Resumen
A large-eddy simulation (LES) study of vertical-axis wind turbine wakes under uniform inflow conditions is performed. Emphasis is placed on exploring the effects of subgrid-scale (SGS) modeling on turbine loading as well as on the formation and development of the wind turbine wake. In this regard, the validated LES framework coupled with an actuator-line parametrization is employed. Three different SGS models are considered: the standard Smagorinsky model, the Lagrangian scale-dependent dynamic (LSDD) model, and the anisotropic minimum dissipation (AMD) model. The results show that the SGS model has a negligible effect on the mean aerodynamic loads acting on the blades. However, the structure of the wake, including the mean velocity and turbulence statistics, is significantly affected by the SGS closure. In particular, the standard Smagorisnky model with its theoretical model coefficient (i.e., CS∼0.16) postpones the transition of the wake to turbulence and yields a higher velocity variance in the turbulent region compared to the LSDD and AMD models. This observation is elaborated in more detail by analyzing the resolved-scale turbulent kinetic energy budget inside the wake. It is also shown that, unlike the standard Smagorinsky model, which requires detailed calibrations of the model coefficient, the AMD can yield predictions similar to the LSDD model for the mean and turbulence characteristics of the wake without any tuning.