Condition Monitoring of an Induction Machine using a Stochastic Search Technique

This paper describes the principles of a new technique for electrical condition monitoring of the stator and rotor circuits in an induction machine. Using time-domain terminal voltages and currents with rotor position data, a stochastic search technique (simulated annealing) is able to confirm that the machine windings are healthy, or indicate that there is a fault and provide information about the location and nature of the fault. The technique is demonstrated using experimental data from a laboratory machine with both armature and secondary winding faults.