a bi-mechanism approach for the detection of inter-turn faults in bldc machines using autocorrelation and harmonic indices

This paper proposes a method for detecting inter-turn fault (itf) in brushless dc motors (bldc) by integrating two novel fault detection mechanisms in parallel. the first mechanism is a novel approach that detects itfs in real-time using two indicators: autocorrelation of a quasi-torque signal named teq and autocorrelation of the reference current. to identify faults in steady-state, the second mechanism uses the fast fourier transform (fft) to extract two novel fault indicators: the 2nd-order harmonic of teq, a variable which is defined in the paper, and the 3rd-order harmonic of iβ in αβ stationary reference frame. a mathematical model is required to model the itf accurately and easily. therefore, a precise model of itf in qd synchronously rotating reference frame is proposed. the simulation results demonstrate that the proposed method yields minimal delay, high accuracy, and low computational burden, and requires no additional hardware, indicating promising performance.