Experimental Validation of CFD Models Capturing the Thermal-Hydraulics in Liquid Metal Cooled Reactor Plena

Abstract

High fidelity velocity field experimental data in a liquid metal plenum is presented and compared with numerical simulations. While work has already been established for fluids like air and water, research on low Pr fluids (Pr ≪ 1) (e.g. liquid metals) has fewer experimental data sets with validation-quality data. Work in advanced reactors using liquid metal coolant requires validated numerical simulations for safety analyses. The Gallium Thermal-hydraulic Experiment (GaTE) facility is outfitted with acoustic backscattering measurement techniques to generate the high fidelity distributed flow field data in a liquid metal plenum (a 1/20th scale of the Department of Energy’s sodium cooled Advanced Burner Test Reactor design). The high spatial and temporal resolution of the sensors are required to capture the fluctuations of velocity to allow a more direct comparison to the numerical simulations. For these simulations the coupled mass and momentum equations under the large eddy simulation (LES) framework were solved with the wall-adapting local eddy-viscosity (WALE) model for sub-grid scale formulations. Since the temperature transients of interest for reactor safety have a period of about a minute in the GaTE system, there may not be enough time to allow statistical tools to check one-to-one correspondence. So the data collection period for both data sets was extended to allow convergence of the mean and a larger sample size for other statistics during system steady-state, isothermal tests. Two characteristic velocities of the plenum inlet barrel were investigated (U = 40, 60 mm/s; Re = 7,000, 11,000). Probability distributions show good agreement between experiment and simulation with the difference only in the low-probability tails that LES is not expected to simulate. The time averaged mean axial distribution of the vertical velocity also shows good agreement between the two setups.