Key points are not available for this paper at this time.
Ribbed channel flows play a crucial role in various engineering systems where turbulence enhancement and improved heat transfer are required, such as in turbine blade cooling and combustor wall applications. Numerical simulations have emerged as an essential tool for analyzing the intricate turbulence dynamics and flow separation phenomena, which are fundamental for optimizing the channel performance. This work presents a comparative analysis of turbulence modeling approaches, specifically Reynolds-Averaged Navier–Stokes (RANS) and Large Eddy Simulations (LES), using OpenFOAM across Reynolds numbers of 4000, 8000, 12000, 16000, 22000, and 24000. Multiple turbulence models, subgrid-scale models, and mesh resolutions are examined to assess their influence on the accuracy of flow and thermal transfer predictions. The numerical results, particularly in terms of turbulence characterization and its effect on thermal performance, are validated against the experimental data of Wang (2007). The mesh sizes vary between 1 and 25 million cells, capturing around 90% of the turbulent energy, suggesting that the LES meshes provide sufficient resolution. Overall, LES results exhibit stronger agreement with experimental observations compared to RANS predictions, with WALE and LDKM subgrid models demonstrating superior performance relative to SMG and OEEVM. Additionally, to deepen the understanding of turbulence mechanisms governing flow separation, reattachment, and eventually redevelopment, we present extensive analyses of flow parameters, e.g., mean velocity, friction coefficient and Reynolds shear stress. The anisotropic characteristics of turbulence at all scales are examined using anisotropic invariant maps, revealing substantial variations in anisotropy across different near-wall regions between consecutive ribs.
Garg et al. (Mon,) studied this question.