Reliability assessment of ZPW-2000A track circuit using Bayesian network

Abstract

Nowadays, railway systems are expected to have high reliability, availability, maintainability, and safety. Thousands of ZPW-2000A track circuits are in operation on the mainline and high-speed railway network in China. This paper presents a scheme for reliability assessment of ZPW-2000A track circuit using Bayesian network. First, a fault tree is built based on expert experience and failure investigation. Then, fault tree model of the ZPW-2000A track circuit is converted into Bayesian network. Finally, quantitative analysis based on the Bayesian network is performed, including forward analysis and backward analysis, to calculate the system reliability and recognize vulnerabilities in the system. The validity of the proposed method is fundamental and meaningful for the real-time reliability assessment and condition based maintenance of ZPW-2000A track circuit.