Sensorless speed estimation and diagnosis of induction motors based on purified space vectors

Abstract

A new scheme based on the demodulation technique and interpolated fast Fourier transform (IFFT) is proposed to construct purified space vector for the purpose of sensorless speed measurement and fault diagnosis for induction motors. The phase and amplitude fluctuation associated with the eccentricity harmonics is demodulated first. Interpolated fast Fourier transform is then performed to estimate the speed of induction motors running under constant speed and transient conditions. Experimental studies have further confirmed good agreement between the estimated motor speed using the proposed scheme and the measured speed using an encoder. The effectiveness of proposed purified space vector has been further demonstrated in the detection of coupling misalignments and stator faults.