Review of Advanced Approaches and Vibration Signature Analysis for Fault Detection, Diagnosis, and Prognosis of Rolling Element Bearings

It is prevalent for rolling element bearing to fail in rotating machinery, and their failure can often result in the failure of the entire machine. This paper review techniques for fault detection, diagnosis, and prognosis of rolling element bearing. The methods include condition-based monitoring, energy entropy ratios (EER), wavelet transforms (WT), artificial neural networks (ANNs), support vector machines (SVMs), fuzzy logic (FL), and vibration signature analysis. Researchers have developed various techniques based on vibration signature analysis to detect the fault and its severity. It has been proven that vibration signature analysis is a reliable method for non-destructive health monitoring. Moreover, fault detection methods have been developed based on time, frequency, time-frequency domain, and wavelet transform techniques to capture specific fault signatures in the vibration data. In addition, we have explored prognosis techniques, which estimate the remaining service life of faulty bearings. Transfer learning and hybrid-based approaches are investigated to improve the remaining useful life estimation accuracy. We expect to integrate emerging technologies into future research directions to enhance remaining service life estimation. We have emphasized the importance of condition-based maintenance techniques and vibratory signal analysis in maximizing maintenance practices. This can lead to improved operational efficiency and cost-effectiveness. The review provides insights into state-of-the-art rolling element-bearing fault analysis techniques that may be useful to researchers and technicians involved in predictive maintenance and condition monitoring.