Diagnosis of multi-phase turn faults of induction motor stator-windings

In an AC motor, the quick detection of an initially small fault is important for preventing any consequent large fault. Various detection approaches have been proposed in previous papers, for example, by the Park vector (PV), AI techniques, wavelet analysis, and negative-sequence analysis. This paper proposes a method for diagnosing the stator-winding faults of an induction motor by the direct detection of its negative-sequence current. Before starting the diagnosis, the asymmetry admittances for the considered fault cases are obtained by analysis or simulation. The amplitude and phase of the positive-sequence voltage, V_p, and of the positive-sequence current, I_p, are extracted from the voltage PV and current PV, respectively. The amplitude and phase of the negative-sequence, I_n, are extracted from the residue. The asymmetry admittance, Y_a, is calculated from I_n and V_p. When the positive-sequence admittance is known, Y_a can also be calculated from Y_p, I_p, and I_n. These steps are repeated for each sample time and the motor condition is diagnosed according to the variations in the Y_a values. The simulation and experimental results are also shown and the proposed method is investigated and validated.