The deep geological repository for high-level radioactive waste requires careful consideration due to its exceptionally long-term implications, making long-term impact assessments essential. However, evaluating the long-term effects of deep geological repositories using performance assessment models is accompanied by various sources of uncertainty, including uncertainties about the future, model uncertainties, and uncertainties associated with input data. These multifaceted uncertainties arise from factors such as a lack of current knowledge, contributing to a complex web of unpredictability. Managing, mitigating, and ultimately eliminating these uncertainties is crucial for ensuring the performance and safety of deep geological repositories. Currently, the Korea Radioactive Waste Agency (KORAD) is developing a complex behavior model that incorporates Thermal-Hydraulic-Mechanical-Chemical (THMC) phenomena within the disposal system using PFLOTRAN. To address model uncertainties and furthermore input data uncertainties for this intricate model, an automated sensitivity analysis system has been developed. This automated system operates without human intervention, facilitating tasks such as automatic parameter adjustments and the quantification of uncertainties. Furthermore, this system aids in identifying key factors characterized by substantial uncertainties. Through this system, it is possible to examine concentration distributions in each components of the deep disposal facility in response to changes in input data and to identify factors with significant uncertainties. The sensitivity results and key uncertainty factors obtained through this system are intended to be used for optimizing uncertainties in future research and development.

Properties of bentonite, mainly used as buffer and/or backfill materials, will evolve with time due to thermo-hydro-mechanical-chemical (THMC) processes, which could deteriorate the long-term integrity of the engineered barrier system. In particular, degradation of the backfill in the evolution processes makes it impossible to sufficiently perform the safety functions assigned to prevent groundwater infiltration and retard radionuclide transport. To phenomenologically understand the performance degradation to be caused by evolution, it is essential to conduct the demonstration test for backfill material under the deep geological disposal environment. Accordingly, in this paper, we suggest types of tests and items to be measured for identifying the performance evolution of backfill for the Deep Geological Repository (DGR) in Korea, based on the review results on the performance assessment methodology conducted for the operating license application in Finland. Some of insights derived from reviewing the Finnish case are as follows: 1) The THMC evolution characteristics of backfill material are mainly originated from hydro-mechanical and/or hydrochemical processes driven by the groundwater behavior. 2) These evolutions could occur immediately upon installation of backfill materials and vary depending on characteristics of backfill and groundwater. 3) Through the demonstration experiments with various scales, the hydro-mechanical evolution (e.g. advection and mechanical erosion) of the backfill due to changes in hydraulic behavior could be identified. 4) The hydro-chemical evolution (e.g. alteration and microbial activity) could be identified by analyzing the fully-saturated backfill after completing the experiment. Given the findings, it is judged that the following studies should be first conducted for the candidate backfill materials of the domestic DGR. a) Lab-scale experiment: Measurement for dry density and swelling pressure due to saturation of various backfill materials, time required to reach full saturation, and change in hydraulic conductivity with injection pressure. b) Pilot-scale experiment: Measurement for the mass loss due to erosion; Investigation on the fracture (piping channel) forming and resealing in the saturation process; Identification of the hydro-mechanical evolution with the test scale. c) Post-experiment dismantling analysis for saturated backfill: Measurement of dry density, and contents of organic and harmful substances; Investigation of water content distribution and homogenization of density differences; Identification of the hydro-chemical evolution with groundwater conditions. The results of this study could be directly used to establishing the experimental plan for verifying performance of backfill materials of DGR in Korea, provided that the domestic data such as facility design and site characteristics (including information on groundwater) are acquired.

