The engineered barrier system (EBS) is an indispensable element of a deep geological repository (DGR) designed to prevent the discharge of radioactive materials into the environment. The buffer material is a vital component of the EBS by creating a physical and chemical barrier that prevents the migration of radioactive materials. In the disposal environment, gases can be generated from the corrosion of the canister. When the gas generation rate exceeds the diffusion rate, the buffer material’s performance can deteriorate by the physical damage induced by the increase in pore pressure. Therefore, understanding the EBS’s behavior under gas generation conditions is crucial to guarantee the longterm safety and performance of the DGR. Lab-scale and field-scale experiments have been conducted to examine the stability of the buffer material concerning gas generation and movement by the previous researchers. To evaluate long-term stability for more than 100,000 years, it is essential to assess stability using a numerical model verified by these experiments. This study investigated the effect of interfacial characteristics on the numerical modeling accuracy of experimental simulation while verifying a numerical model through field-scale experimental results. The findings of this study are expected to furnish fundamental data for establishing numerical analysis guidelines for the longterm stability assessment of disposal systems.

The analysis of uranium migration is crucial for the accurate safety assessment of high-level radioactive waste (HLW) repository. Previous studies showed that the migration of the uranium can be affected by various physical and chemical processes, such as groundwater flow, heat transfer, sorption/ desorption and, precipitation/dissolution. Therefore, a coupled Thermal-Hydrological-Chemical (THC) model is required to accurately simulate the uranium migration near the HLW repository. In this study, COMSOL-PHREEQC coupled model was used to simulate the uranium migration. In the model, groundwater flow, heat transfer, and non-reactive solute transport were calculated by COMSOL, and geo-chemical reaction was calculated by PHREEQC. Sorption was primarily considered as geo-chemical reaction in the model, using the concept of two-site protolysis nonelctrostatic surface complexation and cation exchange (2 SP NE SC/CE). A modified operator splitting method was used to couple the results of COMSOL and PHREEQC. Three benchmarks were done to assess the accuracy of the model: 1) 1D transport and cation exchange model, 2) cesium transport in the column experiment done by Steefel et al. (2002), and 3) the batch sorption experiment done by Fernandes et al. (2012), and Bradbury and Baeyens (2009). Three benchmark results showed reliable matching with results from the previous studies. After the validation, uranium 1D transport simulation on arbitrary porewater condition was conducted. From the results, the evolution of the uranium front with sequentially saturating sites was observed. Due to the limitation of operator splitting method, time step effect was observed, which caused the uranium to sorbed at further sites then it should. For further study, 3 main tasks were proposed. First, precipitation/ dissolution will be added to the reaction part. Second, multiphase flow will be considered instead of single phase Darcy flow. Last, the effect of redox potential will be considered.