煤矿瓦斯突出主主宰性信息提取技术研究.doc

Capturing Dominating Parameters of Gas Outburst Based on Bayesian Network( SUN Lianying1,2 PENG Suping2 (1 Institute of Information Technology of Beijing Union University, Beijing 100101, China; 2 The State Key laboratory of Coal Resources and Safe Mining, China University of Mining and Technology, Beijing 100083, China) Abstract: Gas outburst is a combined result of many geologic and mining engineering factors. In the process of mining, it’s impossible to capture all analysis parameters. Analyzed the characteristics of forecasting gas data, a method for capturing the dominating parameter of the gas outburst based on Bayesian network is presented, and a parameter sensitivity graded network model of gas forecast is established, by which can directly show structural relations between parameters and their effect in the quantities forecast of gas outburst. Its efficiency is proved by case. Keywords: gas outburst; dominating parameter; Bayesian network; sensitivity analysis 1 Introduction The gas outburst is one of the greatest mine safety problems faced by coal mine, which can not only damage the tunnel facilities and destroy the mine’s ventilation system, but also kill a great deal of people. So, effectively controlling and forecasting the gas becomes a dominating factor of mine’s normal safety production. Recent years, the gas forecasting technology has become a hot research point. The traditional gas forecast method is that studies the natural law of gas presence and outburst in the view of geology and believes that the coal seams gas is a resultant of geologic process and that the formation, transportation, presence and enrichment of gas have a closing relation with the geologic condition[1,2]. The technology index system that indicate gas outburst is presented. With mining deeply, the geologic conditions are more and more complex and the relation between geologic indexes is more and more fuzzy and difficult to describe. In resent years, by combining the basic t