Uncertainty evaluation of Charpy ductile-to-brittle transition temperature using Bayesian inference

ベイズ推定を用いたシャルピー延性脆性遷移温度の不確かさ評価

高見澤 悠 ; 西山 裕孝 

Takamizawa, Hisashi; Nishiyama, Yutaka

原子炉圧力容器の脆化評価に関して、シャルピー衝撃試験によって評価される延性脆性遷移温度の不確かさを分析するため、日米の様々な鋼種の未照射材と照射材の約1900件のデータセットを基にシャルピー吸収エネルギーの温度依存性をモデル化し、ベイズ統計を用いて試験数の違い、試験温度条件の違いを考慮可能な評価手法を整備した。整備した手法を用いた分析から、照射の有無、日米の材料の違い、鋼種の違いは延性脆性遷移温度の不確かさに影響しないものの、溶接熱影響部における延性脆性遷移温度の不確かさが母材や溶接金属に比べて有意に大きいことを定量的に明らかにした。

Irradiation embrittlement of reactor pressure vessel steels is usually predicted based on ductile-to-brittle transition temperature shift obtained by Charpy impact tests for unirradiated and irradiated specimens. It is important to understand the uncertainty in the evaluation by Charpy impact test when giving margins for structural integrity assessment using embrittlement trend curves and when discussing the need for surveillance specimens obtained from the weld heat affected zone. In this study, we modeled the temperature dependence of Charpy absorbed energy based on about 1900 datasets of unirradiated and irradiated Japanese and U.S. materials. Then we used Bayesian statistics to construct an evaluation model that could take into account the variation in absorbed energy, the number of tests, and test temperatures for each dataset. In my presentation, we will discuss the results of analyses for the effects of irradiation, difference between Japanese and U.S. materials, and product form on the uncertainty of the Charpy ductile-to-brittle transition temperature.