Ex-vessel stabilization of corium: An analysis of corium-concrete interaction with top flooding for siliceous concrete

In case of severe accident without possibility of in-vessel retention, the corium must be stabilized outside the vessel, either in a designed core-catcher or directly on the concrete basemat of the reactor building, after spreading. In this last case, the stabilization strategy must be efficient enough to avoid the progression of corium through the concrete basemat and to the environment.

In order to ensure corium stabilization, the physical processes leading to quench some part of the liquid corium must be fast enough, compared to the erosion of concrete. One of the possible strategies consists in spreading the corium over a rather large area (i.e. the reactor pit plus one or two adjacent rooms) and flood it with water on top. Two processes have been identified experimentally for a potential quenching of corium: melt eruption and water ingression. Melt eruption is a process that is driven by the flow of gases (CO${}_2$ and H${}_2$O) generated by concrete erosion: the gas flow is likely to entrain droplets of corium into the overlying water, quenching the droplets quickly. Water ingression is a process where cracks are generated by the contraction of the top crust due to cooling: this process enables the flow of water into the cracks and the propagation of the cracking front deep into the corium crust. When the basemat is made of siliceous concrete, stabilization depends mostly on water ingression because there is little gas content in such concrete and melt eruption may only have a limited effect.

This paper presents the analysis of a situation where corium is spread over a rather large area (around 80 m${}^2$) in order to reduce the corium thickness after spreading. Of course, this thickness depends on the reactor design (total volume of corium divided by total spreading area). After spreading, corium is flooded with water on top. First, the melt eruption process is investigated: it is shown that it has a limited efficiency in case of siliceous concrete but it plays an important role in triggering the water ingression earlier. Then, the process of water ingression is examined, as it is the process leading to complete stabilization. The maximum heat flux extracted by water ingression (CHF) is the most important parameter in this process and needs to be evaluated accurately. Therefore, a theoretical determination of the CHF is made, as a function of the concrete content in the corium, based on the existing Lister-Epstein model, on data obtained earlier at Argonne National Laboratory and on data measured in volcanic rocks. In addition, a simple modeling of transient heat conduction through the concrete basemat is proposed, to determine a criterion of stop of erosion that can be used in a lumped-parameter code. As a conclusion, it is shown that the quenching of corium may be divided in two main stages. The first stage is characterized by a fast erosion of concrete and a significant eruption and quenching of debris. This stage lasts from half a day to a few days, depending on the type of concrete and on the model parameters. The second stage is characterized by a slower erosion of concrete and the start of water ingression. This stage may last from half a day up to one week depending on the type of concrete and on the model parameters. Finally, a sensitivity analysis is made to determine the effects on the erosion depth and on the time of stabilization. It appears that there is a non-trivial dependence of water ingression efficiency on the occurrence of melt eruption at the very beginning of corium concrete interaction. Melt eruption, even if limited, is likely to significantly enhance the triggering of water ingression and favor a fast stabilization of corium.