Development of a New Crust Model for Analyzing VULCANO VBS-U3 MCCI Experiment With MPS Method

Abstract In the postulated severe accident of a light water reactor, the molten core material (corium) may be released to the reactor containment floor and the Molten Core - Concrete Interaction (MCCI) may take place. This complex phenomena may involve dehydration, thermal decomposition, and ablation of concrete, redox reactions of oxidic and metallic phases of corium, solidification and remelting of corium, mixing of molten concrete with corium and so on. Evaluation of the resulting concrete floor ablation is important as it may threaten the containment boundary. The Moving Particle Semi-implicit (MPS) method discretizes fluids and solids with calculation points (particles) in a Lagrangian manner and has been extensively developed for analyses of various MCCI tests with consideration of crust formation at the melt-concrete interface. Due to its Lagrangian nature, there is always some randomness to each particle position and Particle Shifting (PS) is one of the numerical techniques, which corrects such randomness by slightly shifting the particle position from a particle dense region to a coarse region to sustain numerical accuracy of the MPS method. Such slight correction does not usually have any physical impacts on the calculation unless the particle velocity becomes extremely small and duration of the phenomena becomes extremely long. In this study, a new crust model is being developed to analyze MCCI, which involves continuous concrete ablation with presence of the crust layer between the corium and the concrete walls, which may gradually move with the slow concrete wall ablation process over long time. The new crust model must enable accumulation of physical displacement of the crust particle over long time (i.e., enable physical creeping) while preventing accumulation of numerical displacement of the crust particles over long time (i.e., preventing numerical creeping), Hence, in the new crust model, the PS has been effectively disabled for the crust particles. Qualitative validity of such numerical modeling was confirmed through some trial analyses of VULCANO-VBS test using a set of tentative calculation conditions and parameters, which should be carefully revised for future quantitative discussions including validation of the analysis results with experimental results.