Search

In this study, the SAMPSON code was modified to evaluate severe accidents in a spent fuel pool (SFP). Not only the SFP but also upper spaces of the SFP, walls of the reactor building, and the blowout panel were included. Air oxidation models obtained by the Zircaroy-4 cladding (ANL model) and the Zircaroy-2 cladding (JAEA model) were included in the modified SAMPSON code. Experiments done by Sandia National Laboratory using simulated fuel assemblies equivalent to those of an actual BWR plant were analyzed by the modified SAMPSON code to confirm the function of fuel temperature calculation in the event of loss of fuel cooling in the SFP. The rapid fuel rod temperature rise due to the Zr air oxidation reaction could be reasonably evaluated by the SAMPSON analysis for the radial propagation experiment. The effect of the oxidation reaction model was evaluated by the analysis of the SFP assuming no initial water level. There was almost no difference in the maximum temperature transient of the fuel rod surface between the ANL and JAEA models since the extent of the oxidation reaction was limited by the amount of oxygen supplied in the current analysis conditions. The analysis was conducted with different initial water levels which were no water, water level at bottom of active fuel, and water level at half of active fuel. The present analysis showed that the earliest temperature rise of the fuel rod surface occurred when there was no water in the SFP and natural circulation of air became possible.

After the Fukushima Daiichi nuclear power plant accident due to the pacific coast of Tohoku earthquake and Tsunami, we have been developing an analysis method considering creep damage mechanisms based on three-dimensional analysis in order to estimate condition of the fuel debris and failure behavior of the reactor due to severe accidents for the early completion of the decommissioning of nuclear power plants. We have also been obtaining material properties that are not provided in existing databases or literature for the analysis. In this study, we measure the tensile and creep properties of low alloy steel, Ni-based alloy, and stainless steel at high temperatures near the melting points. Using experimental data, some parameters for the creep constitutive law and creep failure evaluation method are determined. It is confirmed that the estimated rupture time by the failure evaluation method agree well with the experimental one.

In existing severe accident codes such as MELCOR and THALES2, rupture of reactor pressure vessel (RPV) by relocated molten core is judged using simple models such as temperature and/or stress criteria. However, it is difficult to assess rupture behavior of the lower head of RPV in boiling-water-type nuclear power plants due to severe accident like Fukushima Daiichi. One reason is that boiling water reactors (BWRs) have geometrically complicated structure with a lot of penetrations. Another one is that BWR lower head is composed of various types of materials of RPV, weld-overlay cladding, control rod guide tubes, stub tubes, welds, etc. Therefore, we have been developing an analysis method to predict time and location of RPV lower head rupture of BWRs considering creep damage mechanisms based on coupled analysis of three-dimensional thermal-hydraulics (TH) and thermal-elastic-plastic-creep analyses. The detailed three-dimensional model of RPV lower head with control rod guide tubes, stub tubes, and welds are constructed. TH analysis is performed to obtain three-dimensional temperature distribution in relocated debris. Using TH analysis results, structural analysis is carried out to evaluate creep damage distributions using four types of damage criterions of “considere”, strain, Kachanov, and Larson-Miller-parameter (LMP) criteria. Creep damage evaluation based on Kachanov and LMP models is made by using experimentally determined parameters. From comparison of damage criterions, it is shown that failure regions of BWR lower head are only penetrations under simulated conditions, although there is a large difference in failure time.