In the Sodium-cooled fast reactor (SFR) safety analysis, the Core Disruptive Accident (CDA) is considered as the most dangerous severe accident. During the CDA of SFR, the molten core materials are probably discharged and released from the core area, and then become solid debris or particles after the rapid quenching and fragmentation as a result of the interaction with liquid sodium. Finally, the solid debris will sediment and form the so-called “debris bed” on the lower plenum of reactor vessel. If the decay heat generated from the high-temperature debris cannot be effectively removed, the eutectic reaction between the debris bed and the lower plenum of reactor vessel is possible to occur, resulting in serious consequences. As a result, to improve the assessment for the heat removal capacity of debris bed, it is necessary to deeply and comprehensively investigate and understand the mechanism of the formation of debris bed. Focusing on this aspect, it is believed that relative numerical studies based on Computation Fluid Dynamics (CFD) tools can stimulate the better understandings on the mechanism and characteristics of debris bed formation behavior.

Considering that simulations with the particle methods in CFD can accurately catch the detail behavior occurred on the interfaces of different phases, the Moving Particle Semi-implicit (MPS) method is employed in present study for the simulations of debris bed formation behavior. However, currently, the MPS method is still in the preliminary stage of its development, and it is required to further improve and mature. Therefore, in present study, for the purpose of obtaining more reasonable simulation results, improvements on the numerical stability and accuracy of the MPS method are performed by introducing the Passively Moving Solid (PMS) model and Discrete Element Method (DEM) to respectively evaluate the solid-fluid and solid-solid interactions and modifying the judgement of free surface, the source term of pressure Poisson equation, the discretization accuracy of differential operator models and the non-uniform distributions of particles. Owing to the comparison between the experimental results and the simulation results based on the improved MPS method, the effectiveness of the improvements for MPS method applied in present study is validated, and the applicability of current improved MPS method on the simulations of debris bed formation behavior is confirmed. Finally, extended numerical studies on the debris bed formation behavior are carried out under more actual SFR accident conditions (i.e. liquid sodium conditions) to attain more useful knowledge on this phenomenon. In addition, on the basis of simulation results, the reliability of the predicted model developed from the past experimental studies is investigated and discussed.

The improvements performed in present study for MPS method is believed to be valuable for further numerical investigations on debris bed formation behavior under more complicated conditions, and to provide useful reference for the numerical simulations of other multiphase-flow-related phenomena that might emerge in the CDA of SFR (such as molten-pool sloshing motion and debris bed self-leveling behavior), Knowledge from present study is suggested to be helpful for the deeper understandings of the characteristics of the debris bed formation behavior, and is of significance for optimizing the SFR structural design and promoting the Research and Development (R&D) and application of the SFR.