Benchmark of Advanced Burner Test Reactor model using MCNPX 2.6.0 and ERANOS 2.1

Abstract Significant research is currently being performed whereby fast reactor cores have been designed to burn transuranic materials reducing the volume and long-term radiotoxicity of spent nuclear fuel. These core and depletion models depend on various computer codes. This research used MCNPX 2.6.0 and ERANOS 2.1 to model a standard 250 MW Advanced Burner Test Reactor (ABTR) core. The intent was to benchmark criticality and burnup results from a stochastic Monte Carlo code and a deterministic depletion code using a standard ABTR model created by Argonne National Laboratory. Because each of these codes solves the transport and burnup problem differently, there is a need to benchmark the core models in order to verify results and identify root causes for significant differences in results between codes. Flux calculations in ERANOS were performed using diffusion theory, Legendre polynomial approximations (using the VARIANT module) and discrete ordinates methods. The k-effective for the higher order transport models remained within 1000 pcm of the MCNPX model. The difference between the total heavy nuclide mass balance in ERANOS using the various flux calculations and the MCNPX depletion model was less than 0.4% out to a burnup of 1095 days (67.45 GWd/MTHM). The percent delta between the codes as a fraction of the fissioned mass was 1.34%. For the isotopes with large concentrations, such as 238 U and 239 Pu, the mass differences were 0.38% and 0.01% respectively. The mass difference for 241 Am was also small at 0.42%. Notable isotopes in small quantities with larger mass differences were 242 Am, 242 Cm, 243 Cm and 246 Cm where differences ranged from 0.1 to 0.2% after 26 days and increased to 11–136% at 1095 days.