Abstract:
Distribution-level signals are often contaminated with harmonics and noise, along with dynamic conditions like frequency deviations. This paper proposes an empirical wavelet transform (EWT) based M-class distributionlevel phasor estimation technique under such polluted and dynamic conditions. The capability of the EWT of extracting different frequency components present in the signal makes it suitable for distribution-level phasor estimation. However, the empirical wavelet filters are designed on the basis of the Fourier spectrum of the signal, which, inturn suffers from a spectral leakage problem at off-nominal frequencies. To avoid the errors arising due to spectral leakage,asamplevalueadjustmentbasedprefilteringtechnique isemployed.Theeffectivenessoftheproposedestimatoris demonstrated on various simulated signals, field-data, and real-time signals obtained from the hardware implementation setup. The proposed algorithm is found to be least affected by the presence of dc components, harmonics, and noise. The use of a fixed window size and fixed sampling frequency makes it suitable for synchronization with GPS clocks. The proposed scheme is also able to provide an accurate harmonic phasor estimation.
