Modelling of flue gas injection and collaborative optimization of multi-injection parameters for efficient coal-based carbon sequestration combined with BP neural network parallel genetic algorithms

Cooperative equalization among the injection parameters can control the economic objective function to ensure the efficiency of gas -mixture -enhanced coalbed methane recovery (GM-ECBM). This study first proposed and validated a refined thermal -hydro -mechanical -chemical (THMC) coupling model with a convenient permeability model. Based on GM-ECBM numerical calculation results under different combinations of injection parameters, different types of cost -income objective functions, including different costs and incomes, were optimized by parallelly applying a genetic algorithm and backpropagation neural network (GA -BP). Using the recovery rate as the index, a small inter -well distance and a large injection pressure can increase the optimal CO2 proportion. The effectiveness of enhancing the injection temperature was significantly correlated with both inter -well distance and the injection pressure. Cooperative equalization among the injection parameters can explain different viewpoints of different scholars on the temperature effect. Large Henry's law constants resulted in low carbon sequestration and coalbed methane (CH4) production but did not affect the optimal CO2 proportion. High carbon sequestration and CH4 prices corresponded to small inter -well distances and large injection pressure. Based on the objective function, the optimal inter -well distance and the injection temperature were 175-220 m and 305.5-311.5 K, respectively. The volumetric adsorption -induced expansion coefficient (VAEC) of CO2 determined the optimal CO2 proportion, the optimization feasibility of the injection scheme, and the applicability of GM-ECBM.