Prediction Model of Coal Spontaneous Combustion Based on PCA-GWO-SVM

Abstract

To effectively prevent coal spontaneous combustion disasters, the paper proposes coal spontaneous combustion prediction model based on PCA-GWO-SVM. The principal component analysis (PCA) method is used for attribute approximation and reduction of feature indicators with correlation. The grey wolf optimization algorithm (GWO) algorithm is introduced to optimally select the penalty parameter C and kernel parameter g of support vector machines (SVM). Taking 40 sets of historical data of coal spontaneous combustion in Xuandong No.2 coal mine as the research object, 30 sets of data are selected as training samples and the remaining 10 as prediction samples, the PCA-GWO-SVM model is trained and tested, and the predicted results were compared with those of Fisher and back propagation neural network (BPNN) models. The results show that the PCA method can eliminate the interaction between indicators to reduce the complexity of the model. The GWO algorithm can effectively improve the learning ability of the SVM algorithm. The coal spontaneous combustion prediction model based on PCA-GWO-SVM has higher prediction accuracy and good stability.