Prediction of coal and gas outburst: A method based on the BP neural network optimized by GASA

Abstract

With the development of computer science, many machine learning algorithms such as the back propagation (BP) neural network algorithm have been applied to the prediction of coal and gas outburst. However, the BP neural network algorithm can easily fall into the local minimum and has a low convergence rate. In this study, the genetic algorithm (GA) with strong global search ability and the simulated annealing algorithm (SA) with strong local search ability were integrated into a new GASA algorithm. This algorithm, which boasts strong spatial search ability and serves to optimize the initial weight and threshold of BP neural network, succeeds in solving the problem of easily falling into the local minimum. Furthermore, an improved GASA-BP prediction model was established by introducing an adaptive learning rate into the original BP neural network algorithm. The model is applied to coal and gas outburst prediction in Weicun Coal Mine in Jiaozuo City, Henan Province, China. Experimental results show that compared with the BP and GA-BP models, the GASA-BP model can predict coal and gas outburst accurately and quickly.