Validation and improvement of Fe, Cr, Ni nuclear data in bulk shield benchmark experiments

To check shielding calculation routes and nuclear data libraries against experiments, a database of international Fe benchmarks has been built up. Accurate reaction rate measurements inside a stainless steel block, performed in the recent FNG European experiment, were also used. Two-dimensional sn calculations and three-dimensional Monte Carlo calculations, based on the multigroup matxs87/eff1 library, were carried out. The calculation-experiment comparison demonstrated the overall validity of the European EFF-1 library for shielding design calculations; however, flux underestimation with penetration depth was pointed out. A sensitivity study of the 70 measured response functions to the Fe partial cross-sections was achieved. In order to interpret the calculation-experiment discrepancies in the most efficient way, a “trend analysis” method was implemented, accounting for both microscopic and integral data uncertainties. The Fe elastic cross-section is overestimated by 5% in the 100 keV-2 MeV range, probably because of resonance self-shielding. A clear trend to reduce the inelastic scattering was found out, in agreement with the ENDF/B-6 evaluation.