基于随机子空间的异步电动机转子断条早期故障检测

Method for Early Fault Measure of Broken Bars in Induction Motors Based on Stochastic Subspace Identification

针对早期转子故障时定子电流信号微弱不易检测及电网频率波动易引起频率混淆的问题,提出了基于随机子空间的转子断条早期故障检测方法。将采集到定子电流信号构成Hankel矩阵,经矩阵QR分解和特征值分解,得到实时电网频率和故障特征频率,实现早期故障检测。仿真和实验结果证明该方法有效、可行。

Broken rotor bar fault is a common cage induction motor faults. Broken rotor bar fault de- tection of early stage is to ensure safety in production, improve production efficiency, and reduce eco- nomic losses significantly. Because early failure of the rotor for the stator current is not easy to detect weak signals, and frequency of grid frequency fluctuations lead to confusion, early fault detection method is proposed based on stochastic subspace identification of broken rotor bars. The research gath- ered the stator current signal to constitute the Hankel matrix, and obtained the real-time main frequen- cy and the breakdown characteristic frequency after the decomposition and the characteristic value de- composition after matrix QR, realizing the incipient failure examination. The simulation and the experimental results indicate this method has validity and feasibility.