Power System Generator Coherency Identification for Large Disturbances by Koopman Modes Analysis

This paper presents an effects analysis of large disturbances on power system generator coherency. The analysis is based on a comprehensive parametric study combining fault parameters with system operating conditions. Generator coherency is identified via nonlinear Koopman modal analysis technique applied on generators rotor speed dynamics following large disturbances. The Koopman operator captures the highly nonlinear spatiotemporal dynamics that cannot be assessed with standard linear coherency identification techniques. Faults are short-circuits of variable durations and locations. System operating parameters include loading levels and on line generators. The study methodology is demonstrated on the Tunisian electric network comprising 153-buses 26-machines with a comparison to the conventional slow coherency method. The results show that large disturbance location and system loading levels, alter the coherent grouping determined by weak connections topology. Slow coherent areas may degenerate into smaller coherent groups under large disturbances. Such information is particularly useful for emergency control where intentional islanding actions ought to be taken to mitigate large blackouts.