Bentonite, primarily composed of montmorillonite, plays a crucial role as one of the engineering barrier materials in a deep geological repository (DGR). The advantageous properties of montmorillonite, such as its swelling capacity, low permeability, and low thermal conductivity, make it a key component as a buffer material for isolating high-level radioactive waste from the surrounding natural environment. It has been known that the stability of montmorillonite is influenced by a wide range of pressure-temperature-composition (P-T-X) conditions relevant to the DGR environment. When considering potential geological events, of notable concerns are its interactions with groundwater or seawater at elevated temperatures, leading to safety hazards within the system. In this study, therefore, we investigated the hydration and dehydration reactions of Ca-montmorillonite with saline fluids such as NaCl and KCl solutions at elevated pressures and temperatures by conducting in-situ X-ray diffraction experiments using both a capillary sample heating cell and a resistive-heated diamond anvil cell. As a result, we observed different hydration states of montmorillonite depending on the chemical composition of fluids, i.e., tri-hydrated layers in NaCl and bi-hydrated layers in KCl solutions, respectively. Furthermore, we identified that montmorillonite undergoes distinct stepwise dehydration with increasing temperature, and the dehydration temperature of montmorillonite significantly increases with increasing water pressure. Consequently, this study would provide insights into the stability of hydrated montmorillonite in a seawater-dominated fluid environment and its implications for the long-term safety of the disposal system.

The seven-year research project entitled “Development of workflow for integrated 3D geological site descriptive modeling” is being carried out from 2023. This research is funded by Ministry of Trade, Industry, and Energy (MOTIE). Progress of the research is discussed here. The integrated 3D geological SDM (site descriptive model; GSDM hereafter) consists of three part; 1) three dimensional representation of geologic elements, 2) database for material properties and modeling results from SDMs of other disciplines (e.g., rock mechanics), and 3) a visualization tool for geology, material properties and modeling results. The GSDM is comparable to the GDSMs of SKB and POSIVA in its representation of geology by volume of geologic elements. However, our GSDM is different in that extra information of material properties and an extra tool for visualization is included in the GDSM. The rationale for incorporating material properties and a visualization tool into the GSDM is to expedite the development of the GSDM and SDMs of other disciplines by allowing single institution to integrate database and visualization with the GSDM. SKUA-GOCAD is used for representation of geologic surfaces for ductile and brittle shear zones, and also for surfaces for delineation of volumes of rock units. We have adopted SKUAGOCAD because the program offers powerful functions of interpolation including borehole data and geophysical prospecting. So far, we have tested the program for five different geologies, including sedimentary, high-grade metamorphic, and intrusive igneous geology. The test results are promising. Incorporation of data and modeling results for the SDMs of other disciplines is at conceptual stage. The working conceptual model involves the following steps, 1) to provide the modeler of other disciplines with surface information representing geologic elements, 2) the modeler returns not only material properties but the results of numerical analysis, and 3) incorporation of material properties and modeling results into database. Since the numerical codes in other disciplines adopt different types of formats for 3D geology, we plan to adopt the widely used FEM format prepared by Gmsh. The visualization tool will also adopt Gmsh for graphical representation of 3D geology as well as database for material properties and modeling results. When the working model of GSDM becomes available, rapid and significant progress is expected in the SDMs of other disciplines and related areas, for example, geotechnical investigation for deep geological repository.

With the importance of permanent disposal of high-level radioactive waste (HLW) generated in Korea, the deep geological disposal system based on the KBS-3 type is being developed. Since the deep geological repository must provide the long-term isolation of HLW from the surface environment and normal habitats for humans, plants, and animals, it is essential to assess the longterm performance of the disposal facility considering thermal-hydraulic-mechanical-chemical (TH- M-C) evolution. Decay heat dissipated from HLW contained in the canister causes an increase in temperature in the adjacent area. The requirement for the maximum temperature is established in consideration of the possibility of bentonite degradation. Therefore, when designing the repository, the temperature in the region of interest should be identified in detail through the thermal evolution assessment to ensure that the design requirement is satisfied. In the thermal evolution analysis, it is needed to evaluate the temperature distribution over the entire area of the disposal panel to consider the heat generated from both a single canister and adjacent canisters. Computational fluid dynamics (CFD) codes are widely used for detailed temperature analysis but are limited to simulating a wide range. Accordingly, in this study, we developed an analytical solution-based program for efficiently calculating the temperature distribution throughout the deposition panel, which is based on threedimensional heat conduction equations. The code developed can assess the temperature distribution of engineered and natural barrier systems. Principal parameters to be inputted are as follows: (a) geometry of the panel (e.g. width, length, height, spacing between canisters), (b) geometry of the canister (e.g. diameter, height), (c) thermal properties of bentonite and host-rock, (d) initial conditions (e.g. residual heat, temperature), and (e) time information (e.g. canister emplacement rate, time-interval, period). Through the calculation for the conceptual problem of a deposition panel capable of accommodating 900 (i.e. 30×30) canisters, it was confirmed that the program can adequately predict when and where the maximum temperature will occur. It is expected that the overall temperature distribution within the panel can be obtained by the evaluation of the entire region using this program reflecting the detailed design of the repository to be developed in the future. In addition, the thermal evolution analysis considering the influence of other canisters can be performed by applying the results as boundary conditions in the CFD analysis.

