JEFF-3.2 library validation - Impact of the new library for fast neutrons SFR benchmarks - 15150

The JEFF-3.2 library was released by OECD/NEA in March, 2014. The work presented here deals with the validation of this library for Sodium Fast Reactors (SFR) and attempts to highlight the improvements obtained for this class of reactors. As CEA is involved in the design of a new SFR prototype ASTRID (Advanced Sodium Technological Reactor for Industrial Demonstration) that should be built around 2020, foreseen reactor safety analyses will be performed using both new neutronics computational codes and new nuclear data libraries. So, there is a strong interest in JEFF-3.2 benchmarking for fast neutron spectrum applications. In addition, some thermal neutron spectrum configurations are also tested to check whether JEFF-3.2 does not degrade the performances shown by JEFF-3.1.1. A huge effort has been made in the JEFF-3.2 library to improve nuclear data for many heavy isotopes either in the high energy domain (²³⁵U, ²³⁸U) or in the whole energy domain (²³⁹Pu, ²⁴⁰Pu, ²⁴¹Am). Concerning moderator materials, the ²³Na evaluated data has also been revisited in the whole domain. Some structure materials like Cr or some fission products have also been modified, but their impact on the class of benchmarks of interest remains low. Across this paper, we report the results of a CEA benchmarking study of the library. Benchmarks originate from the criticality database ICSBEP (about 100 cases are used), from the reactor physics database IRPHE (SNEAK7A, SNEAK7B and ZPPR-10A), from the CEA EOLE Light Water Reactor mock-up (EPICURE and MISTRAL experimental programs) or from the MASURCA Sodium Fast Reactor mock-up (MASURCA1B, PRE-RACINE, RACINE and CIRANO experimental programs), from Superphenix Commissioning Tests or finally from the radiation shielding database SINBAD (ASPIS, REPLICA and JANUS-8). As the basis of JEFF-3.2 library is the JEFF-3.1.1 library, comparisons are made between the two libraries on one hand, and with experimental values on the other hand. Main physical parameters are the effective multiplication factor and the sodium void effect for criticality and reactor physics benchmarks, and the dosimetry reaction rates for radiation shielding benchmarks. All the benchmarks are simulated with the CEA Monte Carlo code TRIPOLI-4. Some perturbation analyses are performed using the CEA ERANOS/SNATCH code package for some particularly interesting cases. Reactivity effect breakdowns are shown to highlight the origin of the main differences between JEFF-3.1.1 and JEFF-3.2 calculations for some SFR-type benchmarks. Results clearly show that new ²³⁵U, ²³⁸U, ²³⁹Pu, ²⁴⁰Pu, ²⁴¹Am and ²³Na evaluations have a huge impact on k_eff values for SFR type configurations and generally improve the calculated values compared to experimental ones. Moreover, the revisited ²⁴¹Am data improve results for EOLE thermal cases. Uranium, plutonium or mixed uranium and plutonium thermal configurations are still accurately predicted with the new JEFF-3.2 library. (authors)