Resumen
In spite of the tremendous advances in computing power and continuous improvements in simulation software made in recent decades, the accurate estimation of wind turbine performance using numerical methods remains challenging. Wind turbine aerodynamics, especially when operating outside of the design envelope, is highly complex: blade stall, laminar-to-turbulent boundary layer transition, rotational effects (lift augmentation near blade root), and tip losses are present. The scope of this research is to show that the classic Reynolds-Averaged Navier?Stokes (RANS) modeling approach, although extensively tried and tested, is not yet exhausted. The NREL Phase VI rotor was used as a basis for numerical methodology development, verification and validation. The numerical model results are compared in detail with the available measured data, both globally (turbine torque and thrust, and blade bending moment) and locally (pressure coefficient distributions and aerodynamic force coefficients at several locations on the blade) over the entire experimental wind speed range. Stall initiation and spread over the blade span are well captured by the model, and rotor performance is predicted with good accuracy. RANS still presents significant value for wind turbine engineering, with a great balance between accuracy and computational cost. The present work brings potential impact on all applications of wind turbines, especially targeting offshore wind energy extraction for which great development is expected in the near future.