UC Berkeley

A repository for used nuclear fuel (UNF) must employ several barriers to reduce the risk of a lethal dose to the biosphere over geological periods of time. It is presumed that waste canisters can be engineered to prevent a critical configuration of fissile material by design. However, given the quantity of UNF in interim storage in the United States intended for final deposition, a comprehensive failure scenario involving the repository-wide leaching of fissile isotopes from used fuel, radionuclide transport in groundwater, and re-concentration in the far-field may eventually present a concern for criticality. Such a criticality event may create a pathway for fission products to the environment if the host rock is compromised.

This dissertation explores the impact of far-field criticality on repository performance over long periods of time. It attempts to integrate all necessary processes and considerations under a notion of conservatism favoring the formation of a critical mass from used nuclear fuel. A critical deposition is hypothesized to pose a threat to the material integrity of the host rock through the steady release of heat over time, as opposed to the explosive releases of energy considered in previous studies. In particular, the role of thermal creep is evaluated as a steady-state failure metric, which emphasizes the combined role of reactivity feedback mechanisms and heat transfer in determining the temporal extent of chain reactions underground.

The failure scenario is isolated among other relevant features, events, and processes as the primary concern of the performance assessment for a crystalline granitic repository. Repository loading cases are established based on the uncertainty in the current inventory of used nuclear fuel. These source terms are used in a radionuclide transport model that incorporates the Latin hypercube sampling method to probe the uncertainty in the total accumulation of uranium from advection in fractures. The configurations of fissile material in host rock required for criticality are evaluated based on representative precipitate compositions obtained from the transport analysis, where the role of a reducing and reflecting region of shale is emphasized.

Given the nature of the critical depositions as both open reactor systems and porous media, a thermo-hydrological analysis is employed to evaluate the coupled heat and mass transfer in the geology when energy is released from fission reactions. This analysis is motivated by the assumption that reactivity behavior with evolving water content and temperature must be fully coupled, and boundary conditions are imposed to evaluate a maximal extent of desaturation. The observed changes in fluid content and densities in the system over time are used to guide an integrated neutronics evaluation incorporating the Doppler effect.

Feedback coefficients covering the simultaneous effects of Doppler broadening, the loss of moderator, and the arriving plume of uranium solutes from the repository array are applied to a quasi-steady-state heat transfer model. This model provides insight on the dynamic evolution of system temperature over time parametrized on the source term of incoming fissile material, and results are applied to a thermal creep model to evaluate total integrated strain and system failure. Given the time dependence of the creep phenomenon, optimal source terms of fissile material may be conducive to failure via creep. However, these mass fluxes exceed those observed in the transport analysis, which employed assumptions to maximize these quantities.

The conditions under which repository performance can be impacted by steady-state criticality prove to be heavily dependent on the repository inventory and the specific choice of host rock. This mandates that a site intended for final UNF disposition be subject to increasing scrutiny in proportion to the amount of waste that is loaded. Based on the results of the study, it is strongly proposed that the threat of a far-field criticality event can be dramatically reduced by the introduction of depleted uranium to waste canisters. Overall, given the compounded effect of utilizing many heavily conservative assumptions in the study, the far-field criticality phenomenon is not deemed to be of especial importance to repository engineering, especially with additional enhancements to existing engineered barrier systems.