基于神经网络的机车牵引系统故障诊断研究

Locomotive Traction Malfunction Diagnosing System with Neural Network

为实现故障的准确定位,针对"神华号"电力机车牵引系统,提出了一种基于径向基函数神经网络的故障诊断方法:分别采集机车牵引系统正常与故障时的数据,分类处理后作为训练样本,建立用于机车牵引系统故障诊断的径向基函数神经网络模型。故障定位的实验结果验证了该诊断方法的可行性,现场应用也达到了预期效果。

In order to locate faults accurately, a diagnosing method applying neural network was proposed for the malfunction diagnosis of Shenhua electric locomotive traction system. Locomotive traction system data was collected and preprocessed, including those under normal and failure mode, which was used as samples to build RBFNN for locomotive traction malfunction diagnosis. Results of the location of fault in lab verified the feasibility of the diagnosis method, and application in the locomotive maintenance also achieved the expected result.