The accuracy of large-eddy simulation (LES) of stratified flows is significantly influenced by sub-grid scale (SGS) stress and scalar flux models. In this study, two-dimensional high-resolution velocity and scalar (density) data (simultaneously obtained using a combined Particle Image Velocimetry and Planar Laser Induced Fluorescence technique) in a horizontal turbulent stratified jet are used to examine the SGS parameters and the performance of SGS models. The profiles of SGS dissipation of kinetic energy indicate that the flow has more capability to sustain its structure in the stable region (upper mixing layer) of stratified jet. The backscatter is observed from the components of the SGS dissipation of kinetic energy and SGS dissipation of scalar variance in the stable stratification region of jet in the high-Ri case. The SGS dissipation of kinetic energy and of scalar variance are shown strong dependence on the stability status of local flow field, which experience the ascending and descending as the stability parameter increases. In the SGS model tests, the scale-invariant dynamic model shows better performance of predicting Cs2 than classic Smagorinsky model and scale-dependent dynamic model. From the current study, the SGS turbulent Prandtl number is suggested as constant (e.g., Pr ≃ 0.46) to a achieve a good simulation of scalar field in engineering applications to economize the computational cost.

This content is only available via PDF.
You do not currently have access to this content.