Development of a New Criticality Safety Benchmark with HEU Metal and Thick Graphite Reflectors to Support Advanced Reactor Development

Since the era of the Manhattan Project, practitioners working to ensure the safety of fissile material processes have relied on the results of critical experiments to predict subcriticality. As the nuclear industry’s computational abilities have improved over the decades, criticality safety now frequently depends on the validation and verification of neutron transport codes to accurately represent conditions of real fissile material systems in the field. The International Criticality Safety Benchmark Experiments Project, since 1992, has worked to compile data from critical and subcritical experiments in the International Handbook of Evaluated Criticality Safety Benchmark Experiments. This handbook serves as a standardized source of critical experiment benchmarks from all over the world, allowing criticality safety engineers to apply accurate, reliable, and appropriate margins of safety to the processes they oversee. The evaluations also enable the tuning of nuclear data in order to accurately model larger applications like nuclear reactors; continuous improvement of nuclear data and subsequent lowering of uncertainty on commercial power reactor designs are two ways to support the development of new nuclear in the United States.

Today, performing critical experiments requires far more lead time and monetary investment than it did in the past. A handful of evaluations covering new experiments are added to the handbook every year, but there remain many historical experiments performed at now-defunct facilities like the Oak Ridge Critical Experiments Facility (ORCEF) which have not yet been evaluated. This project sought to evaluate a set of experiments performed by John Mihalczo at ORCEF in the 1960s for inclusion in the ICSBEP handbook. These experiments utilized highly-enriched uranium metal (93.14 weight-% ²³⁵U [U-235]) and 10-inch-thick graphite reflectors. The addition of experiments with significant amounts of graphite to the handbook presents the opportunity for reduction in nuclear data uncertainty of advanced reactor models, and in turn may improve the economics of deploying these new reactor designs. This project has focused on the development of a new ICSBEP benchmark and the application of nine new benchmark models to the nuclear data associated with two advanced reactor designs.