Simultaneous faults identification of rolling element bearings and gears by combining kurtogram and independent component analysis

A combination of kurtogram and independent component analysis (ICA) is proposed in this paper to identify the faults of rolling element bearings (REBs) and gears existing simultaneously in a gearbox. In the proposed scheme, multi-channel vibrations are picked up from the gearbox at first. Then, the fast kurtogram algorithm is employed to extract the envelopes of each vibration from different channels. Subsequently, the envelopes are separated by an ICA algorithm into independent envelope components according to different sources. Finally, the characteristic frequencies of both the faulty REB and the faulty gear can be exposed simultaneously in the envelope spectral plots. A simulation and an experimental test are introduced to show the effectiveness of the proposed method.