Signal processing for time-frequency representation of synchro-phasor data

Abstract

With increased number of phasor measurement units (PMUs) deployment in the power system, there is growing interest in real-time event detection. The study in this paper presents post-processing approach to map the event occurrence for large sets of synchro-phasor data available from different regions/buses in the power network. The time–frequency representation (TFR) is applied to analyse the transients that contain rapid variations in amplitude or phase during the event occurrence against post-event conditions. This is based on reassigned smoothed-pseudo-Wigner–Ville distribution. The capability to correctly localise TF regions, where signals are locally coupled, is assessed on synthetic signal and PMUs signals. Important information such as time marginals is estimated to determine any event present in the analysed signal segment. Furthermore, the event localisation is enhanced applying Hough transform on TFR with calculation of group delay and reassignment vector. The developed procedure gives reasonable results for highlighting the distinct representation of events in study. The proposed approach can be easily implemented in real-time monitoring for event detection in the power system.