamics of CO₂-EOR and storage during the cyclic injection process. Case 1 served as the control model that only considered CO₂–crude oil interactions, while Case 2 incorporated all the CO₂ storage mechanisms, including dissolution trapping, residual trapping, and mineral trapping. Based on the history matching from primary depletion, CO₂ cyclic injection was conducted with the operation parameters given in Table 6. The minimum bottomhole pressure (BHP) was maintained above the MMP to facilitate CO₂ mixing with oil. A total of ten cycles with identical parameters were executed to explore the dynamic response of oil production and CO₂ storage. Fig. 5 shows the well performance over time for multiple cycles.

Case 2 exhibited an oil production rate close to that of Case 1, and the difference in cumulative oil production was scarcely noticeable during the first three cycles. However, as the number of cycles increased, the effect of CO₂ storage on the oil-displacement efficiency gradually emerged, and the cumulative oil production of Case 2 decreased by 2.2% after ten cycles, as illustrated in Figs. 5(a) and (d). In order to figure out the decline in oil recovery factor (RF), the movement of CO₂ in the oil phase throughout this process was continuously monitored, as shown in Fig. 6. The CO₂–oil mixing zone was defined as the region where the CO₂ mole fraction in the oil phase exceeded 0.1. It was observed that the mixing zone in Case 1 was notably larger than that in Case 2, and the gap became more significant as the number of cycles increased. By the end of production, the volume of the mixing zone resulting from CO₂ storage had decreased by 25.9% compared with that in Case 1. This reduction was attributed to the dissolution of the injected CO₂ in water and the consumption by the water–rock geochemical reactions in Case 2, which diminished the impact of CO₂ interaction with crude oil. Therefore, CO₂-EOR simulations would overestimate the RF if CO₂ storage in the formation were not incorporated. The BHP curves for both cases were nearly consistent and less than the original formation pressure, avoiding any potential risk of CO₂ leakage. Both the GOR and the cumulative CO₂ production were notably lower in Case 2 than in Case 1 because of the incorporation of CO₂-trapping mechanisms in the modeling. This gap widened as the number of cycles increased.

The RF of the primary depletion after 980 days was 6.7% of the OOIP, and ten cycles of CO₂ injection further improved the RF by 3.42% OOIP, reaching an ultimate RF of 10.12% OOIP. Fig. 7 indicates the correlation between CO₂-EOR and storage efficiency over cycles. The oil exchange ratio was defined as the volume of oil produced per tonne of CO₂ under surface conditions, while the CO₂ storage efficiency referred to the ratio of CO₂ trapped underground relative to the injected CO₂. The incremental oil production and CO₂ storage efficiency of these two schemes gradually decreased during each cycle due to the fact that the high-pressure CO₂ pushed the crude oil deeper into the matrix while mixing with it. Although Case 1 excluded the interactions among CO₂, water, and rock, a significant amount of CO₂ was still trapped, mainly resulting from the dissolution and mixing of CO₂ in crude oil. After ten cycles, the total storage rates for Case 1 and Case 2 were 40.6% and 48.3%, respectively, demonstrating the technical feasibility of cyclic injection for CO₂ storage in unconventional reservoirs. The temporary retention of CO₂ in crude oil complicated the storage process with some uncertainties because a fraction of CO₂ would be co-produced with oil. This is an inevitable issue for collaborative CO₂-EOR and storage during the cyclic injection process. Hence, it is necessary to identify the storage forms of CO₂ in the reservoir and understand the dynamics governing CO₂ migration and transformation under multi-phase and multi-field coupling conditions. This is a crucial prerequisite to accurately evaluate the potential of CO₂-EOR and geological storage in tight reservoirs.

4.2. CO₂ migration and transformation

The complex subsurface rock–fluid system usually leads to diverse forms of CO₂ in a reservoir, including supercritical, dissolved, aqueous-phase ionic, or mineralized states. In contrast to conventional CO₂ flooding or deep saline formation sequestration, cyclic CO₂ injection is conducted in an intermittent mode, which further complicates the CO₂ storage process. In our work, six possible forms of CO₂ in the reservoir were considered: CO₂ dissolved in oil, CO₂ dissolved in water (physically dissolved), aqueous ions (HCO₃^–, CO₃^2–, etc.), mineralized CO₂ (precipitation of carbonate minerals), trapped gas (due to gas relative permeability hysteresis), and supercritical mobile gas.

