Bearing Misalignment and Eccentric Wear: A Study on Condition Monitoring

Abstract

This study investigates ball bearing failures and evaluates the performance of various methods for detecting such failures. Experimental tests were conducted on a motor-generator system supported by ball bearings, where several types of real-time faults were intentionally introduced. Vibration signals were collected to capture the early stages of bearing faults. The monitored vibrations successfully predicted impending failures in the outer race, inner race, and ball of the bearings. To assess the effectiveness of the prediction methods, different stages of fault development were considered. Additionally, the feasibility of implementing these techniques in real-world systems and real-time applications was examined. The vibration monitoring technique proved to be suitable for analysing various defects in bearings and providing early indications of malfunction. Vibration measurements were recorded at each housing of the rotor bearings. The results indicated that a deteriorating or faulty ball bearing had a significant impact on the vibration spectra. To identify different defects in ball bearings, time domain, frequency domain, and Fast Fourier Transform (FFT) analysis were employed. The findings demonstrated the usefulness of each of these techniques in detecting problems associated with ball bearings.