Multiphysics analysis of RANS-based turbulent transport of solid fission products in the Molten Salt Fast Reactor

The analysis of innovative reactor concepts such as the Molten Salt Fast Reactor (MSFR) requires the development of new modeling and simulation tools. In the case of the MSFR, the strong intrinsic coupling between thermal-hydraulics, neutronics and fuel chemistry has lead to the adoption of the multiphysics approach as a state-of-the-art paradigm. One of the peculiar aspects of liquid-fuel reactors such as the MSFR is the mobility of fission products (FPs) in the reactor circuit. Some FP species appear in form of solid precipitates carried by the fuel flow and can deposit on reactor boundaries (e.g., heat exchangers), potentially representing design issues related to the degradation of heat exchange performance or radioactive hotspots. Other precipitates might be present in the primary system as well, e.g. due to oxidation of fissile species which might lead to local criticality issues. The integration of transport models for solid particles in multiphysics codes is therefore relevant for the prediction of deposited fractions. To this aim, a previously developed Eulerian single-phase transport model is employed to analyze the distribution of solid FPs in two simplified MSFR-related geometries. The effect of physical parameters and of distributed particle sources on the numerical requirements needed to resolve particle concentration gradients at reactor boundaries in the considered geometries is investigated with the aid of analytical results. Analytical estimates of concentration gradients, even though not in full agreement with simulations, prove useful to drive the choice of adequate mesh refinements. Furthermore, the influence of different RANS turbulence modeling approaches on the prediction of particle distributions and deposition is tested. Results show a limited influence of the choice of turbulence models and parameters on the deposited fraction and on concentration gradients. As a result, it is found that deposition rates are scarcely affected by the choice of turbulent Schmidt number, with lower diffusivities being compensated by larger gradients. Large deposited fractions are indeed predicted, suggesting the efficiency of FPs transport mechanisms in the reactor and the need for the integration of adequate FPs transport models within state-of-the-art MSFR codes.