Wall-resolved large eddy simulations of a pitching NACA0012 incurring in deep dynamic stall

Giacomo Baldan,Alberto Guardone
2024-05-20
Abstract:This study investigates the flow evolution past a pitching NACA0012 airfoil undergoing deep dynamic stall using a wall-resolved large eddy simulation approach. Numerical results are validated against experimental data from Lee and Gerontakos at Reynolds number Re = 135 000 and a reduced frequency k = 0.1. A comprehensive analysis of the computational grid span size is presented, highlighting the critical requirement for a span-to-chord ratio of at least one. Furthermore, a comparative assessment with state-of-the-art RANS and hybrid RANS/LES turbulence models is conducted to identify the sources of potential discrepancies between the numerical simulations and the experimental data. All the numerical models demonstrate the same behavior and exhibit similar differences with the experiments.
Fluid Dynamics
What problem does this paper attempt to address?
The paper mainly discusses the Large Eddy Simulation (LES) study of the NACA0012 airfoil during the unsteady stall condition when it is flipped. The unsteady stall is an instability phenomenon that occurs when the angle of attack of an aircraft changes rapidly. Unlike static stall, it involves the destruction of the boundary layer and the formation of dynamic stall vortex (DSV). The study validates the experimental data through numerical simulation and analyzes the influence of different mesh sizes on the results. In the paper, the researchers analyze the influence of different span-to-chord ratios on the flow evolution and point out that a span-to-chord ratio of at least 1 is required for accurate simulation of DSV. They also compare high-accuracy Reynolds-Averaged Navier-Stokes (RANS) and hybrid RANS/LES turbulence models to determine the source of discrepancies between numerical simulation and experimental data. All models exhibit similar behavior but show some differences compared to the experiments. Furthermore, the study highlights the importance of dynamic stall in fields such as helicopter rotor, wind turbine, and high-performance aircraft maneuvering, and introduces the five main stages of dynamic stall. The paper uses a wall-resolved large eddy simulation method with a span-to-chord ratio of 1.2c for numerical simulation and compares it with experimental data from Lee and Gerontakos (2004). The results show that despite some differences in certain aspects, all models exhibit similar dynamic behavior. Lastly, the paper discusses the influence of mesh size on dynamic stall phenomena such as dynamic stall vortex generation and interaction with the trailing-edge vortex, as well as a comparison with RANS and hybrid RANS/LES methods. Although there are deviations between numerical simulation and experimental data in certain angles, the overall trends are consistent with the experiments. Through these studies, the authors aim to enhance the understanding of dynamic stall and subsequently optimize aircraft design and control strategies.