Study and modelling of the in-pile densification of the UO2 and MOx nuclear oxides

Amongst the many phenomena which take place in the course of the irradiation of UO2 or (U, Pu)O2 nuclear fuels, one of them involves the elimination of a fraction of the as-fabricated porosity. In-pile densification or sintering can reach 2.5%, i.e. approximately half the initial volume of pores is likely to disappear. Our literature survey indicates that the amplitude and kinetics of the phenomenon are both heavily dependent on the initial fuel microstructure. Micro-structural characterisation techniques of oxide fuels have therefore been developed in conjunction with quantitative image analysis methods. The ensuing methodology enables a quantitative comparison of micro-structural features in different fuels and has been applied to ascertaining the influence of the local fission rate and temperature on in-pile densification. It is thus revealed that in-pile operation eliminates a significant fraction of pores smaller than 3 microns in diameter. The experimental data generated has been used to set up a semi-empirical and a mechanistic model. The former is based on experimental results and is not essentially predictive. The inability of this model to predict the in-pile densification of oxide fuels is illustrated by the fact that the maximum fraction of pores that disappears is proportional to an empirical function of fission rate, and temperature. The proportionality factor appears to be difficult to correlate quantitatively to any given micro-structural feature. The model has however been applied to the interpretation of an in-pile densification experiment carried out in the Halden reactor (Norway). The latter model is mechanistic, i.e. it is based on the solution to a set of equations that describe the coupled temperature and radiation induced phenomena which occur in-pile. These can broadly be broken down into three categories: the fission fragment-pore interaction, the creation of point defects as the fission fragments slow down, and the diffusion of these point defects to sinks. The model calculates the evolution of the pore size distribution and has successfully been applied to modelling the in-pile densification behaviour of a fuel pellet characterised before and after irradiation. (author)