基于动态贝叶斯网络的列控中心可靠性及可用性评估

为解决共因失效、动态失效及恢复机制等问题，本文基于动态贝叶斯网络对联锁车站和中继站列控中心可靠性及可用性进行评估.在分析列控中心系统结构的基础上构建系统动态故障树，并将动态故障树转化为动态贝叶斯网络，实现结构学习和参数学习.通过动态贝叶斯网络正向推理得到两种类型列控中心可靠度和可用度并进行比较分析.通过动态贝叶斯网络反向推理得到列控中心系统薄弱环节.研究系统敏感性因素并讨论恢复机制对系统可靠性及可用性的影响.结果表明：考虑共因失效的联锁车站和中继站列控中心的稳态可用度分别为0.999 960和0.999 977；电源及驱动采集单元为系统薄弱环节，需要重点关注.该方法能有效提高列控中心智能维修维护水平.

Reliability and Availability Evaluation of Train Control Center Based on Dynamic Bayesian Network

This paper presents a systemic approach to evaluate the reliability and availability of train control center in interlocking station (ITCC) and relay station (RTCC) based on dynamic Bayesian network, aiming to solve the problems such as common cause failure, dynamic failure behavior, and recovery mechanism. Taking account of the systemic architecture, the dynamic fault tree is constructed. By mapping dynamic fault tree to dynamic Bayesian network(DBN), the structure and parameter modeling are conducted. Based on the forward inference of DBN, the reliability and availability of ITCC and RTCC are obtained and compared. Based on the backward inference of DBN, the vulnerabilities of train control center are recognized. Moreover, the sensitivity analysis and the effect of recovery mechanism on reliability and availability are researched. The results show that the availabilities of ITCC and RTCC with common cause failure are 0.999 960 and 0.999 977, respectively; the DY and PIO are the vulnerabilities, which should be paid more attention. The proposed approach can improve the intelligent maintenance level of train control center.