Research On Bayesian Network-based Scenarios Analysis Of Gas Accident In Coal Mine

Gas accident is the most dangerous disaster in coal mine, and it has became the mainhidden danger which constrains the safety production of coal mine. It is one of the importantissues in the field of coal mine safety to construct validity model and make scenario analysis.The paper synthetically analyzed each element of coal accident system in coal mine accordingto the theory of systematology, setting up the chain of gas explosion in coal mine, andelaborated the evolution theory of coal accident. On the existing foundation, it established themodel based on Bayesian network of gas accident in coal mine, and then calculated themathematical expectation of node variable in the output layer by the way of Bayesian networkinference algorithm. According to this research, the managers could preliminary judge theevolutionary trend of accident by the current status, and then made decisions quickly in orderto prevent the occur of accident again.Firstly, the paper identified and analyzed the system of coal accident in coal mine byusing the theory of systematology, and discussed deeply these elements in accident system.On this basis, it analyzed the form of gas accident and its construction, and then made a stepanalysis for gas accident in coal mine, which could provide theoretical basis for constructingthe model of Bayesian network.Secondly, the paper preliminary structured the network model of gas accident in coalmine including the study of network structure and parameters on the base of deeplyresearching Bayesian network. According to the method of researching experts’knowledge inthis field and learning historical material, it obtained the network nodes and its tables ofprobabilities about the network of gas accident in coal mine. This method verified the validityof data and provided analytical basis for the below scene.Thirdly, the paper discussed deeply on scene evolution, which was divided into happenedscene, developing scene and evolved scene and disappeared scene, of gas accident in coalmine by combining the theoretical knowledge of scenarios analysis. On this basis, mathematical expectation of coal mine gas explosion were calculated according to the modelof Bayesian network, which helped to reason state in key scenes. The greater the value ofmathematical expectation, the larger the harm which caused in this path.In the paper, the Bayesian network-based scenarios analysis of gas accident in coal mineprovided the accordance for decision maker who should make the reliable, real time andeffective decision when the gas occurred the burst in coal mine, and reduced the lose causedby the accident, preventing the appearance of gas accident again.