Evaluating dynamic stall models for thick wind turbine airfoils

Abstract

The accuracy of the Beddoes–Leishman and Risø dynamic stall models is evaluated against experiments on thick wind turbine airfoils with a relative thickness of 35% and trailing edge thicknesses of 10% and 2%, both with and without vortex generators. The dynamic lift, drag, and pitching moment coefficients simulation results are compared with the measurements, obtained in the TU Delft LTT wind tunnel at a Reynolds number of Re=1×10⁶ and dynamic reduced frequency of 0.064. The study revealed that while the aforementioned models successfully predicted the direction of the dynamic cycles, they inaccurately captured the dynamic stall behavior of thick flatback and non-flatback airfoils in all configurations, particularly in separated flows. There was no significant difference observed in the performance of the two models. The reasons for modeling failure are thoroughly examined from both fundamental and mathematical perspectives, and suggestions for improvements are provided. The findings raise concerns regarding the accuracy and reliability of the dynamic load assessment and aeroelasticity analysis for modern large wind turbines, using current dynamic stall models and underscore the necessity for enhancing the existing models.