Fault Diagnosis of Ball Bearings Using Support Vector Machine and Adaptive Neuro Fuzzy Classifier

Bearing faults are one of the major sources of malfunctioning in machinery. A reliable bearing health condition monitoring system is very useful in industries in early fault detection and to prevent machinery breakdown. This paper is focused on fault diagnosis of ball bearing using adaptive neuro fuzzy classifier (ANFC) and support vector machine (SVM). The vibration signals are captured and analyzed for different types of defects. The specific defects consider as inner race with spall, outer race with spall, and ball with spall. Statistical techniques are applied to calculate the features from the vibration data and comparative experimental study is carried using ANFC and SVM. The results show that these methods give satisfactory results and can be used for automated bearing fault diagnosis.