Spectral identification of forced oscillation in PMU signal using mode decomposition

This article gives a new method to identify the frequency of Forced Oscillation (FO) in phasor measurement units (PMUs) signals. The processing of the signal by Empirical Mode Decomposition (EMD) technique leads to effective detection with improvement in accuracy. After testing some improved versions of EMD, it is observed that `improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN-2014)' is the best version to use in processing. The developed methodology is able to spot the frequency of FOs even if there exist multiple modes with mode mixing and multiple events. In such cases, self-coherence method fails to spot FO. Spectral analysis is conducted on the decomposed Intrinsic Mode Functions (IMFs) in order to determine the frequency of FO. The analysis with proposed methodology is done on simulated signal as well as on real-time PMUs signal taken from North American SynchroPhasor Initiative (NASPI).