A low frequency rolling bearing fault diagnosis method based on local projective noise reduction

The local projective noise reduction and demodulated resonance method for fault diagnosis of low speed rolling bearing are presented in this paper. First,one would reconstruct the time series using delay coordinates. In the high dimensional phase space,the local projective algorithm is used to separate it into ground signal and noise signal in the time series by dividing the phase space into orthogonal sub-spaces. Then the signal removed noise is processed by demodulated resonance method. By using the method which combines the advantages of local projective noise reduction and demodulated resonance technology,fault on the low speed bearing is successfully diagnosed.