基于RBF神经网络和小波包的电动机故障诊断研究

针对传统的电动机故障诊断存在很难准确提取故障时的特征信号及对故障作出准确预测的问题,提出了一种基于RBF神经网络和小波包的电动机故障诊断的方法。该方法采用小波包分析技术提取电动机典型轴承故障、转子故障和绝缘故障振动信号的特征频段能量并组成向量作为RBF神经网络的输入,用于诊断电动机的故障。实验和仿真结果表明,使用RBF神经网络对电动机故障诊断是非常有效的,对电动机早期故障的发现及维修有积极意义。

Research of Fault Diagnosis for Motor Based on RBF Neural Network and Wavelet Packet

In order to solve the problem that it is difficult to extract fault characteristic signals and make an accurate fault prediction exsisted in traditional fault diagnosis for motor,a diagnosis method of motor faults based on RBF neural network and wavelet packet was put forward.The method can extract energy of special frequency bands of vibration signals of typical faults of bearing,rotor and insulation of motor with wavelet package analysis technology,and serve the energy as a group of vector to be input of RBF network for diagnosing motor fault.The experiment and simulation results showed that the method is very effective to diagnose motor,which has positive significance to find early fault and maintain for motor.