Evaluation of embrittlement models for regulatory use: Comparing RG 1.99 Rev. 2 and an intercept-adjusted ASTM E900-15 approach

This study evaluated the regulatory applicability of embrittlement trend curve models for reactor pressure vessel (RPV) steels by comparing RG 1.99 Rev. 2 with an intercept-adjusted ASTM E900-15 model. A statistically grounded additive adjustment was applied to the E900-15 model to improve its prediction accuracy using plant-specific surveillance data. The analysis utilized a combined dataset of 1,034 data points from Korean and U.S. nuclear power plants, encompassing a broad range of product forms, notch orientations, and fluence levels. Quantitative comparisons with and without model-specific margins were performed to assess the predictive performance and conservatism of each model within the established regulatory upper-bounds for RPV embrittlement. The adjusted E900-15 model significantly reduced the residual bias and root-mean-square deviation, particularly at high fluence levels. Its marginal behavior remained stable and statistically conservative without excessive overestimation, whereas RG 1.99 Rev. 2 tended to underestimate the embrittlement in the long-term fluence regime. Even though RG 1.99 Rev. 2 is widely adopted for its simplicity, the findings indicate that the intercept-adjusted E900-15 model provides a more accurate and adaptable framework for regulatory applications, particularly for long-term operations and life-extension assessments.