Research on mechanical equipment fault feature extraction method based on CEEMD and RobustICA

In order to effectively extract the fault features of rolling bearing under complex background noise, a method based on complementary ensemble empirical mode decomposition(CEEMD) and robust independent component analysis (RobustICA) is proposed. Firstly, the fault signal is decomposed by CEEMD and several signal components with different frequencies are obtained. Then, according to the constructed combined weight index system, the effective signal components are screened and reconstructed, and the virtual noise channel is introduced. Finally, the signal and noise are separated by RobustICA, and the de-noising signal is demodulated by envelope. The results show that the proposed method not only has good denoising effect on strong noise interference, but also can extract fault features accurately.