Prediction Model of Gas Disaster Based on Quasi-Newton Algorithm BP Neural Network

The forecasting of gas outburst and explosion accidents is one of the most pressing problem in current China’s coal mine safety production.Introducing the Quasi-Newton optimization algorithm of BP neural network,this paper discusses the mathematical model,network architecture and programming design of establishing the gas disaster forecasting model of Quasi-Newton optimization algorithm BP neural network under the premise of keeping the relationship among the spatial entities and their distributions,and an instance of Jining No.2 coal mine is tested.The result shows that this model is stable,fast and high prediction accuracy,which can simulate the mine gas outburst and explosion accidents characteristics and make more accurate predictions on the gas disaster.