Data Driven Modal Decomposition of the Wake behind an NREL-5MW Wind Turbine

The wake produced by a utility-scale wind turbine invested laminar, uniform inflow is analyzed means of two different modal decompositions, the proper orthogonal decomposition (POD) and dynamic mode (DMD), in its sparsity-promoting variant. considered NREL-5MW at tip-speed ratio λ=7 diameter-based Reynolds number order 108. flow simulated through large eddy simulation, where forces exerted blades are modeled using actuator line method, whereas tower nacelle employing immersed boundary method. main structures identified both decompositions compared some differences emerge that can be great importance for formulation reduced-order model. In particular, high-frequency directly related to tip vortices found methods, but it ranked differently. other dominant modes composed large-scale low-frequency structures, with frequency content spatial structure. most energetic 200 POD account ≈20% only kinetic energy. While same DMD modes, possible reconstruct field within 80% accuracy. Despite similarities between set comparison these modal-decomposition techniques points out an energy-based criterion such as used may not suitable formulating model wakes, while appears able perform well reconstructing few modes.