Fault Diagnosis of Rolling Bearings based on SVD-LMD Joint De-noising and TEO

Abstract

Aiming at the difficulty extracting the local damage information of rolling bearings under the background of random noise，a new feature extraction method based on singular value decomposition(SVD) and local mean decomposition(LMD) joint de-noising combined with Teager energy operator(TEO) is proposed. Firstly，by using the SVD method，the fault vibration signal of rolling bearings is processed to eliminated the background noise preliminarily. Then，the signal which is denoised by using LMD method is reconstructed after the sensitive product function(PF) is screened out according to the correlation coefficient index. Finally，the reconstructed signal is analyzed by TEO demodulation，the frequency component which amplitude prominent in demodulation spectrum is compared with the theoretical value of fault characteristic frequency to extract fault information. The experimental results demonstrate that the method can effectively extract the characteristic frequency of the local damage information of rolling bearings and the fault diagnosis is realized.

Keywords

• Rolling bearing;Singular value decomposition(SVD);Local mean decomposition （LMD）;Teager energy operator(TEO);Fault diagnosis