Modelling dissolution phenomena during the early phase of a serious accident of nuclear reactor

This thesis focuses on the modeling of dissolution phenomena during the early phase of a serious accident of nuclear reactor. The study focuses on the dissolution of solid ceramic (ZrO2 and UO2) by a liquid metal (Zr). In case of hypothetical severe accident in a pressurized water reactor, the phenomena of dissolution play a role in the further decay and melting oxides at temperatures sometimes significantly below the normal melting temperature. This concerns in particular the main constituents of fuel rods: pellets of UO2 and ZrO2 layer formed on the outer surface of the sheath that will suffer by the dissolution of zircaloy sheaths from 2100 K (the melting temperature of these oxides being greater than 2800 K). In the present state of knowledge, we can assume that dissolution phenomena are responsible for, on the one hand, the premature collapse of the fuel rods in the heart and, on the other hand, the rapid formation of a liquid bath in the bottom of the tank where iron oxides are present. Many experimental studies have been conducted on this subject but modeling is not yet satisfactory today. Current models are essentially 1D models that do not explicitly account for natural convection nor the presence of a biphasic transition zone at the interface metal / ceramic. A 2D model describing the transfer of momentum, heat and mass, determined by taking average density of microscopic transport equations has been developed. This model is based on assumptions of local thermal equilibrium and non-local mass balance and it is based on a model of solidification of binary mixtures (P. Bousquet-Melou, 2000). In theory, the approach is identical. However, the dissolution introduced physico-digital extra constraint that had to be taken into account. Taking into account the two aspects mentioned (convection and two-phase zone) is a novelty in comparison to existing models. The model was then extended to the ternary case with new developments to resolve certain difficulties (addition of an equation of this case, non-bijective equilibrium relations, several diffusion coefficients to be determined, uncertainty on flows in interface, density of the solid non-constant...) to study the dissolution UO2 / Zr. Validation of models is obtained from experimental results of dissolution of crucibles of UO2 and ZrO2 by liquid zircaloy and show the quality of our predictive model. (author)