Experiments of water-moderated plutonium metal arrays: part II - analysis

As part of the International Criticality Benchmark Evaluation Project, 26 experiments are analyzed for use as benchmark experiments for validation of computer codes and cross sections. The experiments consisted of water-moderated plutonium metal arrays of size 2x2xN, where N=2-5 or 3x3x3. Of the 26 experiments, 22 were chosen to develop benchmark models which are described in detail, including geometry and material data. A sensitivity study was performed to determine the uncertainty on the keff value due to measurement uncertainties or limits of accuracies. For all benchmark models, keff values were found using MCNP with continuous-energy ENDF/B-V cross sections, MCNP with continuous energy ENDF/B-VI cross sections, and KENO-V.a with 27-group ENDF/B-IV cross sections. For one case, the codes overestimated keff by more than 1%. The experiments were grouped for three parameters: plutonium mass, horizontal spacing, and vertical spacing. For each parameter, a bias value was found. In all cases, MCNP with ENDF/B-VI cross sections had the smallest difference in the average keff. This same result was found when all experiments were grouped together. However, if the one experiment that was significantly overestimated was excluded, the smallest difference in the average keff was found for KENO-V.a with ENDF/B-IV cross sections