ANN-based system for a discrimination between unbalanced supply voltage and phase loss in induction motors

Abstract

It is documented that almost 98% of all voltage generated by electric utilities has up to 3% unbalance. Single phasing fault deserves special attention since phase loss is considered the worst case of unbalanced supply voltage. This paper focuses on unbalanced supply condition diagnosis and discrimination between an unbalance in the supply and phase loss fault. The discrimination will be based on the ratio of third harmonic to fundamental Fast Fourier Transform (FFT) magnitude components (RTHF-FFT) of the three-phase stator line currents and supply voltages under different load conditions and using artificial neural network (ANN). The proposed approach achieves high accuracy in detecting the unbalanced supply voltage condition in induction motor and identifying the level of severity of the fault. In addition, the proposed algorithm will discriminate between the effects of unbalanced supply voltage and those due to phase losses fault. The paper proposed a reliable approach for detection and diagnosis of unbalanced supply voltage condition. Possible loss of winding insulation under different percentages of unbalanced supply voltages will be predicted which could help preventing sudden failure of the motor during operation. The approach will be proved through experimental validation.