Pattern Mining and Anomaly Detection based on the Power System Synchrophasor Measurements
Real-time monitoring of power system dynamics using phasor measurement units (PMUs) data improves situational awareness and system reliability and potentially can help prevent blackouts due to early anomaly detection. The study presented in the paper is based on the actual PMU measurements of the U.S. western interconnection system. Given the nonlinear and nonstationary PMUs data, we developed a robust anomaly detection framework. Wavelet-based multi-resolution analysis, combined with moving-window-based outlier detection and anomaly scoring, were deployed to identify potential PMU events. The candidate events were evaluated using spatiotemporal correlation analysis, and then classified for a better understanding of the event types. The results demonstrated successful anomaly detection and classification, compared with the recorded real-world events.