Effect of wall stress models and subgrid-scale models for flow past a cylinder at Reynolds number 3900

Abstract

The wall-modeled large eddy simulation has achieved some success in simulating wall-bounded flows. However, their predictive accuracy in separated flow still requires further validation. In this study, we employ the open-source computational fluid dynamics toolbox OpenFOAM to numerically investigate the flow past a cylinder at subcritical Reynolds numbers ReD=3900. At two different sampling heights, h = 2nd and h = 4th, three wall stress models: the algebraic model, the equilibrium wall model (EQWM), and the non-EQWM (NEQWM), and five subgrid-scale (SGS) models: the Smagorinsky (SMAG) model, the k-equation SGS (KSGS) model, the wall-adapting local eddy viscosity (WALE) model, the dynamic SMAG (DSMAG) model, and the dynamic KSGS (DKSGS) model, are selected for comparative study. Various physical quantities, including statistical flow quantities, wall pressures, time-averaged wake velocity profiles, and Reynolds stresses, are extracted and compared with the experimental data. Power spectral analyses for wake velocity are conducted, and the three-dimensional vortex structures are illustrated. The results indicate that for small sampling height, all wall models yield favorable numerical simulation results. However, for larger sampling height, the NEQWM is preferred over the other two wall models. In terms of SGS models, the DKSGS model and WALE model perform better than other SGS models. The SMAG and KSGS models, due to inherent model limitations, struggle to accurately predict the flow separation angle and the Reynolds stresses in the free shear layer.