Diagnosis of stator winding inter-turn shorts in induction motors fed by PWM-inverter drive systems using a time-series data mining technique

An effective technique for diagnosis of stator winding inter-turn shorts in induction motors fed by PWM-inverter drive systems is proposed. This is done through the use of a time-series data mining technique which identifies and extracts hidden and inherent patterns (characteristics) in the machine phase currents, that can be used for fault identification. This technique can effectively detect and determine the severity of stator inter-turn faults in motors by analyzing the extracted fault signatures of these faults in comparison with the healthy performance signatures. In addition, it will be seen from the experimental results that the proposed technique is immune to motor "non-idealities" such as inherent manufacture-based motor asymmetry due to motor structural imperfections, performance measurement imperfections, and supply voltage unbalances, which result in departure of performance results from those of an ideally balanced 3-phase machine. In this paper, the case-study under investigation is a 230-volt, 60-Hz, 2-pole, 2-hp, squirrel-cage three-phase induction motor-drive system. The experimental results will demonstrate the soundness and robustness of this technique for reliable fault diagnostics.