摘要
分析定子电流频谱特性诊断感应电动机转子故障,为了避免故障特征的边频分量被基频"旁瓣"淹没,利用扩展Park矢量法提取故障特征。在此基础之上,对所得到的故障特征矢量进行归一化,作为BP神经网络的输入量,同时以故障类型作为BP神经网络的输出量。设计BP神经网络结构,利用输入输出量对BP神经网络进行训练,进而实现感应电动机转子故障的自动识别。
In order to avoid the inundation of the edge frequency components of the fault characteristics by the lobes of basic frequency, while the stator current spectrum characteristics are analyzed for rotor fault diagnosis of the induction motor, the extended Park vector method is adopted to extract the fault characteristics. On this basis, the structure of BP neural network is designed, taking the normalized fault characteristic vectors as its input and the fault style as its output. The network is trained by these input and output to realize the auto-identification of the induction motor＇ s rotor fault.
出处
《煤矿机电》
2014年第3期89-91,共3页
Colliery Mechanical & Electrical Technology
关键词
电流频谱分析
PARK矢量
BP神经网络
故障诊断
current spectrum analysis
Park vector
BP （Back Propagation） neural network
fault diagnosis
