LES Model Comparison in AMR Turbulence Simulations

Large-eddy simulation (LES) is a powerful technique for modeling turbulent flows, particularly in complex geometries. Adaptive mesh refinement (AMR) further enhances LES by dynamically adjusting the grid resolution to focus computational resources where needed most, typically in regions of high turbulent activity. This allows for efficient simulation of a wider range of scales. However, the choice of LES model significantly influences the accuracy and efficiency of the simulation. This article explores a comparison of several popular LES models within the context of AMR turbulence simulations.

One key aspect to consider is the subgrid-scale (SGS) model, which parameterizes the effects of unresolved small scales on the resolved larger scales. Common choices include the dynamic Smagorinsky model, the scale-similarity model, and various variations of these. The selection of an appropriate SGS model depends on the specific application and flow characteristics.

For example, in simulating the atmospheric boundary layer Turbulence Modeling in Atmospheric Boundary Layers, a model capable of accurately representing the near-wall region's complexities might be favoured, while a different approach might be best for simulations of free shear flows. In complex scenarios involving shock interactions, special considerations may apply. See this great discussion on Shock-Turbulence Interactions.

AMR itself presents additional challenges and opportunities. The dynamic refinement strategy – how and when the mesh is refined – must be carefully considered, since its impact will strongly influence simulation cost and accuracy. An overly aggressive refinement may introduce numerical artifacts while inefficient refinement reduces gains. Careful consideration for balancing the SGS model parameters in conjunction with AMR methods needs attention. For a deep-dive into these trade-offs, refer to the Optimization of AMR Parameters for LES article.

This comparison across LES models, using AMR, aims to investigate some key numerical attributes that differ in each numerical technique. Many of these effects become prominent during more complicated applications such as Multiphase Flows.

Choosing the right model requires careful consideration. A comparative analysis, exploring different cases in an organized, methodical approach will bring substantial advancement.

Further reading: More on AMR methods