No abstract

The reliability and availability of the onboard high-speed train control system are important to guarantee operational efficiency and railway safety. Failures occurring in the onboard system may result in serious accidents. In the analysis of the effects of failure, it is significant to consider the operation of an onboard system. This article presents a systemic approach to evaluate the reliability and availability for the onboard system based on dynamic Bayesian network, with taking into account dynamic failure behaviors, imperfect coverage factors, and temporal effects in the operational phase. The case studies are presented and compared for onboard systems with different redundant strategies, that is, the triple modular redundancy, hot spare double dual, and cold spare double dual. Dynamic fault trees of the three kinds of onboard system are constructed and mapped into dynamic Bayesian networks. The forward and backward inferences are conducted not only to evaluate the reliability and availability but also to recognize the vulnerabilities of the onboard systems. A sensitivity analysis is carried out for evaluating the effects of failure rates subject to uncertainties. To improve the reliability and availability, the recovery mechanism should be paid more attention. Finally, the proposed approach is validated with the field data from one railway bureau in China and some industrial impacts are provided.

Application of Dynamic Bayesian network to reliability and availability evaluation of CTCS-3 onboard system Highlights:1. Using DBN-based approach to evaluate the reliability and availability, taking account of dynamic failure behavior, recovery mechanism, and temporal effect. 2. Combining some actual situation and analyzing the hierarchical architecture and field data of CTCS-3 onboard system. 3. Comparing the reliability and availability of three kinds of onboard system and analyzing the degraded state.Findings:1.The CDD, TMR and HDD onboard system possess high reliability and availability. Besides, the low occurrence probabilities of degraded state are almost the same. 2.The VDX, RLU, C3CPU, C2CPU, and Power should be given more attention in physical level. 3.The effects of failure rate on the onboard systems follow the order: WP > TD > KN > PB> LP, and KN> WP> TD > PB> LP for CDD and HDD system, respectively. However, the order is KN>TD>WP > PB> LP for TMR system. 4.The recovery mechanism should be paid more attention to improve the reliability and availability of CTCS-3 onboard system. 5.The results of availability are validated by the field data from one railway bureau.Application of Dynamic Bayesian network to reliability and availability evaluation of CTCS-3 onboard system Abstract This paper presents a systemic approach to evaluate the reliability and availability of Chinese Train Control System Level 3 (CTCS-3) onboard system based on dynamic Bayesian network (DBN), aiming to solve the problems such as dynamic failure behavior, recovery mechanism, and temporal effect. Taking account of the actual operational situation, the hierarchical architecture and field data of the CTCS-3 onboard system is analyzed. Classified by the redundancy strategy of vital computer (VC), three kinds of train control systems corresponding to the triple modular redundancy VC (TMR), hot standby double dual VC (HDD), and cold standby double dual VC (CDD) are presented. By mapping the dynamic fault tree into DBN, the structure and parameter modeling are conducted in a case study. Adopting the forward inference, the results show that the reliability and availability of CDD onboard system are higher than HDD, while TMR is between them. And, the low occurrence probabilities of degraded state are almost the same. Through the backward analysis, the difference between posterior and prior probability shows the VDX, RLU, C3CPU, C2CPU, and Power should be paid more attention. The results of sensitivity analysis demonstrate that the effects of failure rate on the onboard systems follow the order: WP > TD > KN > PB> LP, and KN> WP> TD > PB> LP for CDD and HDD system, respectively. However, the order is KN>TD>WP > PB> LP for TMR system. The recovery mechanism should be paid more attention to improve the reliability and availability of CTCS-3 onboard system. The validation of the proposed approach is demonstrated by the field data from one railway bureau.