In 2012, POSIVA selected a bentonite-based (montmorillonite) block/pellet as the backfilling solution for the deposition tunnel in the application for a construction license for the deep geological repository of high-level radioactive waste in Finland. However, in the license application (i.e. SC-OLA) for the operation submitted to the Finnish Government in 2021, the design for backfilling was changed to a granular mixture consisting of bentonite (smectite) pellets crushed to various sizes, based on NAGRA’s buffer solution. In this study, as part of the preliminary design of the deep geological repository system in Korea, we reviewed history and its rationale for the design change of Finland’s deposition tunnel backfilling solution. After the construction license was granted by the Finnish Government in 2015, POSIVA conducted various lab- and full-scale in-situ tests to evaluate the producibility and performance of two design alternatives (i.e. block/pellet type and granular type) for backfilling. Principal demonstration tests and their results are summarized as follows: (a) Manufacturing of blocks using three types of materials (Friedland, IBeco RWC, and MX-80): Cracking and jointing under higher pressing loads were found. Despite adjusting the pressing process, similar phenomena were observed. (b) 1:6 scale experiment: Confirmation of density difference inhomogeneity due to the swelling of block/pellet backfill and void filling due to swelling behavior into the mass loss area of block/pellet. (c) FISST (Full-Scale In situ system Test): Identification of technical unfeasibility due to the inefficient (too manual) installation process of blocks/pellets and development of an efficient granular in-situ backfilling solution to resolve the disadvantage. (d) LUCOEX-FE (Large Underground Concept Experiments – Full-scale Emplacement) experiment: Confirmation of dense/homogeneous constructability and performance of granular backfilling solution. In conclusion, the simplified granular backfill system is more feasible compared to the block/ pellet system from the perspective of handling, production, installation, performance, and quality control. It is presumed that various experimental and engineering researches should be preceded reflecting specific disposal conditions even though these results are expected to be applied as key data and/or insights for selecting the backfilling solution in the domestic deep geological repository.

In order to ensure the long-term safety of a deep geological repository, the performance assessment of the Engineered Barrier System (EBS) considering a thermal process should be performed. The maximum temperature at the side wall of a disposal canister for the technical design requirement should not exceed 100°C. In this study, the thermal modelling was conducted to analyze the effects of the thermal process from a disposal canister to the surrounding near-field host rock using the PFLOTRAN code. The mesh was generated using the LaGriT code and the material properties were assigned by applying the FracMan code. Initial conditions were set as the average geothermal gradient (25.7°C/km) and an average surface temperature (14.7°C) in Korea. The highest temperature was observed at the middle of the canister side wall. The temperature of the buffer was lower than that of the canister, and the temperature increase of the deposition tunnel and the host rock was insignificant due to the lower effect of the heat source. The result of the thermal evolution of the EBS represented the highest thermal effects in the vicinity of the canister. In addition, the thermal effects were largely decreased after 10 years of the entire simulation period. It demonstrated that the model took 3 years to heat up the buffer around the canister. The temperature at the canister side wall increased until 3 years and then decreased after that time. This is because that the radioactive decay heat from the heat source was emitted enough to raise the overall temperature of the EBS by 3 years. However, the decay heat rate of the canister decreased exponentially with the disposal time and then its decay heat was not emitted enough after 3 years. In conclusion, the peak temperature results of the EBS were lower than 70°C to meet the technical design requirement.