Fig. 8 shows the transformation of CO₂ storage in different forms in the tight reservoir with well production. It is evident that the CO₂ forms were notably impacted by pressure changes, particularly the dissolved CO₂ in oil and water, and the mobile gas. The CO₂ content in the forms of aqueous ions and mineralization gradually increased and was less affected by the operations. The trapped gas form showed an overall increasing trend during each cycle, owing to the frequent reversals of non-wetting phase drainage and wetting phase imbibition. To further elucidate the dynamic transformation of CO₂ in the subsurface formation, the process was divided into three stages according to the cyclic injection operations: injection, soaking, and production.

(1) In the injection stage, CO₂ was injected at a rate of 50 000 m³·d⁻¹. The migration of CO₂ from the fractures into the matrix was mainly driven by pressure gradients. As a result, the content of CO₂ in the oil, gas, and water quickly increased, with a considerably higher amount of CO₂ dissolved in oil than in water. The contribution of each form to CO₂ storage was defined as the ratio of the moles of CO₂ in a specific form relative to the total moles of subsurface CO₂, as shown in Eq. (30).

(30)

C O_{2} storage contributio n_{(dc)} = \frac{N_{s d c}}{N_{inj c}}

where N_s_dc represents the moles of CO₂ captured by the dth storage mechanism at the cth cycle, and N_inj_c represents the total moles of subsurface CO₂ at the cth cycle.

Fig. 9(a) shows the contribution of each form to CO₂ storage after CO₂ injection with each cycle. As indicated, more than 50% of the injected CO₂ was mixed with crude oil, while less than 10% of the CO₂ existed in a flowable supercritical state. The high connate water saturation created a favorable condition for CO₂ dissolution and geochemical reactions. The contribution of mineralization to CO₂ storage increased steadily and accounted for 10% of the subsurface CO₂ after ten cycles.

(2) In the soaking stage, CO₂ accumulated in the SRV and then diffused into deeper regions, driven by concentration gradients. The movement of subsurface CO₂ during the shut-in period was tracked, as shown in Fig. 9(b). To elucidate the CO₂ transformations, a conversion rate was proposed based on the cumulative injected moles of CO₂. The primary sources of transformation were CO₂ that dissolved in oil and CO₂ that was trapped in gas. The high pressure generated in the SRV during the injection stage propagated to the low-pressure matrix in the soaking stage. Once the formation pressure stabilized, the excess CO₂ that had dissolved in oil and water was released and converted into free gas, and then accumulated in the fractures. The free gas was further transported deeper into the matrix and transformed into other CO₂ forms. In the first soaking period, 7.2% of the CO₂ in the oil and 2.0% of the trapped gas were transformed into 5.9% of the mobile gas and 3.3% of other relatively stable sequestered forms. The regions of changes in the CO₂ mole fractions in the oil and water phases during the initial soaking were captured to support these findings, as shown in Fig. S2 in Appendix A. The release of CO₂ from the fractures and its dissolution in the matrix can be readily observed. However, as the cycle number increased, 90% of the released CO₂ became mobile gas collecting in the fractures, suggesting that the fluid in the vicinity of the fractures tended to be saturated. Hence, certain measures, such as increasing the injection rate or extending the soaking time, need to be implemented to facilitate CO₂ penetration into the matrix.

(3) In the production stage, the minimum BHP was maintained at 35 MPa. The GOR was observed to increase at the end of production, and the mole proportion of CO₂ in the produced gas increased from 24.62% (first cycle) to 73.69% (tenth cycle). Fig. 9(c) indicates the CO₂ conversion rate in each production stage. The loss of the mobile gas and the dissolved CO₂ in oil and water resulted in the gain of produced gas, trapped gas, aqueous ions, and minerals. In particular, the dissolved CO₂ in the oil along with the mobile gas significantly contributed to CO₂ production. The loss of the dissolved CO₂ was attributed to the formation pressure reduction and fluid production. The residual gas was mainly trapped in the non-fracture zone of the SRV and gradually expanded from the near-well to the far-well non-fracture zone as the number of cycles increased.

