Automated diagnosis of rolling bearing faults in electrical drives

This paper deals with a new automated diagnosis method for detecting bearing faults. The diagnosis is carried out by frequency response analysis of the mechanical system of the drive. The detection of bearing damages of a ball bearing on motor and load side of the mechanical drive system is addressed. The mechanics is assumed to be a two-inertia-system with one dominant resonant frequency The actual frequency response of the monitored system can be compared to the one obtained during the commissioning of the plant which can therefore serve as reference. The present paper addresses the investigation of two types of bearing faults: generalized roughness and single point defects. As concerned the latter ones, inner and outer race bearing faults are topic of the research. Furthermore, based on the calculation of two tolerance bands around the calculated frequency responses, an algorithm for the automated diagnosis of both fault categories by pattern recognition is presented.