Engineered Barrier Systems (EBS) are a key element of deep geological repositories (DGR) and play an important role in safely isolating radioactive materials from the ecosystem. In the environment of a DGR, gases can be generated due to several factors, including canister corrosion. If the gas production rate exceeds the diffusion rate, pore pressures may increase, potentially inducing structural deterioration that impairs the function of the buffer material. Therefore, understanding the hydraulic-mechanical behavior of EBS due to gas generation is essential for evaluating the longterm stability of DGR. This study employed X-ray computed tomography (CT) technology to observe cracks created inside the buffer material after laboratory-scale gas injection experiments. After CT scanning, we identified cracks more clearly using an image analysis method based on machine learning techniques, enabling us to examine internal crack patterns caused by gas injection. In the samples observed in this study, no cracks were observed penetrating the entire buffer block, and it was confirmed that most cracks were created through the radial surface of the block. This is similar to the results observed in the LASGIT field experiment in which the paths of the gas migration were observed through the interface between the container and the buffer material. This study confirmed the applicability of high-resolution X-ray CT imaging and image analysis techniques for qualitative analysis of internal crack patterns and cracks generated by gas breakthrough phenomena. This is expected to be used as basic data and crack analysis techniques in future research to understand gas migration in the buffer material.

• Jung-Tae Kim(Korea Atomic Energy Research Institute (KAERI)) Corresponding author
• Minhyeong Lee(Korea Atomic Energy Research Institute (KAERI))
• Jin-Seop Kim(Korea Atomic Energy Research Institute (KAERI))