Rolling bearings time and frequency domain fault diagnosis method based on Kurtosis analysis

摘要

When rolling bearing has mechanical localized faults in its various components, pseudo-cyclostationary transient impact signal will be generated. But when in transient faults, this weak impact signal may submerge in strong background noise and gear vibration signal. In this paper, adaptive threshold wavelet de-noising method is used to reduce background noise, then Kurtosis of the noise-reduction signal is calculated in time domain. By comparing the calculation result with a given threshold, a conclusion can be made that whether the bearing is healthy or having mechanical localized faults. When the bearing is diagnosed as faulty in time domain, Kurtogram is used to find out a most suitable pass-band in frequency domain by maximizing the Kurtosis value. Band-pass filtering and do demodulation to this band-pass signal, the faulty component can be positioned precisely and reliably.