An envelope-based method with second order generalized integrator adaptive notch filter for diagnosis of rotor bar breakage at very low slips

Abstract

Detection of rotor bar breakage in squirrel cage induction motors (SCIMs) via motor current signature analysis (MCSA) is difficult when the motor is working at very low slip operating condition. The fast Fourier transform (FFT) spectrum of the stator current signal is affected by the fundamental frequency component whose leakage masks the fault characteristic frequency. To address this problem, a new method utilizing the second order generalized integratoradaptive notch filter (SOGI-ANF) is proposed. Two signal processing layers which effectively use the SOGI-ANF are used to extract the envelope of the so-called total harmonic signal (THS) which is free from the dominant effect of the main component. The SOGI-ANF is capable of extracting both the input signal envelope and main frequency component. Analysis of the THS envelope reveals significant advantages of the SOGI-ANF over the previous envelope-based analysis techniques using Hilbert transform (HT) in terms of spectral resolution and hardware implementation using embedded devices. The theoretical development of the proposed method and its experimental verification in the laboratory are investigated in this paper. Hardware implementation of the proposed method is achieved through an STM32F4 discovery microcontroller board which makes it a powerful diagnostic tool.