New technique for identifying bearing faults in three-phase induction motors

Abstract

This paper presents a novel technique based on evolutionary algorithms for identifying severity of ball bearing faults in three-phase induction motors. The proposed technique can be used to determine if the induction motor bearings are still in healthy condition or there is some defect at their inner or outer raceway. These defects cause basic motions on the air gap distribution that can be recognized as temporary air gap eccentricity. The particle swarm optimization algorithm is used to estimate severity of the resultant temporary air gap eccentricity. At this end, a rather exact model of the induction motor is required with the proposed bearing faults at variable severities. Winding function approach is extended to develop such model. This is a non-invasive technique because it uses only the line voltages and line currents of the stator. Other advantages of this method include: insensitive to unbalanced supply voltage, load level variation and temperature variation. The proposed technique is evaluated using simulation data obtained for a 2.2 kW three-phase induction machine. The simulation results demonstrate the effectiveness of the proposed technique for the diagnostics of the bearing faults.