Towards a Data Driven Robust Event Detection Technique for Smart Grids

Increasing levels of power system complexity, due to growing volumes of renewable generation with an associated influx of power electronics, require enhanced operational procedures. Consequently, phasor measurement units (PMUs) are being rapidly deployed in order to improve situational awareness for power system operators. This paper proposes a novel real-time event detection technique based on random matrix theory (RMT) and Kalman filtering. A dynamic Kalman filtering technique has been developed in order to condition PMU data. RMT provides the theoretical basis of the event detection technique, and mean spectral radius (MSR) is employed as an event indicator. Simulation results indicate that the proposed technique is robust when applied to applications involving power system data that includes significant noise levels.