Corrections to the 148Nd method of evaluation of burnup for the PIE samples from Mihama-3 and Genkai-1 reactors

Abstract

The value of the burnup is one of the most important parameters of samples taken by post-irradiation examination (PIE). Generally, it is evaluated by the Neodymium-148 method. Precise evaluation of the burnup value requires: (1) an effective fission yield of 148Nd; (2) neutron capture reactions of 147Nd and 148Nd; (3) a conversion factor from fissions per initial heavy metal to the burnup unit GWd/t. In this study, the burnup values of the PIE data from Mihama-3 and Genkai-1 PWRs, which were taken by the Japan Atomic Energy Research Institute, were re-evaluated using more accurate corrections for each of these three items. The PIE data were then re-analyzed using SWAT and SWAT2 code systems with JENDL-3.3 library. The re-evaluation of the effective fission yield of 148Nd has an effect of 1.5–2.0% on burnup values. Considering the neutron capture reactions of 147Nd and 148Nd removes dependence of C/E values of 148Nd on the burnup value. The conversion factor from FIMA(%) to GWd/t changes according to the burnup value. Its effect on the burnup evaluation is small for samples having burnup of larger than 30 GWd/t. The analyses using the corrected burnup values showed that the calculated 148Nd concentrations and the PIE data is approximately 1%, whereas this was 3–5% in prior analyses. This analysis indicates that the burnup values of samples from Mihama-3 and Genkai-1 PWRs should be corrected by 2–3%. The effect of re-evaluation of the burnup value on the neutron multiplication factor is an approximately 0.6% change in PIE samples having the burnup of larger than 30 GWd/t. Finally, comparison between calculation results using a single pin-cell model and an assembly model is carried out. Because the results agreed with each other within a few percent, we concluded that the single pin-cell model is suitable for the analysis of PIE samples and that the underestimation of plutonium isotopes, which occurred in the previous analyses, does not result from a geometry model.