A validation study of existing neutronics tools against ZPPR-21 and ZPPR-15 critical experiments.

A study was performed to validate the existing tools for fast reactor neutronics analysis against previous critical experiments. The six benchmark problems for the ZPPR-21 critical experiments phases A through F specified in the Handbook of Evaluated Criticality Safety Benchmark Experiments were analyzed. Analysis was also performed for three loading configurations of the ZPPR-15 Phase A experiments. As-built core models were developed in XYZ geometries using the reactor loading records and drawer master information. Detailed Monte Carlo and deterministic transport calculations were performed, along with various modeling sensitivity analyses. The Monte Carlo simulations were carried out with the VIM code with continuous energy cross sections based on the ENDF/B-V.2 data. For deterministic calculations, region-dependent 230-group cross sections were generated using the ETOE-2/MC-2/SDX code system, again based on the ENDF/B-V.2 data. Plate heterogeneity effects were taken into account by SDX unit cell calculations. Core calculations were performed with the TWODANT discrete ordinate code for the ZPPR-21 benchmarks, and with the DIF3D nodal transport option for the ZPPR-15 experiments. For all six ZPPR-21 configurations where the Pu-239 concentration varies from 0 to 49 w/o and the U-235 concentration accordingly varies from 62 to 0 w/o, the core multiplication factor determined with a 230-group TWODANT calculation agreed with the VIM Monte Carlo solution within 0.20 %{Delta}k, and there was no indication of any systematic bias. The quality of principal cross sections generated with the MC-2 code was comparable to that of VIM cross sections. The overall reactivity effect due to the errors in the 230-group principal cross sections was estimated to be less than 0.05 %{Delta}k. The statistics of the differences between calculated values and specified benchmark experimental values showed similar bias (from -0.28 %{Delta}k to 0.33 %{Delta}k) for MC{sup 2}-2/TWODANT and VIM. This result suggests that the criticality prediction accuracy of MC{sup 2}-2/TWODANT is comparable to VIM. Investigation of group collapsing methods showed that direct generation of broad-group cross sections from MC{sup 2}-2 calculations was not adequate for analysis of ZPPR-21 assemblies. Scalar flux weighting for all cross sections, including anisotropic cross sections, was not sufficiently accurate, either. The use of higher flux moments for anisotropic scattering cross section collapsing reproduced the fine-group results with broad-group calculations. The ZPPR-15 analyses, starting from detailed as-built plate geometries, showed that the plate heterogeneity effect was as large as 1.3 %{Delta}k. Through a series of sensitivity studies, a procedure to generate effective cross sections was developed based on one-dimensional SDX unit cell calculations. With this procedure to account for the plate heterogeneity effect, the core multiplication factor determined with a 230-group DIF3D nodal transport calculation agreed with the VIM Monte Carlo solution within 0.12 %{Delta}k. It was however observed that the calculated values consistently underestimated the criticality by 0.32 %{Delta}k to 0.43 %{Delta}k. The sodium void worths determined from VIM Monte Carlo and DIF3D nodal transport calculations were also very close to each other, but both predictions overestimated the measured void worth by {approx}0.1 %{Delta}k, which amounted to {approx}40% of the measured value. Further investigation is needed to identify the reason for this discrepancy between calculated and measured sodium void worths. In summary, for all nine core configurations of ZPPR-21 and ZPPR-15 analyzed in this study, the deterministic transport solutions showed good agreement with Monte Carlo results. These results indicate that the existing deterministic methods for multigroup cross section generation and core calculation are adequate to use in the initial design stage of Advanced Burner Reactors, for which the startup fuel is expected to be conventional plutonium fuel. However, further verification/validation studies need to be performed for transmutation fuel to assess the impact of relatively large amount of minor actinides. In addition, to take into account the multi-dimensional heterogeneity effects properly, the current homogenization scheme based on one-dimensional cell calculation needs to be improved