UC Berkeley

Damaged fuels originated from the accident at the Fukushima-Daiichi Nuclear Power Station, and the spent nuclear fuels from commercial light water reactors (LWRs) in Japan are considered to be disposed of in deep geological repository. For a prospective repository, as part of generic performance assessment, a criticality safety assessment (CSA) should be performed to ensure that the repository system including the engineered barriers and host geological formations remains sub-critical for tens of thousands to millions of years. For various repository concepts, CSA is considered to include three major stages in chronological order: (1) the stage before package failure, (2) the stage after package failure, while fissile nuclides remain within the engineered barrier system (EBS) and in the near-field region, and (3) the stage in which fissile nuclides originated from multiple packages are deposited in far-field host rocks. Defining the model for neutronics calculations plays a central role in CSAs, where conservative assumptions are usually made to cope with various uncertainties and to simplify the model. The aim of this dissertation is to develop neutronics models for different stages in the criticality safety study, and provide basic understandings for the long-term criticality safety for the disposal of spent nuclear fuel in geologic repository.

In the near-field analysis, a neutronics model has been developed for a system consisting of a canister containing fuel debris from Fukushima reactors and the surrounding buffer, in a water-saturated deep geological repository. The fuel debris has been modeled as a hexagonal lattice of spherical fuel particles. Following key observations have been concluded from the numerical results: (a) the calculated neutron multiplicity (keff) is sensitively dependent on assumptions related to moderation, (b) the carbon steel canister plays an important role in reducing the potential for criticality, (c) the maximum keff of the canister-buffer system could be achieved after a fraction of fissile nuclides been released from the canister, and (d) under several assumptions, the maximum keff of the canister-buffer system could be principally determined by the dimension and composition of the canister, not by the initial fuel loading. Based on the preliminary results and findings, a parametric study has been made to identify the optimized lattice parameters for criticality. And the critical mass of damaged fuels for a single canister has been calculated. If this critical mass is used as the maximum canister mass loadings, roughly a thousand canisters are needed to contain the damaged fuels from the three damaged cores. For the LWR spent fuels, a parametric study has been performed to examine spent fuels with different designs and burnup histories. The numerical results indicate that, under the conditions assumed, for all UO2 spent fuels and most of the MOX spent fuels, the single canister model will always be subcritical.

The far-field study has been focusing on neutronic analysis to examine the criticality conditions for uranium depositions in geological formations which result from geological disposal of damaged fuels from Fukushima reactors. Neutronics models are used to evaluate the keff and critical mass for various combinations of host rock and geometries. The present study has revealed that the planar fracture geometry applied in the previous criticality safety assessment for geological disposal would not necessarily yield conservative results against the homogeneous uranium deposition. It has been found that various far-field critical configurations are conceivable for given conditions of materials and geological formations. Prior to knowing the site location, some important points for selecting a site for criticality safety can be suggested. These include: (a) iron existing in the host rock reduces the likelihood of criticality significantly; (b) low host rock porosity is preferred for criticality safety; (c) the conservatism could change when comparing heterogeneous geometries for different fracture apertures; and (d) the importance of the mass of the deposition increases when it is smaller.

As part of the improvement for the models developed in the far-field analysis, preliminary works on uranium depositions in randomly fractured rocks have been presented. The randomly fractured geometry could fundamentally influence the far-field criticality, because the system’s keff value sensitively depends on the fracture aperture and the depositions at fracture intersections. No previous work has been made to study the effect of random geometry in the context of the long-term criticality safety in a geologic repository. Different numerical schemes have been developed and compared for the direct sampling of uranium depositions in randomly fractured rocks using MCNP. A general literature review of existing methods for neutron transport problems with random processes has been made. And the analytical Feinberg-Galanin-Horning (FGH) method has been derived and tested for a numerical example.