A General Method for Large Eddy Simulation in Different Range of Turbulent Flows

TL;DR

This paper introduces a general constraint-based approach for large eddy simulation that adapts the model coefficient across different turbulence scales, improving prediction accuracy in non-inertial ranges.

Contribution

It proposes a new principle of constraint and develops a constrained dynamic Smogorinsky model for better energy spectrum prediction across various turbulent flow regimes.

Findings

Constrained models outperform classical models in energy spectrum prediction.

The new models show strong correlation with real stress in different turbulence ranges.

The approach effectively handles non-inertial range turbulence simulations.

Abstract

Traditional large eddy simulation is based on Kolmogrov's hypothesis, and done in the inertial range. In inertial range the LES model coefficient is scale-invariant. In many cases, such as computing in the boundary layer, the filter scale is not in the inertial range, and the model coefficient changes drastically with the filter scale, thus the assumption of scale-invariance is not proper. In this paper, we propose a general principle of constraint, and then we deduce two concrete constraints based on the existing spectrum theory and priori data. The constrained dynamic Smogorinsky model can predict energy spectrum well in different range of forced and decayed turbulence. Dynamic mixed models with one of the constraints shows more efficient than some classical models in predicting energy spectrum and SGS dissipation, and also displays a strong correlation with the real stress in different range of the turbulent flows.