贝叶斯网络在轨道交通屏蔽门系统安全分析中的应用  

The Application of Bayesian Network in Safety Assessment of the Platform Screen Door System of Rail Transit System

作　　者：商晖[1] 燕飞[1] 李开成[1] 

机构地区：[1]北京交通大学轨道交通控制与安全国家重点实验室,北京硕士研究生100044

出　　处：《城市轨道交通研究》2011年第11期47-52,共6页Urban Mass Transit

基　　金：国家科技支撑计划项目(2009BAG14B01)

摘　　要：安全是城市轨道交通的重要因素。随着轨道交通系统复杂度的不断提高、系统动态性和可维修性的不断增加,典型的故障树法和事件树法已经不能满足系统的故障预测与故障定位分析。探讨了使用贝叶斯网络进行轨道交通屏蔽门的安全评价。这是首次将贝叶斯网络引入轨道交通系统的安全分析中。建立了轨道交通屏蔽门系统模型,构造了轨道交通屏蔽门的贝叶斯网络,分析造成屏蔽门夹人这一危险源的原因和后果;同时比较了贝叶斯网络安全分析法与经典分析法的结果。结果表明,静态的贝叶斯网络在分析上可以作为经典故障树法和事件树法的综合分析法,其双向推理可以在相应的证据输入的支持下获得系统的故障预测概率,并判断出故障发生后系统部件对故障的贡献概率,从而为系统的设计和维护提供建议。This Paper explores using Bayesian Networks for safety evaluation of Platform Screen Door System of Rail Transit System. This is the first time to introduce Bayesian Network into safety analysis of rail transit system. Firstly, the system model of PSD System is established and the Bayesian Network of the PSD System of Rail Transit System is constructed. The paper analyzes the reason and consequence of this hazard source of ＂platform screen door＇s clipping passengers＂ and the author, meanwhile, compares the result of Bayesian Network security analysis method with that of the classic analysis method. The results show that static Bayesian Network can be used as the synthetic analysis method of classic fault tree method and event tree method in the analysis. Supported by the appropriate evidence input, its bi-direction inference can gain access to the system failure prediction probability and determine the contribution probability of system component to the fault after failure, thus providing recommendations for system design and maintenance.

关 键 词：贝叶斯网络 城市轨道交通 屏蔽门 安全分析 

分 类 号：U239.5[交通运输工程—道路与铁道工程]