Abstract
The nonlinear wavelet-morphological wavelet is introduced into the field of vibration signal processing to explore methods that can process effectively both Gauss noise and non-Gauss noise based on the routine of bi-spectrum analysis. Morphological wavelet packet was developed and its analysis and synthesis steps were described,to perfect the morphological wavelet theory. On the basis of excellent de-noising effects of morphological wavelet packet,three kinds of cascaded algorithms were put forward based on the morphological wavelet packet,Hilbert envelope analysis and bi-spectrum analysis,that is morphological wavelet packet-bi-spectrum,morphological wavelet packet-Hilbert-bi-spectrum and Hilbert-morphological wavelet packet-bi-spectrum.These three algorithms showed good inhibition for Gauss noise and non-Gauss noise. Hilbert-morphological wavelet packet-bi-spectrum algorithm especially unfolds best adaptability. Results in rolling bearing signal processing show that Hilbert-morphological wavelet packet-bi-spectrum algorithm can tell bearing fault of different components and different fault degree from each other clearly.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call