Research on gas extraction borehole inversion optimization system based on reinforcement learning
Liu Zhi 1, Li Junfang 2, Xu Shiying 1, Wang Zehua 2, Sun Kelei 3 *
1. Technical Research Institute of Yimei Company, Sanmenxia 472300, China ; 2. Xinyi Coal Mine of Yimei Company, Sanmenxia 472300, China; 3. National Key Laboratory of Digital Intelligence Technology for Coal Unmanned Mining, Anhui University of Science and Technology, Huainan 232001, China
Aiming at the risk of poor coal permeability, high gas content and coal and gas outburst in high gas mine, taking 12030 working face of Xinyi Mine as the object, the inversion optimization of gas drainage borehole based on reinforcement learning is carried out. Based on AutoCAD graphics platform and secondary development technology, a gas extraction design and inversion system including parametric automatic design, real drilling trajectory inversion and blank zone quantitative identification is constructed. The dynamic adjustment of borehole space layout is modeled as a Markov decision process, and the proximal strategy optimization algorithm is used to optimize the number of boreholes, extraction coverage and trajectory deviation in the simulation environment. The experimental results show that the design efficiency of the system is more than 80% higher than that of manual work, and the deviation calculation meets the engineering requirements. Under the premise of ensuring 98% effective extraction coverage, the number of borehole construction is reduced by 15%～20% compared with the original scheme.
刘 治1，李俊方2，徐世英1，王泽华2，孙克雷3*. 基于强化学习的瓦斯抽采钻孔反演优化系统研究[J]. 煤炭与化工, 2026, 49(10): 98-104.
Liu Zhi 1, Li Junfang 2, Xu Shiying 1, Wang Zehua 2, Sun Kelei 3 *. Research on gas extraction borehole inversion optimization system based on reinforcement learning. CCI, 2026, 49(10): 98-104.