Fig. 9(d) shows the contribution of different forms of CO₂ to the storage following each cycle. The proportion of produced CO₂ increased with the number of cycles, and more than 50% of the injected CO₂ was produced after eight cycles. The trapped CO₂ in the subsurface was in different forms, among which the oil-dissolved CO₂ was the most significant, followed by water-dissolved CO₂ and aqueous-phase ions. The contributions of mineralization and free gas gradually increased as the number of cycles increased, while the trapped CO₂ induced by gas relative permeability hysteresis tended to stabilize. At the end of ten cycles, 48% of the injected CO₂ (∼28 000 t) was sequestered in this reservoir. Given the re-release of the dissolved CO₂ from oil along with production, an estimated 20% of the CO₂ (approximately 12 000 t) was projected to be stored over the long term in the forms of carbonate mineralization (6%), dissolution in water (6%), aqueous ions (4%), and trapped gas (4%). It is noteworthy that the evolution of the storage mechanism in an unconventional reservoir presents notable differences from that in conventional reservoirs and saline aquifers, especially regarding trapped gas. The high connate water and the low flowable water saturation in the tight matrix attenuate the imbibition process, resulting in a reduction of trapped CO₂.

4.3. Assessment of geochemical reactions

Quartz and feldspar are the predominant minerals in this tight sandstone reservoir. In the presence of CO₂, the dissolution of primary minerals leads to the precipitation of carbonate minerals, thereby achieving permanent sequestration. Our simulation indicates that approximately 6% of the injected CO₂ precipitated as carbonate minerals within ten years. The changes in the mineral contents in moles of the primary and secondary minerals in the reservoir are shown in Figs. 10 (a) and (b). Positive values indicate mineral precipitation, whereas negative values mean mineral dissolution. As indicated, the mineral contents were not strongly dependent on the production dynamics because they were significantly governed by kinetic models. Among all the primary minerals, anorthite experienced a significant dissolution, followed by chlorite, while K-feldspar showed a minimal dissolution. In contrast, kaolinite, albite, and quartz underwent substantial precipitation, while illite demonstrated slight precipitation. As for the secondary minerals, dolomite showed notable precipitation, followed by dawsonite and calcite, while anhydrite and magnesite were only slightly precipitated. The process of mineral dissolution or precipitation depended on the ratio of the activity product to the chemical equilibrium constant. The former significantly relied on the concentration of aqueous ions, whereas the latter was a temperature-dependent parameter. The dissolution of anorthite, chlorite, and K-feldspar released substantial amounts of Al³⁺, Ca²⁺, and Mg²⁺ ions, facilitating the precipitation of carbonate minerals. Figs. 10(c) and (d) show the changes in the concentrations of the cations and anions in the aqueous phase. When the rate of ions released from mineral dissolution exceeded the rate of ions depleted from mineral precipitation, the concentration of aqueous ions, such as Al³⁺ and Mg²⁺ ions, increased. The formation water was abundant in Na⁺ ions, which were consumed by the precipitation of albite and dawsonite. Although the dissolution of anorthite released substantial amounts of Ca⁺ ions, its concentration remained stable due to the significant precipitation of dolomite and calcite. The average pH of the brine significantly varied with the well operations, as indicated in Fig. 10(e). The pH fluctuations indicated the intensity of the geochemical reactions. The injected CO₂ primarily accumulated in the fracture zone and then slowly diffused into the matrix. This made SRV the dominant area for geochemical reactions.

4.4. Porosity and permeability change

For CCUS-EOR in an unconventional reservoir, any changes in reservoir properties (e.g., stress and mineral reactions) would significantly impact oil recovery and storage efficiency. This section focuses on the impact of mineral dissolution/precipitation on reservoir properties. The changes in reservoir porosity and permeability were characterized by the difference in the petrophysical properties in the absence or presence of mineral reactions, as given in Eq. (31). The average porosity and permeability fluctuations across the reservoir matrix, non-fracture, and fracture zones of the SRV were analyzed, as depicted in Fig. 11. The change rate in the rock’s physical properties was influenced by the mineral reactions, as indicated in Fig. 12.

(31)

\{\begin{matrix} Δ ϕ = ϕ^{*} \cdot m_{ϕ} - ϕ \\ Δ k = k^{*} \cdot m_{k} - k \end{matrix}

where

Δ ϕ

and

Δ k

are the changes of reservoir porosity and permeability induced by mineral reactions, respectively;

m_{ϕ}

is the pressure-dependent porosity multiplier; and

m_{k}

is the pressure-dependent permeability multiplier.

