A novel dynamic subgrid-scale model based on partial-averaged velocity for large eddy simulation

Abstract

This study presents an innovative dynamic subgrid-scale model for large eddy simulation based on the partial-averaging method. In this method, the partial-averaged velocity fluctuations, known as drift velocities, were used to formulate subgrid-scale turbulent viscosity. This feature enables the preservation of more comprehensive first-order turbulence information so that the anisotropic turbulent flow can be better predicted. Dynamic coefficient adjustment was adopted to effectively address the sharp change of subgrid-scale viscosity in complex flows. The proposed model was programed and employed to simulate a plane jet, a circular jet, a fully developed pipe flow, and a swirling jet based on OpenFOAM. The results indicate that the model exhibits high predictive accuracy, adapting well to free shear flows and wall-bounded shear flows. The distribution of drift velocities provides direct statistical information of turbulent flows and reveals anisotropic characteristics of turbulent fluctuations.