This study aimed to provide better understanding of the bedrock aquifer bacterial communities and their functions in deep geological repository (DGR) environment. Two study sites of uranium deposits in the Ogcheon Metamorphic Belt were selected: Boeun and Guemsan. From two study sites, six groundwater samples were obtained with different boreholes and depths: OB1 (Boeun, 25 m), OB3 (Boeun, 80 m), GS1 (Guemsan, 25 m), GS2 (Guemsan, 85-90 m), GS3-I (Guemsan, 32- 38 m), GS3-II (Guemsan, 70-74 m). The physicochemical properties of groundwater were analyzed by multi-parameter sensors, ion chromatography (IC), and inductively coupled plasma optical emission spectroscopy (ICP-OES). Illumina Miseq sequencing was performed to investigate bacterial community in six groundwater samples. In addition, the number of sulfate-reducing bacteria (SRB) was quantified by a quantitative PCR (qPCR). Bacterial community composition varied in response to boreholes and depths. A total of 14 different phyla and 36 classes were detected from six groundwater samples. Overall, Proteobacteria, Actinomycetota, and Bacteroidota were dominant in the phylum level. SRB and iron-reducing bacteria (IRB) were detected in all groundwater samples even though organic carbon sources were not abundant (0.7-3.3 mg-total organic carbon/L). This result shows a potential to immobilize uranium in DGR environment. In particular, SRB, Desulfosporosinus fructosivorans and Humidesulfovibrio mexicanus were mainly detected in GS1 and GS2 groundwater samples, which attributed to higher dissimilatory sulfite reductase functional gene copy number in GS1 and GS2 groundwater samples. Statistical analysis was performed to understand the correlation between environmental factors and core bacterial species. Dissolved oxygen (DO), Fe, and Mn concentrations were positively correlated with Curvibacter fontanus while Undibacterium rivi had a negative correlation with pH. These results indicate that bacterial community could be changed in response to environmental variation. Further study with a greater number of samples is necessary to obtain statistically reliable and meaningful results for a safe DGR system.

The thermal evaluations for the conceptual design of the deep geological repository considering the improved modeling of the spent fuel decay heat were conducted using COMSOL Multiphysics computational program. The maximum temperature at the surface of a disposal canister for the technical design requirement should not exceed 100°C. However, the peak temperature at the canister surface should not exceed 95°C considering the safety margin of 5°C due to several uncertainties. All thermal evaluations were based on the time-dependent simulation from the emplacement time of the canister to 100,000 years later. In particular, the heat source condition was set to the decay heat rate and axial decay heat profile of the PLUS7 fuel with 4.0wt% U-235 and 45 GWD/MTU. The thermal properties of the granitic rock in South Korea were applied to the host rock region. For the reference design case, the cooling time of the SNF was set to 40 years, the distance between the deposition holes 8 meters and that between the deposition tunnels 30 meters. However, the peak temperature at the canister surface at 10 years was 95.979°C greater than 95°C. This design did not meet the thermal safety requirement and needed to be modified. For the first modified case, when the distance between the deposition tunnels was set to 30 meters, three cooling time cases of 40, 50 and 60 years and five distances of 6, 7, 8, 9 and 10 meters between the deposition holes were considered. The design with the distances of 9 and 10 meters between the deposition holes for the cooling time of 40 years and all five distances for 50 and 60 years were less than 95°C. For the second modified case, when the distance between the deposition holes was set to 8 meters, three cooling time cases of 40, 50 and 60 years and five distances of 20, 25, 30, 35 and 40 meters between the deposition tunnels were considered. The design with the distances of 35 and 40 meters between the deposition tunnels for the cooling time of 40 years, the distances of 25, 30, 35 and 40 meters for 50 years and all five distances for 60 years were less than 95°C. As a result, the peak temperature at the canister surface decreased as the cooling time and the distance between the deposition holes and the tunnels increased.

Spent nuclear fuels (SNFs) are stored in nuclear power plants for a certain period of time and then transported to an interim storage facility. After that, SNFs are finally repackaged in a disposal canister at an encapsulation plant for final disposal. Finland and Sweden have already completed the design of the spent nuclear fuel encapsulation plant. In particular, Finland has begun the construction of the encapsulation plant and is on the verge of completion. Korea Radioactive Waste Agency (KORAD) is conducting a conceptual design of a deep geological repository for SNFs. Conceptual design of the encapsulation plant is part of the research activity. It is highly required to draft an operation process of the encapsulation plant before an actual design activity. As part of the activity, Finnish design concept of the encapsulation plant and experience were thoroughly reviewed. Finally a preliminary concept of the operation process was proposed considering Korean unique situations such as the volume of SNFs estimated to be disposed of, types of transportation cask and other considerations.

