Cross-Evaluation of the PU-MET-FAST criticality benchmark experiments in ICSBEP handbook (by the end of 2022)

•Cross-evaluation method has been tested which could evaluate the biases of criticality benchmark experiments.•PU-MET-FAST criticality benchmark experiments in ICSBEP handbook have been cross-evaluated systematically.•The results indicate that there are many contradictions between the retrospective evaluation experiments in PU-MET-FAST category. Criticality experiments play an important role in the fields of criticality safety, neutron transport program validation and nuclear data validation. ICSBEP handbook collects about 5000 criticality benchmark experiments. Most of the criticality benchmark experiments in ICSBEP handbook are retrospective evaluation experiments, they are evaluated long after the experiments and their evaluations may be based on inaccurate or incomplete data. When these experiments are used in nuclear data validations and cross section adjustments, contradictions between some criticality benchmark experiments have been found. Therefore, it is necessary to analyze whether these retrospective evaluation criticality benchmark experiments are applicable in the fields of nuclear data validations (including cross section adjustments) and validations of nuclear system design. To solve this problem, we have developed cross-evaluation method of criticality benchmark experiments, which could evaluate the biases of criticality benchmark experiments from the point of statistics. In this article, we used this method to cross-evaluate the PU-MET-FAST criticality benchmark experiments in ICSBEP handbook. The results indicate that there are many contradictions between the retrospective evaluation experiments in the PU-MET-FAST category. Because of the limited number of experiments in this category, the number of experiments that are similar to each other after cluster analysis is small, which reduces the effectiveness of the cross-evaluation. There are only 1 or 2 experiments in some clusters. Even if there are contradictions in these clusters, it is impossible to judge which experiment has a higher quality. Based on these results, it is recommended to do new PU-MET-FAST criticality benchmark experiments to resolve contradictions between these retrospective evaluation experiments and to help validate the nuclear data more effectively.