Applicability comparison of three algorithms for electromechanical mode identification

Prony method, the autoregressive moving average (ARMA) method and the stochastic subspace method are three typical signal processing approaches for identifying the electromechanical mode properties in power grids. Based on the review of their elementary principles, the applicability of the three algorithms to different kinds of signals is comparatively assessed from the perspective of near real-time oscillation characteristic estimation. Their performances are evaluated using synthetic ringdown data and ambient data obtained from a 36-node benchmark system. Moreover, the methods are employed to process ambient signals with different SNR levels in order to systematically analyze their application in practical power systems. Some general conclusions are drawn from the analysis for several simulation cases.