Considerations in modelling the melting of fuel containing fission products and solute oxides

It is well known that the oxidation of a defected fuel element by steam gives rise to an increase in O/U ratio with a consequent lowering of the incipient melting temperature. Concurrently, the hyperstoichiometry reduces the thermal conductivity thereby raising the centerline fuel pellet temperature for a fixed linear power. The development of fission products soluble in the UO2 phase or, more important, the deliberate introduction of additive oxides in advanced CANDU fuel bundle designs further affects and generally lowers the incipient melting temperature. For these reasons, the modeling of the molten (hyperstoichiometric) UO2 phase containing several solute oxides (ZrO2, Ln2O3 and AnO2) is advancing in the expectation of developing a moving boundary heat and mass transfer model aimed at better defining the limits of safe operating practice as burnup advances. The paper describes how the molten phase stability model is constructed. The redistribution of components across the solid-liquid interface that attends the onset of melting of a non-stoichiometric UO2 containing several solutes will be discussed. The issues of how to introduce boundary conditions into heat transfer calculations consistent with the requirements of the Phase Rule will be addressed. The Stefan problem of a moving boundary associated with the solid/liquid interface sets this treatment apart from conventional heat and mass transfer problems. (author)