基于贝叶斯网络的CTCS-3级列控系统车载子系统可靠性评估  

Dependability Assessment of CTCS-3 On-Board Subsystem Based on Bayesian Network

作　　者：苏宏升[1] 车玉龙[1] 张友鹏[1] 

机构地区：[1]兰州交通大学自动化与电气工程学院,甘肃兰州730070

出　　处：《中国铁道科学》2014年第5期96-104,共9页China Railway Science

基　　金：铁道部科技研究开发计划项目(2012X003-B)

摘　　要：根据CTCS-3级列控系统车载子系统的功能结构构建其故障树,将故障树转换为贝叶斯网络;考虑维修性、共因失效和多故障模式等因素,分别建立基于贝叶斯网络的二状态(考虑正常运行和系统故障2种故障模式)和三状态(考虑正常运行、系统故常和降级运行3种故障模式)车载子系统可靠性评估模型;计算二状态和三状态情况下车载子系统的可用度,对车载子系统的可靠性进行评估。结果表明:利用贝叶斯网络的反向推理和正向推理便于查找车载子系统的薄弱环节;若不考虑车载子系统冗余结构中的共因失效,则得到的可靠性指标会偏于乐观;二状态和三状态情况下车载子系统的可用度分别为99.999 2%和99.997 7%,可见,减少车载子系统运行时的故障降级切换,可提高其运行时的可靠性。The fault tree of CTCS-3 on-board subsystem was constructed on the basis of its function and structure, and then it was transformed into a Bayesian network. Taking into account some factors, such as maintainability, common cause failure, multiple failure mode and so on, the two states （considering two kinds of failure modes of normal operation and system failure） and three states （considering three kinds of failure modes of normal operation, system failure and degraded running） dependability assessment models of on-board subsystem were established separately based on Bayesian network. The dependability of on- board subsystem was evaluated by calculating the availability of the subsystem in the cases of two states and three states. The results show that the backward and forward reasoning of Bayesian network can easily find the weaknesses in subsystem. The reliability index is somewhat optimistic without considering the common cause failure in the redundancy structure of on-board subsystem. The availability of on-board subsystem under two states and three states is 99. 999 2% and 99. 997 7% respectively, showing that reducing the failure degraded switching of on-board subsystem operation can improve its reliability.

关 键 词：列车控制系统 车载子系统 共因失效 多故障模式 可用度 可靠性评估 

分 类 号：U284.48[交通运输工程—交通信息工程及控制] X913.4[交通运输工程—道路与铁道工程]