基于SVD与ESPRIT的异步电动机转子断条故障检测新方法  

New Detection Method of Broken Rotor Bar Fault in Induction Motors Based on Singular Value Decomposition and Estimation of Signal Parameters Via Rotational Invariance Techniques

作　　者：许伯强[1] 白旭[1] 

机构地区：[1]华北电力大学电气工程学院,河北保定071003

出　　处：《电机与控制应用》2014年第1期42-47,59,共7页Electric machines & control application

基　　金：国家自然科学基金项目(51277077)

摘　　要：提出一种基于奇异值分解(SVD)滤波和旋转不变信号参数估计技术(ESPRIT)的异步电动机转子断条故障检测方法。通过性能对比,说明SVD滤波可以理想地滤除待分析信号中的供电谐波及其他噪声,因而优于当前在转子断条故障检测中广泛使用的自适应滤波。进一步利用SVD估计经SVD滤波信号的谐波个数,以此改进ESPRIT的频率分辨性能。在此基础之上,将SVD与ESPRIT结合而进行异步电动机转子断条故障检测,试验结果表明这是行之有效的。Based on estimation of signal parameters via rotational invariance techniques（ESPRIT） and singular value decomposition（SVD） filtering technique,a new method of detecting broken rotor bar fault （BRB） in induction motors was proposed. Firstly, SVD filtering technique was compared with adaptive filtering technique commonly used in BRB detection, and the results showed that SVD filtering was completely superior over adaptive filtering, in te^xns of ~moving the power harmonics and other noises in the analyzed signal. Secondly, SVD was further used to improve ESPRIT by pre- estimating the number of harmonics in the SVD-filtered signal. Finally, a new BRB detection method was presented by combining SVD and ESPRIT seamlessly, and test results on an induetion motor showing promising results.

关 键 词：异步电动机 转子断条 奇异值分解 旋转不变信号参数估计技术 

分 类 号：TM307.1[电气工程—电机]