Comparing modelling approaches for a generic nuclear waste repository in salt

This paper contains a comparison of five modelling approaches for a simplified nuclear waste repository in a domal salt formation. It is the result of a four-year collaboration between five international teams on Task F of the DECOVALEX-2023 project on performance assessment modelling. The primary objectives of Task F are to build confidence in the models, methods, and software used for performance assessment (PA) of deep geologic nuclear waste repositories, and/or to bring to the fore additional research and development needed to improve PA methodologies. This work demonstrates how these objectives are accomplished through staged development and comparison of the models and methods used by participating teams in their PA frameworks. Participating teams made a wide range of model assumptions, ranging from compartmentalized networks to full 3D models of the salt formation and repository. Despite differences in the modelling strategies, all models indicate that salt compaction and diffusion of radionuclides in brine are key processes in the repository. For the isothermal spent nuclear fuel and vitrified waste scenario with multiple early failures considered, all models indicate little of the disposed radionuclides will migrate beyond the repository seal over the 100,000-year simulations. In general, the model output quantities have the largest differences over the short term and near the waste. Disparities between the models are believed to be due to differing simplifications from the conceptual model. • Five teams of international experts developed a simplified model concept for a subsurface nuclear waste repository in a salt dome. • Team numerical models for long-term chemical transport have a wide range of modelling assumptions and simplifications. • Results indicate that creep closure of salt and chemical diffusion drive transport of tracers for the scenario considered.