Center for Turbulence Research Annual Research Briefs 2014 193 Componentality-based wall-blocking for RANS models By M. Emory AND G. Iaccarino 1. Motivation and objectives Although there has been increased adoption of large-eddy simulation (LES) in aca- demic and industrial applications, the affordability of the Reynolds-averaged Navier- Stokes (RANS) approach ensures it will continue to be one of the most popular methods for solving these complex flows (O’Sullivan et al. 2012). There are still many shortcom- ings with most common RANS closures (i.e., turbulence models), for example difficulties due to both the physical assumptions and the numerical stiffness, or the inability to re- produce turbulence near solid walls. Proper prediction of the near-wall region is critical for accurately capturing many flow characteristics of engineering interest, e.g., friction, heat transfer, or separation. The presence of a solid wall influences turbulent flow primarily through two mecha- nisms. The first is through viscous effects which require that the velocity in all directions to be zero at the wall. The second mechanism, known as the blocking effect, is due to the impermeability of the solid boundary. This effect suppresses fluctuations primarily in the wall normal direction, creating highly anisotropic turbulence structures in the near-wall region (Manceau & Hanjali´ c 2002). The blocking effect is sensitive to the topology of the wall and is further complicated by the reflection of pressure fluctuations off the wall, which can reduce the turbulence anisotropy. This anisotropy is often not captured or only poorly represented by common turbulence models. Incorporating the wall-blocking effect in RANS closures has been approached in several ways, the most popular of which is through empirical damping functions, essentially a correction to the eddy-viscosity, which fits the near-wall turbulence behavior to either theory or direct numerical simulation (DNS). These corrections often suffer from a lack of physical justification and poor performance in flows with complex geometry (Billard 2007). A different approach was taken by Durbin (1991) where two additional equations are added to a standard k−ǫ model. The first is a transport equation for v 2 , representative of the wall-normal velocity fluctuations, a term which damps the eddy-viscosity near walls. The second is an elliptic equation which describes the generation of v 2 due to pressure redistribution. While this approach has shown good results in several flows, there are numerical stiffness issues related to the coupling of the additional elliptic and transport equations. This elliptic relaxation approach has also been adapted for complex RANS modeling frameworks. The algebraic structure based model (ASBM) introduces a blockage tensor, based on a similar parameter to v 2 , to include proper wall-blocking (O’Sullivan et al. 2012) while Durbin (1993) and Manceau & Hanjali´ c (2002) have extended the approach to Reynolds-stress transport models. The complexity of these RANS approaches, however, limits adoption of these models by industry. Capturing wall-blocking effects is an important challenge at all levels of RANS model