Rolling element bearing is one of the most widely used components in almost all kinds of machinery, which directly influences on working safety of machinery in industry. Based on analyzing the rolling element bearing frequency spectrum characteristic, Since symmetry or approximately symmetry of working parts of a bearing and its uniquely periodical rotating working mode, vibration of a destroyed bearing generally exhibits strong periodicity. These periodical pulses bring amplitude modulation (AM) characteristic. Considering the above reasons, a new signal processing method for bearing condition monitoring is studied in this thesis.AM vibration signal of rolling element bearing's cyclostationarity mechanism is discussed. It has been proved theoretically that the signal exhibits first-order cyclostationarity. The conclusion will be useful at processing amplitude modulation vibration. For a cyclostationary signal, its cycloergodicity ensures the correlation for analyzing a single measured data. The difficulty of signal processing would be reduced greatly. Cyclostationarity is a kind of inherent property for a bearing, which comes from its physical configuration and working mode. So cyclic statistics theory would be able to identify truly condition of a bearing.A new conclusion is introduced that cyclic autocorrelation function possesses demodulation capability. Comparing with envelope demodulation technique, it is proved that the two methods have a similar performance on extracting feature frequencies of a bearing. However the two methods also have essential difference.Based on studying the noise elimination ability of wavelet transform, a analytic method of wavelet cyclic autocorrelation function is set up.The simulation proves this method's validity. |