Most of available 2D MCCI experiments with real material from MCCI-OECD and VULCANO programs have been analysed and an interpretation of outstanding MCCI tests is presented. This interpretation has been performed with the help of the 'MEDICIS' module of ASTEC code. The analysis is focussed on the 2D ablation behaviour, the thermal resistance of the pool/concrete interfaces, and indirectly the structure of the pool/concrete interface versus the concrete type.
The present work describes the improvement of the modelling of the pool/concrete interface structure compared to the previous approach presented in MCCI seminar in September 2007. The basic assumption of the formation of a mushy boundary layer without a stable crust at the pool/concrete interface is kept. Indeed no proof of a stable crust could be derived from available post-test examinations (PTE) of real material MCCI experiments. However, the evaluation of the inner temperature of this boundary layer (called solidification temperature in the previous approach) from precise thermochemistry data for the average pool composition is not consistent with the complex structure of phases and the absence of any uniform zone near the pool/concrete interface, as deduced from PTE for VBU5 and VBU6 tests. Therefore, it is proposed to model the heat transfer from the bulk pool to the ablation interface by taking into account in series only the convective heat transfer in the bulk pool and the heat conduction transfer across a resistive boundary layer (such as a slag layer) without imposing any boundary temperature condition between convective and conductive zones.
Since the thermodynamic equilibrium at the pool/upper crust interface is likely not reached in real material experiments because of transient conditions, high gas bubbling and high liquid corium viscosity due to the presence of ablated silica, the upper crust is assumed to be in a mushy state : the interface temperature with the lower pool convective zone is determined from a threshold solid fraction below which convective heat transfer is replaced by conduction and consequently is much lower than the pool liquidus temperature. This assumption is supported by the very thin upper crust thickness observed in several MCCI experiments.
The heat convection transfer within the homogeneous pool is supposed to be isotropic. With this assumption the 2D ablation is governed by the thermal resistance profile along the pool interfaces, which should depend on the concrete type. It is proposed here to relate the influence of concrete on the 2D ablation mainly to the existence in case of a siliceous concrete of a stable refractory solid accumulation causing an increased thermal resistance at the bottom corium/concrete interface. As far as possible the same assumptions are kept in case of both types of concretes for the other pool/concrete interfaces and the evaluation of the pool/upper crust interface temperature. The chosen set of assumptions permits to reproduce rather satisfactorily the temperature evolution, the axial and lateral ablation kinetics, and the final eroded cavity shape in most of CCI and VULCANO 2D MCCI experiments, while keeping as much as possible the same equivalent transfer coefficient at the pool/concrete interface whatever the concrete type.
This interpretation work of existing MCCI-OECD and VULCANO experiments confirmed and completed by the analysis of future tests should help to build more reliable models for predicting MCCI behaviour in the reactor case either for an homogeneous pool or for the oxidic layer in case of a stratified pool.