基于统计回归技术的铁路货运脱轨事故风险建模
发布时间：2018-10-16 10:13
【摘要】：对铁路货物运输生产中事故风险的有效控制一直是铁路运输企业及政府相关部门所追求的目标。不断提升铁路货物运输的安全性具有重要意义。列车脱轨事故由于其事故率极高、事故后果严重,一直是世界范围内铁路货物运输中的一个重大安全隐患。本文以数理统计方法为基础,以回归建模技术为核心,综合运用概率统计、回归分析、方差分析、相关分析、组合优化、遗传算法等方法和工具,分别建立起了针对路网、线路区段以及站内编组过程的货物列车脱轨风险分析模型。 在路网层面,以美国铁路网2001-2010十年期间的货物列车脱轨事故为研究对象,通过对路网上的线路区段分级、对事故原因分类、用统计方法估计事故率以及建立基于能量转化原理的脱轨严重度非线性回归模型,完成了对不同等级线路区段上主要风险因子的识别与定量分析。应用实例中对不同既有线升级改造方案的风险评估表明：既有线升级改造对各主要风险因素所产生的影响差异很大——将降低绝大多数与线路有关的风险因子的风险值,但同时将使某些与轨上设备有关的风险因子的风险值升高。在作方案选择时,应综合考虑对象路网的主要风险因子及其具体运营指标。 在线路区段与站内作业层面,以货物列车在途脱轨风险受具体车列编组方案、具体径路特性及列车自身特性三方面因素影响为基本假设,通过建立脱轨率的解析模型、车列脱轨点POD的概率模型以及脱轨严重度的截尾几何回归模型,形成了一套针对具体径路及具体车列编组方案的货物列车在途风险建模策略。应用举例中通过综合考虑一条径路上涉及危险品运输的一列货物列车的在途风险与站内作业过程,将问题抽象为组合优化问题并用遗传算法对模型进行了求解分析。分析的结果验证了模型的可行性与实用性,并为铁路危险品运输在具体径路上的风险控制提供了新的思路。 与之前的相关研究相比,本研究的进步性体现在以下三方面：(1)将面、线、点结合的“三位一体”的铁路货物运输风险建模策略进行了详细而深入的研究；(2)与目前流行的铁路货运安全建模方法相比,本文的模型方法客观性较强、稳定性较好；(3)与目前相关研究中偏于理论研究的现状相比,本文将理论与应用较好地结合了起来。
[Abstract]:The effective control of accident risk in railway freight transportation is always the goal pursued by railway transport enterprises and government departments. It is of great significance to improve the safety of railway freight transportation. Because of its high accident rate and serious consequences, train derailment accident has been a major safety hazard in railway freight transportation all over the world. On the basis of mathematical statistics and regression modeling technology, this paper uses the methods and tools of probability statistics, regression analysis, variance analysis, correlation analysis, combinatorial optimization, genetic algorithm and so on. The risk analysis model of freight train derailment in line section and station marshalling process. At the road network level, taking the freight train derailment accident of the United States railway network 2001-2010 decade as the research object, through the route section classification on the road network, the cause of the accident is classified. The statistical method is used to estimate the accident rate and the nonlinear regression model of derailment severity based on the principle of energy conversion is established. The identification and quantitative analysis of the main risk factors on different grade lines are completed. The risk assessment of different existing railway upgrading schemes in application examples shows that the impact of existing railway upgrading on each major risk factor is very different-it will reduce the risk value of most of the risk factors related to the line. But at the same time, it will increase the risk value of some risk factors related to rail equipment. The main risk factors of the target road network and its specific operation index should be considered comprehensively when making scheme selection. Based on the assumption that the risk of derailment of freight train is affected by three factors such as specific train formation scheme, specific path characteristics and train's own characteristics, the analytical model of derailment rate is established. The probability model of train derailment point POD and the truncated geometric regression model of derailment severity have formed a set of risk modeling strategy of freight train in transit for specific path and train formation scheme. By taking into account the in-road risk and in-station operation process of a freight train involved in dangerous goods transportation on a path, the problem is abstracted as a combinatorial optimization problem and the genetic algorithm is used to solve the model. The results of the analysis verify the feasibility and practicability of the model and provide a new idea for the risk control of railway dangerous goods transportation on specific roads. Compared with the previous research, the progress of this study is reflected in the following three aspects: (1) the "Trinity" railway freight transport risk modeling strategy is studied in detail and deeply; (2) compared with the popular railway freight safety modeling method, the model method in this paper is more objective and stable; (3) compared with the current research situation, this paper combines the theory with the application.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：U298
本文编号：2274031
[Abstract]:The effective control of accident risk in railway freight transportation is always the goal pursued by railway transport enterprises and government departments. It is of great significance to improve the safety of railway freight transportation. Because of its high accident rate and serious consequences, train derailment accident has been a major safety hazard in railway freight transportation all over the world. On the basis of mathematical statistics and regression modeling technology, this paper uses the methods and tools of probability statistics, regression analysis, variance analysis, correlation analysis, combinatorial optimization, genetic algorithm and so on. The risk analysis model of freight train derailment in line section and station marshalling process. At the road network level, taking the freight train derailment accident of the United States railway network 2001-2010 decade as the research object, through the route section classification on the road network, the cause of the accident is classified. The statistical method is used to estimate the accident rate and the nonlinear regression model of derailment severity based on the principle of energy conversion is established. The identification and quantitative analysis of the main risk factors on different grade lines are completed. The risk assessment of different existing railway upgrading schemes in application examples shows that the impact of existing railway upgrading on each major risk factor is very different-it will reduce the risk value of most of the risk factors related to the line. But at the same time, it will increase the risk value of some risk factors related to rail equipment. The main risk factors of the target road network and its specific operation index should be considered comprehensively when making scheme selection. Based on the assumption that the risk of derailment of freight train is affected by three factors such as specific train formation scheme, specific path characteristics and train's own characteristics, the analytical model of derailment rate is established. The probability model of train derailment point POD and the truncated geometric regression model of derailment severity have formed a set of risk modeling strategy of freight train in transit for specific path and train formation scheme. By taking into account the in-road risk and in-station operation process of a freight train involved in dangerous goods transportation on a path, the problem is abstracted as a combinatorial optimization problem and the genetic algorithm is used to solve the model. The results of the analysis verify the feasibility and practicability of the model and provide a new idea for the risk control of railway dangerous goods transportation on specific roads. Compared with the previous research, the progress of this study is reflected in the following three aspects: (1) the "Trinity" railway freight transport risk modeling strategy is studied in detail and deeply; (2) compared with the popular railway freight safety modeling method, the model method in this paper is more objective and stable; (3) compared with the current research situation, this paper combines the theory with the application.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：U298
【参考文献】
相关期刊论文 前2条
1 王卓;贾利民;秦勇;杨凯淳;;铁路行车事故预测方法分析与比较[J];中国安全科学学报;2009年08期
2 王卓;刁朋娣;贾利民;秦勇;;铁路事故致因与风险分析[J];中国安全科学学报;2012年06期
相关博士学位论文 前2条
1 向俊;列车脱轨机理与脱轨分析理论研究[D];中南大学;2006年
2 海涛;铁路危险货物罐车运输安全理论与关键技术研究[D];北京交通大学;2010年
相关硕士学位论文 前2条
1 姜姗姗;铁路危险货物运输安全综合分析技术研究[D];北京交通大学;2009年
2 刘广武;铁路危险货物办理站运输安全综合评价研究[D];西南交通大学;2009年
，本文编号：2274031
本文链接：https://www.wllwen.com/kejilunwen/anquangongcheng/2274031.html
