3D transient CFD simulation of an in-vessel loss-of-coolant accident in the EU DEMO WCLL breeding blanket

Safety is one of the most important aspect in the design of the EU DEMO fusion reactor, therefore it is fundamental to study the Design-Basis Accidents (DBA) already in its pre-conceptual design phase. The in-vessel Loss Of Coolant Accident (LOCA) is one of the most relevant DBA, since it has the potential to seriously damage the components inside the vacuum vessel. Usually this kind of transients are modeled with system-level codes, which are typically employed to analyze the behavior of the entire system and for this reason are based on a lumped approach. However they are unable to evaluate localized quantities such as pressure peaks, which cannot be neglected during this kind of accidents since they can directly lead to a failure. To account also for the local parameters, it is necessary to develop a 3D transient model. In this work, an in-vessel LOCA from a water-cooled breeding blanket has been modeled, considering a rupture area of about 〖1 m〗^2 . The implementation was performed in the commercial CFD code STAR-CCM+. The model simulates the propagation of a pressurized water jet (water in the same thermodynamic conditions as in a pressurized water reactor) in the vacuum chamber, starting from the pipe break. In this way it was possible to analyze the whole evolution of the jet, accounting also for the phase-change phenomena expected at the pipe exit due to the huge difference in pressure between the cooling water and the chamber atmosphere. Being the pressure ratio equal to 1550, supersonic flow conditions are expected, as well as strong shock waves, which may propagate throughout the whole torus and bounce back from a wall to another. For this reason, adopting a static mesh is not the best choice, since a huge number of cells would be required to properly solve the jet evolution, as the whole domain must be finely refined. A better approach is to employ an adaptive mesh refinement (AMR) algorithm which, being capable of adapting the mesh to the jet evolution, allows to reduce a lot the number of cells and, as a consequence, to spare a considerable amount of computational time. The physical models adopted have been validated against a 2D reference problem with similar features before being applied to the more complex 3D EU DEMO-relevant case. The simulation results show that the shock wave impacting on the wall in front of the vessel is not dangerous, as the resulting pressure peak is below the design limit. Also the temperatures reached during the transient are not of concern for the integrity of the materials. In order to compare the average pressure evolution with that computed with a system-level code, the shock wave has been followed until the impact on the other side of the torus, where it could potentially cause the opening of the burst disks and then be reflected back. Finally a comparison with the in-vessel LOCA from a helium cooled blanket is performed, highlighting the main differences and showing how the water jet is less severe and much slower than the helium one, being affected also by phase change phenomena.