Failure And Reliability Analysis Of Urban Rail Train Braking System

Urban rail transit,a major means of urban public transportation,the good operation of urban rail trains is the key to ensure the safe operation of urban rail transit.The braking system is one of the core subsystems of urban rail vehicles,it plays a decisive role in safety and punctuality of the train.The stability and reliability of braking system are significant for the safe operation of urban rail transit.Due to the braking system has many components,and its structure is complex,the requirements for reliability analysis are high,and traditional reliability analysis methods are difficult to meet the requirements.According to the characteristics of the braking system,this paper proposes a GO-DBN model to analyse the reliability of the urban rail train braking system,which mainly includes the following four aspects:(1)Taking the domestic computer-controlled straight-through electro-pneumatic braking system developed by China Academy of Railway Sciences as an example,the system function and structure composition are studied in depth,and a comprehensive fault analysis of the braking system is performed.(2)A GO-DBN model is proposed,and the conversion rules are introduced in detail.The specific steps of mapping the GO method to a dynamic Bayesian network are given,and the calculation method of the prior state probability and conditional probability table of the nodes are also given.(3)Considering the maintenance,a GO-DBN model was established based on the braking system schematic diagram.According to the conversion rules,the GO method of the braking system is mapped to the DBN model.Based on the stationary and Markov assumptions,the prior probability distribution and conditional probability table of the nodes are calculated to obtain the reliability data of each component.The Bayesian modeling software Ge NIe 2.3 is used to solve the model and the effectiveness of the braking system is obtained.Inferring the dynamic Bayesian network model of the braking system,the posterior probability of each component of the system is obtained,and sort them.The weak links in the system unit are identified as brake cylinders,brake hoses,speed sensors,anti-skid vent valves,plug doors,and air spring pressure sensors.The result can help maintenance personnel quickly find the source of the fault,improve maintenance efficiency,and enhance the system’s reliability.(4)Considering the unit correlation and closed-loop feedback structure,the GO-DBN optimization model of the braking system is established.The equivalent units are used to represent the shutdown correlation in the series structure of the system.The new type 18 operator is used to represent the closed-loop feedback structure.The prior probability distribution of the equivalent units are obtained by solving the linear equations.Similarly,the Bayesian modeling software Ge NIe 2.3 is used to solve the brake system optimization model,and the effectiveness of the optimization model is obtained.Comparing the reliability results before and after the optimization,it is found that the efficiency of the system after optimization is higher,that is,the probability of the system working normally is higher,indicating that the optimization is in line with the actual working conditions.Although the result of the original model and the optimized model changes slightly,the optimized model further improves the accuracy of the reliability analysis.