Abstract
The kurtosis-based indexes are usually used to identify the optimal resonant frequency band. However, kurtosis can only describe the strength of transient impulses, which cannot differentiate impulse noises and repetitive transient impulses cyclically generated in bearing vibration signals. As a result, it may lead to inaccurate results in identifying resonant frequency bands, in demodulating fault features and hence in fault diagnosis. In view of those drawbacks, this manuscript redefines the correlated kurtosis based on kurtosis and auto-correlative function, puts forward an improved correlated kurtosis based on squared envelope spectrum of bearing vibration signals. Meanwhile, this manuscript proposes an optimal resonant band demodulation method, which can adaptively determine the optimal resonant frequency band and accurately demodulate transient fault features of rolling bearings, by combining the complex Morlet wavelet filter and the Particle Swarm Optimization algorithm. Analysis of both simulation data and experimental data reveal that the improved correlated kurtosis can effectively remedy the drawbacks of kurtosis-based indexes and the proposed optimal resonant band demodulation is more accurate in identifying the optimal central frequencies and bandwidth of resonant bands. Improved fault diagnosis results in experiment verified the validity and advantage of the proposed method over the traditional kurtosis-based indexes.
Highlights
Rolling bearings are one of the most common but the most vulnerable parts in rotating mechanical systems
As to the drawbacks of kurtosis detectingcorrelated repetitivekurtosis transient impulses and diagnosingfunction rolling bearing faults, this manuscript firstlyinredefines based on auto-correlation bearing thisThen, manuscript firstly correlated kurtosis based onbasis auto-correlation function and the faults, kurtosis
It proposes anredefines improved correlated kurtosis on the of squared envelope and the kurtosis. It proposes an improved correlated kurtosis on the basis of squared envelope spectrum, which has been proved more effective in detecting repetitive transient impulses
Summary
Rolling bearings are one of the most common but the most vulnerable parts in rotating mechanical systems. Some artificial intelligence algorithms are proposed to detect bearing faults [16,17,18] In view of these deficiencies, Wang [10] proposed an enhanced kurtogram by calculating kurtosis based on the envelope spectrum of the wavelet package transform filtered signal, which performs well in determining resonance bands, and in demodulating the fault features of rolling bearings. The manuscript redefines the correlated kurtosis, on the basis of kurtosis and auto-correlative function, and puts forward an improved correlated kurtosis based on the squared envelope spectrum of rolling bearing vibration signals This manuscript proposes an optimal resonant frequency band demodulation by combining the complex Morlet wavelet band-pass filter and the Particle Swarm Optimization algorithm, which can adaptively identify the optimal resonant frequency bands and demodulate transient features of rolling bearings.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
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
