Modeling the Transport of Corrosive Species through Highly Compacted Bentonite Clay: Optimizing Boundary Conditions

Abstract:

Canada’s plan for the long-term management of spent nuclear fuel in a deep geologic repository (DGR) designed by the Canadian Nuclear Waste Management Organization (NWMO) consists of a steel used fuel cannister (UFC) with a copper coating surrounded by a highly compacted bentonite (HCB) block. Under some conditions, sulphide, produced through microbial processes at or beyond rock-bentonite interface could diffuse through the HCB and corrode the UFC’s copper coating. The HCB is designed to limit the flux of corrosive species from the surrounding subsurface environment. A thorough understanding of sulphide transport through the HCB is required to estimate UFC corrosion rates and ensure the long-term safety of the DGR. However, sulphide transport through the HCB under DGR conditions is driven by a strongly coupled, complex suite of processes including: heat transport from the UFC, multiphase flow from groundwater re-saturation, liquid-vapour water phase changes, water vapour transport, chemical reactions between sulphide and bentonite compounds, and aqueous mass transport.

Implementing these processes in a numerical model often requires simplifying assumptions. Additionally, model results can be sensitive to imposed boundary conditions (BCs) (e.g., saturation BC, model’s total depth). Optimizing BCs is important for predicting DGR environment behaviour. In this study, a 2-D, multiphase, non-isothermal model has been developed to examine sulphide transport through the HCB under anticipated conditions of Canada’s DGR. This model is then tested with various BCs as a part of verification and validation process. Among all cases considered, saturation BC and model depth are found to be sensitive factors impacting DGR re-saturation and UFC temperature respectively. These impacts can affect predictions for full saturation time and UFC corrosion rates therefore careful consideration of BCs is needed during DGR model development and implementation.