Search

To deepen our understanding for the current debris status and investigate the debris-structure interactions in the pedestal region of Fukushima Unit-3, the Moving Particle Semi-implicit (MPS) method is further developed for simulation of multi-component liquid/solid relocation with solid-liquid phase changes aiming for plant-scale practices. The improvement of the existing MPS method mainly consists of two parts, namely 1) the improvement of numerical stability and accuracy, including a) applying second-order corrective matrix to the particle interaction models and b) particle shifting that can optimize particle configurations; and 2) improvement of calculation efficiency, including a) hybrid OpenMP and MPI parallelization and b) particle type-dependent speed-up algorithms to reduce calculation costs for particles with extremely high viscosity and low velocity. In the current study, the improved MPS method is validated against the experiments carried out at Waseda University, in which molten salt droplets were released to interact with aluminum pillars and solidify on them. Good agreement of the total height of the solidified salt and its distribution on the pillars has been achieved. The successful validation has shown the capability of the current MPS method for simulations of Unit-3 pedestal region.

Interpretation for the two-week long Unit 3 ex-vessel debris cooling behavior was conducted based on 1F plant and site data such as pressure, temperature, gamma ray level and live camera pictures. It was estimated that the debris relocated to the pedestal was in partial contact with liquid water for about initial two days. With the reduction of the sea water injection flowrate, the debris, both in-vessel and ex-vessel ones, became “dry”, in which the debris was only weakly cooled by vapor and this condition lasted for about four days until the increase of the sea water injection. During this dry period, in-vessel and ex-vessel debris were heat up and it took another four days to re-flood the heated up debris.

Molten corium-concrete interaction (MCCI) is an important ex-vessel phenomenon that could happen during the late phase of a hypothetical severe accident in a light water reactor. When the molten corium, which is generally comprised of UO2, ZrO2 and metals such as zircalloy and stainless steel, is discharged into a dry reactor cavity, a stratified molten pool configuration with two immiscible oxidic and metallic phases can be expected to form and lead to MCCI. Compared to a homogenous oxidic molten pool configuration, the metallic phase in the stratified molten pool might influence the crust formation on the corium-concrete interface and consequently cause different concrete ablation behavior to evaluate MCCI progression concerning containment failure. In terms of this issue, past experimental studies, such as COMET-L, VULCANO VBS and MOCKA test series, have been carried out to investigate the influence of such oxidic and metallic stratified pool configuration on MCCI. The experimental results have shown that the metallic phase can have a significant impact on the axial and radial ablation kinetics that could influence the ablation patterns of reactor pit. As regards numerical studies, past numerical modeling of MCCI was generally based on Eulerian methods and simplified empirical approach to simulate solid/liquid phase change and evolving of corium/crust/concrete interface. Such modeling might be efficient but have shown deficiencies and inadequacies due to its Eulerian and empirical nature, which has suggested a necessity to seek for a more mechanistic approach for modeling of MCCI. In this sense, Moving Particle Semi-implicit (MPS) method is considered suitable for MCCI analysis for its advantages of tracking interfaces and modeling phase change accurately as a Lagrangian particle method. In the present study, a three-dimensional (3-D) numerical study has been performed to simulate COMET-L3 test carried out by KIT with a stratified molten pool configuration of simulant materials with improved MPS method. Solid/liquid phase change was simulated with types of solid and liquid particles with thermal and physical properties including temperature and solid fraction, which enabled tracking of the solid/liquid status of each particle to achieve accurate free surface and corium/crust/concrete interface capturing. The heat transfer between corium/crust/concrete was modeled with heat conduction between particles. Moreover, the potential influence of the siliceous aggregates was also investigated by setting up two different case studies since there was previous study indicating that siliceous aggregates in siliceous concrete might contribute to different axial and radial concrete ablation rates. The simulation results have indicated that metal melt as corium in MCCI can have completely different characteristics regarding concrete ablation pattern from that of oxidic corium, which needs to be taken into consideration when assessing the containment melt-through time in severe accident management.

Authors are developing an experimental technology to realize experiments simulating Severe Accident (SA) conditions using simulant fuel material (ZrO2 with slight addition of MgO for stabilization) that would contribute not only to Fukushima Daiichi (1F) decommissioning but also to enhance the safety of worldwide existing and future nuclear power plants through clarification of the accident progression behavior. Based on the results of the prototype test, improvement of plasma heating technology was conducted. The Core Material Melting and Relocation (CMMR)-1/-2 experiments were carried out in 2017 with the large-scale simulated fuel assembly (1 m × 0.3 mϕ) applying the improved technology (higher heating power and controlled oxygen concentration). In these two tests, heating history was different resulting basically in similar physical responses with more pronounced material melting and relocation in the CMMR-2 experiment. The CMMR-2 experiment is selected here from the viewpoint of establishing an experimental technology. The CMMR-2 experiment adopted 30-min heating period, the power was increased up to a level so that a large temperature gradient (> 2,000 K/m) expected at the lower part of the core in the actual 1F accident conditions. Most of the control blade and the channel box migrated from the original position. After the heating, the simulated fuel assembly was measured by the X-ray Computed Tomography (CT) technology and by Electron Probe Micro Analyzer (EPMA). CT pictures and elemental mapping demonstrated its excellent performance with rather good precision. Based on these results, an excellent perspective in terms of applicability of the non-transfer type plasma heating technology to the SA experimental study was obtained.