Data assimilation of post-irradiation examination data for fission yields from GEF
Laboratory for Reactor Physics and Systems Behavior, École Polytechnique Fédérale de Lausanne,
2 Nuclear Criticality Safety Division, Lawrence Livermore National Laboratory, Livermore, CA 94550, USA
3 Nuclear Energy and Safety Research Division, Paul Scherrer Institut (PSI), Villigen PSI, Switzerland
4 Department of Physics and Astronomy, Applied Nuclear Physics, Uppsala University, Uppsala, Sweden
5 Nuclear Data Section, International Atomic Energy Agency, Vienna 1400, Austria
* e-mail: email@example.com
Received in final form: 26 June 2020
Accepted: 16 July 2020
Published online: 28 August 2020
Nuclear data, especially fission yields, create uncertainties in the predicted concentrations of fission products in spent fuel which can exceed engineering target accuracies. Herein, we present a new framework that extends data assimilation methods to burnup simulations by using post-irradiation examination experiments. The adjusted fission yields lowered the bias and reduced the uncertainty of the simulations. Our approach adjusts the model parameters of the code GEF. We compare the BFMC and MOCABA approaches to data assimilation, focusing especially on the effects of the non-normality of GEF’s fission yields. In the application that we present, the best data assimilation framework decreased the average bias of the simulations from 26% to 14%. The average relative standard deviation decreased from 21% to 14%. The GEF fission yields after data assimilation agreed better with those in JEFF3.3. For Pu-239 thermal fission, the average relative difference from JEFF3.3 was 16% before data assimilation and after it was 12%. For the standard deviations of the fission yields, GEF’s were 100% larger than JEFF3.3’s before data assimilation and after were only 4% larger. The inconsistency of the integral data had an important effect on MOCABA, as shown with the Marginal Likelihood Optimization method. When the method was not applied, MOCABA’s adjusted fission yields worsened the bias of the simulations by 30%. BFMC showed that it inherently accounted for this inconsistency. Applying Marginal Likelihood Optimization with BFMC gave a 2% lower bias compared to not applying it, but the results were more poorly converged.
© D. Siefman et al., published by EDP Sciences, 2020
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.