基于贝叶斯网络的铁路旅客运输安全评价研究

发布时间：2018-02-27 05:02

本文关键词： 铁路旅客运输致因分析风险分析 TOPSIS方法安全评价贝叶斯网络　出处：《北京交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着我国铁路大发展,给人们带来了更加舒适、便捷的出行,铁路旅客运输迎来一次又一次的高峰,2016年我国铁路旅客发送量达到28亿人次之多。尤其是高铁的发展在带来新机遇的同时,也伴随着安全运营等方面新的挑战,对铁路旅客运输的安全生产提出了更新更高的要求。因此,本文在铁路旅客运输安全生产现状背景及其生产“安全第一”的目标原则下,运用先进的安全科学理论与方法及贝叶斯网络法对铁路旅客运输安全生产进行了深入地分析与评价研究,具有重大的理论意义与现实意义。首先,本文在文献查阅学习的基础上,结合铁路旅客运输生产特点及其安全生产目标,运用风险理论和安全系统工程的方法对铁路旅客运输生产安全进行了深入分析,包括事故致因分析和客运安全风险分析。从人员、设备、环境和管理四个方面对铁路客运安全进行系统分析来识别风险,并运用基于改进的TOPSIS方法对风险因素进行了排序。其次,依据铁路客运安全风险分析与排序的结果,结合评价指标构建原则、研究文献及专家咨询,从人员安全、设备安全、环境安全和管理安全四个层次建立了铁路客运安全评价指标体系。再次,基于因果关系下,将铁路客运安全评价指标向贝叶斯网络节点转化构造了贝叶斯网络模型的拓扑结构并给出了各节点的变量值域,结合贝叶斯网络理论与专家决策,研究了基于贝叶斯网络集结专家意见推理更新确定根节点先验概率的算法,结合实际情况,研究改进了非根节点条件概率的确定方法,构建了铁路旅客运输安全评价的贝叶斯网络模型。最后,论文以X铁路局为实例进行分析,得出X铁路局客运安全等级为“比较安全”。通过诊断推理分析,根据根节点风险的发生概率(高于0.05)得出了主要风险因素,根据根节点的重要度得出了重要风险因素。通过与模糊综合评价结果的对比分析,验证了贝叶斯网络模型的合理性与优越性。
[Abstract]:With the great development of railway in our country, it has brought people more comfortable and convenient travel. In 2016, the volume of railway passenger transport in China reached as many as 2.8 billion people. In particular, the development of high-speed rail has brought new opportunities as well as new challenges in terms of safe operation and so on. Therefore, under the background of the present situation of railway passenger transport safety production and the principle of "safety first" in production, this paper puts forward a higher demand for the safety of railway passenger transportation. The advanced theory and method of safety science and Bayesian network method are used to analyze and evaluate the safety of railway passenger transportation, which is of great theoretical and practical significance. On the basis of literature review and study, combined with the characteristics of railway passenger transport production and its safety production objectives, this paper makes a deep analysis of railway passenger transport production safety by means of risk theory and safety system engineering method. Including accident cause analysis and passenger transport safety risk analysis. From the personnel, equipment, environment and management of four aspects of railway passenger safety system analysis to identify the risks, Secondly, according to the results of risk analysis and sequencing of railway passenger transport safety, combined with the evaluation index construction principle, research literature and expert consultation, from personnel safety, equipment safety, The evaluation index system of railway passenger transport safety is established on the four levels of environmental safety and management safety. Thirdly, based on causality, The topological structure of Bayesian network model is constructed by transforming the evaluation index of railway passenger transport safety to Bayesian network node, and the variable range of each node is given, combined with Bayesian network theory and expert decision. This paper studies the algorithm of determining the priori probability of root node based on Bayesian network, and improves the method of determining the conditional probability of non-root node. The Bayesian network model of railway passenger transport safety evaluation is constructed. Finally, taking X Railway Bureau as an example, the paper concludes that the grade of passenger transport safety of X Railway Bureau is "relatively safe". According to the occurrence probability of root node risk (above 0.05), the main risk factors are obtained, and the important risk factors are obtained according to the importance degree of root node. The rationality and superiority of Bayesian network model are verified.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U298

【相似文献】