Dynamic mode decomposition of wind turbines wake: a path to data-driven reduced order modeling

In this study, the DMD method is applied to high-fidelity simulations of a wind turbine wake to analyze the intricacies of flow dynamics. LES of the wind turbine wake is conducted. The high-fidelity simulation results were validated against field measurements. The flow field is examined and found that large structures are identifiable, gradually decreasing in size as the mode number increases. Attempts are made to reconstruct the flow features using DMD. The first 40 modes captured [Formula: see text] of the total energy and were sufficient for reconstructing the flow field, closely aligning with the results from LES. The modes are ranked according to frequency and analyzed. This finding indicates that the dominant coherent structures are consistent across different cases, potentially enabling the development of a data-driven ROM based on such generic building blocks. The fundamentals of a ROM are outlined for predicting the wind turbine power coefficient using DMD modes.