基于神经网络的轨道电路补偿电容故障诊断研究

发布时间：2018-06-07 23:33

本文选题：轨道电路 + 补偿电容　；参考：《西南交通大学》2017年硕士论文

【摘要】：轨道电路是列车控制系统的重要基础设备,其工作状态直接关系到列车的运行效率与行车安全,保障轨道电路的健康运行是我国铁路事业蓬勃发展的关键环节。然而现阶段轨道电路故障诊断技术已无法满足我国高速铁路飞速发展的需求,如何及时准确地判定轨道电路故障是当前急需解决的难题,因此面向轨道电路故障诊断技术开展相关研究具有重要的现实意义与应用价值。本论文以ZPW-2000A轨道电路作为研究对象,对补偿电容故障定位和诊断方法进行研究,论文主要工作总结如下:首先根据ZPW-2000A轨道电路的基本结构,基于传输线理论建立轨道电路调整状态和分路状态短路电流模型,并将模型仿真数据与轨道电路模拟盘实验平台测量数据进行对比,以验证模型的准确性,同时对误差产生原因进行分析和讨论。其次根据短路电流曲线的形状与特征,基于改进夹角链码算法描述正常状态和补偿电容不同故障模式的短路电流曲线,通过计算正常和故障曲线的差异度,实现故障定位和故障类型判定。接着在不同的道砟电阻下,研究单个和两个补偿电容故障对轨道电路调整和分路状态数据的影响,通过相关性分析确定并提取了调整状态和分路状态的特征参数。最后建立了轨道电路调整和分路状态补偿电容故障诊断模型,分别利用概率神经网络、粒子群参数优化的支持向量机以及BP神经网络算法对轨道电路补偿电容故障进行诊断,从预测准确率和预测用时对三种模型的诊断效果进行对比,得出结论:概率神经网络具有预测准确率高和预测用时短的优点,具有一定的实际应用价值与意义。
[Abstract]:Track circuit is an important basic equipment of train control system. Its working state is directly related to the running efficiency and safety of the train. Ensuring the healthy operation of the track circuit is the key link of the vigorous development of the railway industry in China. However, at the present stage, the fault diagnosis technology of track circuit can not meet the requirement of the rapid development of high-speed railway in our country. How to judge the fault of track circuit in time and accurately is a difficult problem that needs to be solved urgently. Therefore, the research on fault diagnosis technology of track circuit has important practical significance and application value. In this paper, ZPW-2000A track circuit is taken as the research object, and the fault location and diagnosis method of compensating capacitor is studied. The main work is summarized as follows: firstly, according to the basic structure of ZPW-2000A track circuit, Based on the transmission line theory, the short-circuit current model of track circuit adjustment state and shunt state is established, and the model simulation data are compared with the measured data of track circuit simulator platform to verify the accuracy of the model. At the same time, the causes of errors are analyzed and discussed. Secondly, according to the shape and characteristics of the short-circuit current curve, based on the improved angle chain code algorithm, the short-circuit current curve of the normal state and different fault modes of compensating capacitance is described, and the difference between the normal and the fault curve is calculated. Fault location and fault type determination are realized. Then the influence of single and two compensating capacitor faults on track circuit adjustment and shunt state data is studied under different ballasting-resistance conditions. The characteristic parameters of adjusting state and shunting state are determined and extracted by correlation analysis. Finally, the fault diagnosis model of track circuit adjustment and shunt state compensation capacitor is established. The fault diagnosis of track circuit compensation capacitor is made by using probabilistic neural network, particle swarm optimization support vector machine and BP neural network algorithm, respectively. By comparing the diagnostic results of the three models with the prediction accuracy and the prediction time, it is concluded that the probabilistic neural network has the advantages of high prediction accuracy and short prediction time, and has certain practical application value and significance.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP183;U284.92

【相似文献】