基于小波分析和神经网络的异步电机早期故障诊断

Early Fault Diagnosis in Asynchronous Motors Based on Wavelet Analysis and Neural Network

针对异步电机早期定子故障诊断,根据电机定子故障的特点,采用小波变换极大模分析法检测故障信号突变点的位置;利用小波包各个频带能量的变化完成能量特征提取,采用BP神经网络故障识别算法识别电机的各种运行状态来诊断电机早期故障.仿真实验结果表明,小波分析和神经网络算法的结合能有效定位并检测异步电机的早期故障.

Aiming at early fault diagnosis to asynchronous motor,a detection fault method of fault signal mutating position is proposed by wavelet analysis,the energy features are extracted based on the change of band energy by means of the wavelet packet for training BP neural network,and the motor running states are discriminated to diagnose all kinds of early motor faults by fault identification algorithm of BP neural network.The simulation results show that the combining wavelet analysis with neural network algorithm can efficiently locate and detect early fault of the asynchronous motor.