Assessment of advanced RANS models for sodium in the SUPERCAVNA experiment

Thermal stratification in a sodium fast reactor (SFR) leads to an accumulation of hot sodium at the top of some of the reactor structures and alters heat exchanges. The modification of these heat exchanges could lead to significant thermal gradients on the metallic structures that are harmful to the conservation of the integrity of the mechanical properties of these elements. Hence, thermal stratification of the flow should be minimized or avoided for safety reasons. 

SUPERCAVNA was an important experimental setup at CEA Grenoble to study thermal stratification in sodium. In the 1980s, several interesting experiments have been performed, which are currently being used for the validation of Computational Fluid Dynamic (CFD) simulations. Whereas state-of-the-art Reynolds-Averaged Navier-Stokes (RANS) simulations give relatively accurate results for steady-state simulations of SUPERCAVNA [1], there are significant discrepancies for the transient behavior in highly-stratified conditions [2]. One of the particular problems of the use of a RANS model for sodium is the fact that Reynolds analogy is violated due to a strong decorrelation between the instantaneous velocity and temperature fields [3]. To this end, several advanced RANS approaches have been developed for turbulent heat flux closure models in the framework of multiple European projects. The von Karman Institute for Fluid Dynamics has implemented several advanced RANS models in the OpenFOAM software.

In this thesis, the student will compare OpenFOAM simulation results with different RANS approaches for the SUPERCAVNA setup to the experimental data and previous CFD simulations obtained with the TrioCFD code [1] and Code_Saturne [2]. One student will focus on the assessment of the different RANS closures, while the second student will consider accurate time-dependent simulations. The main research question is if improving the thermal turbulence model has a significant effect on reducing simulation-experiment discrepancies. By varying the Richardson number, the accuracy of the models can be validated for forced and mixed convection regimes. The validation of RANS models in presence of buoyancy is an important expected outcome, as the state of the art of mixed and natural convection RANS validation for fluids with low Prandtl number is currently limited.

References

[1] A. Genty, C. Roy, C. Geffray, “Numerical simulation of steady-state mixed convection sodium flow experiments”, Nuclear Engineering and Design 381, art. nr. 111363, 2021.

[2] C. Geffray, A. Genty, “Numerical simulation of transient forced and mixed convection sodium flow experiments”, International Journal of Heat and Mass Transfer 251, art. nr. 127277, 2025.

[3] M. Duponcheel, L. Bricteux, M. Manconi, G. Winckelmans, Y. Bartosiewicz, “Assessment of RANS and improved near-wall modeling for forced convection at low Prandtl numbers based on LES up to Re(pi)=2000”, International Journal of Heat and Mass Transfer 75, pp. 470-482, 2014.

The minimum diploma level of the candidate needs to be

• Master of sciences in engineering

The candidate needs to have a background in

• Physics
• It is recommended that the student(s) obtained a master with sufficient focus on fluid mechanics and heat transfer. CFD experience is highly recommended, preferentially with OpenFOAM.

Estimated duration