Research On Safety Evaluation Of Railway Passenger Transport Based On Bayesian Network

With the rapid development of China's railway,railway transportation is more comfortable and convenient for people travel,it results that rail passenger transport usher in again and again peak,Date dispay that China's railway passenger shipments reached more than 2.8 billion last year.In particular,the development of high-speed rail not only bring a lot of new opportunities,but also accompanied by new challenges of safe operation and other aspects.Meanwhile,it put forward newer and higher requirements for the safety production of railway passenger transport.Therefore,under the background of the security of railway passenger transport and the principle of "safety first",the paper utilize the advanced theory of safety science and Bayesian network to analyze and evaluate the security of railway passenger transport,with great theoretical and practical significance.First of all,on the basis of literature review,the paper utilize risk theory and the method of system safety engineering to analyzes the production safety of railway passenger transport in combination with the characteristics of railway passenger transport production and its safety production objectives,including the cause analysis of accident and the risk identification and risk analysis.To distinguish the risk from four aspects of personnel,equipment,environment and management,and use the improved TOPSIS method to sort the risks.Secondly,according to the results of the identification and analysis of risk and combined with the principle of index's construction,literature and expert consultation,the safety evaluation system of railway passenger transport is established from the four levels of personnel safety,equipment safety,environmental safety and management safety.Thirdly,in order to construct the topological structure of the Bayesian network model,by the index of safety evaluation is transfornned into the Bayesian network node based on the causal relationship.The range of each node is given.Combining Bayesian network and expert's decision,the paper studies the algorithm of determining the prior probability of the root node by collect expert opinion resoning update with Bayesian network's method.According to the actual situation,the paper studies the method to improve the conditional probability of non-root node.The paper constructs the model of railway passenger transport safety evaluation based on Bayesian network.Finally,the paper select a railway administration named X as a case to study.As a result,the safety level of X Railway Administration is "safer".Through diagnostic reasoning analysis,the main risk factors obtained according to the probability of root node's risk(above 0.05),and the important risk factors obtained according to the importance degree of root node.By compared with the result of fuzzy comprehensive evaluation,the paper verified the rationality and superiority of the model of railway passenger transport safety evaluation based on Bayesian network.