长江干线水上交通安全多因素耦合预警模型研究

发布时间：2018-07-24 11:33

【摘要】：加快长江航运发展己上升为国家战略,而发展长江航运必须以安全为前提和保障,水上安全预警管理工作必将加强。现阶段对长江干线水上交通安全预警管理研究较多,但其中的多因素耦合的安全预警模型研究则相对较弱,适应性较差。因此,多因素耦合预警模型则成为进一步研究的方向,并结合对通航环境、航行船只、动态交通流安全状态的实时数据采集,实现对长江通航的实时动态监管,将安全监管从事后向事前、从被动向主动、从经验向科学的转变,对于促进内河航运安全畅通,提高整体运输效率,具有重要意义。本文是笔者参与交通运输部交通运输建设重大科技专项“长江黄金水道重点航段安全保障关键技术研究”(201132820190)部分研究工作的成果。将现代应用数学方法、信息论、概率统计与安全系统工程等理论和方法,引入水上交通安全预警管理研究中,从大系统角度出发,重点开展基于多因素耦合的长江干线水上交通安全的预警技术研究,主要研究内容包括三部分。 1)根据统计的历年长江干线水上交通安全事故,分类总结事故特点,探究长江干线水上交通安全风险因素,为预警指标体系的构建提供依据。结合长江干线安全监管工作实际,分析安全监管与预警管理的现状与存在的问题。 2)分析水上交通安全风险因素,设计长江干线水上交通安全预警指标调查问卷并开展调查,在此基础上,构建4大类11亚类共33个指标、具有实时动态性的长江干线水上交通安全预警指标体系,并确定各指标的阈值。根据不同航段特征,从事故诱因与事故发生次数间的关系出发,采用ABC分析法,分别构建长江干线浅险和繁忙航段水上交通安全预警指标体系。 3)研究基于多因素耦合的预警模型。构建了基于贝叶斯网络的多因素耦合预警模型,并进行了算例验证。构建了基于结构方程加权、BP人工神经网络的长江干线水上交通安全多因素耦合预警模型,前者以显态的指标权重进行水上交通危险程度的预算,后者则以非显态指标权重进行水上交通危险程度的预测。针对长江干线浅险航段特征,在100个事故统计基础上,进行了实证研究,两模型预警结果一致,可相互验证。研究显示,两种模型方法相辅相成,可提高水上交通安全预警的准确性和可靠性。本文构建的具有实时动态性的长江干线水上交通安全预警指标体系,基于结构方程加权多因素耦合和基于BP人工神经网络多因素耦合的长江干线水上交通安全预警模型,及实证研究,具有创新性和较强的实用价值。
[Abstract]:Speeding up the development of the Yangtze River shipping has risen to a national strategy, and the development of the Yangtze River shipping must be based on security as the premise and guarantee, and the early warning management of water safety must be strengthened. At the present stage, there are many researches on the early warning management of the traffic safety on the Yangtze River main line, but the research on the multi factor coupled safety early warning model is relatively weak and the adaptability is poor. Therefore, the multi factor coupling early warning model is the direction of further research, and combined with the real-time data collection of navigation environment, navigation ships and dynamic traffic flow safety state, the real-time dynamic supervision of the navigation of the Yangtze River is realized, and the safety supervision is engaged in the forward, from the initiative, from the experience to the science, to the promotion of the inland river. Shipping is safe and smooth, and improving overall transport efficiency is of great significance.
This article is the result of the author's part of the study on the key technology of the key section of the Yangtze golden waterway (201132820190), which is a major scientific and technological special special project of the transportation and transportation construction of the Ministry of transportation. In the study of police management, from the perspective of large system, the research on early warning technology of water traffic safety on the Yangtze River trunk line based on multi factor coupling is carried out. The main research contents include three parts.
1) according to the statistics of the traffic safety accidents on the Yangtze River trunk line in the past years, the characteristics of the accident are classified and summarized, and the risk factors of the water traffic safety on the Yangtze River trunk line are explored to provide the basis for the construction of the early warning index system. The current situation and problems of the safety supervision and pre police management are analyzed in combination with the actual safety supervision work of the Yangtze River trunk line.
2) analyze the risk factors of the water traffic safety, design the survey questionnaire of the early warning index of the Yangtze River trunk line water traffic safety, and carry out the investigation. On this basis, we build 4 categories and 11 subcategories with 33 indexes, and have the real-time dynamic early warning index system of the Yangtze River trunk line water traffic safety, and determine the threshold of each index. Based on the relationship between the number of incentives and the number of accidents, the ABC analysis method was used to build an early warning index system for water traffic safety in shallow and dangerous sections of the Yangtze River.
3) an early warning model based on multi factor coupling is studied. A multi factor coupling early warning model based on Bayesian network is constructed, and an example verification is carried out. A multi factor coupling early warning model of water traffic safety on the Yangtze River trunk line based on structural equation weighting and BP artificial neural network is constructed. The former uses the explicit index weight to carry on the water traffic danger. On the basis of 100 accident statistics, the two model early warning results are consistent and can be verified each other. The study shows that the two model methods complement each other and can improve the safety of water traffic safety. The accuracy and reliability of early warning.
The early warning index system of the Yangtze River trunk traffic safety with real-time dynamic characteristics is based on the weighted multi factor coupling of structural equation and the water traffic safety early warning model of the Yangtze River trunk line based on the multi factor coupling of BP artificial neural network, and the empirical study, which is of innovation and strong practical value.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：U698

【参考文献】