A unified wall treatment for seamless PANS simulations across all resolution levels

This paper investigates a scalable wall treatment for Partially-Averaged Navier–Stokes (PANS) simulations, with emphasis on its applicability across all resolution levels. While PANS enables a seamless transition from Reynolds-averaged (RANS) to scale-resolving regimes, its predictive accuracy may be sensitive to the consistency of near-wall modeling. The proposed hybrid wall model, which has previously been validated for the RANS k-ε-ζ-f model, provides a single, unified formulation that is valid for any value of the resolution parameter f k , representing the ratio of unresolved to total turbulent kinetic energy. In contrast to conventional approaches, it eliminates the need for switching between wall functions and near-wall integration, as well as case-dependent tuning. In addition, the influence of variable values of f k is examined with respect to near-wall behavior. The model is assessed for turbulent channel flow and flow over periodic hills at Reynolds number 10,595. The results demonstrate that the wall treatment adapts consistently to varying resolution, ensuring stable and physically sound predictions in both RANS-like and scale-resolving regimes. For the benchmark cases considered, the results indicate limited sensitivity to the specific f k formulation when a physically sound parent RANS model such as k-ε-ζ-f is employed. Overall, the proposed approach enables PANS to function as a truly seamless bridging method, providing a robust and resolution-independent wall modeling strategy for practical engineering applications.