235U elastic cross-section adjustment in criticality benchmarks - Comparison between JENDL-4.0 and ENDF/-VII.1

The main aim of this study is to estimate the uncertainty on the multiplication factor (k(eff)) caused by the elastic cross-section uncertainty of the U-235 nucleus, and the adjustment of this cross-section using the generalized linear-least squares method. For this purpose, the sensitivity matrix of keff with respect to U-235 elastic cross section in the two nuclear evaluations ENDF/B-VII.1 and JENDL-4.0, has been assessed in several thermal, intermediate and fast benchmarks; which have been taken from the International Handbook of Evaluated Criticality Safety Benchmark Experiments (IHECSBE). The selection of these benchmarks is based on chi-square statistic test (chi(2)) that measures the statistical weight of deviations between calculus and experience results of the integral parameters caused by uncertainty in differential data. In this statistical test the sensitivity and covariance matrices are used. The sensitivity matrix has been calculated by the adjoint-weighted perturbation technique of the multi-purpose Monte Carlo code MCNP6.1. The U-235 elastic covariance matrix has been generated in 44 neutron energy groups by the ERRORJ module of the nuclear data processing system NJOY99. The prior and posterior uncertainties on the keff, the adjusted elastic cross-section of the U-235 and covariance matrix have been estimated using FORTRAN 95 code developed in our laboratory based on the generalized linear-least squares method. The results obtained show that: the U-235 elastic cross-sections taken from JENDL-4.0 requires an adjustment which can reach 6% in 0100 keV, 4.5% in 100 keV625?keV and 2.5% in 625 keV0.5 MeV, while those from ENDF/B-VII.1 needs an adjustment which can reach 5% in 0100 keV, 3% in 100 keV-625 keV and 1% in 625 keV0.5 MeV interval. The adjusted cross-section in the two evaluations ameliorates the posterior calculated k(eff) with respect to the experimental k(eff). Also the posterior covariance matrix decreases the nuclear uncertainty of the adjusted k(eff). (C) 2018 Elsevier Ltd. All rights reserved.