A Dynamic Subgrid-Scale Model Based on Liutex Theory for Wall-Bounded Turbulent Flows

TL;DR

This paper introduces a novel subgrid-scale model for large eddy simulation of wall-bounded turbulence, utilizing Liutex theory to improve near-wall predictions without empirical tuning.

Contribution

A new SGS model based on Liutex theory with a dynamic, physics-based length scale and coefficient that adapts locally and vanishes at walls, validated for turbulent channel flow.

Findings

Enhanced accuracy in near-wall Reynolds stress predictions

Model adapts dynamically without empirical parameters

Computational overhead is minimal at 4.21%

Abstract

Accurate subgrid-scale (SGS) modeling remains a major challenge in large eddy simulation (LES), particularly for wall-bounded turbulent flows with strong near-wall anisotropy. This study proposes a novel SGS model based on Liutex theory, featuring a dynamically adaptive model coefficient and a physics-based length-scale formulation. The magnitude of the Liutex vector is employed as the characteristic velocity scale, enabling a direct and objective quantification of local vortical intensity. The length scale is determined from local flow properties and reflects the physical nature of turbulent diffusion, which occurs predominantly in directions perpendicular to the rotation axis. The dynamic model coefficient adapts locally to variations in vortical structures and naturally vanishes at the wall. Importantly, this coefficient is free of empirical tuning, as it is derived rigorously from the triple decomposition. The model is validated through LES of turbulent channel flow at the friction Reynolds number of 544. Results show improved predictions of both normal and shear Reynolds stresses, particularly in near-wall regions, compared with other SGS models. The proposed model demonstrates competitive computational efficiency, with only a 4.21% overhead.