Computational Fluid Dynamics (CFD) simulations with fully resolved turbine geometries are increasingly popular for wind turbine aerodynamic analysis. However, the results obtained from the CFD simulations are sensitive to many numerical settings, such as grid spacing, time increment, turbulence models, etc., and thus rigorous quantification of the numerical uncertainties is necessary to be able to have confidence in the results.
In the present study, a modern verification procedure is adopted to assess the numerical uncertainties in the CFD simulations. The NTNU Blind Test 1 turbine is analyzed in the CFD simulations by using ReFRESCO. To quantify the spatial and temporal discretization uncertainties, a simulation matrix consisting of four computational grids with different level of resolutions combined with three different time increments is established. In all the computational grids, the turbine geometry is fully resolved including the blades, hub, nacelle, and tower. Moreover, the computational domain for CFD simulations is constructed such that it exactly matches the wind tunnel in the NTNU Blind Test 1 experiment. Unsteady Reynolds Averaged Navier-Stokes (URANS) simulations with the k-ω SST turbulence model are performed. The Moving-Grid-Formulation (MVG) approach with the sliding interface technique is leveraged to handle the rotating wind turbine and stationary tower and nacelle. By using a modern verification procedure, the numerical uncertainties on the thrust coefficient (CT) and power coefficient (CP) are determined for an inlet velocity of 10 m/s at a tip speed ratio (TSR) of 6.
In order to achieve carbon neutrality in the coming decades, electricity generation from renewable sources, including wind, will have to increase several folds. To meet this requirement, more wind farms need to be deployed worldwide, and more efficient wind turbines and wind farms need to be designed. Computational Fluid Dynamics (CFD) methods have become a very powerful tool in the predictions of wind turbine performance with the continually decreasing computational cost and have been widely used in the wind engineering industry with different levels of fidelity. Blade element method (BEM) based approaches coupled with RANS or LES equations have been predominant in the wind turbine simulations, examples can be found in Troldborg (2009), Wu and Porté-Agel (2015), and Xie and Archer (2017). CFD simulations with fully resolved rotor geometries are also becoming increasingly popular in the predictions of wind turbine performance. Li et al. (2012) performed CFD simulations for the NREL phase VI wind turbine by leveraging the dynamic overset technique. In the study, the geometries of the turbine including the tower are resolved. Both the RANS and the detached eddy simulation (DES) approaches were used in the simulations. Lynch and Smith (2013) also investigated the NREL phase VI turbine by using the CFD code FUN3D with unstructured overset grids. Tran and Kim (2016) studied the performance of the NREL 5 MW turbine under prescribed platform motions by using Star-CCM+. In the study, RANS equations with the k − ω SST turbulence model are solved, and the overset grid technique is adopted. Ye et al. (2021) performed CFD simulations with fully resolved rotor geometry for two tandemly arrayed wind turbines, in which steady-state RANS simulations were carried out by using the absolute formulation method (AFM).
