For the precise prediction of coal spontaneous combustion (CSC) temperature and effectively prevent coal mine fires, this study proposes a temperature prediction model integrating the Sparrow Search Algorithm (SSA) and Random Forest (RF). Firstly, the gas production characteristics during CSC were analyzed via coal temperature-programmed experiments, and the correlation intensity between temperature and indicator gases at each stage was quantified using the grey relational analysis method. Secondly, SSA was applied to optimize the hyperparameters of the RF model, thus constructing the SSA-RF CSC temperature prediction model. Under the same experimental conditions, the prediction performance of the proposed model was compared with that of five other models. In addition, the applicability of the model was verified using field data collected by the borehole bundle monitoring system. The results show that moisture content exerts a dual effect on the CSC process, an appropriate amount of moisture can promote CSC, while excessively high moisture content will inhibit this process. The mean absolute error (MAE), root mean square error (RMSE) and coefficient of determination (R2) of the SSA-RF model are 1.63 °C, 2.64 °C and 0.9974, respectively, indicating that its prediction accuracy is superior to that of the other five comparative models. Meanwhile, the results of feature importance evaluation of the SSA-RF model are highly consistent with those of the grey relational analysis, which verifies the reliability of the model in screening key indicators. Further verification with field data shows that the SSA-RF model still maintains high prediction accuracy, with MAE, RMSE and R2 values of 0.35 °C, 0.45 °C and 0.9898, respectively, demonstrating good engineering applicability.
扫码关注我们
求助内容:
应助结果提醒方式:
