【作者】 杨柳；

【导师】 高仕斌；

【作者基本信息】 西南交通大学 ， 电气工程（专业学位）， 2018， 硕士

【摘要】 近年来大规模的高速铁路建设使高速铁路进入发展的黄金时代,其安全性和可靠性是最不容忽视的两方面。接触网是高速铁路牵引供电系统的重要组成部分,为了保证高速列车能平稳运行,必须保障接触网系统的良好性能。然而接触网的结构复杂且故障原因及种类很多,加之接触网的工作环境在室外,不配备备用装置,外界环境对其性能影响较大,因此接触网是牵引供电系统中的薄弱环节。如果可以正确识别接触网系统的状态,并对其维修时间进行预测估计,不但能减少平时的系统维护费用,还能大大提高高速铁路接触网的可靠性,从而减少事故发生的概率。本文以清远至广州北区间下行试验高铁线路的接触网作为研究对象,利用6C系统采集该区间的静动态参数,对此线路的接触网系统的状态识别和维修时间的估计展开了深入的研究:首先研究了接触网的特征数据和性能退化的特点,通过对现场实际数据的跟踪与调研,得到接触网系统的静动态参数,并提取了表征接触网性能退化的特征值;由现场收集的全线路故障报告、工作日志构建的接触网系统故障数据库,建立了接触网故障演化曲线,确定采用四种状态表征接触网的退化过程。其次根据接触网特点,结合时变隐半马尔可夫模型(CHSMM)的相关理论,建立了基于时变CHSMM的接触网系统级退化模型与维修时间估计的理论框架;考虑到接触网系统的实际情况,对时变CHSMM模型中的参数做出了相应的修改。然后建立了接触网维修时间的估计方案,针对接触网性能退化过程中不会保持相同的退化速度的特点,引入了状态转移系数来描述不同退化阶段的退化速度,利用该系数修改描述系统内部状态变化的概率矩阵,从而使模型更符合实际退化过程。最后将表征接触网退化过程的特征数据导入模型中,实现了对接触网系统健康状态的识别和维修时间的估计,对比模型计算出的结果与实际检修时间,验证该模型具有更高精度的预测能力。更多还原

【Abstract】 In recent years,massive construction across the country has lead the high-speed train to its golden age.In the meantime,the safety and the reliability are the last two aspects that we should ignore.As the catenary system is one of the most important parts of the power traction system,the catenary system should always be in its level of excellent performance in order to make sure the steadiness of the high-speed train.However,the complexity of the catenary system makes the causes of malfunction various.Additionally,the catenary system is placed in open space without any back-up devices,which result in its vulnerability to the environment.Thus,the catenary system is the weakest component of the power traction system.If the healthy conditions of the catenary system can be accessed,and predict its residual life,not only we are able to improve the reliability of high speed train,but also reduce the chance of accidents.This thesis is based on the catenary system of Qingyuan-Guangzhou high speed railway,of which the data are used to extract characteristic parameters of the catenary system.Deeper researches about the performance degradation assessment and estimation of repair time have conducted.Firstly,the characteristic data of the catenary system and performance degradation patterns are analyzed.After treating the raw data,the dynamic and statistic parameters are obtained,from which the characteristic parameters of the catenary system are extracted.According to the typical curve of system performance degradation built up with failure reports and work logs of catenary system,the number of healthy states is fixed.Secondly,the property of the catenary system has combined with the theory of the TimeVarying State Transition Probability of Continuous Hidden Semi-Markov Model(CHSMM)in order to build up the framework of the catenary system’s degradation model and estimation of repair time.Considering the special properties of the catenary system,the parameters of CHSMM have been modified accordingly.Then,the estimation of catenary system’s repair time is planned.To make the model fit to the real conditions,new parameters are introduced to describe different degradation paces in different stages because the rate of degradation of the catenary system is not always the same.Lastly,the characteristic parameters of the catenary system are regarded as input of the model,which makes the assessment and the estimation of repair time possible.And the model has finally been proven to output results of higher precision through comparing the result given by the model and the real one.更多还原