A Novel Bio-Inspired Technique for Rapid Real-Time Generator Coherency Identification

Abstract

Generator coherency identification is establishing itself as an important task to aid in the resistance of cascading failures within wide-area power systems and as a necessary preprocessing stage in real-time control for transient stability. Inspired by flocking behavior in nature, we propose a novel multiflock-based technique to identify generator coherence rapidly within a short observation window. Our measurement-based approach transforms generator data from the observation space to an information space, whereby the generator frequencies and phases characterize the movement and dynamics of boids within multiple flocks. Analysis of the boids' trajectories enables the discrimination of multiple flocks corresponding to coherent generator clusters. We demonstrate the effectiveness of our technique to identify generator coherency rapidly while exhibiting robustness to environmental noise and cyber attack on the 39-bus New England test system and a modified IEEE 118-Bus test system.