Bayesian Monte Carlo Evaluation Framework for Cross Sections Nuclear Data and Integral Benchmark Experiments~1

摘要

A Bayesian MC framework for evaluation of differential cross-section data and IBE data has been outlined, and an updated progress report on implementation of its modules has been provided. The framework would enable simultaneous evaluation of differential cross-section data and IBE data for the first time, although options for performing separate eval-uations of differential cross-section data or IBE data would certainly be available. Evaluation of IBE data will calculate posterior values of responses and their covariances, including contributions from differential cross-section data and geome-try/composition IBE data for the first time. One key feature of our BMC framework is that it accounts for nonlinear effects and it does not assume normal (i.e., Gaussian) PDFs. In the case of ~233U, we have demonstrated that nonlinear effects of resonance parameters on cross sections are significant and are not accounted for by the linear approximation. In anticipa-tion of the evaluation of a large number of model parameters (e.g., R-matrix resonance parameters), we have implemented and tested a Metropolis Hastings MCMC algorithm. This framework could also be used to prioritize measurements from among many possible differential cross sections and/or IBE data measurements to achieve a particular user-specified goal (e.g., response uncertainty of some nuclear application).