A new scheme in model-based fault detection in three-phase induction motors

In this paper, an induction motor under mixed air-gap eccentricity condition is simulated using modified winding function approach. Then the simulated stator line currents and voltages are used as outputs and inputs to obtain a linear model. Modified recursive least square is used to estimate the model parameters for an ARX model of the induction motor. Variations in estimated parameters show the occurrence and amount of mixed eccentricity in this motor. Also it is proposed as a method for other faults that may occur in three-phase induction motor alone or together, as in this method, faults can be detected and isolated very well. Simulation and experimental results confirm our claims.