Prediction of the height of water-conducting fracture zone using the FA-ALO-SVR model

Accurately predicting the height of the water-conducting fracture zone (WCFZ) on the coal seam roof is of great significance for ensuring the safe and efficient mining of coal seams at the mining face. To enhance the prediction accuracy of the WCFZ height, a multi-factor comprehensive analysis based on measured data from several mines in the Huainan Mining Area and Huaibei Mining Area has been conducted. A prediction index system for the height of WCFZ was established, incorporating factors such as coal seam roof type, mining method, mining depth, coal seam inclination, mining thickness, length of the working face slope, and the presence of faults in the working face. The primary component of their prediction model is the Support Vector Regression (SVR) model. Factor analysis (FA) was utilized to optimize the original data structure, and the ant lion optimization (ALO) algorithm was used to optimize the penalty factor "C" and the kernel function parameter "g" of the SVR model. Consequently, a prediction model for the WCFZ height was established based on FA-ALO-SVR. Subsequently, the predictive performance of the model was tested using new samples, and a comprehensive evaluation was conducted from three perspectives: prediction accuracy, prediction ability, and generalization ability. Five indicators, namely, mean absolute error (MAE), root mean square error (RMSE), mean relative error (MRE), Willmott's index of agreement (WIA), and Theil inequality coefficient (TIC), were used for the comparison with the traditional SVR model, the FA-SVR model, the ALO-SVR model and gray wolf optimization (GWO)-SVR model. The results indicated that the model achieved the lowest values for MAE, RMSE, MRE, and TIC, as well as the highest WIA value. Specifically, these values were 3.7857, 4.5393, 7.9066%, 0.1099, and 0.9370 respectively. This performance demonstrates a strong predictive capability of the model. Finally, the model was utilized to predict the height of WCFZ at the 3220 working face of Xutuan Mine in the Huaibei Mining Area. The prediction results indicated that the heights of WCFZ in both fault-affected and unaffected areas were 48.0120 m and 40.7602 m, respectively. These findings are largely consistent with the physical similarity simulation test results of 45.4 m and 40.0 m, thereby further validating the effectiveness and practicality of the model. This study introduced a novel approach for accurately predicting the height of the WCFZ in the coal seam roof of mines.