Research on intelligent regulation strategy of negative pressure in gas drainage parallel boreholes based on LSTM prediction model

The problem of air leakage in borehole can be effectively reduced by negative pressure regulation. However, excessive dependence on manual regulation will lead to inaccurate and untimely regulation of negative pressure. Intelligent regulation of drainage negative pressure can effectively solve this problem. Therefore, this paper used numerical simulation to theoretically analyze the air leakage and gas drainage effect under different negative pressures. At the same time, the advantages of LSTM (Long-Short Term Memory) compared with other prediction algorithms were compared. On account of the LSTM prediction model, the intelligent regulation strategy of NPGDPB (negative pressure in gas drainage parallel boreholes) was proposed. Finally, a 120d field test was also conducted to confirm the feasibility of the proposed intelligent regulation strategy. Research findings show that changing the drainage negative pressure can reduce the coal seam gas content effectively, reduce air leakage, enhance the gas drainage effect and make gas cascading utilization more efficient. Compared with other prediction algorithms, LSTM has higher adaptability to the prediction of drainage negative pressure and the best prediction performance. The LSTM prediction model can accurately predict the corresponding gas drainage negative pressure based on the gas concentration and flow. After two times of regulation, the gas flow attenuation rate in the parallel boreholes is significantly lower than that of the unregulated parallel boreholes, and the concentration of gas in the parallel boreholes has been obviously improved. The proposed strategy of intelligent regulation of NPGDPB is evidenced to be feasible.