The impact of thermal non-equilibrium and large-scale 2D/3D effects on debris bed reflooding and coolability

Abstract During a severe nuclear accident, a part of the molten corium resulting from the core degradation may relocate down to the lower plenum of the reactor vessel. The interaction with residual water in the lower plenum leads to a fragmentation of the corium and formation of particles (characteristic length-scale: 1–5 mm). In order to predict the safety margin of the reactor under such conditions, the coolability of this porous heat-generating medium and the possibility to reflood the particle bed are studied in this paper and compared with other theoretical or experimental results. A quick overview of the existing experimental results and models is provided to identify the remaining uncertainties on some modelling issues and the lack of understanding of some of the physical processes involved. It also justifies the approach chosen by Institut de RadioProtection et de Surete Nucleaire (IRSN) to deal with the issues of debris coolability and reflooding. The detailed description of two-phase flow in a debris bed is addressed in IRSN by a special module of the ICARE/CATHARE code. This thermalhydraulic module is designed to deal with a non-homogeneous debris bed of any shape. The momentum balance equation for each fluid phase is an extension of Darcy's law. This extension takes into account the capillary effects between the two phases, the relative permeabilities and passabilities of each phase, the interfacial drag force between liquid and gas, and the porous bed configuration (porosity, particle diameter, …). The model developed is three-dimensional, which is important to better predict the flow in configurations such as natural convection co-current flows in large beds or to emphasize the impact of preferential paths induced by porous geometry (existence of regions with lower or higher porosity and permeability). The energy balance equations of the three phases (liquid, gas and solid phase) are obtained by a volume averaging process of the local conservation equations. In this method, the local thermal non-equilibrium between the three phases is taken into account and the heat exchange coefficients as well as the thermal dispersion coefficients are calculated as a function of the local geometry of the porous medium and the local phase distribution. Numerical estimations of these thermal properties can be performed, which is quite convenient, on a practical point of view, since they are very difficult to determine experimentally. This feature is a great advantage of this approach. Examples of numerical determination of effective properties are given in the paper, with analytical solutions for a simple geometry. The phase change rate is also naturally determined without additional phenomenological equation. One-dimensional predictions of critical dryout fluxes are presented and compared with results from the literature. Reasonable agreement is obtained. Calculations of one-dimensional reflooding (from top or bottom) are compared with experimental data. The results show the importance of using a non-equilibrium model for temperatures. They also indicate that channeling effects may exist and should be taken into account in the model for further improvements. Two-dimensional calculations are presented and show the influence of the porous medium characteristics. As expected, water circulation is improved considering multi-dimensional flow in the bed and the dryout heat flux is larger than predicted by 1D modelling. Conditions for reflooding are also more favourable if large-scale non-homogeneities exist in the debris bed. This leads to a flow pattern where steam can exit the debris bed in preferential channels and there is less limitation by counter-current flow.