Spatial–Temporal Data Analysis-Based Event Detection in Weakly Damped Power Systems

Focusing on tackling the challenging issue of detecting events in weakly damped power systems, this letter proposes a novel event detection approach via spatial-temporal data analysis. Based on the intrinsic spatial-temporal correlations in regional PMU measurements, it efficiently profiles spatial-temporal nearest neighbors (STNNs) of multiple buses’ featured time series. With such discriminative STNN profiles, it is able to detect the occurrence of various events even in the presence of significant frequency oscillations. Numerical test results on the Nordic test system as well as several IEEE test systems demonstrate the efficacy of the proposed approach.