Improving Waste Repository Simulations with a Fracture-Enhanced Conforming Reproducing Kernel Method

The safe disposal of spent nuclear fuel in deep geologic repositories requires accurate predictions of fracture in repository geomaterials. This is of significant interest to the Department of Energy and nuclear energy providers as they seek to ensure the long-term isolation of spent nuclear fuel and prevent the migration of radionuclides into the environment. Given the dynamic nature of repositories over hundreds to thousands of years and the potential for fracture during deformation, reliable simulation methods are crucial for predicting permeability and porosity changes over time, thereby impacting the safety assessment of disposal systems. The overall objective of Phase I is to address the need for robust and accurate fracture prediction in repository geomaterials by enhancing the Conforming Reproducing Kernel method. This will be achieved through collaboration with researchers at Sandia National Laboratories, focusing on demonstrating proof-of-concept with an analysis of a storage room in the Waste Isolation Pilot Plant. Specific objectives include mitigating mesh dependencies and efficiently simulating material separation while preserving mass and volume, which are common issues in fracture simulation. The versatility of the Conforming Reproducing Kernel method makes it a top candidate for solving these challenges. A reproducing kernel strain regularization approach will be implemented to address mesh dependencies, and conforming kernels can adapt to provide clean separation of materials during fracture. Success will be demonstrated by achieving mesh-insensitive fracture predictions while accurately simulating repository collapse and closure scenarios. The successful completion of Phase I will pave the way for significant advancements in repository performance analysis, facilitating the expansion of nuclear energy as a sustainable energy source both domestically and internationally. Furthermore, it will position us to provide valuable expertise and solutions for future repository projects. Beyond repository analyses, the developed capabilities will find applications in diverse fields such as mining, petroleum engineering, and hydrogeology. Additionally, integrating fracture simulation into the Conforming Reproducing Kernel method will streamline the design-to-analysis workflow, leading to substantial reductions in analysis efforts for dynamic problems.