基于DSP的交流电机故障诊断  

AC electric machines fault diagnosis based on DSP

作　　者：孟敬[1] 陈在平[1] 

机构地区：[1]天津理工大学自动化学院,天津300191

出　　处：《天津理工大学学报》2008年第2期29-31,共3页Journal of Tianjin University of Technology

基　　金：天津市自然科学基金(043601511);天津市自然科学基金重点项目(07JCZDJC09600);天津市科技创新基金(2004BA08)

摘　　要：交流异步电机是广泛使用的能量转换装置.对鼠笼异步电动机的常见故障进行分析及诊断,提高电机故障的监测诊断水平,具有重大实用价值.本文应用小波包-神经网络算法实现了基于DSP的交流电机故障诊断.利用小波包分解算法提取出电机故障特征向量,并针对轴承和偏心故障特征,提出对数故障特征向量的算法.将故障特征输入基于任务分解的神经网络,实现基于DSP的电机定子、偏心以及轴承故障诊断.The electric machines as driving equipment are used widely in industry. It is significant in practice to improve the electric machines diagnosis technique. The scheme AC electric machines fault diagnosis based on DSP using the wavelet packet-neural network algorithm is presented in this paper. The eigenveetors of fault are extracted by wavelet packet decomposition algorithm. To extract the bearing of eccentricity fault eigenvectors, this paper gives a novel algorithm based on logarithm fault eigenvectors extraction. The neural networks based on assignments decomposition realize AC electric machines stator,the bearing of eccentricity fault diagnosis based on DSP, using the fault eigenvectors.

关 键 词：DSP 电机故障诊断 故障特征 小波包 神经网络 

分 类 号：TM307[电气工程—电机]