High-performance computing for the investigation of the flow past an airfoil with trailing-edge stall

Abstract

The paper is concerned with the prediction and analysis of the turbulent flow past an unswept NACA-4415 airfoil at high angle of attack, for which largeeddy simulations (LES) were carried out. In order to resolve as much as possible of the turbulent scales in the flow, and thus to allow for a detailed investigation of the flow phenomena, high-performance computers such as the SMP cluster Hitachi SR8000-F1 used are inevitable. The Reynolds number investigated is Rec = 105 based on the chord length c of the airfoil. Admittedly, this Reynolds number is considerably lower than under real flow conditions. However, in combination with the inclination angle of α = 18° chosen, the present case yields a trailing-edge stall which is commonly observed in the real flow at higher Rec and slightly lower angles of attack. Hence, the strategy behind the present investigation enables us to analyze some real world phenomena occurring at the flow past highly inclined airfoils based on well-resolved LES results. These phenomena include the formation of a thin separation bubble, transition to turbulence in a realistic scenario, separation of the turbulent boundary layer, and large-scale vortical structures in the wake. Applying a grid with approximately 23.56 million control volumes, the flow is very well resolved, especially in the interesting near-wall region of the airfoil. Therefore, the processes within the thin separation bubble are captured, illuminating the instantaneous features of the flow in this restricted area. For the purpose of validation, the numerical results are compared with corresponding experimental data obtained from measurements at DLR Göttingen. Since in this case transition to turbulence occurs naturally, i.e., without any forcing from outside, the uninfluenced mechanisms of transition are covered in the simulation results. Hence, the most contributing components of the velocity fluctuations to the generation of turbulence are clearly identified.