Dynamic Stall of Thick Wind Turbine Airfoils

Abstract

Over recent decades, the wind energy sector has undergone rapid expansion, driven by escalating global energy demand and the widespread transition toward sustainable power generation. This development has been accompanied by a substantial increase in wind turbine rotor size. However, scaling up turbine dimensions introduces significant engineering challenges, particularly in achieving high aerodynamic performance while maintaining cost effective structures. The resulting trend toward longer and more slender blades has amplified both static and dynamic loading, making structural integrity a central concern in contemporary turbine design. To address these challenges, modern rotor concepts incorporate thicker inboard and mid-span airfoils, including flatback configurations, which improve structural efficiency and aerodynamic performance while reducing overall blade mass. At the same time, enlarging blade radii leads to greater elastic deformations, which further increase the dynamic loads experienced by the rotor.

Ensuring reliable turbine operation requires accurate load assessment, which depends heavily on aeroelastic and dynamic load simulations across a broad range of operating conditions. As blades become larger and more flexible, they are subjected to stronger inflow turbulence, more pronounced unsteady variations in angle of attack, and an increased likelihood of transient stall phenomena. These factors make accurate unsteady aerodynamic modeling essential. In this context, dynamic stall models are frequently employed to predict unsteady aerodynamic forces. Historically, however, most dynamic stall models were developed for thin airfoils in helicopter applications, and despite subsequent adaptations for wind turbine use, their suitability for thick airfoils remains limited. The aerodynamic response of thick airfoils, particularly those typical of large modern blades, still lacks comprehensive validation, resulting in uncertainties in load predictions. Moreover, thick mid‑span airfoils experience large unsteady angles of attack and are more susceptible to early stall, necessitating active or passive flow‑control strategies. Flatback airfoils and vortex generators have emerged as promising solutions, offering increased lift, delayed flow separation, and improved stall characteristics. Nevertheless, their combined impact on unsteady aerodynamics is not yet fully understood, revealing a critical gap in current aeroelastic modeling capabilities.

This knowledge gap compromises the reliability of current aeroelastic models and poses risks for the design, optimization, and structural assessment of large modern turbines. Consequently, this dissertation is motivated by the need to advance the understanding and modeling of dynamic stall phenomena in thick wind turbine airfoils, with particular emphasis on the combined effects of vortex generators under unsteady flow conditions. By addressing these gaps, the work aims to contribute improved physical insight and enhanced predictive capability for the aerodynamic behavior of next generation wind turbine blades.

Therefore, several wind tunnel campaigns were conducted to investigate the unsteady aerodynamic characteristics of two $35\%$ thick airfoil configurations, comprising a conventional airfoil and a flatback airfoil, tested in different setups in Low-speed, low-Turbulence wind Tunnel (LTT) of the Delft University of Technology.

The results show that flatback airfoils exhibit inherently more stable hysteresis loops, as evidenced by reduced force fluctuations, larger lift overshoots, and delayed separation compared to sharp trailing edge airfoils, primarily due to their more favorable suction side pressure gradients under the studied operating conditions. Dynamic stall on these thick airfoils is governed not by leading edge vortex formation, as seen in thin airfoils, but by a gradual trailing edge separation mechanism that produces moderated load transients. Reduced frequency strongly influences dynamic stall, i.e, higher frequencies delay separation, enhance lift overshoot, and stabilize flow evolution for both airfoils. The addition of vortex generators substantially improves flow attachment during the upstroke phase, increases lift overshoot, and suppresses force fluctuations, particularly when placed near the leading edge. Their effectiveness, however, depends sensitively on chordwise placement, the underlying airfoil geometry, and the oscillation regime. While flatback airfoils generally benefit from improved stall delay, they experience larger force excursions during fully separated flow, indicating both the potential and the risks of employing blunt trailing edge designs. The evaluation of the Beddoes–Leishman and Risø dynamic stall models demonstrates that both models are not capable to predict separated flow behavior on thick airfoils, especially regarding lift overshoot, lift drop, post-stall loads, drag, and pitching moment. Lack of modeling accuracy lead to systematic discrepancies in model predictions, even when tuning time constants. These findings highlight the need for new or revised separation and reattachment modules tailored specifically for thick airfoils for reliable load prediction in the inboard and mid-span regions of large modern wind turbine blades.