APro, developed in KAERI for the process-based total system performance assessment (TSPA) of deep geological disposal systems, performs finite element method (FEM)-based multiphysics analysis. In the FEM-based analysis, the mesh element quality influences the numerical solution accuracy, memory requirement, and computation time. Therefore, an appropriate mesh structure should be constructed before the mesh stability analysis to achieve an accurate and efficient process-based TSPA. A generic reference case of DECOVALEX-2023 Task F, which has been proposed for simulating stationary groundwater flow and time-dependent conservative transport of two tracers, was used in this study for mesh stability analysis. The relative differences in tracer concentration varying mesh structures were determined by comparing with the results for the finest mesh structure. For calculation efficiency, the memory requirements and computation time were compared. Based on the mesh stability analysis, an approach based on adaptive mesh refinement was developed to resolve the error in the early stage of the simulation time-period. It was observed that the relative difference in the tracer concentration significantly decreased with high calculation efficiency.

For the performance analysis of deep geological repository systems, numerical simulation with multi-physics is required, which specifically covers Thermal (T), Hydraulic (H), and Mechanical (M) behaviors in the disposal environment. Numerous simulation models have been developed so far, each of which varies in the approach and methodology for solving THM problems. Fully-coupled THM simulation codes such as ROCMAS, THAMES, and CODE_BRIGHT were mainly developed in the initial stage of DEvelopment of COupled models and their VALidation against EXperiments (DECOVALEX), with the advantage of thorough calculations consisting of correlated several variables on different physics. Due to the difficulty of solving the complex Jacobian Matrix and the following burden for the computational calculation, weakly-coupled THM models have been suggested in recent researches: TOUGH2-MP with FLAC3D, TOUGH2 with UDEC and OpenGeoSys with FLAC3D. This methodology of loose coupling allows the practical use of computational code optimized for each physics, thereby increasing the efficiency in simulation. However, these suggested models require two different numerical codes to calculate THM behaviors, which leads to several inherent issues: compatibility during maintenance, updating and dependency between two codes. In this study, therefore, the authors build a unified code for simulating THM behaviors in the deep geological repository. The concept involves the iterative sequential coupling between TH and M for calculation efficiency. As having developed the simulation code, High-level rAdiowaste Disposal Evaluation System (HADES), to describe TH behavior based on Multi-physics Object-Oriented Simulation Environment (MOOSE) software, the authors make a milestone to develop and couple the MOOSE-based new code for M behavior as Sub-app, with the previous HADES set to be Main-app. New model for M behavior will be verified with the benchmark case of DECOVALEX-THMC Task D, comparing the mechanical simulation results: stress evolution over time, profiles of stress and vertical displacement. The existing simulation results from HADES will also be updated with the coupled calculations, with regard to temperature and saturation. Additionally, the effective stress evolution can be assessed in terms of repository’s stability with Spalling Strength and Mohr-Coulomb failure criterion. This concept for new simulation model has its meaning in that it aims to demonstrate the specific methodology of loosely coupling multi-physics in unified simulation code and analyze THM complex interactions with considering mutual influence on various physics. It is expected that HADES can be renewed as an integral simulation model for deep geological repository systems by possessing the capacity for analyzing and assessing mechanical behavior.

