Research On Comprehensive Evaluation Methon Of Reliability Of Existing Railway Bridage Based On Bayes Theory

With the rapid development of national transportation construction,the surge of traffic volume and the increase of train load both pose new challenges to the reliability evaluation of existing railway bridges.The reliability evaluat ion of existing railway bridges is directly related to traffic safety.Based on the background of beam bridge,this paper conducts a theoretical study on the comprehensive reliability evaluation of existing railway bridges,summarizes and summarizes the work carried out in the reliability evaluation of bridges at home and abroad,and analyzes the limitations of current bridge evaluation methods in evaluation methods,evaluation angles and evaluation theories.The purpose is to evaluate the reliability of railway bridge structures under various disease factors through Bayesian network model,and establish a comprehensive reliability evaluation method of existing railway bridges based on Bayesian theory.Firstly,the risk factors of railway bridge structure are analyzed,the selection principle of risk factors is put forward,the identification process of existing risk factors is clarified,the index system including three risk subsystems of safety,applicability and durability is constructed,the grading eval uation standard of each risk factor index is put forward,and the correlation between risk factors of railway bridge is analyzed.Based on the fault tree theory,a multi-level and key Bayesian network structure model is established.By introducing the con cept of ’ fuzzy ’,the prior probability of root nodes is determined,and the conditional probability between non-root nodes of Bayesian network structure is determined by analytic hierarchy process.Secondly,the Bayesian network model can be used for bid irectional reasoning and sensitivity analysis to predict the occurrence of accidents and determine the causes of accidents.It can continuously integrate the health monitoring information of railway bridges during service,update the Bayesian network param eters,and reflect the health status of railway bridge structures in time.The Bayesian network evaluation system of existing railway bridge structure is written based on Python language,which realizes the programming of the Bayesian evaluation system of existing railway bridge structure,and can efficiently and accurately carry out the reasoning and analysis of Bayesian network.Finally,the reliability evaluation of Songhua River Bridge on Binbei Line is used to verify the reliability comprehensive evaluation method of existing railway bridges,and the failure probability of each risk factor is evaluated.With the diagnostic reasoning ability of Bayesian network model,the main causes of structural failure are analyzed,and the most likely risk chain is judged.The sensitivity analysis is carried out to determine the sensitivity value of each node,and the targeted suggestions are given according to the failure probability of each risk factor,which provides a theoretical basis for the maintenance and reinforcement of Songhua River Bridge on Binbei Line.