SVD滤波技术与Prony谱线估计算法相结合的异步电动机转子故障检测新方法

A Detection Method for Broken Rotor Bar Fault in Asynchronous Motors Based on SVD and Prony Spectrum Line Estimation

提出了一种奇异值分解(SVD)滤波技术与Prony谱线估计算法相结合的异步电动机转子断条故障检测的新方法。SVD滤波技术可以理想地滤除电机定子电流信号中的基频分量与有色噪声,从而凸显转子断条故障特征频率分量。Prony谱线估计算法可以准确计算出各特征分量的频率和幅值,且不失高频谱分辨能力。将二者结合即可在短时采样信号条件下准确、有效地提取转子断条故障特征频率分量。对一台异步电机进行试验,结果表明:基于SVD滤波技术与Prony谱线估计算法的异步电动机转子断条故障检测方法效果理想。

A new method for detecting broken rotor bar fault （ BRB） in asynchronous motors was proposed which was based on singular value decomposition （ SVD） filter and Prony spectrum line estimation algorithm. Simulations of BRB showed that the SVD filter could effectively filter the fundamental component and the color noise from motor stator current signals, which highlights the BRB fault feature component. The results showed that Prony spectrum line estimation could accurately calculate the frequencies and the amplituded of the broken rotors bar feature components with the ability of high spectral resolution. By combining the two methods listed above cluld extract fault feature frequencies accurately and effectively in short sampling signals. Thus, the experimental results of an induction motor showed that the method based on SVD and Prony spectrum line estimation had a desirable performance.