The Forecast Methods Of Mining Risk Of The Coal Resources Affected By Dynamic Disaster

Coal mine's dynamic phenomena is the phenomena that the coal or rock in the high-stress state which have accumulated a large number of elastic energy, suddenly damage, fall or thrown out, and release a lot of energy.The coal is the main energy of our country, coal account for more than 70% in the energy structure. However, coal mine disasters often occur, for example, in Fushun, Beijing, Datong, Zaozhuang, Xinwen, Kailuan and other coal mines all have rock burst phenomenon, and the rock burst have caused a great harm. In coal mines, coal and gas outburst is even more frequent, according to statistics, there 107 coal mines have coal and gas outburst disaster in our country, the number of coal and gas outburst disaster account for 35% in the total number of the world's, the coal and gas outburst have caused many casualties and economic losses. Rock burst and coal and gas outburst are the typical dynamic disasters, so the coal mine dynamic disasters have seriously affected our country's coal mining.Scientific and effective forecast the coal mine's dynamic disasters, can reduce the probability of occurrence of dynamic disaster.The rough sets-artificial neural network and rough sets-support vector machine technology are used to establish the rock burst and the coal and gas outburst prediction model in this paper. The main contents are as follows:The rock burst and coal and gas outburst mechanism have been analyzed; The rough sets-neural networks and rough sets-support vector machines prediction models of the rock burst risk have been established; The rough sets-neural networks and rough sets-support vector machines prediction models of the coal and gas outburst have been established; The above models have been tested, and the models' prediction results of rough sets-neural networks and rough sets-support vector machines have been compared, the results show:In the rock burst risk prediction and the coal and gas outburst prediction, the prediction accuracy rate of rough sets-support vector machine models are higher than the rough sets-neural networks, so the rough sets- support vector machine approach is more suitable for coal mine's dynamic disaster forecast than the rough sets- neural network method.