Bearing faults detection in induction machines based on statistical processing of the stray fluxes measurements

Frequent defects of induction machines are due to diverse bearing faults. The detection of such faults in their incipient phase can decisively contribute to the prevention of unplanned breakdowns in industrial plants. In this paper the detection of three types of bearing faults by means of statistical processing of the stray fluxes measurements is detailed. The developed noninvasive method requires only both simple probes and easy computations. Numerous measurements had been performed for all the combinations of bearing faults, loads and stray flux probes taken into study. All the results emphasized the effectiveness of the applied simple fault diagnosis method.