Wall-model integrated computational framework for large-eddy simulations of wall-bounded flows

The wall-modeled large-eddy simulation (WMLES) computational framework generally includes a wall-model solver outside the large-eddy simulation (LES) infrastructure, with the two solvers communicating only at the matching location and the wall. Having a wall-model solver outside the LES jeopardizes the performance of WMLES: first, the wall-model solver adds significant computational overhead; second, the LES solution in the wall-adjacent cell is ambiguous; and third, it is very difficult to utilize the emerging high-order numerical schemes. This paper addresses the above issues by abandoning wall-model solvers altogether and integrating wall models into LES solvers. We will employ a set of physics-inspired bases for LES solution reconstruction in the wall-adjacent cell. The methodology gives rise to a computational framework that effortlessly accounts for non-equilibrium effects in a high-order code without a stand-alone wall-model solver. We consider channel flow for a proof of concept and periodic hill for validation.

Automated summary

The paper addresses performance issues of wall-modeled large-eddy simulation (WMLES) by integrating wall models into LES solvers, abandoning stand-alone wall-model solvers. By employing physics-inspired bases for LES solution reconstruction in the wall-adjacent cell, the computational framework effortlessly accounts for non-equilibrium effects in high-order codes, with channel flow for proof of concept and periodic hill for validation.