Wide-Area Monitoring of Power System Using Dynamic Mode Decomposition on Nonlinear Observables

Modern interconnected electric grids face a challenge from the oscillatory behavior and associated stability of the system. The system will never have a steady-state as small and continuous fluctuations in load occurs. Moreover, in the interconnected grid structure, a disturbance in one location can spread and affect the rest of the parts. The knowledge about the nature of dynamical behavior will really help in monitoring and identifying the systems temporal evolution. To employ good control strategies for better stability, information regarding poorly damped modes must be known. Recent literature has shown the effectiveness of dynamic mode decomposition (DMD) based approach as a faster method for wide-area monitoring (WAM) in power systems. The modeling using DMD presumes that the dynamics are linear. However in many cases, the behavior will be nonlinear. The concept of nonlinear observable allows one to extend the DMD based approach to capture the dynamics. The present paper is an attempt to explore and evaluate the usefulness of explicit mapping methods to the nonlinear behavior of the power system when subjected to fast-rate disturbances. From the simulated experiments, it is found that the DMD with nonlinear observables using explicit mapping approach is able to capture the dynamics in real-time.