Poster
in
Workshop: Tackling Climate Change with Machine Learning: Fostering the Maturity of ML Applications for Climate Change
Identifying Complex Dynamics of Power Grid Frequency
Xinyi Wen · Ulrich Oberhofer · Leonardo Rydin Gorjão · G.Cigdem YALCIN · Veit Hagenmeyer · Benjamin Schäfer
The energy system is undergoing rapid changes to integrate a growing number of intermittent renewable generators and facilitate the broader transition toward sustainability.As millions of consumers and thousands of (volatile) generators are connected to the same synchronous grid, no straightforward bottom-up models describing the dynamics are available on a continental scale comprising all of these necessary details.Hence, to identify this unknown power grid dynamics, we propose to leverage the Sparse Identification of Nonlinear Dynamics (SINDy) method. Thereby, we unveil the governing equations underlying the dynamical system directly from data measurements.Investigating the power grids of Iceland, Ireland and the Balearic islands as sample systems, we observe structurally similar dynamics with remarkable differences in both quantitative and qualitative behavior.Overall, we demonstrate how complex, i.e. non-linear, noisy, and time-dependent, dynamics can be identified straightforwardly.