Verification and validation plus uncertainty quantification of heat transfer simulation for liquid metal in wire-wrapped rod assembly

Abstract

A complete CFD best practice process is proposed, including Phenomena Identification and Ranking Table (PIRT), Verification and Validation (V&V), and Uncertainty Quantification (UQ). The heat transfer of liquid metal within a rod assembly under both normal and blocked conditions is used as an example to demonstrate the full CFD process. Key flow and heat transfer phenomena are identified, and model sensitivity analyses are performed based on these phenomena. Turbulent Prandtl models, porous blockage models and different meshing strategies sensitivity analysis are conducted. Uncertainty quantification is performed. The discretization uncertainty is calculated using the Grid Convergence Index (GCI). The convergence orders are greater than unity for all cases except the fluid temperature in the blocked condition using meshing strategy 2. The input parameter uncertainty is estimated using finite difference. The model uncertainty is derived. Under normal condition, the model uncertainty of the wall and fluid temperatures are 2.88 K and 3.37 K, respectively. Under blocked condition, the model uncertainty of wall temperature is slightly higher than fluid temperature. The wall temperature uncertainty for strategies 1 and 2 are 8.46 K and 9.03 K, while the fluid uncertainty of strategy 1 and 2 are 3.57 K and 3.06 K, respectively.