Research On Case-based Reasoning Early-Warning Support System Of Gas In Coal Mine

Gas accident is the most dangerous disaster in coal mine, and it has become the main contradiction which constrains the safety production of coal mine. It is one of the important issues in the field of coal mine safety to identify and analyze the dangerous factors that lead to gas accidents, and establish an effective early-warning support system of gas in coal mine at present. The paper synthetically considered and analyzed a variety of dangerous factors that lead to gas accident by using the theory of three types of hazard, setting up the fault tree of gas explosion in coal mine, and established an initial case-based reasoning early-warning support system of gas in coal mine.Firstly, the paper identified and analyzed the dangerous sources of accident in coal mine according to three types of hazard theory, and probed deeply into the contents of three types of hazard. On this basis, it fully analyzed the dangerous factors that lead to gas accident including physical factors (the factors lead to gas concentration exceeding limits, the fire sources cause a gas explosion) and the factors of management, and established the fault tree of gas explosion in coal mine.Secondly, the paper designed primary structure of early-warning support system of gas in coal mine based on CBR study. The system included a gas explosion early warning subsystem and a case-based reasoning gas explosion subsystem. The gas explosion early warning subsystem was composed of gas concentration exceeding limits early warning management and fire early warning management. The case-based reasoning for gas explosion subsystem provided decision-making support quickly when gas explosion happened by searching the similar cases, and help to control gas accidents in coalmines.Thirdly, the paper researched deeply on the key technology of case-based reasoning for gas explosion subsystem, such as case representation, case retrieval, case revision, case learning and maintenance. For the characteristics of gas explosion case retrieval, the weights of retrieval features were obtained clearly by using of cloud model, the similarity algorithm based on semantic class dictionary was proposed, and gave the detailed calculation process and formula about the word level, sentence level and paragraph level. Two methods provided a meaningful way for the implementation of gas explosion case-based reasoning system.In the paper, the case-based reasoning early-warning support system of gas in coal mine provided reference for the integrative emergency management and early-warning of gas accidents effectively in coal mine.