d-SyncAED: Distribution Synchrophasor Anomaly and Event Detection Tool in Real-Time

Timely and precise identification of anomalies and events in cyber-power distribution systems is critical. This work introduces d-SyncAED, a real-time bad data anomaly and event detection, and a broader event classification tool based on phasor measurement units in distribution systems. d-SyncAED builds on the integration of lightweight base detectors including Chebyshev, Linear regression, and DBSCAN to detect bad data anomaly and utilizes DBSCAN, Hierarchal Clustering, and K-means for event detection as well as IP-based cyber anomaly detection. The tool also offers an intuitive, user-friendly graphical interface for continuous monitoring and analysis of anomalies by power system operators and engineers. Front-end and back-end of the developed prototype tool have been discussed, which also incorporates multi-threading to process large volumes of real-time d-PMU data for higher efficiency and scalability. The tool is validated using a hardware-in-the-loop cyber-power testbed for multiple scenarios including power system events, cyber-induced outages, and data falsification attacks on real-time d-PMU data, obtaining an average precision of 98.3%.