A Near-Wall Methodology for Large-Eddy Simulation Based on Dynamic Hybrid RANS-LES

Area of Science:

• Computational Fluid Dynamics
• Turbulence Modeling
• Fluid Mechanics

Background:

• Large-eddy simulation (LES) is computationally expensive in near-wall regions.
• Existing methods like hybrid RANS/LES (HRL) and wall-modeled LES (WMLES) have limitations.
• Accurate near-wall modeling is crucial for resolving turbulent flows.

Purpose of the Study:

• To develop a novel LES wall treatment method combining strengths of HRL and WMLES.
• To reduce computational cost while maintaining accuracy in near-wall turbulence simulation.
• To present a meshing strategy and evaluate performance for turbulent channel flow.

Main Methods:

• A hybrid approach using WMLES-like streamwise/spanwise mesh and RANS-like wall-normal mesh.
• Implementation of a mixing length model for eddy viscosity, limited by local mesh size.
• Smooth blending of RANS and LES regions using the dynamic hybrid RANS-LES (DHRL) framework.

Main Results:

• The proposed method reasonably predicts mean and fluctuating quantities for turbulent channel flow.
• No log-layer mismatch was observed when compared to direct numerical simulation (DNS) data.
• The method requires approximately twice the grid points of traditional WMLES but avoids wall function complexities.

Conclusions:

• The developed LES wall treatment offers a viable alternative for near-wall turbulence simulation.
• It balances computational cost and predictive accuracy effectively.
• The approach simplifies implementation by avoiding analytical or numerical wall functions.