小波与倒频谱分析法在异步电机故障诊断中的应用

Application of Wavelet and Cepstrum Analysis in Fault Diagnosis of Induction Motors

小波具有优良的时频局部化特性,其多分辨率性质可逼近细化频谱,在故障信号提取方面具有突出作用。倒频谱对多成分边频的频谱图分析非常有效,具有解卷积的作用,将边频带谱线简化为更易于观察的单根谱线,实现电机耦合故障分离和故障特征提取。本文对小波分解基本思想和过程进行了详细的论述,系统论述了倒频谱分析的基本原理以及在时频分析中的优势,结合两者特点提出了运用小波与倒频谱分析相结合的电机故障诊断方法。通过实验验证了该方法在电机复合故障诊断中的特点及优越性。理论和实验分析表明该方法可有效地判定故障类型,为电机故障诊断提供了新的思路。

Wavelet has good time-frequency localization properties. The refined spectrum can be approximated by its multi-resolution nature. Therefore, it plays a prominent role in fault signal extraction. Cepstru＇m is very effective for the multi-component side-band frequency spectrum analysis, and has a function of deconvolution, side-band spectrum which can simplify the to more easily observed single spectrum, realizing coupling fault separation and fault feature extraction of induction motors. The basic idea of wavelet decomposition and process is discussed in detail in the paper. The basic principles of cepstrum analysis and the advantage in time-frequency analysis are systematically discussed. Combinating the both features, cepstrum analysis combined with wavelet is recommended for the fault diagnosis of induction motors. Experimental results have demonstrated the characteristics and advantage of the proposed method in composite fault diagnosis of induction motors. Theoretical and experimental analysis indicate that the method can determine fault type effectively, providing a new way for fault diagnosis of motor.