Locus-Based Fault Indicators and a Cumulative Sum Algorithm to Detect Induction Machine Stator Interturn Short-Circuit Faults

In this paper, a detection algorithm for stator winding interturn short-circuit faults in induction machines is presented. A two-step procedure is implemented with first the computation of fault indicators, and secondly detection of the fault based on the application of the cumulative sum algorithm. Locus-based fault indicators using electric measurements are defined along with conventional fault indicators. With the use of the cumulative sum algorithm, these fault indicators are assessed regarding their capability of detecting interturn short-circuit faults for multiple load conditions for a field-oriented controlled induction machine. Experimental results obtained on a 5.5kW induction machine setup show that a higher sensitivity is obtained with the presented fault indicators compared to the conventional ones, and that better performances are obtained with the cumulative sum algorithm compared to a direct thresholding of the fault indicators.