PMU Algorithms Based on Adaptive Taylor–Fourier Models: A Three-Phase Compressive Sensing Method for Spectral Support Recovery

Phasor measurement units (PMUs) are the backbone of transmission grid monitoring and, in a likely future scenario, their use will be extended to distribution systems. This context stimulated the design of PMU algorithms that are more accurate and responsive, capable of tracking fast dynamics while rejecting the interference due to harmonics and interharmonics. An attractive approach relies on the identification of a Taylor-Fourier multifrequency (TFM) signal model that shall include all the significant narrowband components. In this respect, a proper detection of these components, representing the spectral support of the signal, has vital impact on accuracy. The present article improves the spectral support recovery technique adopted by the recently proposed compressive sensing weighted TFM (CS-WTFM) algorithm for synchrophasor, frequency, and rate of change of frequency (ROCOF) estimation. The underlying idea is considering the three-phase signal as a whole and leveraging its quasi three-phase symmetry. The developed measurement algorithm, called CS3-WTFM, searches for a shared spectral support among the three phases. It is analytically proven that, thanks to the aforementioned symmetry, it ensures a more robust and effective retrieval of the signal components with respect to a per-phase approach. Simulation results highlight that CS3-WTFM enables a remarkably higher accuracy in the presence of interharmonic interference. A dedicated selection of the window weights used in CS3-WTFM, leveraging the benefits of the three-phase support recovery, further improves performance.