Simulation Research on Rolling Element Bearing Feature Extraction Based on Recursive Least-Squares Lattice-Ladder Algorithms

Abstract

In order to extract the weak fault information from complicated vibration signal of rolling element bearing, the Recursive Least-Squares (RLS) Lattice-Ladder Algorithms is introduced into the field of rolling bearing feature extraction. An adaptive feature extraction method is proposed. The RLS Lattice-Ladder algorithms and its adaptive filter property in the process of feature extraction were discussed. The rolling bearing vibration signal was refined by the RLS Lattice-Ladder filter method, and the refined vibration signal was demodulated by square envelope, then the rolling bearing’s characteristic fault frequency was identified by enveloped normalized amplitude-frequency spectrum. Simulation results show that compared with the LMS filter method, this method can identify fault frequency more quickly and more effectively.