Fault Diagnosis of Roller Bearing Using Dual-Tree Complex Wavelet Transform, Rough Set and Neural Network

In a complex field environment for modern mechanical equipment, how to identify all kinds of operational status of the rolling element bearings fastly and accurately is very important and necessary. A novel approach to automated diagnosis is introduced, which is based on feature extraction with the Dual-Tree Complex Wavelet Transform (DT-CWT), then attribute reduction with rough set theory and finally pattern recognition with Artificial Neural Network. In our experiment, 4 kinds of states on a rolling element bearing test table, including normal, pitting on inner ring, pitting on outer ring and pitting on rolling element, are adopted. The experimental results indicate that the proposed feature extraction and automated diagnosis method can extract significant feature sets from signal, and can accurately distinguish many fault pattern, and has some practical value for the on-line condition monitoring of modern industrial demands. KeywordsDual-Tree Complex Wavelet Transform; rough set theory; Neural Network; rolling element bearings; fault diagnosis