A new approach of multi-degree damage modal identification for motors

Abstract

The bearing was the key part of asynchronous motors, and the bearings would be damaged with multi-degree modal from slightly to badly if they were used for a long time. This paper proposed a method for multi-degree damage modal identification of motors based on wavelet and neural networks. the vibration signals of asynchronous motor were considered to be researched, and the feature vectors were extracted by wavelet-packet transform, and the feature vectors were computed as the form of energy spectrum of the wavelet-packet coefficient, and combined with the classifying tool RBF neural networks for identification of many damaged modals of bearings for motors. The experiment results had proofed its' rationality and validity, and provided a new method for fault analysis and system identification of asynchronous motors.