Enhancing sustained oscillation detection by data pre-processing using SSA

Abstract

Detecting sustained oscillations at their early stages can improve power system's situational awareness (SAW), and allows operators take remedial reactions to potential threatens to system's stability. Previously coherence spectrum based detection methods have been proposed to detect sustained oscillation in low signal-to-noise ratios (SNRs) using phasor measurement unit (PMU) data. However, their performance suffers long detection delays. In this paper, a pre-processing method based on Singular Spectrum Analysis (SSA) theory is proposed, which enables the self-coherence method to achieve shorter delay and lower false alarm rate using less operating data. Through extensive experimental study, the effectiveness of the proposed pre-processing procedure has been verified.