ZPW-2000轨道电路故障诊断方法研究

Research on Fault Diagnosis Method of ZPW-2000 Track Circuit

针对ZPW-2000轨道电路故障的多样性、复杂性、诊断难等问题,提出基于PCA-PSO-PNN的ZPW-2000轨道电路智能故障诊断方法。首先,对影响因素进行主成分分析,提取了主要的影响因素,将输入维数降低,然后建立8种常见故障的概率神经网络诊断模型,其次采用PSO算法优化PNN模型参数,最后采用某电务段提供数据进行故障划分和诊断,得到较好的诊断效果,因此该法能够为维护人员提高诊断效率及正确率,很好的解决只依赖维护人员现场维护经验诊断轨道电路设备故障,效率低可靠性差诊断难的问题,提高了运行效率。

Aiming at the diversity, complexity and difficulty of fault diagnosis of ZPW-2000 track circuit, an intelligent fault diagnosis method of ZPW-2000 track circuit based on PCA-PSO-PNN is proposed. Firstly, the main influencing factors are extracted by kernel principal component analysis, and the input dimension is reduced. Then the probabilistic neural network diagnostic model of eight common faults is established. Secondly, the parameters of PNN model are optimized by PSO algorithm. Finally, the data provided by a certain power station are used to classify and diagnose the faults, and a better diagnostic effect is obtained. Therefore, this method can be put forward for maintenance personnel. It has high diagnostic efficiency and correct rate. It solves the problem that it is difficult to diagnose the fault of track circuit equipment only depending on the maintenance experience of maintenance personnel on site. It has low efficiency and poor reliability, and improves the operation efficiency.