It was observed that the porosity and permeability of the reservoir increased in the first cycle but declined in subsequent cycles. This indicated that mineral precipitation became more significant than dissolution in terms of pore volume. The initial lower RSA of the secondary minerals corresponded to a slower precipitation process, which led to the predominance of mineral dissolution in the initial period. However, the dissolution of minerals, including anorthite and chlorite, offered the necessary cations for carbonizing minerals. In addition, this process promoted the precipitation of quartz and various silicate minerals such as kaolinite and albite, which accordingly reduced the pore volume of the reservoir.

Fig. 11, Fig. 12 demonstrate that the SRV was the main area in which porosity and permeability changes occurred. At the end of production, the porosity decreased by 0.86% and 0.81% in the non-fractured and fractured zones, respectively, while the permeability was reduced by 2.51% and 2.39%, respectively. It was also observed that the change rate of porosity and permeability experienced a slight increase during CO₂ injection, owing to the enhanced mineral dissolution induced by the CO₂. With the cessation of CO₂ injection, the pH gradually returned to normal, leading to a decrease in the mineral dissolution reaction. This phenomenon was particularly significant in the fracture zone. After multiple cycles, the physical properties of the reservoir matrix were barely changed. The porosity and permeability ultimately decreased by 0.61% and 1.62%, respectively. In a short view, the impact of geochemical reactions on rock properties was less significant than that of geomechanics. However, CO₂ storage resulted in a 2.2% decrease in the cumulative oil production. Interactions among the CO₂, brine, and rock decreased the efficiency of CO₂-EOR. Therefore, the CCUS-EOR process in an unconventional reservoir must be carefully designed and optimized with respect to oil production and CO₂ storage.

5. Limitations of this study

Artificial fractures in unconventional reservoirs after fracturing are strongly stress sensitive, which has a significant impact on the response of CCUS-EOR under cyclic CO₂ injection. In the simulation, we simplified the geomechanical module by utilizing the multiplier correlations of pressure-dependent porosity and permeability. This simplification provides a more visual description of the complex strain and hysteresis behaviors of rock under repeated fluctuations in stress. Nevertheless, it comes with the limitation of being unable to comprehensively capture the dynamic behavior of fractures. For instance, the effect of fracture propagation may demand more refined and intricate geomechanical models for proper representation. Geochemical reactions are another crucial factor influencing CO₂–water–rock interactions, with the reaction processes being governed by the reaction kinetics models. The RSA is a critical parameter with high uncertainty. In the calculation process of RSA, we used the limited published data in an attempt to minimize the effect of the uncertainty regarding the mineral characteristics and composition on the RSA. Nevertheless, the development of more accurate and precise RSA test methodologies remains an urgent necessity.

6. Conclusions

This work presented a modeling assessment of cyclic CO₂ injection for CCUS-EOR in a specific tight reservoir. The impacts of multiphase flow, fluid phase behavior, geomechanics, and geochemistry were integrated into the modeling to accurately predict the fate of the injected CO₂. Subsurface CO₂ migration and transformation, mineral reactions, and porosity and permeability changes were thoroughly investigated. After ten cycles of CO₂ injection, the oil recovery of this reservoir was increased by 3.4% OOIP, and 48.3% of the injected CO₂ was stored. The CO₂ storage process diminished the CO₂–oil interactions, leading to a 25.9% reduction in the volume of the CO₂–oil mixing zone and a 2.2% decline in the cumulative oil production. The evolution of the CO₂ storage mechanism under cyclic injection showed significant differences compared with conventional CO₂ flooding. Among all forms of CO₂ storage, dissolved CO₂ in oil and water, mobile gas, and residual gas were notably affected by the production dynamics, while aqueous phase ions and mineralization were less affected by operations. The residual gas was also affected by the cycles and was predominantly trapped in the non-fracture zone of the SRV. Although dissolved CO₂ in oil accounted for over half of the total storage, it should be removed during the short-term CCUS-EOR process, since it is an unstable form with the potential to be released with production. By the end of production, 20% of the injected CO₂ was sequestered permanently by carbonate mineralization (6%), dissolution in water (6%), aqueous ions (4%), and trapped gas (4%). Dolomite, dawsonite, and calcite emerged as the primary CO₂-trapping minerals. The reservoir porosity induced by geochemical reactions decreased 0.86% and 0.81% in the non-fractured and fractured zones, respectively, while the permeability decreased by 2.51% and 2.39%, respectively, after ten cycles.

