小波包与神经网络在电机故障诊断中的应用研究  

Research on Application of Wavelet Packet and BP Neural Network in Fault Diagnose of Electromotor

作　　者：朱文[1] 侯北平[1] 

机构地区：[1]浙江科技学院自动化与电气工程学院,浙江杭州310023

出　　处：《电气自动化》2006年第1期10-11,16,共3页Electrical Automation

摘　　要：文章对电机的故障特点进行分析,根据小波包变换能将信号按任意时频分辨率分解到不同频段的特性,结合小波包的能量特性,提出了故障信号在不同分解频段的能量特征概念及算法,并将其与 BP 神经网络相结合,提出了一种新的电机故障诊断方法,实验结果证实了该方法的正确性和有效性。This paper analyzes the fault characteristics of electromotor. Based on wavelet packet transform theory which can decomposes the signal into several frequency sections according to the different resolution, the energy characteristics and the algorithm of the fault signal in different frequency sections are presented, a new approach of electromotor fault diagnose is presented to combine with BP neural network. The experiment result proves its validity.

关 键 词：小波包变换 故障诊断 BP网络 

分 类 号：TM307.1[电气工程—电机] TP183[自动化与计算机技术—控制理论与控制工程]