Optimization of Corner Gas Accumulation Processing Method Based on Improved Wavelet Threshold and BiLSTM

Nowadays, mine gas management is facing increasing challenges, especially the accumulation of corner gas, which poses a great threat to the safety production of miners. This study proposes a Bidirectional Long Short Term Memory (BiLSTM) model based on improved wavelet thresholding for optimizing the detection and processing of mine corner gas. Firstly, by improving the wavelet thresholding method to denoise the gas concentration data, the quality and processing efficiency of the data have been enhanced; next, the BiLSTM model is used to learn and predict the processed data, in order to accurately identify the risk areas of gas accumulation. During the 6-month experimental period, the prediction accuracy of the BiLSTM model remained within the range of 92.5% to 93.3%, with little fluctuation. This article not only optimizes the treatment method of mine corner gas, but also provides a new technical solution for gas safety management.