Optimizing high-level nuclear waste disposal within a deep geologic repository

Abstract

Many countries produce significant quantities of nuclear waste which will have to be permanently and safely placed in a repository. We develop a mixed integer program that determines where to place each waste package of a specific waste type in a given time period with the goal of minimizing heat load concentration within a repository. Operational constraints include: (1) heat load limitations, (2) location and time at which waste packages can be placed, and (3) the number of waste packages that must be placed based on type and time period. Although applicable to other settings, we use the Yucca Mountain repository in Nevada as a case study. Each of the three objectives used for minimizing heat load concentration improves upon existing greedy and sequential filling methods. Existing filling methods give at least a 17 % to an 873 % higher, i.e., worse, heat load concentration in the repository with respect to these objectives than do optimal methods. Enhancements, i.e., symmetry reduction constraints, perturbations, and heuristics, increase the size of solvable problem instances. This research can be applied to any deep geologic repository planned for operation around the world with slight modifications to incorporate site-specific objectives and constraints.