Research on Gas Emission Prediction and Risk Identification of Yuqia Coal Mine in Qinghai Province from the Perspective of Information Fusion

Abnormal gas emissions are one of the main risk factors evoking coal mine gas accidents. How to accurately and efficiently predict gas emissions and identify the risk of gas anomalies has become a key issue in coal mine safety management. This study takes the Yuqia No.1 Mine of Qinghai Energy Group as the research object, collecting environmental variable data such as the gas emission quantity of the mining face, coal seam depth, coal seam thickness, coal seam gas content, temperature, wind speed, and respirable dust concentration. Multi-parameter data fusion, gray correlation degree analysis, least square support vector machine (LS-SVM), random forest (RF), back propagation neural network (BPNN), and other methods were adopted in this paper to explore the prediction accuracy and risk factors of mine gas emissions. The results show the following: (1) The correlation coefficients between coal seam depth, coal seam thickness, coal seam gas content, daily progress, daily output, and wind speed and gas emission quantity are 0.955, 0.975, 0.963, −0.912, 0.983, and 0.681, respectively, showing a significance level of 0.01, and are used as external input characteristic quantities for gas emission quantity prediction. (2) For the LS-SVM model, the root mean square error (RMSE) and mean absolute error (MAE) values on the training set were 0.015 and 0.012, respectively, while the corresponding test errors were 0.216 and 0.094, which represent the lowest among all models. The R2 values for the training and test sets were 0.993 and 0.951, respectively, indicating higher predictive accuracy compared to the other three benchmark models. (3) According to the comprehensive correlation degree, the top five factors that had a greater impact on the amount of gas emissions were, successively, as follows: coal seam thickness (0.8896), coal seam gas content (0.8849), daily output (0.6456), coal seam depth (0.6258), and wind speed (0.5578). The research results can provide a reference for high-precision prediction of gas emissions.