Nuclear data uncertainty for criticality-safety: Monte Carlo vs. linear perturbation

详细信息    查看全文

• 作者：D. Rochman^a ; ; A. Vasiliev^a ; H. Ferroukhi^a ; T. Zhu^b ; S.C. van der Marck^c ; A.J. Koning^d ; ^e
• 关键词：Criticality benchmark ; Nuclear data ; Uncertainty ; Sensitivity
• 刊名：Annals of Nuclear Energy
• 出版年：2016
• 出版时间：June 2016
• 年：2016
• 卷：92
• 期：Complete
• 页码：150-160
• 全文大小：1165 K

文摘

This work is presenting a comparison of results for different methods of uncertainty propagation due to nuclear data for 330 criticality-safety benchmarks. Covariance information is propagated to k_eff using either Monte Carlo methods (NUSS: based on existing nuclear data covariances, and TMC: based on reaction model parameters) or sensitivity calculations from MCNP6 coupled with nuclear data covariances. We are showing that all three methods are globally equivalent for criticality calculations considering the two first moments of a distribution (average and standard deviation), but the Monte Carlo methods lead to actual probability distributions, where the third moment (skewness) should not be ignored for safety assessments.