Grey Relational Analysis On Coal And Gas Outburst In Mine

This study adopts an improved grey relational analysis comprehensive evaluation method,based on the accident cause index system and the 2-4 model,comprehensively and multi-dimensionally study the causes of coal and gas outburst accidents in my country in the past 20 years.Constructed an index system for the causes of coal and gas outburst,and the system is applied to specific engineering examples.Based on coal seam gas data in Furong mining area,the random forest algorithm and AHP are combined to determining the weight of each index in the coal and gas outburst index system in Furong mining area.The weight of each index in the outburst index system is used to predict the risk of coal and gas outburst in Furong mining area by using the classic algorithm of machine learning.On this basis,the improved GRA-AHP analysis method is used to evaluate the correlation between coal and gas outburst ash and predict the coal and gas outburst risk of coal mines in Xingwen County,provide more accurate control measures for coal mining enterprises to prevent coal and gas outburst.The specific work contents are as follows:(1)Through the analysis of the major and extraordinarily serious accidents of coal and gas outburst in China from 2001 to 2020,the time and space rules of the accident are obtained.The research shows that: the main accident month is December,the main occurrence time are 1:00 and 18:00,the main shift is evening shift;the main accident provinces are Guizhou,Hunan,Henan,Chongqing,and Yunnan,the main accidents occurred in the coal mining face and the excavation face.(2)According to the Accident Cause Index System and the 2-4 Model,the causes of coal and gas outburst accidents are analyzed.Take the overall accident risk as the target layer from five aspects: gas factor,coal body factor,geological factor,location factor and human factor classification and identification,12 primary indicators and 18 secondary indicators are proposed to construct the cause index system of coal and gas outburst accidents.(3)The improved GRA-AHP analysis method is used in constructing the weights,combining the objective weights determined by the Random Forest algorithm with the subjective weights determined by the expert scoring method,to improve the drawbacks of the traditional grey relational evaluation method being too subjective.(4)Based on the gas geological data of the Furong mining area,the Adaboost ensemble algorithm,SVM support vector machine,and NBM naive Bayes are used to learn the sample data.The learning results show that the Adaboost algorithm has a high accuracy in predicting coal and gas outburst accidents in the Furong mining area,and it is used to verify the risks of coal and gas outburst in Furong mining area,which can more accurately predict the possibility of coal and gas outburst in mines.(5)Through actual cases,the feasibility of improving the gray correlation evaluation methods and the effectiveness of the coal and gas outburst prediction model were verified.On this basis,the gray correlation evaluation of the Sichuan Xingwen Jianshe Coal Mine was carried out.According to the evaluation results,improvement measures are put forward.