关键词: 感应电动机, 转子断条, 多重信号分类, 支持向量回归机

Abstract:

A novel technique which combines multiple signal classification(MUSIC) and support vector regression machine (SVRM) was studied, this technique can improve the accuracy and rapidity of tracking the broken rotor bar characteristic frequencies and amplitudes in induction motors. MUSIC has better detection precision and higher frequency discernment, and it can track the frequencies of sidebands as well as the fundamental frequency component with a very high accuracy even with a short-time sample. MUSIC is a powerful tool extracting meaningful frequencies from the signal, but it is not able to estimate the amplitudes and phase angles of those components. As a solution for this problem, SVRM was introduced to determine the amplitudes and phase angles of the signal and the results have a superior performance. A Y132M-4-typed induction motor was used to conduct an experiment, experimental results illustrate the validness and superiority of the proposed method.

Key words: induction motor, broken rotor bar, multiple signal classification (MUSIC), support vector regression machine (SVRM)