Research on the Strength Prediction Method of Coal and Rock Mass Based on the Signal While Drilling in a Coal Mine

To study the response relationship between drilling signal and rock mass geomechanical parameters, accurately and quickly perceive and predict the strength of coal and rock mass, guide the optimization of drilling control parameters and the design of the support scheme, and improve the efficiency of roadway excavation, the prediction of rock uniaxial compressive strength based on drilling signal was carried out. Based on the 112,206 return air chute in the Xiaobaodang No.1 Coal Mine as the engineering background, through the drilling data obtained from the roof anchor cable support, data processing, and feature selection, this paper establishes a coal and rock mass strength prediction model based on the AdaBoost integrated algorithm, optimizes the hyperparameter of the model, and analyzes and evaluates the prediction results. The results show that in the AdaBoost integration model, the R2 of SVM is the highest, 0.972, and the values of RMSE, MAE, MAPE, and other error indicators are the lowest. The prediction accuracies of the SVM model, tree model, and linear model are 98.8%, 85.4%, and 75.6%, respectively. The experimental results show that the AdaBoost integrated algorithm using a based learning machine has higher prediction accuracy. At the same time, compared with the current advanced model, it further verifies the effectiveness of the model in the coal mine.