Study On Safety Risk Identification And Evaluation Of Coal Mine Gas

Gas accident threatens the work safety in coal mines.Hence,systematic analysing the causes leading to gas accidents and realization of early warning of such incidents is one of urgent demands to address the safety in underground mines.In order to effectively prevent gas accidents,the causing factors need to be further investigated,which will provide theoretical basis for objective,scientific and correct evaluation of coal mine safety risks.In this regard,study on the gas safety risks can not only enhance safety management of coal mining enterprises,but is of realistic significance and theoretical value to minimize the safety risks,strengthen disaster prevention capacity and enrich gas disaster prevention theory for the enterprises.This dissertation adopts such approaches as literature research,theoretical study,field investigation,algorithm improvement and demonstrative analysis to conduct the study on coal mine gas safety risk.Starting from the causing mechanism,the dissertation probes into the shortage in the gas disaster prevention technologies and measures,and further explores three scientific issues:identification of causing factors,pre-evaluation of safety risks and safety risk evaluation.Using associated rule theory and data mining technique,it carries out the research on the identification of the causing factors of gas safety risks.Using Markov Chain model,it predicts and analyzes the gas safety risks.By optimizing algorithm and improving wolf colony algorithm and combining BP neural network,it establishes coal mine gas safety evaluation model,which enhances the accuracy of the evaluation.Meanwhile,the dissertation conducts demonstrative analysis of the gas safety risk prediction model and safety evaluation model,to further validate its correctness.Main research conclusions include:(1)Literature summary is made for the identification of coal mine gas safety risk factors,study on the gas accident causing mechanism and gas safety risk evaluation method.It identifies such problems as incorrect and incomplete understanding in identification of gas risk factors,pre-evaluation of risks and safety evaluation,which provides a new way of thought for gas disaster prevention study(2)Based on coal mine disaster prevention theory and safety evaluation theory,etc.,the dissertation,in respect of the shortage in current gas accident control technologies and measures,points out that further improvement and study is needed in the accuracy and completeness of identifying the gas safety risk factors,gas safety risk measurement and prediction correctness,and gas safety risk evaluation correctness.These are also the theoretical bottleneck and technical difficulties in lowering gas safety risks and preventing gas accidents.(3)According to 72 gas accident cases and 85 causing factors,the dissertation constructs the causing factor model of coal mine gas safety risks,and establishes gas safety risk network model.At the same time,adopting associated rule theory,it establishes causing factor relationship model accordingly.Finally,30 evaluation indicators influencing coal mine gas safety risks are chosen.Thus,major risk factor collection leading to gas accidents is achieved,which lays a basic indicator system for the following prediction and safety evaluation of coal mine gas safety risks.(4)According to the situation of gas disasters in China's coal mines,the dissertation counts and analyzes the gas accidents in the ten years,and adopts Markov Chain model for safety risk predication and analysis.The evaluation results comply with the site conditions,proving the applicability and advantages of Markov Chain model in coal mine gas accident prediction,and providing reference to the multi-factor accident safety risk prediction.(5)By comparative analysis of algorithm,it demonstrates the advantages of WPA in global search and function optimization,and has sound robust characteristics.Based on related theories such as BP neural network and wolf colony algorithm,it presents a combination of belief learning model optimization-based wolf colony algorithm and BP neural network,and designs the computation flow of the model.Several samples demonstrate the feasibility of IWPA-BP model in function fitting.Then,the improved model is applied to coal mine gas safety risk evaluation.The model testing results indicate a good agreement of simulation classification results with expert classification ones,proving the reliability of the established safety risk evaluation model in coal mine gas safety risk evaluation.(6)In order to further confirm the correctness and reliability of the gas safety risk prediction model and safety evaluation model,a demonstrative analysis is conducted,where Liyazhuang Coal Mine is taken for example.Through identifying the gas safety risk factors of the mine,the main factors influencing the mine's gas safety are determined.Combining the established Markov Chain and IWPA-BP neural network gas risk evaluation model,a demonstrative analysis is made for the gas safety risk and safety of the mine.The results validate the model is scientific and objective in the mechanism analysis and evaluation of coal mine gas safety risks.The research renders a way of thought and method for the risk control in similar coal mines,and has good promotion and application value in coal mine enterprises.By systematic study on the coal mine gas safety risks,the dissertation achieves innovative outcomes in identification of gas safety risk factors,safety risk level prediction and analysis and safety evaluation,as follows:(1)Using risk-causing theory and the associated rule,it establishes the network model and relationship model for coal mine gas safety risk causing factors and determines the main risk factor set,laying the indicator system for the coal mine gas safety risk level prediction and safety evaluation.(2)By referring to the coal mine safety risk evaluation method and concept,Markov Chain model is adopted for evaluation of the occurrence probability and risk damage of coal and gas outburst.Accordingly,the evaluation method is established,which conducts quantitative measurement and prediction of gas safety risk level.(3)Based on related theories such as BP neural network and wolf colony algorithm,the belief learning-based model is optimized to improve the wolf colony algorithm,and combining BP neural network,it constructs IWPA-BP neural network as coal mine gas safety risk evaluation model,which enhances the efficiency and correctness in scientific evaluation of coal mine gas safety risks.