On the use of (U)RANS and LES approaches for turbulent incompressible single phase flows in nuclear engineering applications

Abstract

The present paper gives some ideas and guidelines in order to run high quality (U)RANS or LES computations. The paper starts with (U)RANS approaches, advocating the use of Reynolds Stress Models for complex flows and recommending further work on modeling of turbulent heat fluxes, which remains today too basic in industry. The superiority of wall-resolved models vs. wall-modeled in RANS is recalled and the use of adaptive wall treatment is suggested. The concept of Unsteady RANS is finally questioned.Then, important issues around LES are raised. The mesh refinement criteria are recalled for wall-resolved LES and the use of wall models addressed. The production of DNS and wall-resolved LES calculations for flow understanding and RANS validation is encouraged.