Eigenvalue analysis of small and micro gas-cooled reactors based on different evaluated nuclear data libraries and RMC code

Background With the continuous advancement of nuclear data processing codes and evaluated libraries, legacy continuous-energy libraries (e.g., those derived from ENDF/B-VII.1 using NJOY99) frequently fail to satisfy the stringent accuracy demands of modern advanced reactor designs.

Purpose This study aims to generate updated continuous-energy neutron and thermal scattering law (TSL) libraries using modern processing workflows, validate their reliability for Monte Carlo reactor simulations, and systematically assess how different major evaluated nuclear data libraries influence key neutronic parameters, particularly the effective multiplication factor, for both fast- and thermal-spectrum reactor systems.

Methods ACE-format continuous-energy neutron and TSL libraries were processed using the NJOY2016 code based on three primary evaluated nuclear data files: ENDF/B-VII.1, ENDF/B-VIII.1, and JENDL-5.0. These custom-generated libraries were integrated into the Monte Carlo code RMC to perform criticality simulations for two reference models: a self-defined small gas-cooled fast reactor (SGFR) and a micro high-temperature gas-cooled reactor (MHTGR). The official MCNP libraries served as the baseline reference for validation and inter-library comparison.

Results Validation against the MCNP official libraries confirmed that the k_eff deviations for both reactor models using the self-generated libraries remained within ±100×10⁻⁵, demonstrating the correctness and feasibility of the NJOY2016 processing pipeline. Inter-library comparisons revealed distinct system behaviors: for the SGFR, the b71 library yielded k_eff values approximately 400×10⁻⁵ lower than b81, while j50 produced values roughly 500×10⁻⁵ higher. In contrast, the MHTGR exhibited smaller discrepancies, with b71–b81 deviations within ±100×10⁻⁵ and j50–b81 deviations within ±200×10⁻⁵. Sensitivity analysis indicated that the SGFR is significantly more responsive to library selection, primarily driven by cross-section variations in the high-energy regions of ²³⁹Pu, ²³⁵U, ⁹⁰Zr, ⁹¹Zr, and ⁹²Zr, as well as the intermediate-energy region of ²⁴⁰Pu.

Conclusions The successfully validated NJOY2016 library generation workflow provides reliable, modern ACE libraries for advanced reactor neutronics analysis. The pronounced sensitivity of fast-spectrum systems like the SGFR to evaluated data libraries underscores the necessity of carefully selecting and verifying nuclear data, particularly in the medium- and high-energy ranges, to ensure the computational accuracy and safety margins of next-generation fast reactor core designs.