Comparison of Prony and ARMA Methods for Oscillation Mode Identification in Distribution Systems Based on μPMU

Abstract

Prony method and auto regressive moving average (ARMA) method are two typical methods used in the low-frequency oscillation mode identification of large-scale power system. This paper applies the two methods to distribution systems with multiple DGs to identify oscillation modes. The applicability of the methods to different signals is accessed by employing them to ringdown and ambient signals simulated from a 4-DG islanded distribution system and comparing to the eigenvalues calculated by eigen-analysis of the system’s small-signal model. The results indicate that the ARMA method has better applicability in this situation while Prony method is simpler to implement in case of ringdown signals.