Skip to Main Content
The monitoring and rejection of voltage and current harmonics in power electronics applications such as power quality conditioners or distributed generation systems require correct estimation algorithms especially if the harmonic amplitudes are time varying. Power signal decomposition in multiple synchronous rotating reference frames (MSRFs) is considered one of the best solutions. The most commonly employed implementations of this signal transformation are based on phase-locked loops (PLLs), recursive discrete Fourier transforms (RDFT), or discrete Kalman filtering (DKF). In this paper, a rigorous analysis of the performance of these implementations has been carried out. Complete tests have been performed to evaluate the computational burden, the frequency domain response, and the tolerance to low frequency amplitude variations. The results make it possible to select the proper method depending on the requirements of each application.