Comparison of RANS Turbulence Models for the Simulation of Smooth Wall Boundary-Layers in Pressure Gradients at Moderate and High Reynolds Numbers

In this work, our focus is to assess the effect of the Reynolds number on the performance of six different RANS turbulence models in the simulation of viscous flows. To this end, we have selected the experiments performed in a wind tunnel at Virginia Tech that measured boundary-layers in favorable and adverse pressure gradients. A rectangular wing with a NACA 0012 section placed in the middle of the wind tunnel at angles of attack ranging between -10 and 12 degrees is used to induce the pressure gradient.
The two-dimensional computational domain proposed for the comparison with the experimental data is used to simulate the ows at Reynolds numbers of Re = 2 x10⁶ (model scale, applicable to the wind tunnel tests) and Re = 10⁹ (representative for full scale conditions) with the wing at angles of attack of -10, 0 and 12 degrees. Calculations are performed with a RANS solver using five eddy-viscosity models and a Reynolds-stress model. Differences between the solutions obtained with the six turbulence models are quantified for several quantities of interest including force coefficients, shear-stress and pressure coefficients, mean velocities, turbulence kinetic energy and Reynolds stresses using point-wise and global Validation metrics.
For the present test cases, discrepancies between solutions of the six selected turbulence models at model scale Reynolds number are significally larger than those observed at full scale Reynolds number. Therefore, modeling error assessments (validation) performed at model scale Reynolds number cannot be extrapolated to full scale.