The Intelligent Failure Diagnosis Of Rolling Bearings Based On Wavelet Packet

The research on failure diagnosis of rolling bearings, which are the most ordinary parts in mechanical equipments, is very significant. When locally scathed, the bearings would bump the other parts periodically with the result that the seasonal impulses come into being. With band enough broad to overcast each connatural frequency of the whole bearings, the impulses necessarily arose each connatural vibrations and thereby the stationary vibrations turn into transients.Although well localized in frequency, the Fourier transform was localized none too well in time which makes it a cumbersome tool for transients. Moreover, a local time-frequency composition as the windowed Fourier transform, the short-time Fourier transform namely, it has the same resolution across the time-frequency plane because of the same spread of the window on which the resolution depends. Having a multiscale resolution in time and frequency, the Wavelet transform is endowed with the indisputable hegemony in signal processing especially for transients.Aiming at the failure diagnosis of the rolling bearings, this paper would devote most of its efforts to the following:1 .Introducing systematically the vibration mechanism of the rolling bearings and the vibration character of typical failure;2.Expatiating on the essential theories and the applications in signal processing of the Wavelet transform and proposing the frequency-shift algorithm for wavelet packets to overcome the frequency aliasing existing in Mallat algorithm and make the decomposed WP sequences correspond to the linearly divided frequency bands;3.Diagnosing the typical failure of the rolling bearings respectively through the envelope demodulation and the delayed correlated envelope demodulation whose principles are also set forth in this paper, and confirming the effectiveness of the latter by experiments which can reduce noise further compared with the envelope demodulation;4.Specifmg how to apply the BP neural network to the failure diagnosis andproposing the improved algorithm to overcome the shortages of the BP neural network. Moreover, appling the BP neural network to identify the failure of the rolling bearings which shows the feasibility and the effectiveness of this method.