ナトリウム冷却高速炉の炉心崩壊事故における炉心残留燃料の冷却性評価手法の開発に関する試験と解析

Experiment and analysis for development of evaluation method for cooling of residual core materials in core disruptive accidents of sodium-cooled fast reactors

今泉 悠也 ; 青柳 光裕 ; 神山 健司  ; 松場 賢一  ; Akayev, A. S.*; Mikisha, A. V.*; Baklanov, V. V.*; Vurim, A. D.*

Imaizumi, Yuya; Aoyagi, Mitsuhiro; Kamiyama, Kenji; Matsuba, Kenichi; Akayev, A. S.*; Mikisha, A. V.*; Baklanov, V. V.*; Vurim, A. D.*

高速炉の炉心崩壊事故での炉心からの燃料流出後の残留炉心物質の冷却性は、炉心物質の配置に大きな影響を与え、それは炉容器内保持(IVR)達成のための重要な要素である。残留炉心物質の冷却は「インプレース冷却」と呼ばれ、その評価のため、実機の炉心での挙動をSIMMER-IIIで解析し、その解析結果に基づき重要度ランクテーブル(PIRT)の手法を適用した。そして、PIRTによって抽出された熱流動現象に着目した実験をEAGLE-3のフレームワークで実施した。実機のインプレース冷却の段階で発生し得る継続的な液位振動が実験では観察され、SIMMER-IIIで解析を実施した。解析結果の調査から、実験結果と解析結果の差異は、ナトリウム液位の上での非凝縮性ガスの残存と占有によることが分かったが、これは実験では非現実的だと考えられる。この問題を解決するため、非凝縮性ガスとナトリウム蒸気のガス混合モデルを開発し、その新モデルにより実験結果と解析結果の一致が大幅に改善された。

The cooling of the residual core materials after the fuel discharge from the SFR core in the core disruptive accident can significantly affect the distribution fraction of the core materials which is an important factor for the in-vessel retention (IVR). The cooling of the residual core materials is called "in-place cooling". For the evaluation of the in-place cooling, behavior in a SFR core was simulated by SIMMER-III, and method of phenomena identification and ranking table (PIRT) was applied based on the analysis result. Experiment which focuses on the thermal-hydraulic phenomena which were extracted by the PIRT was conducted in the framework of EAGLE-3 project. Continuous oscillation of sodium level which can occur in the phase of in-place cooling of SFRs was observed in the experiment, and analysis by the SIMMER-III was conducted. By investigation of the analysis result, difference between the experiment and analysis results was revealed to be due to remaining and occupation of non-condensable gas above the sodium level which would be unrealistic in the experiment. Gas mixture model between non-condensable gas and sodium vapor was developed to solve this problem, and coincidence between experiment and analysis results was largely improved by this new model.