Research On Prediction Of Coal And Gas Outburst Based On KPCA And Improved Extreme Learning Machine

Coal and gas outburst is one of the most destructive and most dangerous mine safety accidents in coal mining,and its essence is an extremely complex mine dynamic phenomenon.How to effectively predict gas disasters has always been an important research topic in the field of coal mine safety.However,in view of the fact that the nature of coal and gas outburst mechanism has not been fully understood,and the prominent disaster itself is affected by a variety of complex factors,it is a high-dimensional problem with strong nonlinearity and complexity,which is difficult to use traditional prediction methods to achieve accurate forecasting.Therefore,in order to improve the prediction accuracy and prediction efficiency of gas disasters,based on the analysis of the mechanism and influencing factors of coal and gas outburst,this thesis draws on the two working faces and regional outburst predictions described in China's Regulations on Prevention of Coal and Gas Outburst.The method and the rules for establishing the predictive index system are analyzed from the perspective of data analysis.A new prominent predictive index system is proposed,and a coal and gas outburst prediction model based on KPCA and IQGA-ELM is established.It is used in the prediction of coal and gas outburst prediction in coal mining face.This thesis aims to accurately and quickly achieve coal and gas outburst prediction,and mainly works on the following aspects:(1)Summarize the current mainstream coal and gas outburst mechanism hypothesis.By comparing different coal and gas outburst prediction methods,analyze the research status and current deficiencies in this field,and discuss the necessity and realistic meaning of establishing coal and gas outburst prediction model.(2)The related factors affecting coal and gas outburst are analyzed,and the prediction system of surface area prediction index is established.The kernel principal component analysis method is used to analyze the weight of coal and gas outburst prediction indicators,and it is used as the selection principle of the input variables of the prominent prediction model.The kernel principal component analysis method can simplify the input variables of the prediction model,accelerate the prediction rate,and improve the prediction accuracy.(3)Through the in-depth analysis of the principle of extreme learning machine,based on the shortcomings of randomly generating the input layer weight and the hidden layer threshold,the improved quantum immune genetic algorithm is used to improve the extreme learning machine,and the ELM coal and gas prediction model based on IQGA algorithm is established.(4)The established KPCA and IQGA-ELM prediction model is verified by examples.The prediction experiment was carried out using the data of Qianjiaying Mine,and the prediction results verified the accuracy of the prediction model.And as well as comparing this model with other models for predictive performance highlights the good predictive performance of the model.