The Risk Research Of Subway Construction In Lanzhou Based On Bayesian Network

With the fast development of urbanization, urban subway construction has become the first choice to solve increasingly prominent problem of traffic congestion. Metro has a large capacity, little space occupation, security, comfort and environmental protection and other comparative advantages, but because subway construction belongs to underground closed operations, complex technology, a tight schedule, affected by the surrounding environment and other unpredictable subject, which brings challenges to the smooth development of the project. If a safety accident is happen, it will have disastrous effects on casualties, property damage and social environment. So to identify the risk factors in subway construction and formulate corresponding countermeasures to reduce or even eliminate the hidden danger of accidents is the imminent task.According to accident-causing theory and combining with the inquiries of experts, the risk factors in subway construction are divided into five dimensions of personnel, environment, materials, equipment and system 52 risk index. In order to reduce the workload and avoid unnecessary operations, after preliminary screening of repeated, with a correlation between the risk index, finishing 42 risk index. Firstly, use the improved TOPSIS method and combined with questionnaire to calculate the corresponding properties of each risk index(risk probability, risk degree of loss and risk control) of weights, secondly, calculating positive ideal solution and negative ideal solution of each risk indicators for the square sum of distance, which is used as the standard to evaluate the risk index, finally, the risk index is pre-sorting. On this basis, use entropy theory to the screening index. The smaller the entropy, the greater the amount of information provided by the index, the smaller the uncertainty, the larger the role of the index in the comprehensive evaluation, finally, the 18 more overall reasonable risk factors of the subway construction are extracted and analyzed.For subway construction has the own characteristics of the dynamic and uncertainty, this paper established Bayesian network model to analyze risk, and combined with a bid project of Lanzhou, Metro Line 1 for empirical analysis. Firstly, the prior probability of the evidence node and the conditional probability of the intermediate layer node of the Lanzhou Metro Line 1 were obtained by questionnaire, secondly, choose three forms of Bayesian network reasoning including causal reasoning, diagnosis reasoning and support reasoning to calculate the risk probability level of a bid project of Lanzhou, Metro Line 1 and find out the most broadly for chains, the reasoning results confirmed that the Bayesian network model in this paper is reasonable, effective and accurate, and it can be used in practical engineering risk analysis. Finally, in order to ensure the smooth completion of the subway construction in Lanzhou, according to the construction situation of Lanzhou, Metro Line 1, putting forward the main risk response measures and other risk responses, while also providing guidance and help for the construction of similar projects in future.