Research On Reliability Analysis Of Urban Rail Transit Signaling System Based On GO-Bayesian Method

Urban rail transit signal system is a critical system to the safe operation of rail transit,it has a direct impact on the operational safety and traffic efficiency of entire urban rail transit system and personal safety of passengers.Urban rail operation system with the characteristic of low speed,large operating capacity,short distance between stations and operating complex environment,therefore,the signal system must have high reliability in order to ensure the safe operation of urban rail transit system.If the signal system equipment stays at low reliability for a long-term,there have high possibility that endanger the traffic safety of urban rail system.After reviewing the existing models of signal system reliability analysis,based on the affect factors of the signal system,this paper studies the reliability analysis methods as below:(1)This paper adopts a reliability analysis method GO-Bayesian network modeling method,which is suitable for urban rail transit signal system.The GO-Bayesian network method not only possesses the GO method modular modeling,multi-state simulation characteristics,but also combined with the advantages of Bayesian network system inference feature,at the same time,conversion process from GO method to dynamic Bayesian network can make up the universal procedure immature for the GO method,so that GO-Bayesian network method has more practicality and can use in urban rail transit signal system reliability analysis.(2)This paper defines three types of basic unit operators of GO-Bayesian network modeling,presents the operation rules and its applicable scope of each operator.Based on the basic reasoning model and signal flow and dynamic Bayesian network,network conversion rules from GO method to the dynamic Bayesian network have been presented,with the example of axle counter,this paper describes the details of model modeling steps of improved system GO-Bayesian network.(3)Based on Communication-Based Train Control(CBTC)and the geographical distribution of rail transit signal system,this paper divides the signal system into 4 sub-system.The GO-Bayesian network reliability analysis model of each sub-system of urban rail transit signal system is established according to the proposed GO-Bayesian network modeling method.Combining with the historical fault data of signal system,GO-Bayesian network reliability analysis algorithm and modeling software,the reliability of signal system is calculated and analyzed with the example of trackside subsystem and vehicle-mounted subsystem.The reliability status prediction of components and systems in a long-term operation of the system was also explored and predicted.(4)In this paper,the GO-FLOW reliability analysis model of on-board subsystem in urban rail transit signal system is further studied.The comparison analysis between GO-Bayesian network modeling and GO-FLOW method have been done based on the analysis results and analysis method.