基于小波神经网络的牵引电机齿轮故障诊断研究

Research of Fault Diagnosis System of Traction Motor Gears Based on Wavelet Neural Network

以小波变换、小波神经网络为工具,采用定子电流对HXD1B型机车的YQ1633异步牵引电动机开展牵引电机齿轮故障诊断研究。定子电流法相对于振动法更容易实施,有效克服了振动信号中包含的复杂干扰。由小波分析完成齿轮故障的特征量提取,通过神经网络对故障类型进行判断,实际测试表明,该方法具有较好的故障诊断性能。

It researches fauh diagnosis of the traction motor gear of the YQ1633 asynchronous traction motor on HXD1B locomotive, taking wavelet transform and wavelet neural network as a tool,using stator current.The method of stator current is easier to implement relative to vibration and effectively overcomes the complex interference contained in the vibration signal. Feature extraction of gear failure is completed by wavelet analysis and then the type of fault is judged by the neural network, in the end,though actual tests which shows that the method has better performance of fault diagnosis.