RANS Analysis of Turbulent Natural Convection in BALI Configuration for IVR-ERVC SAM Strategy

Abstract

During severe accident, the relocation of molten core to the vessel lower head would progress to release decay heat that can threaten the vessel integrity if no effective cooling mechanism is provided. Thus, to mitigate the progression, the in-vessel retention by external reactor vessel cooling (IVR-ERVC) is a strategy that has been adopted to low-to-intermediate power reactors, but many uncertainties can exist if applied to high-power reactors. To take this into account, the total thermal loading on the vessel of these reactors must be evaluated through the analysis of the turbulent natural convection occurring in the molten pool. The main idea was to numerically model the turbulent natural convection of oxide pool using CFD models simulated at high internal Rayleigh number. These turbulence models include: RNG k-ε, SST, v2f, and modified v2f-AFM. The modified v2f-AFM solves the turbulent heat flux using Algebraic Heat Flux Model. In this study, the numerical analysis is performed in configurations such as Differential-Heated Vertical Cavity and BALI homogenous test using OpenFOAM 2.3.1, which is an open-source CFD software. Mesh sensitivity results in both two-dimensional geometrical configurations showed that grid-independent solutions were achieved at y+ ≈ 1.0. The four-equation standard v2f model showed satisfactory performance from the near-wall modeling of DHC and from the numerical profiles of BALI test case. Upon modifying it to v2f-AFM, as well as improving it, there is an improvement in the complex heat transfer phenomena, particularly in the flow separation from unstable homogenous to stable stratified zones of oxide pool. The coefficients, from the sensitivity analysis, alongside the use of buoyancy production term, paved the way to yield around 10% error, with respect to the power-split ratio.