Targeted approach to apply masking signal-based empirical mode decomposition for mode identification from dynamic power system wide area measurement signal data

This study proposes an improved masking signal approach for the effective handling of the mode mixing problem associated with closely spaced mode frequency components of power system signals while applying empirical mode decomposition (EMD). The effectiveness of the proposed technique is shown with the help of synthetic as well as real-timesignals. A multifrequency synthetic signal with frequencies in the range of inter-area oscillation modes of a typical power system is analysed to identify a dominant modal frequency component with the existing and the proposed technique as well. Different modal frequencies are extracted from the wide area measurement signals (WAMS) recorded in the Eastern Interconnect Phasor Project on 20th July 2005. The modal frequency extracted is compared with the extracted frequency using the masking signal approach adopted in the standard EMD technique. Further, the WAMS data recorded on Northern grid of the Indian Power System have also been analysed. The data obtained from four recently installed PMUs after an incident of loss of generation of 2000 MW, which occurred at Rihand Super Thermal Power Station on 1st June 2010, have been analysed in this study. The magnitude and frequency variations of various signals extracted are compared using the Hilbert spectrum.