The understanding of the dynamic CO₂ transport process enabled by this work can provide support for the design and optimization of injection–production schemes. Future research should focus on the co-optimization of CO₂ EOR and storage, life-cycle assessment, and enhancement of storage efficiency, with the aim of developing more efficient injection strategies and management techniques to maximize oil recovery and long-term storage.

CRediT authorship contribution statement

Jinzhou Zhao: Writing – review & editing, Supervision, Resources, Methodology. Lele Wang: Writing – original draft, Software, Methodology, Supervision. Bing Wei: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Valeriy Kadet: Writing – review & editing, Validation, Methodology.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors gratefully acknowledge financial support from the National Key Research and Development Program of China (2023YFE0120700), National Natural Science Foundation of China (52274041), and Distinguished Young Sichuan Science Scholars (2023NSFSC1954).

Appendix A. Supplementary material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.eng.2025.01.005.

References

Publishing order | Descend order by publishing year | Descend order by cited within

[1]	C O2 emissions in 2023.Report.Paris: International Energy Agency; 2024.

[2]	Edouard MN, Okere CJ, Ejike C, Dong P, Suliman MAM.Comparative numerical study on the co-optimization of CO₂ storage and utilization in EOR, EGR, and EWR: implications for CCUS project development.Appl Energy 2023; 347:121448.

[3]	Hepburn C, Adlen E, Beddington J, Carter EA, Fuss S, Dowell NM, et al.The technological and economic prospects for CO₂ utilization and removal.Nature 2019; 575:87-97.

[4]	Liu Y, Rui Z.A storage-driven CO₂ EOR for a net-zero emission target.Engineering 2022; 18:79-87.

[5]	Meng S, Zhang F, Tao J, Jin X, Xu J, Liu H.Carbon storage potential of shale reservoirs based on CO2 fracturing technology.Engineering. In press.

[6]	utilization and storage (CCU S) annual report (2021): China CCUS path study.Report.Beijing: Chinese Academy of Environmental Planning; Wuhan: Insitute of Rock and Soil Mechanics, Chinese Academy of Sciences; Beijing: Administrative Center for China’s Agenda 21; 2021. Chinese.

[7]	Wei B, Song T, Gao Y, Xiang H, Xu X, Kadet V, et al.Effectiveness and sensitivity analysis of solution gas re-injection in Baikouquan tight formation, Mahu sag for enhanced oil recovery.Petroleum 2020; 6:253-263.

[8]	Wei B, Gao K, Song T, Zhang X, Pu W, Xu X, et al.Nuclear-magnetic-resonance monitoring of mass exchange in a low-permeability matrix/fracture model during CO₂ cyclic injection: a mechanistic study.SPE J 2020; 25(1):440-450.

[9]	Yu W, Lashgari HR, Wu K, Sepehrnoori K.CO₂ injection for enhanced oil recovery in Bakken tight oil reservoirs.Fuel 2015; 159:354-363.

[10]	Jacobs T.Shale EOR delivers, so why won’t the sector go big?.J Petrol Technol 2019; 71(5):37-41.

[11]	Ajoma E, Saira ST, Ge J, Le-Hussain F.Water-saturated CO₂ injection to improve oil recovery and CO₂ storage.Appl Energy 2020; 266:114853.

[12]	Hawthorne SB, Gorecki CD, Sorensen JA, Steadman EN, Harju JA, Melzer S.Hydrocarbon mobilization mechanisms from upper, middle, and lower Bakken reservoir rocks exposed to CO2.In: Proceedings of SPE Unconventional Resources Conference; 2013 Nov 5–7; Calgary, AB, Canada. Richardson: OnePetro; 2013.

[13]	Hoffman BT, Reichhardt D.Sensitivity analysis of cyclic gas injection recovery mechanisms in unconventional reservoirs.In: Proceedings of SPE/AAPG/SEG Unconventional Resources Technology Conference; 2020 Jul 20–22; online. Richardson: OnePetro; 2020.

[14]	Cronin M, Emami-Meybodi H, Johns RT.Diffusion-dominated proxy model for solvent injection in ultratight oil reservoirs.SPE J 2018; 24(2):660-680.

[15]	Ma M, Emami-Meybodi H.Multiphase multicomponent transport modeling of cyclic solvent injection in shale reservoirs.SPE J 2023; 29(3):1-20.

