A density based clustering scheme for situational awareness in a smart-grid

Abstract

Synchrophasors are the state-of-the-art measuring sensors that sense voltage, current, or frequency with high data rate. This paper presents an approach to analyze the streaming smart-grid data generated by synchrophasors. A novel unit-circle representation is used to visualize the real-time phasor data. A Density based clustering (DBSCAN) method is proposed to cluster the phasor data to detect bad-data for classification, and identifying various fault anomalies. The paper uses datasets from an IEEE 14 bus system test-bed to aid system operators in carrying various predictive analytics and decisions.