Research On Prediction And Control Of Hidden Dangers In Railway Depot Based On Bayesian Network

Railway is the artery of national economic development.Railway safety is related to people’s life and property safety,and directly affects social stability and railway operating efficiency.Once railway safety accidents are discovered,it will cause extremely serious consequences and have great influence on public opinion and social harmony.Therefore,China State Railway Group Co.,Ltd.has invested heavily in safety management,especially in the safety management system of civil air defense,physical defense and technical defense,such as vigorously promoting CTC equipment,interval logic occupancy inspection equipment and other scientific and technological means to ensure safety,but it still cannot completely prevent accidents.According to Heinrich’s law,no accident happens by accident,and there must be a large number of hidden safety problems behind it.If we can mine the hidden safety problems found in the past and discover the rules,we can accurately predict and effectively control the hidden safety problems,and then the probability of accidents will be greatly reduced.In this paper,Bayesian network method is used to analyze the relationship between the hidden troubles found in the depot by the group company and the hidden troubles found by the depot managers’ self-inspection,and find out which hidden troubles found by the depot self-inspection will lead to the hidden troubles found by the group company.At the same time,under the condition of obtaining a large number of self-inspection problems of the depot,the hidden troubles found by the group company and their probability are predicted,and the methods to control the hidden troubles found by the group company’s inspection are put forward,and auxiliary support is provided for safety decision-making,thus reducing the occurrence probability of hidden troubles.This research is of great significance to improve the overall safety level of the depot,reduce the incidence of railway traffic accidents and ensure the personal and property safety of passengers,cargo owners and employees.First of all,on the basis of consulting and studying a large number of literatures,combined with the characteristics of railway safety management,this paper introduces the methodology of railway double mechanism construction,and makes it clear that hidden dangers are the enemy of safety production and should be eliminated and avoided as much as possible,so it is more valuable to study the relationship between hidden dangers.Collect the hidden troubles found in the self-inspection of Qingdao Railway Depot of Jinan Bureau Group Company in recent years and the safety inspection of Qingdao Railway Depot by Jinan Bureau Group Company,including the notices and instructions of hidden troubles from 2016 to 2020,determine the classification basis of hidden troubles,and classify hidden troubles.Using the expert analysis method,this paper analyzes and studies the internal logical relationship between all kinds of hidden trouble problems found by the group company’s inspection and those found by the locomotive depot’s self-inspection,and based on this,builds the Bayesian network structure model of security hidden trouble problems.Then,according to Bayesian network theory,combining with practical problems,the parameters and calculation methods of Bayesian network of hidden trouble problems are determined.Taking hidden trouble problems from 2016 to 2020 as data samples,the conditional probability is calculated by statistics,and a complete Bayesian network of hidden trouble problems is established.Using Bayesian network’s powerful ability to deal with uncertain problems,this paper studies the causes and results of hidden troubles,deeply analyzes the specific causes of hidden troubles and possible hidden troubles,and analyzes the final results to find out the chain of hidden troubles.On the basis of obtaining the probability of hidden trouble discovered by the self-inspection of the train depot in 2021,the Bayesian network model is used to predict the occurrence rate of hidden trouble discovered by the group company in 2021,and compared with the actual occurrence rate of hidden trouble discovered by the group company,which proves the accuracy of the Bayesian network and the feasibility of the research.Finally,on the basis of verifying the validity of the Bayesian network model,according to the analysis results of the Bayesian network model,the measures to control the hidden problems found in the inspection of group companies are put forward.In the paper,a total of 10 maps,13 tables and 43 references.