[16]	Tovar FD, Barrufet MA, Schechter DS.Enhanced oil recovery in the Wolfcamp shale by carbon dioxide or nitrogen injection: an experimental investigation.SPE J 2021; 26(1):515-537.

[17]

Kumar R, Ali SJ, Mathur A.A rigorous workflow for evaluation of huff and puff recovery efficiency of immiscible and miscible gases in unconventional reservoirs by integrating core tests with NMR and GC analysis.In: Proceedings of SPE/AAPG/SEG Unconventional Resources Technology Conference; 2020 Jul 20–22; online. Richardson: OnePetro; 2013.

[18]	Kim TH, Lee JH, Lee KS.Integrated reservoir flow and geomechanical model to generate type curves for pressure transient responses of a hydraulically-fractured well in shale gas reservoirs.J Petrol Sci Eng 2016; 146:457-472.

[19]	Zheng S, Sharma M.The effects of geomechanics, diffusion and phase behavior for huff-n-puff IOR in shale oil reservoirs.In: Proceedings of SPE Annual Technical Conference and Exhibition; 2020 Oct 26–29; online. Richardson: OnePetro; 2020.

[20]

Hamdi H, Clarkson CR, Ghanizadeh A, Ghaderi SM, Vahedian A, Riazi N, et al.Huff-n-puff gas injection performance in shale reservoirs: a case study from Duvernay shale in Alberta, Canada.In: Proceedings of SPE/AAPG/SEG Unconventional Resources Technology Conference; 2018 Jul 23–25; Houston, TX, USA. Richardson: OnePetro; 2018.

[21]	Gao C, Jia J, Fan W, Chen S, Hu T, Wang X, et al.Optimization of CO₂ flooding under dual goals of oil recovery and CO₂ storage: numerical case studies of the first-ever CCUS pilot in Changqing oilfield.Geoenergy Sci Eng 2024; 240:213063.

[22]	Lashgari HR, Sun A, Zhang T, Pope GA, Lake LW.Evaluation of carbon dioxide storage and miscible gas EOR in shale oil reservoirs.Fuel 2019; 241:1223-1235.

[23]	Bao X, Amaya AF, Aguilera R.Life cycle assessment of improved oil recovery while helping to achieve net zero emissions from shale reservoirs.SPE J 2024; 29(1):554-573.

[24]	Ning Y, Tura A.Economic and operational investigation of CO₂ sequestration through enhanced oil recovery in unconventional reservoirs in Colorado, USA.Geoenergy Sci Eng 2023; 226:211820.

[25]	Rui Z, Zeng L, Dindoruk B.Challenges in the large-scale deployment of CCUS.Engineering 2025; 44:17-20.

[26]	Lu X, Jordan KE, Wheeler MF, Pyzer-Knapp EO, Benatan M.Bayesian optimization for field-scale geological carbon storage.Engineering 2022; 18:96-104.

[27]	Nghiem L, Shrivastava V, Kohse B, Hassam M, Yang C.Simulation and optimization of trapping processes for CO₂ storage in saline aquifers.J Can Petrol Technol 2010; 49(8):15-22.

[28]	Wang X, Yang H, Huang Y, Liang Q, Liu J, Ye D.Evolution of CO2 storage mechanisms in low-permeability tight sandstone reservoirs.Engineering. In press.

[29]	Ershadnia R, Wallace CD, Hajirezaie S, Hosseini SA, Nguyen TN, Sturmer DM, et al.Hydro-thermo-chemo-mechanical modeling of carbon dioxide injection in fluvial heterogeneous aquifers.Chem Eng J 2022; 431:133451.

[30]	Wang F, Ping S, Yuan Y, Sun Z, Tian H, Yang Z.Effects of the mechanical response of low-permeability sandstone reservoirs on CO₂ geological storage based on laboratory experiments and numerical simulations.Sci Total Environ 2021; 796:149066.

[31]	Ramadhan R, Promneewat K, Thanasaksukthawee V, Tosuai T, Babaei M, Hosseini SA, et al.Geomechanics contribution to CO₂ storage containment and trapping mechanisms in tight sandstone complexes: a case study on Mae Moh Basin.Sci Total Environ 2024; 928:172326.

[32]	Pearce J, Raza S, Baublys K, Hayes P, Firouzi M, Rudolph V.Unconventional CO₂ storage: CO₂ mineral trapping predicted in characterized shales, sandstones, and coal seam interburden.SPE J 2022; 27(5):3218-3239.