Chemical environments of near-field (Engineered barrier and surrounded host rock) can influence performance of a deep geological repository. The chemical environments of near-field change as time evolves eventually reaching a steady state. During the construction of a deep geological repository, O2 will be introduced to the deep geological repository. The O2 can cause corrosion of Cu canisters, and it is important predicting remaining O2 concentration in the near-field. The remaining O2 concentration in the near field can be governed by the following two reactions: oxidation of Cu(I) from oxidation of Cu and oxidation of pyrite in bentonite and backfill materials. These oxidation reactions (Cu(I) and pyrite oxidation) can influence the performance of the deep geological repository in two ways; the first way is consuming oxidizing agents (O2) and the second way is the changing pH in the near-field and ultimately influencing on the mass transport rate of radionuclides from spent nuclear fuel (failure of canisters) to out of the engineered barrier. Hence, it is very important to know the evolution of chemical environments of near-field by the oxidation of pyrite and Cu. However, the oxidation kinetics of pyrite and Cu are different in the order of 1E7 which means the overall kinetics cannot be fully considered in the deep geological repository. Therefore, it is important to develop a simplified Cu and pyrite oxidation kinetics model based on deep geological repository conditions. Herein, eight oxidation reactions for the chemical species Cu(I) were considered to extract a simplified kinetic equation. Also, a simplified kinetics equation was used for pyrite oxidation. For future analysis, simplified chemical reactions should be combined with a Multiphysics Cu corrosion model to predict the overall lifetime of Cu canisters.

The mobility of uranium (U) in the environment of a deep geological repository is controlled by various geochemical conditions and parameters. In particular, oxidation state of uranium is considered as a major factor to control the mobility of uranium in most of geological environments. In this study, therefore, we investigated the mobility of uranium in a deep geological repository by a natural analogue approach using a uranium deposit in the Ogcheon Metamorphic Belt (OMB). Uranium contents of rock samples from the study site ranged from 1.3 to 71 ppm (average 17.4 ppm). Uranium minerals found in the study site were mostly uraninite (UIVO2+x) and uranothorite ((UIV, Th)SiO4). The concentrations of U in the groundwater samples were very low (0.025~0.690 ppb) even though redox conditions are weakly oxidizing. Calculation results for U speciation in groundwater samples showed that major dissolved uranium species in the groundwater samples are mainly as calcium uranyl (UO2 2+) carbonate complexes such as Ca2UO2(CO3)3(aq) and CaUO2(CO3)3 2-. However, the activity ratios between 234U and 238U (AR(234U/238U)) showed U behavior in reducing conditions although the groundwater conditions were not reducing conditions and major dissolved U species were U(VI) species. Results from electron microscopic analyses for rock samples showed that major uranium minerals were U(IV) minerals such as uraninite and uranothorite. We could not identify other uranyl minerals and altered minerals from uraninite. This means that the geochemical condition of the study site has been maintained a reducing condition although the groundwater condition was a weakly oxidizing condition. Thus, the dissolution of uranium is strongly limited by the low solubility of uraninite. It is not obvious how the reducing condition of the study site has been maintained. Reducing agents such as pyrite, organic materials, and reducing bacteria might contribute to maintaining the reducing condition although further studies will be necessary. Results from this study imply that uranium mobility will be greatly limited by low dissolution of uraninite into groundwater if the reducing condition is well reserved. This limited mobility of uranium will be also contributed by low possibility of uraninite alteration into uranyl minerals which have a higher solubility than uraninite.

The nuclear criticality analyses considering burnup credit were performed for a spent nuclear fuel (SNF) disposal cell consisting of bentonite buffer and two different types of SNF disposal canister: the KBS-3 canister and small standardized transportation, aging and disposal (STAD) canister. Firstly, the KBS-3 & STAD canister containing four SNFs of the initial enrichment of 4.0wt% 235U and discharge burnup of 45,000 MWD/MTU were modelled. The keff values for the cooling times of 40, 50, and 60 years of SNFs were calculated to be 0.79108, 0.78803, and 0.78484 & 0.76149, 0.75683, and 0.75444, respectively. Secondly, the KBS-3 & STAD canister with four SNFs of 4.5wt% and 55,000 MWD/MTU were modelled. The keff values for the cooling times of 40, 50, and 60 years were 0.78067, 0.77581, and 0.77335 & 0.75024, 0.74647, and 0.74420, respectively. Therefore, all cases met the performance criterion with respect to the keff value, 0.95. The STAD canister had the lower keff values than KBS-3. The neutron absorber plates in the STAD canister significantly affected the reduction in keff values although the distance among the SNFs in the STAD canister was considerably shorter than that in the KBS-3 canister.