Financial support : Region Auvergne Rhône Alpes (FRI Transfert)
Collaboration : P. Benard (CORIA, Rouen)
The aim of this work is to implement more recent and efficient near wall modeling methods in the computational code YALES2. This is done with the long-term objective of providing better simulation tools to support the simulations of the flow in complex geometries. Several wall modeling strategies may be distinguished. Traditional wall models (already implemented in YALES2) attempt to provide the wall shear stress by using an algebraic law of the wall. Important progress have since been made in the framework of hybrid RANS/LES methods: the LES is used to solve the equations of the fluid motion in regions away from the wall while the RANS methods are used in the near wall region. The main drawback of those methods is that they depend upon empirical coefficients in the turbulence (RANS) models, or assumptions about the local state of the boundary layer and the logarithmic velocity profile. Therefore, we chose to focus on a recent method proposed by Bose & Moin[PoF 26, 2014] that is free from a priori assumptions ; the approach is called dynamic slip wall. Briefly, it consists in applying a slip velocity at the wall that is directly computed in the framework of the filtering operation of the Navier-Stokes equations. We implemented this method in the YALES2 code structure and assessed it with simulations of a turbulent channel flow at a friction Reynolds number equal to 2000 and of an airfoil (FFA-W3-241 profile) with different angles of attack. In a general point a view, our results show a satisfying efficiency. In the figure 1, we can see the comparison of the mean velocity profile in the turbulent channel flow between the dynamic slip wall condition, the no-slip condition and the results from DNS. The use of the dynamic slip wall condition avoid to overestimate the mean velocity profile in contrast with the no-slip condition. The figure 2 shows the lift coefficient of the airfoil predicted by the dynamic slip wall condition, the no-slip condition and the standard law of the wall compared with experimental results. In the detached case (high angle of attack), our model succeeds to capture the boundary layer detachment, leading to a better prediction of the lift coefficient. Our future work will be to evaluate the sensitivity and the limits of our implementation of the dynamic slip wall boundary condition.
