基于SVD-ESPRIT与PSO的笼型异步电机转子断条故障检测  

Detection of Broken Rotor Bar Fault in Asynchronous Motors Based on SVD-ESPRIT and Particle Swarm Optimization Algorithm

作　　者：许伯强[1] 董俊杰[1] 

机构地区：[1]华北电力大学电气与电子工程学院,河北保定071003

出　　处：《电机与控制应用》2016年第3期93-99,共7页Electric machines & control application

基　　金：国家自然科学基金项目(51277077)

摘　　要：将基于奇异值分解(Singular Value Decomposition,SVD)滤波的旋转不变信号参数估计技术(Estimation Of signal parameters via rotational invariance technique,ESPRIT)与粒子群(Particle Swarm Optimization,PSO)算法相结合提出一种异步电动机转子断条故障检测方法。利用ESPRIT的高频率分辨力特性,通过SVD滤波准确提取定子电流信号中转子断条故障特征分量及主频分量之频率,但因其对幅值和初相位估计的效果欠佳,进而尝试应用粒子群优化算法确定各频率分量的幅值和初相位。仿真及试验结果表明,基于SVD-ESPRIT与粒子群算法的异步电动机转子断条故障检测方法是有效的,且因算法简单、运行耗时短亦可用于在线检测。A detection method for broken rotor bar（ BRB） fault in asynchronous motors was presented,which combined the estimation of signal parameters via rotational invariance technique（ ESPRIT） based on Singular Value Decomposition（ SVD） and Particle Swarm Optimization（ PSO）. With the high frequency resolution characteristic,the ESPRIT could precisely extract the broken rotor bar fault feature frequency components and power frequency component in stator current signals filtered by SVD. However,it could not estimate the amplitudes and initial phases accurately. And then try to apply the PSO algorithm to determine the amplitude and the initial phase of each frequency component. The simulation and experimental results demonstrated that the broken rotor bar fault detection method based on SVD-ESPRIT and PSO was effectively and it was also suitable for online detection due to its simple algorithm and short runtime.

关 键 词：异步电动机 转子故障检测 奇异值分解 高频率分辨力谱估计技术 粒子群算法 

分 类 号：TM343.3[电气工程—电机]