Prediction of Destroyed Floor Depth Based on Principal Component Analysis (PCA)-Genetic Algorithm (GA)-Support Vector Regression (SVR)

Abstract

In order to prevent water inrush from working face floor, it is an urgent problem to predict destroyed floor depth. In this paper, based on the collected samples of floor failure depth of north China type coal fields, the factors that influence the development of floor failure depth were obtained by principal component analysis(PCA), that is, mine pressure, working face information, tectonic conditions, lithology, and sedimentary conditions of coal seam. Then, the parameters of support vector regression (SVR) were optimized by genetic algorithm (GA), the PCA-GA-SVR prediction model of floor failure depth was established, we put the last 4 samples into the model and verified its excellent generalization ability. Finally, we applied the PCA-GA-SVR model to the floor failure depth prediction of Zhaizhen coal mine, comparing the predicted value of the model with the field measured value using double side seal borehole water injection device, the effectiveness of the model was demonstrated.