Development of numerical simulation method of natural convection around heated porous medium by using JUPITER

JUPITERを用いた加熱多孔体周りの自然対流の数値シミュレーション手法の開発

上澤 伸一郎  ; 山下 晋   ; 柴田 光彦 ; 吉田 啓之  

Uesawa, Shinichiro; Yamashita, Susumu; Shibata, Mitsuhiko; Yoshida, Hiroyuki

福島第一原子力発電所の廃炉における汚染水対策として、注水低減,間欠注水,空冷が検討されている。しかし、格納容器内には燃料デブリの不確実性があるため、燃料デブリの分布状態や燃料デブリ取り出し作業の進捗状況に応じて、最適な冷却方法を事前に検討する必要がある。そのため我々は、燃料デブリの位置,発熱,気孔率の影響を含む空冷中の熱挙動を推定する方法の開発を進めている。多孔体と考えられる燃料デブリの内部構造をモデル化した上で大規模な熱流動解析を行うことは困難であることから、JUPITERに多孔体モデルを追加することにより多孔体の熱流動の解析を可能にした。本研究では、多孔体モデルを導入したJUPITERの妥当性検証結果について報告するとともに、多孔体の伝熱モデルについて直列,平行,幾何平均モデルのどのモデルが最も有効かについて議論する。多孔体周辺の自然対流の検証データについては、多孔体を含む系における自然対流の伝熱流動実験を独自に行った。実験と各モデルでの数値解析と比較を行ったところ、幾何平均モデルを用いた数値結果が実験結果に最も近い結果を得られた。しかしながら、定量的には温度と速度ともに実験結果よりも過大評価しており、特に、多孔体と空気との境界付近の温度は、より過大評価していることを確認した。

For contaminated water management in decommissioning Fukushima Daiichi Nuclear Power Stations, reduction in water injection, intermittent injection water and air cooling are considered. However, since there are uncertainties of fuel debris in the PCV, it is necessary to examine and evaluate optimal cooling methods according to the distribution state of the fuel debris and the progress of the fuel debris retrieval work in advance. We have developed a method for estimating the thermal behavior in the air cooling, including the influence of the position, heat generation and the porosity of fuel debris. Since a large-scale thermal-hydraulics analysis of natural convection is necessary for the method, JUPITER developed independently by JAEA is used. It is however difficult to perform the large-scale thermal-hydraulics analysis with JUPITER by modeling the internal structure of the debris which may consist of a porous medium. Therefore, it is possible to analyze the heat transfer of the porous medium by adding porous models to JUPITER. In this study, we report the validation of JUPITER applied the porous model and discuss which heat transfer models are most effective in porous models such as series, parallel and geometric mean models. To obtain validation data of JUPITER for the natural convective heat transfer analysis around the porous medium, we performed the heat transfer and the flow visualization experiments of the natural convection in the experimental system including the porous medium. In the comparison between the experiment and the numerical analysis with each model, the numerical result with the geometric mean model was the closest of the models to the experimental results. However, the numerical results of the temperature and the velocity were overestimated for those experimental results. In particular, the temperature near the interface between the porous medium and air was more overestimated.