[33]	Hu T, Rui Z.Carbon migration and phase distribution patterns in CO₂ geological utilization and storage.Nat Gas Ind 2024; 44(4):56-67.

[34]	Yu K, Cao Y, Qiu L, Sun P.The hydrocarbon generation potential and migration in an alkaline evaporite basin: the Early Permian Fengcheng Formation in the Junggar Basin, northwestern China.Mar Petrol Geol 2018; 98:12-32.

[35]	Wang L, Wei B, You J, Pu W, Tang J, Lu J.Performance of a tight reservoir horizontal well induced by gas huff–n–puff integrating fracture geometry, rock stress-sensitivity and molecular diffusion: a case study using CO₂, N₂ and produced gas.Energy 2023; 263:125696.

[36]	Win Prop user’s guide.Calgary: Computer Modeling Group Ltd.; 2021.

[37]	Li YK, Nghiem LX.Phase equilibria of oil, gas and water/brine mixtures from a cubic equation of state and Henry’s law.Can J Chem Eng 1986; 64(3):486-496.

[38]	Bethke C.Geochemical reaction modeling.Oxford University Press, New York City (1996)

[39]	GE M user’s guide.Calgary: Computer Modeling Group Ltd.; 2021.

[40]	Jia W, Xiao T, Wu Z, Dai Z, McPherson B.Impact of mineral reactive surface area on forecasting geological carbon sequestration in a CO₂-EOR field.Energies 2021; 14(6):1608.

[41]	Tian W, Lu S, Huang W, Wang W, Li J, Gao Y, et al.Quantifying the control of pore types on fluid mobility in low-permeability conglomerates by integrating various experiments.Fuel 2020; 275:117835.

[42]	Xu T, Yue G, Wang F, Liu N.Using natural CO₂ reservoir to constrain geochemical models for CO₂ geological sequestration.Appl Geochem 2014; 43:22-34.

[43]	Beckingham LE, Mitnick EH, Steefel CI, Zhang S, Voltolini M, Swift AM, et al.Evaluation of mineral reactive surface area estimates for prediction of reactivity of a multi-mineral sediment.Geochim Cosmochim Ac 2016; 188:310-329.

[44]	Palandri JL, Kharaka YK.A compilation of rate parameters of water–mineral interaction kinetics for application to geochemical modeling.Reston: U.S. Geological Survey; 2004.

[45]	Kharaka YK, Gunter WD, Aggarwal PK, Perkins EH, DeBraal JD.SOLMINEQ.88; a computer program for geochemical modeling of water–rock interactions.Reston: U.S. Geological Survey (1988)

[46]	Wilke CR, Chang P.Correlation of diffusion coefficients in dilute solutions.AIChE J 1955; 1(2):264-270.

[47]	Coats KH.An equation of state compositional model.SPE J 1980; 20(5):363-376.

[48]	Land CS.Calculation of imbibition relative permeability for two- and three-phase flow from rock properties.SPE J 1968; 8(2):149-156.

RIGHTS & PERMISSIONS

THE AUTHOR

PDF (3645KB)

Part of a collection:

Supplementary files

Appendix A. Supplementary data

5610

Accesses

Citation

Detail

Sections

Recommended

Browse

Authors & reviewers

About the journal

Abstract

Graphical abstract

Keywords

Cite this article

1. Introduction

2. Reservoir description

3. Methodology

3.1. Reservoir model

3.2. Fluid composition and model

3.3. Rock stress sensitivity

3.4. Geochemical reactions and models

3.4.1. Aqueous reactions

3.4.2. Mineral reactions

3.4.3. Porosity and permeability evolution due to geochemical reactions

3.4.4. Reaction kinetics parameters

3.5. Molecular diffusion

3.6. Relative permeability

4. Results and discussion

4.1. CO2-EOR and geological storage induced by cyclic injection

4.2. CO2 migration and transformation

4.3. Assessment of geochemical reactions

4.4. Porosity and permeability change

5. Limitations of this study

6. Conclusions

CRediT authorship contribution statement

Declaration of competing interest

Acknowledgments

Appendix A. Supplementary material

References

RIGHTS & PERMISSIONS

4.1. CO₂-EOR and geological storage induced by cyclic injection

4.2. CO₂ migration and transformation