关键词: 奇异值分解, Teager能量算子, 故障诊断

Abstract: In order to solve the difficult problem in?extracting?the weak fault information from rolling bearing vibration signal, a?method of feature extraction of bearing vibration signal based on Singular Value Decomposition（SVD） and Teager-Kaiser Energy Operator（TKEO） is proposed. The abrupt information is extracted from satin signal and noise by using the SVD. It calculates the instantaneous?energy of abrupt information by using TKEO, and the energy information is analyzed to gain the feature of the energy frequency spectrum of bearing vibration signal that used for fault detection. The method is used to extract the fault feature of bearing vibration signal with outer and inner circle faults, and the?fault type can be?accurately detected and identified by feature information. The result shows that the method presented here is feasible?and valid.

Key words: Singular Value Decomposition（SVD）, Teager-Kaiser Energy Operator（TKEO）, fault diagnosis