CONDITION MONITORING OF 3-PHASE INDUCTION MOTOR BEARING BASED ON VIBRATION SIGNAL ANALYSIS

Abstract

The present work focuses onto classify the bearing faults based on vibration signal analysis by using online data available on case western reserve university data sets for bearings. In this work predictive condition monitoring for early prediction of faults in the machine have been studied. Time domain analysis and frequency domain analysis methods based on vibration signals of machines for the calculation of faults have been done. In time domain analysis different statistical parameters calculated. The parameters measured when the bearing is brand-new or undamaged can be compared with the calculated parameter, which indicates the current condition. Time domain statistical parameters like RMS, kurtosis and crest factor are gave best results to identify the fault. Frequency domain analysis is computationally efficient and more effective than time-based methods. By this method the machinery condition can be assessed by observing the presence of stationery frequency components. FFT of the signal gives the spectrum information in the frequency domain. The spectrum presents information indicating the level of vibration at a particular frequency. Analyzing the signature of this spectrum can help detect and identify most types of faults.