Data-Driven Fast Transient Stability Assessment Using (Fault-on + 2) Generator Trajectories

Abstract

For transient stability assessment (TSA) in modern power systems, the assessment results should be issued as soon as possible to leave enough time for pre-emptive control. To this end, this paper develops a fast online TSA scheme using fault-on trajectories and their two adjacent data-points in pre- and post-fault stages, i.e., (fault-on + 2) trajectories. First, (fault-on + 2) trajectories of voltage magnitudes, rotor angles, and frequency deviations are acquired from multiple generators via PMUs. With such transient trajectories, a novel anti-noise transient bitmap based descriptor is then strategically designed to comprehensively describe the system-wide transients in bitmap forms. Finally, a convolutional neural network based TSA model is constructed by deep learning of transient bitmaps. Test results on the IEEE 39-bus system demonstrate the effectiveness, adaptability, and robustness of the proposed TSA scheme.