基于小波包频带-能量重构的电机断条故障诊断

Rotor Bar Breaking Fault Diagnosis for Motor Based on Wavelet Packet Frequency Band-Energy Reconstruction

异步电机转子断条故障特征分量(1±2s)f与基波频率f相比,幅值较小、易被主频淹没。在研究小波包分解与重构的特征基础上,讨论了小波包分解树的频带分布规律,确定以sym8小波对电机定子电流信号进行12层分解,并通过对故障特征频带的信号能量重构来判断电机有无转子断条。运用该方法能避免基波频率对故障特征分量的影响,故障特征检测更明显,并能诊断转子轻微断条故障。仿真和试验结果证明了该方法的有效性。

Compared with the main frequency, the feature frequency （ 1 ± 2s）f of the fault signal of broken rotor of asynchronous motor has smaller amplitude and was easily drowned. So based on the study of the characteristics of wavelet packet decomposition and reconstruction, the distribution regularity of the frequency band of wavelet packet decomposition-tree was discussed. Sym8 wavelet was selected as wavelet basis. When the motor stator current signals were decomposed at twelfth frequency scales, whether the rotor bars have been broken or not can be determined by energy reconstruction of the fault characteristic frequency band signal. Used this method, the impact on the component of fault features by the fundamental frequency can be well to avoid. Fault feature were more obviously and minor rotor broken bars fault can be detected. Simulation and experimental results showed that the method was correct and effective.