Condition Monitoring of Rolling Element Bearings Through Statistical Features and Support Vector Machine

In rotating machinery, bearings are essential and critical components, and their failures frequently result in expensive downtime. This paper studies the vibration of an SKF 6206 deep groove ball bearing, focusing on predictive maintenance and checking how the bearing performs at different speeds and defect sizes. From vibration signal, statisticalfeatures suchas root mean square (RMS), crest factor, peak factor, peak value, kurtosis, and skewness were extracted to help identify faults at an early stage. These features were then used to train a Support Vector Machine (SVM) model for classifying the health state of bearings. By comparing the predictions with experimental data, it was clear that the selected features with SVM gave reliable accuracy in telling healthy bearings apart from faulty ones. According to the study, this strategy can enhance machinery reliability and assist predictive maintenance.