共振解调与小波降噪在电机故障诊断中的应用

Application of resonance demodulation and wavelet denosing in fault diagnosis of induction motors

针对异步电机形成复合故障时电流频谱存在的故障频率成分难以准确分离的问题,结合小波降噪算法与共振解调技术,提出一种异步电机复合故障分离方法。依托小波优良的时频局部化特性,有效地区分信号中的突变部分和噪声,实现信号的降噪;利用软件方法实现共振解调,构造带通滤波器提取共振信息。利用Hilbert变换进行解调分析得到包含故障特征信息的低频包络信号,经过低通滤波、频谱分析后实现异步电机耦合故障分离和故障特征提取。实验结果表明,该方法使复合故障情况下的异步电机电流信号的故障特征频率更容易识别和分离。

For the induction motor with composite faults,the fault frequencies component in stator current spectrum is difficult to be separated accurately.Based on the combination of wavelet denosing algorithm and the technique of resonance demodulation,a method of separating composite faults of induction motors is proposed.Depending on good time-frequency localization characteristics of wavelet,it can effectively distinguish mutations of the signals from the noise to achieve signal noise reduction.When implementing resonance demodulation using software,a band-pass filter was constructed firstly to extract information a-bout resonances.And then the Hilbert transformation was applied to the output signal of the band-pass fil-ter for demodulation analysis to obtain fault features including low-frequency envelope signal.Separation of coupling faults and extraction of fault features of induction motors were realized through low-pass filte-ring and spectrum analysis finally.Experimental results suggest that this method makes fault characteristic frequencies more identifiable and separable in induction motors current signal in the case of composite faults.