基于神经网络信息融合的机车牵引电机故障诊断

Fault Diagnosis of Locomotive Traction Motor Based on Neural Network Information Fusion

对信息融合的故障诊断进行了研究,提出了一种基于信息时空融合的故障诊断模型,并将其成功应用于电力机车牵引电机的故障诊断;这种方法是在运用神经网络进行局部的故障诊断的基础上,再运用D-S证据理论进行全局决策的融合,从而实现了这两种算法优势的互补,提高诊断的准确率;实例分析结果表明,该故障诊断模型能够准确地检测出故障发生的位置及其故障发生的原因,适合于电力机车牵引电机的故障诊断,并具有推广的价值。

Fault diagnosis based on information fusion is studied, and a fault diagnosis model based on information temporal-spatial fusion is presented, and successfully applied to fault diagnosis of electric-power locomotive traction motor. It applies neural network to diagnose local faults, then applies D-S evidence theory to carry through the whole decision-making fusion, thereby realizes to eonigate the two algorithms＇ advantages and increases the fault diagnosis accuracy. Instance analysis result indicates that the fault diagnosis model is able to detect the location and the reason of fault generating exactly, and is fit for fault diagnosis of electric-power locomotive traction motor, and is worth to extend.