Despite the ever-increasing capacity of the modern high-performance computing platforms, wall-resolved large-eddy simulation (LES) of practical wall-bounded turbulent flows is still deemed infeasible, owing to the prohibitive grid resolution requirement for resolving the stress-carrying small near-wall eddies. As a result, such rigor is pursued for relatively simple flows, while the vast majority of industrial computational fluid dynamics analyses still rely on cheaper low-fidelity techniques. To overcome this key technology barrier in LES, wall-modeled LES (WMLES) attempts to provide a predictive but at the same time affordable framework of LES, by modeling the computationally demanding inner-layer physics. The grids employed in WMLES are designed to resolve the outer layer only, thereby easily reducing the computational cost to affordable levels.
The purpose of this talk is to convey and expose the basic concepts and methodologies of wall modeling to the general audience in the area of computational flow physics. This talk intends to introduce the basic philosophy of wall modeling in LES, and to walk the audience through the state-of-the-art wall modeling techniques. The wall models to be discussed include the traditional models based on the RANS legacy, as well as a newly emerging paradigm. Practical aspects such as the cost of wall modeling, limitations of some models, accuracy and grid convergence considerations will be discussed. Lastly, open challenges and questions that need to be addressed in the upcoming years to render the method of greater utility will be discussed.