Numerical Analysis of the Performance of a Dynamic Global-Coefficient Subgrid-scale Model for Buoyancy-driven Flow in a Differentially Heated Cavity

During the severe accident of a nuclear power plant, the molten fuel is relocated in the lower head of the reactor pressure vessel. It is important to understand the thermal behavior of the molten pool to assess the integrity of the vessel. Turbulent natural convection is a major phenomenon that determines thermal behaviors. A limited investigation of the thermal behavior of the molten pool has been conducted owing to the experimental difficulties. Recently, a numerical approach has been used with a high-fidelity method such as large eddy simulation (LES). In this study, representative LES subgrid-scale (SGS) models, such as the Smagorinsky model (SM), dynamic Smagorinsky model (DSM), Vreman model (VM), and dynamic global-coefficient model (DVM) were selected to investigate their performance in simulating a turbulent natural convective flow. A comparative study of the LES SGS models was conducted in a turbulent natural convective flow (Ra similar to 10(10)).