Vibration based diagnostic of cracks in hybrid ball bearings

The paper presents a vibration based diagnostic system to detect cracks in balls of hybrid ball bearings. The diagnostic system based on a Bayesian Classifier. It is shown that it is able to separate healthy bearings from damaged balls and races. The system is tested with measurement data from a bearing test bench in different operating points of the bearing. Moreover, a novelty detection approach based on a support vector machine is presented.