Electrical machine winding faults diagnosis using periodogram

Many methods for the fault detection and diagnosis of the electrical machine are found in the relevant literature; however, the periodogram has not really been applied to detect and diagnose faults in an electrical machine. This paper explains periodogram as a non-parametric estimate of the power spectral density (PSD) and also discusses how two types of periodogram (Default-P and DFT-P) can be used to detect stator winding faults in an induction machine. The maximum PSD values of the healthy electrical machine for both Default-P and DFT-P signature indicate the machine is in good condition. However, when there is any fault in the machine, there are discrepancies in the signatures compared to the signature of the machine in good condition. The percentage deviations, and abnormalities observed from an electrical machine with faults give indications about the extent of the severity level of the fault in the machine. The percentage deviations for each state of the machines are similar in the two periodograms. It is approximately 10% for the phase-to-ground fault, around 12% for the shorted-turn fault and about 94% for the coil-to-coil fault. Thus, with periodogram, it is possible to diagnose winding faults of electrical machines.