
基于模糊评判和RCM的CTCS-3级车地通信子系统维修决策.doc
13页基于模糊评判和RCM的CTCS-3级车地通 信子系统维修决策张友鹏张芸芸石磊杨金凤兰州交通大学自动化与电气工程学院摘要:CTCS-3级车地通信子系统是列控系统的重要组成部分么一,且列控系统需求的 列车位置、行车许可、临时限速等安全信息都由其提供,因此,对其进行可靠性 分析具有十分重要的意义采用贝叶斯网络对系统进行可靠性分析,并对维修方 式进行了讨论首先,根据车地通信子系统的功能与结构构建其贝叶斯网络模 型然后,综合考虑维修性、共因失效等因素,对车地通信子系统进行可靠性分 析最后,在可靠性分析的基础上运用模糊综合评判法对其进行维修决策结果 表明:利用贝叶斯网络的双向推理,不仅可以计算出车地通信子系统的可靠度, 还可以有效识别系统的薄弱环节;若不考虑车地通信系统兀余结构中的共因失效, 则得到的可靠性指标会偏于乐观地面GSM-R单元失效是引起车地通信子系统失 效的关键事件,因此针对此薄弱环节进行状态维修检查能够有效降低事故的发 生概率关键词:安全工程;可靠性分析;CTCS-3级车地通信子系统;贝叶斯网络;共因失效; 模糊综合评判; 作者简介:张友鹏,教授,从事系统可靠性研究,ypz2548@163. com0收稿日期:2016-04-20基金:中国铁路总公司科技研究开发计划课题(2015X007-H)Maintenance decision for train-ground communication subsystem of CTCS-3 via the fuzzy comprehensive evaluation and RCMZHANG You-peng ZHANGLYun-vun SHILei YANGJin-fengSchool of Automation and Electrical Engineering,Lanzhou Jiaotong University;Abstract:The paper intends to introduce a dccision-pcrscvcrancc method based on the fuzzy comprehensive evaluation and RCM means for determining the maintenance mode of the railway carriage communication subsystem of CTCS-3 ( short for the Chinese Train Control System-3) . The subject has bee n chosen just for the urgent need that the curre nt main tenance met hods in this way fail to take into account the railway carriage control system as an integrated comprehensive system and the factors involved, such as their maintenance and general failure, whereas the CTCS-3 carriage-ground communication subsystem is by nature one of the key control systems to provide such items of information, as those of the exact location, the driving situeition, the current speed limit and other items of the safcty control. It is just for the above reasons that we have first of all established a faulty tree model for the said communication subsystem of CTCS-3 in terms of the structures and functions. And, secondly, an assessment model based on the Bayesian network has been set up by using a software known as MSBNX to emeilyzc the rcliability of the carriage-ground communication subsystem on account of such factors, as the maintenance errors and dynamic failure, in case of the failure of the basic units and the reliability assessment model. And, finally, the fuzzy comprehensive evaluation has to be made to det ermine the maintenance modes based on the rcliability emeilysis results. The key points of the fuzzy comprehensive evaluation can be summed up to determine the weighting factors and the weighing matrix. The results of our study indicate that, the bidirectional reason ing of Bayesian net work can not only help towork out the reliabi1ity of the carriage-ground communication subsystem of CTCS-3 but cam also identify effeetivcly the weak links of the subsystcm. For example, the deviation of the analysis results can be too serious regardless of the influence of common cause failures on the reliability of the system, whichare likely tolead to the reduction of the failure of the train-ground communication subsystem by giving more regular inspcctions of the working conditions of the weak links of the carriage-ground communication subsystem based on its actual operation. Thus, it can be seen that the study results of the paper can be taken as a guide to enhance the routine maintenance of CTCS-3, as well as a referenee to the scientific method for successful management of the maintenance routine.Keyword:safety cnginccring; reliability amalysis; train-ground communication subsystem of CTCS-3; Bayesian network; common cause failure; fuzzy comprehensive evaluation;Received: 2016-04-200引言CTCS-3 (Chinese Train Control System-3)级列控系统是时速 300 km/h 及以 上客运专线的运行控制装备,应用了大量复杂的电子相关组件和计算机系统, 其规模庞大、关键设备冗余配置,维修性、共因失效等现象普遍存在。
CTCS-3 级列控系统通过GSM-R (GSM for Railway)无线网络实现列车与地面间不间断、 双向的信息传输,若车地间的信息无法正常交换,就会影响到整个列控系统的 可靠性和安全性山目前,国内外学者对CTCS-3级车地通信的研究主要针对 GSM-R网络的业务流程Zimmermann[2]根据欧洲列车控制系统(ETCS)技术规 范,利用确定随机Petri网建立GSM-R的信道传输模型,对GSM-R的可用性进行 分析;徐田华等宜提岀GSM-R分层赋吋有色Petri网模型,研究随机信道恶化、 越区切换、信道连接中断等无线信道失效模型对信息传输延迟的影响高婷婷虫 利用Markov模型方法对CTCS-3级列控系统通信可靠性进行分析,评价了可靠性 影响因素但以往没有将列控系统作为一个整体进行通信可靠性分析和维修决策, 且未考虑维修性、共因失效等因素对车地通信子系统可靠性的影响因此,分析 车地通信子系统的可靠性,找岀系统的薄弱环节,并在可靠性分析的基础上对 其进行维修决策,为进一步改善系统设计、确保列车安全运行以及降低维护成本 具有重要意义贝叶斯网络(Bayesian Network, BN)理论是近年来一种较为常用的可靠性评估 方法,弥补了传统可靠性分析方法gl在可靠性分析时存在的不足。
BN是一种 基于贝叶斯概率公式的可靠性分析方法,用于表达和分析多态性和不确定性的 问题,且可以有效处理共因失效问题既能用于推理,又能用于诊断,非常适用 于可靠性分析宜本文在综合考虑维修性、共因失效等因素对系统可靠性影响 的基础上构建车地通信子系统的贝叶斯网络模型,并对其进行可靠性分析,通 过计算系统故障条件下各单元的不可用度寻求系统的薄弱环节,并运用模糊综 合评判法确定薄弱环节的维修方式,将定性问题转化为定量问题1贝叶斯网络基本概念及建模1.1贝叶斯网络基本概念 贝叶斯网络由两部分组成,一部分是代表随机变量的节点;另一部分是连接这些 节点的有向边,有向边表示了随机变量间的相互关联回设V二{Xi, X2,…,XJ 是值域上的H1个随机变量的集合,并H P (X.) >0, B为V的任意事件,那么,当 P (B) >0时,贝叶斯公式为 在贝叶斯网络中,对于条件独立的n个随机变量Xi, X2, •: Xn,对于任意变量 B有 根据贝叶斯网络的网络结构和先验概率分布,可通过贝叶斯公式(1)和全概率 公式(2)计算任意一个事件发生的概率贝叶斯网络具有正向推理和反向推理 的特性,即正向推理是指在己知各个节点条件概率的基础上,计算出任意节点 发生的概率的推理。
反向推理是指在假设某个节点故障的条件下,分析出造成该 节点故障的原因及原因发生的概率的推理1.2基于故障树的BN建模咽的建模包括网络拓扑结构的确定以及网络中所有节点条件概率分布的确定两 个主要过程目前,贝叶斯网络一般通过三种方式进行建模,即第一种方式是根 据以往的专家知识手动建立贝叶斯网络模型;第二种方式是通过对数据库的学习 而自动获取贝叶斯网络模型;第三种方式是在综合了前两种方式优势的基础上的 两阶段建模方法,首先根据专家知识手动。












