基于改进动态贝叶斯的CTCS-3级列控车载子系统可靠性分析  

Reliability analysis of on-board subsystems in the CTCS-3 level train control system using an enhanced dynamic Bayesian approach

作　　者：赵斌[1] 王敏 张文韬[1] ZHAO Bin;WANG Min;ZHANG Wentao(School of Automation and Electrical Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学自动化与电气工程学院,兰州730070

出　　处：《安全与环境学报》2025年第2期413-421,共9页Journal of Safety and Environment

基　　金：甘肃省自然科学基金项目(23JRRA850);甘肃省科技专员专项(24CXGA020)。

摘　　要：针对传统中国列控车载子系统可靠性分析中存在的认知不确定性和动态性不足问题,为有效识别认知不确定性对可靠性分析的影响,将证据理论和动态贝叶斯网络相融合,分析列控车载子系统的可靠性。首先,根据中国列控车载子系统的结构与功能建立系统失效故障树模型;其次,将故障树模型转化为对应的Dempster-Shafer证据理论-动态贝叶斯网络(Dempster-Shafer evidence theory-Dynamic Bayesian Network,DS-DBN)模型;最后,综合考虑维修、共因失效和覆盖因子等因素,引入认知重要度,通过DS-DBN网络模型双向推理能力,采用区间值的形式来描述中国列控车载子系统可靠性。仿真结果表明:在列控车载子系统可靠性分析中忽略共因失效会导致分析结果偏大,采取维修措施能有效减缓可靠性的降低速度;通过反向推理识别出列车接口单元是中国列控车载子系统的薄弱环节;人机界面的认知重要度和不确定度最高,表明该环节对中国列控车载子系统可靠性影响最大。To address cognitive uncertainty and the lack of dynamic considerations in the reliability analysis of traditional CTCS-3 train control on-board subsystems,we integrate evidence theory with Dynamic Bayesian Network(DBN).This approach enables a more effective identification and analysis of how these factors influence the system's reliability.A Dynamic Bayesian Network model incorporating Dempster-Shafer evidence theory(DS-DBN)for the CTCS-3 train control on-board subsystem has been established.This model is simulated and analyzed to provide a comprehensive evaluation of the subsystem's reliability.First,a dynamic fault tree model for system failure is developed based on the structure and function of the CTCS-3 level column control vehicle-mounted subsystem.Next,this dynamic fault tree model is converted into the corresponding DS DBN network model.Finally,cognitive importance is incorporated by considering factors such as maintenance,co-causal failures,and coverage.The reliability of the CTCS-3 level column control vehicle-mounted subsystem is evaluated through the DS-DBN network model using bidirectional reasoning characteristics.Reliability is described in terms of interval values to provide a comprehensive assessment of the subsystem's performance.The results indicate that neglecting common cause failures in the reliability analysis of the CTCS-3 train control on-board subsystem can result in significantly inflated assessments.Additionally,implementing regular maintenance measures can effectively slow the rate of reliability degradation in this subsystem.Reverse reasoning reveals that the Train Interface Unit(TIU)is a weak link within the CTCS-3 train control on-board subsystem.Furthermore,the analysis of cognitive importance indicates that the Driver-Machine Interface(DMI)is the most critical component of the train control on-board subsystem.The results demonstrate that the DMI exhibits the highest cognitive importance and uncertainty,suggesting that uncertainties associated with the DMI significantly impact th

关 键 词：安全工程 列控车载子系统 证据理论 重要度 认知不确定性 

分 类 号：X951[环境科学与工程—安全科学]