Tunnel and Underground SpaceISSN:1225-1275(Print) 2287-1748(Online)한국암반공학회

With the increasing accumulation of spent nuclear fuel (SNF) in South Korea, establishing a stable disposal method has become essential. This study analyzes international deep geological disposal cases to provide fundamental data for developing a domestic disposal system. The disposal methods and site characteristics of Finland, Sweden, Switzerland, and Canada were compared.The analysis identified three key factors: the quality of buffer and backfill materials, the stability of disposal sites, and the suitability of disposal methods. Finland and Sweden plan to use a combination of KBS-3V and KBS-3H in crystalline rock, while Switzerland utilizes Opalinus Clay in sedimentary rock. Canada is evaluating disposal options for CANDU SNF, applying highly compacted bentonite (HCB) buffer boxes and assessing sedimentary and crystalline rock sites.This study provides essential insights for developing Korea’s deep geological disposal system, emphasizing the need for continuous research on site selection, cost reduction, and material stability, as well as international cooperation.

This study evaluates the hydrogeological properties of deep fractured rock in Korea by applying the diagnostic analysis of head derivative to long-term pumping test data. Pumping tests were conducted in three deep boreholes (Nakmin, Bugok, and Jangjeon) located in granite and volcanic rock formations. The drawdown data measured in these boreholes were analyzed using the derivative method. By simultaneously matching the standard curves of suitable analytical models with both the drawdown curves and their derivatives, the hydraulic conductivity of the fractured rock was determined. The Nakmin borehole was best described by a generalized radial flow model, while the Bugok borehole was best represented by a double porosity model, and the Jangjeon borehole was most consistent with a constant head boundary model. The derivative analysis effectively identified early hydraulic responses (wellbore storage and skin effects), flow regime transitions, and hydraulic boundary effects, which were difficult to detect using conventional drawdown curves alone. The effective application of diagnostic analysis facilitates the optimal selection of analytical models, thereby improving the accuracy and reliability of hydraulic properties derived from pumping tests in fractured rock.

As the demand for mechanized tunneling construction increases, research on TBM excavation methods has been actively progressing. In particular, among mechanized construction methods, the shield TBM method is being widely applied in projects passing through urban areas and rivers. The shield TBM method uses precast concrete segment linings, so the construction process and the loads acting on the structure are very different from those of conventional cast-in-place linings, requiring separate structural stability reviews. However, the current domestic design standards do not clearly specify the structural stability review items for segment linings, and as a result, stability reviews are being conducted based on the items reviewed in previous projects. This study aims to categorize and systematize the review items necessary for the structural stability of segment linings into 11 stages, by referencing overseas design standards that consider the entire process from segment lining production to installation and operation. Additionally, a tunnel where shield TBM is applied will be selected to perform the 11-stage structural stability review of segment linings, and the appropriateness of each review item will be verified through the review at each stage.

Recently, with the development of 400km/h high-speed trains and their application in design, there has been an increase in high-speed railway large-section tunnel design cases aimed at addressing issues such as low pressure inside tunnels and passenger discomfort due to pressure changes. As the design speed of 400km/h is applied, the length of high-speed railway large-section tunnels is expected to increase. However, current design standards do not present a standard support pattern for high-speed railway large-section tunnels. This study proposes a standard support pattern suitable for 400 km/h-class high-speed railway large-section tunnels to ensure excavation stability and improve ground response. After analyzing existing standard support pattern cases for large-section tunnels and design standards, the suitability of the proposed support pattern was verified through numerical analysis. Furthermore, by reviewing the stress safety factor of the support, the support amount was optimized, and the bottom excavation face was increased by verifying the appropriate support length, leading to the proposal of an improved standard support pattern for 400 km/h-class large-section tunnels.

Rock cutting is a critical process in mining and civil engineering applications, with acting forces of cutting tools playing a pivotal role in determining excavation efficiency, tool wear, and energy consumption. This study presents a comprehensive analysis of predicting mean normal force and mean cutting force using an advanced machine learning approach. The research leverages a robust dataset of 195 experimental samples encompassing diverse rock types, cutting tool geometries, and operational parameters. A two-stage modeling methodology was employed, initially establishing baseline correlations through multivariate linear regression, followed by an optimized Random Forest model with systematic hyperparameter tuning via randomized search cross-validation. The linear regression analysis revealed moderate predictive capabilities, with R² values of 0.743 for FNm and 0.674 for FCm. In contrast, the optimized RF model demonstrated exceptional performance, achieving R² values of 0.993 (training) and 0.983 (testing) for FNm, and 0.972 (training) and 0.908 (testing) for FCm. Feature importance analysis highlighted the dominance of uniaxial compressive strength in predicting normal force, while cutting force prediction demonstrated a more complex interaction of multiple parameters.