Data-Driven Analysis of Distributed Generator-Based Power Systems Using Koopman Mode Decomposition

Accurate analysis and simultaneous control of observables, such as frequency and voltage, are necessary for power system stabilization. However, the trigonometric function-based Fourier transform (FT) is insufficient for analyzing power systems with multiple harmonics and nonlinear distortions as it approximates the dynamics with a linear model that assumes periodicity and stability, thereby prompting a need to develop a new method with higher performance and speed. In this study, we used a new nonmodel-based data-driven empirical mode decomposition (EMD) method, known as Koopman mode decomposition (KMD), to analyze nonlinear and unsteady power systems consisting of synchronous generators and distributed generators. The strengths and weaknesses of KMD were discussed by comparing the individual results obtained from KMD and FT.