【作者】 林立新；

【导师】 蒋新华；

【作者基本信息】 中南大学 ， 交通信息工程及控制， 2008， 博士

【摘要】 电力机车作为载运工具,其运行状态直接影响到列车的安全性和舒适性。电气系统是电力机车的重要组成部分,电气系统正常工作才能保证机车的正常运行。当机车某一部件发生故障时,要求能迅速诊断故障部位或部件,能迅速对故障部件进行自动隔离处理或提示司乘人员进行处理,这样才能保证列车的安全正点运行,满足人民群众的基本需要。目前SS7E电力机车只具备极少的故障诊断功能,因此,开展SS7E电力机车电气系统的故障诊断技术研究,具有重要的理论意义和现实意义。论文的主要研究内容如下:(1)对国际国内电力机车及其故障诊断技术进行了综述,主要分析了日本、德国、法国以及美国和加拿大发展的高速列车及其故障诊断技术,阐述了我国电力机车在故障诊断技术方面的发展进程。阐明了电力机车故障诊断的意义。(2)对SS7E电力机车电气系统的主要部分:主电路、辅助电路和控制电路进行详细分析,建立了SS7E电力机车主电路各部件数学模型。(3)通过对SS7E电力机车的控制电路进行详细原理研究,分析了SS7E电力机车电气系统主要部件的故障机理,推导并得到SS7E电力机车电气系统的故障集。在此基础上构建了SS7E电力机车电气故障开关量检测点全集、模拟量检测点全集,以及故障诊断过程中用到的中间点全集,提出了SS7E机车故障检测系统的硬件框架。(4)研究了SS7E机车故障诊断系统的知识表示形式,推导并构建了SS7E电气故障诊断系统的规则库。提出了SS7E机车电气系统故障诊断树,构建了推理机制。通过改善匹配算法,实现了SS7E机车电气系统的专家推理及诊断。(5)详细论述了小波变换理论,研究了变流器故障检测方法,建立了变流器仿真模型,提出了一种基于小波变换的SS7E电力机车变流器输出电压波形能量特征向量提取方法。针对离散数据,提出了实用的计算方法,以实现小波分析在SS7E机车上的实际应用。(6)详细研究了神经网络故障诊断技术,提出了一种基于神经网络的SS7E机车变流器故障诊断方法,研究推导了SS7E电力机车变流器神经网络模型,给出了SS7E机车变流器故障诊断神经网络训练图。通过仿真试验,验证小波变换+神经网络技术用于变流器故障诊断的有效性。(7)从硬件和软件两个方面提出了SS7E电力机车电气故障诊断系统的具体实现方法。针对SS7E机车变流器的故障诊断,提出将神经网络的训练与应用分别实施的方法,将训练好的网络用于车载实时诊断,解决了神经网络故障诊断方法中计算量与实时性的矛盾。论文最后指出了SS7E电力机车电气故障诊断系统中存在的一些问题,探讨了我国电力机车故障诊断技术的发展方向。更多还原

【Abstract】 The locomotive being a means of conveyance, its running state affects directly the security and comfort of the train. The electrical system is the very important composing part of the electric locomotive. Only working well of the electrical system can assures the locomotive in good station. When one component of the locomotive goes wrong, it is required to make certain the error place or error component quickly, and to separate the error component automatically or to tell the drivers how to deal with. So it can assure the train running safely and punctually, and suffices the basic need of the masses. The SS7E locomotive has a few of diagnosis functions presently, so it has very important theory significance and practical significance for fault diagnosis technology and research of the electrical system of SS7E electric locomotives. The main researches are as follows:(1) The paper summarizes the international and domestic electric locomotive and their fault diagnosis technology. It analyses mostly the high-speed trains and their fault diagnosis technology in Japan, Germany, France, America and Canada. It also expounds the development of the fault diagnosis technology of the electric locomotive in China, and illustrates the significance of electric locomotive fault diagnosis.(2) The paper analyses detailedly the main part of electrical system: the main circuit, assistant circuit and control circuit. The mathematical model of each component of the SS7E electric locomotive main circuit is set up in this paper.(3) By a detailed principle study of the SS7E locomotive control circuit, it analyses the main components of the electrical system fault mechanism of the SS7E electric locomotive, and deduces the electrical system fault set of the SS7E electric locomotive. On this basis, it builds the full set of digital check points and the full set of analog check points of the SS7E electric locomotive electrical fault, as well as the full set of the middle points which is used in the process of fault diagnosis. It sets up the hardware system of the SS7E locomotive fault detection.(4) The paper studies the form of knowledge representation in the SS7E locomotive fault diagnosis system, and deduces and constructs the rule library of the SS7E locomotive fault diagnosis system. It puts forward the SS7E locomotive electrical system fault tree, and constructs the reasoning mechanism. By improving the matching algorithm, it achieves the expert reasoning and diagnosis of the SS7E locomotive electrical system.(5) The paper dissertates the wavelet transform theory in detail, and studies the converter fault detection method, and establishes the converter simulation model. It puts forward an energy eigenvector extraction method from the output voltage wave of the SS7E locomotive converter. For discrete data, the paper puts forward a practical calculation method to achieve the practical application of wavelet analysis theory in the SS7E locomotive.(6) The paper studies the neural network fault diagnosis technology in detail, and puts forward a kind of fault diagnosis method of the SS7E locomotive converter based on the neural network. It researches and deduces the SS7E electric locomotive converter neural network model, and gives the SS7E locomotive converter fault diagnosis neural network training figure. By the simulation experiment, the validity of the converter fault diagnose method, which use the wavelet transform and the neural network technology, is proved.(7) From both hardware and software aspect, the paper puts forward the practical application method of the SS7E electric locomotive electrical fault diagnosis system. For the fault diagnosis of the SS7E locomotive converter, the paper puts forward the method, which realizes individually the training and application of the neural network, in order to apply the trained network to real-time diagnosis on the locomotive. So it solves the conflict between the calculation quantity and the real-time ability of the neural network fault diagnosis.Finally, the paper points out some problems of the SS7E electric locomotive electrical fault diagnosis system, and explores the development of electric locomotive fault diagnosis technology in our country.更多还原