The Wolonghe Gas Field is geographically located at the junction of Dianjiang County and Changshou District in Chongqing City, and tectonically, it is situated in the central part of the ancient gentle uplift and steep fault-fold belt of the eastern Sichuan Basin. The formation is a gentle anticline structure, with a steep west and a gentle east, and is sandwiched between the Mingyue Gorge and Goujiachang high-steep anticline belts. The geological structure of the eastern Sichuan area is mainly modified by the Caledonian and Indosinian-Yanshanian movements. The reservoir is deeply buried, and local faults are well-developed, forming complex oil and gas accumulation and migration channels, which provide favorable geological conditions for natural gas accumulation. Fig. 1 shows the geographical location of the gas field.

The main formation of this study, the Lower Permian Maokou Formation, is a set of carbonate reservoirs. In the Wolonghe area, the Maosiduan is absent, and part of the Maosanduan remains. The lithology is mainly gray to dark gray argillaceous limestone, with local interbeds of dolomite. The Maokou Formation’s reservoir type is fracture-vuggy, belonging to an autogenous-autostored gas system, and is one of the main reservoirs of the Lower Permian. The gas reservoir is characterized by abundant hydrocarbon source rocks, well-developed reservoirs, good cap rocks, and preservation conditions. The vertical configuration of source, reservoir, and cap is good, and all accumulation elements are well-matched, providing excellent conditions for gas accumulation.

The average reservoir porosity is mainly concentrated in the range of 4%~9%, and the permeability is often less than 1.0 mD, consistent with the low-porosity and low-permeability characteristics of a tight gas reservoir. The methane content is 94.18%~97.88%, carbon dioxide content is 1.13%~4.03%, nitrogen content is 0.18%~0.38%, and hydrogen sulfide content is 0.11%~1.34%, with an average of 0.41%. It is a dry gas with low to medium H₂S and CO₂ content.

During the simulation of CO₂ geological storage, the injected CO₂ can undergo a series of water-rock reactions with formation water under high-pressure conditions, affecting mineral dissolution/precipitation behavior and ion migration characteristics, thereby having a potential impact on the reservoir pore structure and storage stability. To convert porosity variation into permeability distribution, a power-law relationship is used to estimate permeability as a function of porosity [22]: (1) K i K h = ϕ i ϕ h P where K i and ϕ i are the permeability and porosity of the ith grid, respectively. K h and ϕ h are the corresponding values for porosity and permeability of a homogeneous system, respectively. P is the power exponent. Porosity is the main indicator and the most direct result of geochemical dissolution/precipitation reactions. The changing trend of permeability is completely consistent with that of porosity. Therefore, in the subsequent studies, only the changes in porosity were analyzed.

To more realistically reflect the geochemical environment of the wellbore area and the reaction path after CO₂ injection, the model introduced typical water chemical ion compositions and the dissolution reactions of representative carbonate minerals. To maintain focus on the main CO₂–water–rock interactions and ensure computational efficiency, we omitted the reactions related to H₂S. The water type of the formation water is CaCl₂ type, and the specific ion composition is shown in Table 1.

To simulate the acidification and carbonate dissolution process caused by CO₂ injection, the reaction pathways are shown in Table 2:

Based on logging, geological, and other data, a three-dimensional gas reservoir numerical model was established. To accurately characterize the CCUS-EGR process, the simulation employs a generalized compositional formulation that couples multiphase fluid flow with geochemical reactions. Following the fully coupled geochemical EOS formulation described by Nghiem et al. [23], the component mass conservation equation incorporating the reaction rate is expressed as: (2) ∂ ∂ t ϕ ∑ j ρ j S j x i j + ∇ ⋅ ∑ j ρ j x i j v j = q i − ∑ k = 1 N r x ν i k r k where ϕ is porosity; ρ j , S j , and v j denote the molar density, saturation, and Darcy velocity of phase j (gas, water), respectively; x i j is the mole fraction of component i in phase j; q_i is the source/sink term; and the last term represents the rate of change due to geochemical reactions, where ν i k is the stoichiometric coefficient and r k is the reaction rate of mineral k. The fluid velocity v j is governed by Darcy’s law: (3) v j = − K k r j μ j ( ∇ p j − ρ j g ∇ D ) where K is the absolute permeability tensor, k r j is the relative permeability of phase j, μ j is the viscosity, p j is the pressure, and g is the gravitational acceleration.

The total grid size is 100 × 30 × 10, with a total of 30,000 effective grid cells.

The horizontal grid size is set to 50 m × 50 m, and the vertical direction is divided into 10 layers, with a single layer thickness of about 9 m and a total thickness of about 90 m. The overall model size is 5000 m × 1500 m × 90 m.

To eliminate the influence of grid size on the simulation results, a quantitative mesh sensitivity analysis was conducted. Three grid systems with varying densities were established: Coarse (50 × 15 × 10), Current (100 × 30 × 10), and Fine (200 × 60 × 10). The cumulative gas production was selected as the representative physical variable for comparison. As summarized in Table 3, the results show that the relative deviation in gas production between the Current scheme and the Fine scheme is only 0.2%. However, the computation time for the Fine scheme increases drastically to 3966.9 s, compared to 355.8 s for the Current scheme. Therefore, considering both computational accuracy and efficiency, the Current grid dimension (100 × 30 × 10) was determined to be the optimal choice for this study.

The initial depth of the reservoir model is 3447 m, and the formation temperature is set to 85.0°C. The natural gas is mainly methane, with a content as high as 96.57%. Other components include nitrogen (0.23%), carbon dioxide (2.11%), ethane (0.65%), propane (0.28%), and trace amounts of hydrogen sulfide. The overall gas composition is stable. These parameters are derived from the logging and fluid analysis results of the Wolonghe Gas Field. The average reservoir porosity is 5%, the horizontal permeability is 1.5 mD, and the vertical permeability is 0.15 mD. These data are representative average values of logging and core data for analyzing the target reservoir section.

The model is set with three production wells (P-1, P-2, P-3), two injection Wells (I-1, I-2), the production rate of P-1 and P-2 at 40,000 m³/d, and the production rate of P-3 at 60,000 m³/d. The distribution of well locations is shown in Fig. 2. The simulation time is from 1985 to 2025 for depressurization production. The period from 2025 to 2045 is the predictive simulation stage, during which the initial gas injection rate of the injection well is 50,000 m³/d. The original gas in place is about 3.54 × 10⁹ m³, the cumulative gas production accounts for 59.93% of the original gas in place, and the final remaining reserves are about 1.43 × 10⁹ m³. The relative permeability of gas and water is set as shown in Fig. 3.

Based on the numerical model initialized with 40 years of simulated production history, this study designed multiple simulation scenarios to analyze the impact of different geological and engineering parameters on the CCUS-EGR effect of the Maokou Formation gas reservoir in Wolonghe. The investigated ranges for the injection-production ratio and injection rate were selected based on practical operational constraints. The main evaluation indicators include CO₂ breakthrough time, cumulative natural gas production, and enhanced recovery factor. And in these simulations, breakthrough time is defined as the time (in days) when the mole fraction of CO₂ in the produced gas stream from the three production wells (P-1, P-2, or P-3) first exceeded 1.0%. The simulation injection time is set to 20 years.