A REVIEW ON BEARING CONDITION MONITORING AND DIAGNOSIS THROUGH VIBRATION AND ARTIFICIAL INTELLIGENCE TECHNIQUES

mechanical engg

The important sources of vibration generation considered here are as the non-linear stiffness/damping and localized defects of the contacting surfaces of bearing elements. It also covers non-linear dynamic mathematical modeling of cylindrical roller bearing for localized defects. The numerical integration technique Newark-has been used for solution of the system equations. The parameter of study for cylindrical roller bearings is rotor speed. Results have been analyzed by Fast Fourier Transform (FFT), phase plot and Poincare maps. focus on use of multi-parameters, synthesis of fault features and optimization of feature sets, in order to improve fault diagnostics accuracy, which would thus enhance machinery reliability, availability, safety, and reduce maintenance costs. The complex and non-stationary vibration signals with a large amount of noise make fault detection of rolling element bearings very challenging, especially at the early stage. Daubechies wavelet is popular for smoothing of signals so, it is chosen for reducing the background noise from raw vibration signal. This research investigates the possibilities of improving machinery diagnostics accuracy based on time domain features and various intelligence techniques like artificial neural network, support vector machine.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.