On The Combined Application Of Genetic Algorithm And Neural Network In The Prediction Of Coal/Gas Outbursts

Coal is the main source of energy in china, in current and coming decades,and it still take the most part in primary energy structure.As one of the most damaging nature disaster, coal and gas outburst seriously restrict the economic efficiency of mine and threatened the safety of miners.Considering the factors that influenced coal and gas outburst are complicated, the disciplines of outburst still not completely known,and accidents of coal mine gas outburst becoming even more frequently,this article built a modle combined the genetic algorithm and BP neural network based on previous studies.The main contents are as follows1.The article described Status of coal and gas outburst forecast here and abroad, the necessity and significance to forecast the outburst, and searched the problems of prediction of outburst, confirmed the searching plan that based on coal and gas outburst mechanism and nural network.2.Its mechanism,factors and the general laws of coal and gas outburst have been discussed,the model of coal and gas outburst was designed in virtue of BP neural network and MATLAB programming language.By using this model and the data of coal and gas outburst index that tested by China University of Mining in several mines in the Huainan region, we conduct simulation and prediction, pointing out that in training, BP neural network model has problems such as slow convergence, low accuracy in forcast.3.Based on problems that appeared in BP neural network model of coal and gas outburst, we introduced the concept of genetic algorithm, disscused the advantages and disadvantages of Genetic Algorithm and BP algorithm, proposed an optimization program that use genetic algorithm to improve BP neural network, we established a coal and gas outburst prediction model which combined genetic algorithm and BP neural networks, and using the same data to predict, finding that convergence speed is accelerated, accuracy has been improved.All that proved this program is feasible and effective.4.Combined with mine field data and the practical situation in Sihe Jincheng Coal Industry Group in Shanxi, the article analyzed the main control factors and determined drilling cuttings desorption index of parameters of gas outburst(K1),the maximum amount of drilling(S), drilling gas emission rate(q0),and consistent coefficient of coal(f) of and correlation parameter as Basic indicators influenced Coal and gas outburst. Use of coal and gas outburst prediction system,Re-forecasted risk areas of outburst that have been forecasted using traditional single-target prediction. The final forecast conclusions shows that by using multi-index comprehensive determination it can reduce the total work of preventing coal and gas outburst to 35% under the guarantee of ensuring safety.As the factors that influence coal and gas outburst are complex, and coal geological structure,occurrence conditions,mining methods and other factors are different in different mine,the mine control factors influenced coal and gas outburst are different.To predict it with the method mentioned befor, only need to adjust some prominent indicators. By this way, it can greatly improve accuracy by using the traditional single indicator to predict the risk of coal and gas outburst, making the prediction and prevention of outburst even more scientific and rational. That is the purpose of this article.Wish it Plays an active role in coal and gas outburst forecast.