The Application Study Of Self-organizing Competition Neural Network Based On PSO Algorithm In The Prediction Of Coal And Methane Outburst

As a conventional energy, coal, which occupies an important position in social production activities in our country, amounts to more than two thirds of primary energy consumption.And the energy structure won’t have a qualitative change in the future30years. However, the coal and methane outburst which is one of the most serious disasters, always threat to mine safety of workers’life and coal production. So it is very important to take preventive measures for coal and methane outburst. According to "quaternity" prevent policy in "Prevent Outstanding Regulation",coal and methane outburst prediction is the first and most important step. Completing the coal and methane outburst forecast work, can not only ensure downhole staffs life safety and the property of coal mine safety, ensure the safety of coal production, but also improve the economic benefit of coal mine, safeguard China’s energy supply, make a contribution to the rapid and healthy development contribution strength of the national economy.This thesis summarizes the situation of outstanding disaster and all kinds of outstanding prediction method at home and abroad, and analyzes the mechanism of coal and methane outburst and the general rules of influence factors.On that basis,this thesis points out the deficiency of the existing prediction method, namely can’t use a single index and fixed threshold method to predict coal and gas outburst. In order to solve the coal and methane outburst prediction problem, on the basis of in-depth study the basic concept and algorithm based on the principle of the self-organizing competitive network and PSO algorithm, this thesis analyzes the advantages and disadvantages of self-organizing competitive network, and uses the PSO algorithm to optimize the net. After the optimization of network weights through the PSO algorithm and improve the stability of the network output, the net can achieve the best classification results. Then run the network after the optimization in the MATLAB platform, and establish the coal and methane outburst prediction model, and applied it to the actual production of coal mine. This thesis selects in Shaqu coal mine as the research object, and take the biggest drilling cuttings amount (S), tortuous cinder analytical indexes (K1), Initial velocity of gas emission from borehole (q) and coal rigidity coefficient(f) as prediction index, which is considered as the input of the network after optimization. After the operation, the output results are in keeping with the observation results.So using this method to predict coal and methane outburst is feasible. Compared with the traditional prediction method, using the method of prediction results can effectively save the cost of mine outburst prevention,under the precondition of ensured safety.Because of the complexity of coal and methane outburst mechanism and the difference of each mine geological structure, methane occurrence and mining methods situation, there is a complex relationship between each factors and coal and methane outburst.The method which is introduced in this thesis, not only can accurate forecasting coal and methane outburst, provides a new way of thinking for coal and gas outburst theory research, but also provide a new reference method of coal and methane outburst forecast for mine in practical production management.