Online stator fault diagnosis in inductionmotors

Abstract

This paper presents a noninvasive and online method for detection of stator winding faults in three-phase induction motors from observation of negative sequence supply current. A power decomposition technique (PDT) has been used to derive positive and negative sequence components of measured voltages and currents. A fault detection algorithm has been developed to characterise the effects of supply imbalance and nonlinear motor effects (motor iron saturation, winding imbalance, rotor static eccentricity), which also generate negative sequence current. The effects of motor heating and variation in negative sequence resistance with slip change are minimised by using only motor negative sequence reactance to calculate supply negative sequence current. Change in the motor negative sequence impedance under supply imbalance due to load variation (mainly because of closed rotor slots effect) has been included. Experimental results on several motors show that negative sequence impedance can vary between 10% to 50%. Semiempirical formulas based on theoretical and experimental results have been proposed to eliminate the effects of supply imbalance, load, and voltage variation. Compensation for this negative sequence current before making the fault decision enables a high fault sensitivity to be achieved.