Numerical study of the natural circulation cooling stability of spent fuel pools: Analysis of turbulent natural convection and free surface evaporation driven by internal heating

Details

Supervisors

Papalexandris, Miltiadis

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil mécanicien, à finalité spécialisée

Abstract

enThis thesis focuses on spent nuclear fuel assemblies in deactivation pools. It studies the consequences of a loss-of-cooling accident (LOCA) in such pools, i.e. a scenario where the cold water inlet and hot water outlet stop abruptly, as was the case with the Fukushima disaster in 2011. The physical phenomena studied in this work are, on the one hand, the behaviour of the flow induced by the heating of the spent nuclear fuel and, on the other hand, the evaporation at the free surface of the pool during the early stage of the accident, when the spent fuel assemblies are still under water and not yet exposed to the ambient air. The evaporation at the free surface was modelled on the basis of a non-zero Neumann boundary condition on the temperature field. This model was developed by William Hay in his doctoral thesis, Evaporation-driven turbulent convection in water pools heated from below, in 2021. In this thesis, the analysis of the convective behaviour of the flow was first performed using full-scale two-dimensional geometry. Then, a three-dimensional geometry was developed, but at a reduced scale due to limited computational resources. For each of the two geometries, four cases were investigated, based on different distributions of thermal power at the bottom of the pool (i.e. different arrangements of “hot” and “cold” assemblies). These assemblies were modelled as a porous medium inside of which head losses were taken into account. The simulations were carried out using Code_Saturne, an open source code developed by EDF. These simulations were performed using the Oberbeck-Boussinesq approximation, with the exception of an additional simulation, for each geometry and for comparison purposes, that was carried out using the dilatable flow algorithm of Code_Saturne, which is akin to the low-Mach number approximation. It was found that, for the same total thermal power at the bottom of the pool, assemblies with lower power should be placed at the centre and surrounded by higher-power assemblies. This configuration is advantageous compared to the opposite arrangement, where higher-power assemblies are placed at the centre and lower-power assemblies surround them. The former setup delays the time before assemblies become exposed to ambient air by 5 to 7% compared to the latter.