Subcritical measurements for a HEU storage vault at the Y-12 plant using the ²⁵²Cf-source-driven frequency-analysis method

The tube vaults at the Oak Ridge Y-12 Plant were constructed in 1975 out of massive amounts of concrete to safely store large quantities of highly enriched uranium (HEU) metal. At the time of construction, criticality safety specialists were concerned about the concrete drying out over time. The difference between dry and wet concrete is the absence of unbound water, which results in a change in the hydrogen content of the concrete by a factor of 2. For this system, dry concrete increases the neutron interaction between the uranium metal castings in the array. With the development of the ²⁵²Cf-source-driven frequency-analysis method for the determination of subcriticality, it became feasible to perform a more sensitive, meaningful, and relevant subcriticality measurement of this vault configuration with HEU. For this vault loading, the average mass of the castings was 18.75 kg of uranium, and the ²³⁵U enrichment was 93.2 wt %. These measurements are interpreted with a general Monte Carlo model rather than limited-point kinetics models that were used in previous measurements. The Monte Carlo code, MCNP-DSP, is a major modification of MCNP4a that will better calculate the time and frequency-analysis parameters that are related to the second moment of the neutron population. For this loading of Vault 16 at the Oak Ridge Y-12 Plant, the neutron multiplication factor obtained from these measurements was 0.7899 ± 0.0028. The bias in the calculated neutron multiplication factor for MCNP-DSP and ENDF/B-IV cross section is +0.0002 for the assumption of dry concrete and -0.050 for the assumption of wet concrete. Thus for the purposes of criticality safety assessments, the best assumption for this vault is that it is composed of dry concrete. This measurement and analysis shows how the ²⁵²Cf-source-driven frequency-analysis method with MCNP-DSP can be used to obtain the subcritical neutron multiplication factor and the bias in calculational methods for criticality safety for an in-plant configuration of fissile material. The determination of the neutron multiplication factor from the measured data did not depend on knowing the condition of the vault concrete. The use of a more general Monte Carlo model in place of the limited-point kinetics models extends the application of this method.