Study On The Regional Prediction Of Coal And Gas Outburst Risk Before Development

Coal and gas outburst is one of the most serious natural disasters threatening coal mine safety production.At present,China's coal and Gas Outburst Zone Prediction mainly depends on the underground roadway,measured gas parameters in developing mainly point prediction,site risk before coal and gas outburst prediction problems encountered in developing engineering practice,has not been a feasible solution.Support vector machine is a pattern recognition method can solve the small sample,nonlinear and high dimension problems very well,has been widely applied to many areas of production forecasting,this paper introduces the establishment of learning model to predict the outburst coal mine before the support vector machine,expect to improve the prediction accuracy before opening coal and gas outburst.The No.9 coal seam of a mine in Shanxi,Lvliang province has not been revealed,and the underground gas parameters have not been measured.The regularity of gas occurrence in the mine has been studied,and the gas parameters measured during the geological exploration period have been collected and analyzed.By drilling coal desorption method measured the same geological unit adjacent to the mine gas content,gas content of geological prospecting is modified;according to the relationship between the gas content and pressure,inverse gas pressure in coal seam by using indirect method.There are complicated nonlinear relationships between the various factors of coal and gas outburst,feasibility analysis of coal and gas outburst prediction of support vector machine in this paper,the introduction of particle swarm optimization algorithm(PSO),the parameters of SVM are optimized by particle swarm algorithm,develop the coal and gas outburst prediction model PSO-SVM.Collect belong to a geological unit adjacent mine outburst samples as the training samples,a coal mine in Lvliang geological prospecting revised data as test samples,PSO-SVM algorithm program written by MATLAB in the prediction of coal and gas outburst of coal and gas outburst prediction of No.9 coal seam,the results are consistent with the single index method and comprehensive index method and D K forecast.The support vector machines based on particle swarm optimization(PSO)are feasible to predict the pre coal and gas outburst risk,and provide the direction for prediction of coal and gas outburst before coal mine development.