Underground Space最新文献

Obtaining a comprehensive understanding of solute transport in fractured rocks is crucial for various geoengineering applications, including waste disposal and construction of geo-energy infrastructure. It was realized that solute transport in fractured rocks is controlled by stochastic discrete fracture-matrix systems. However, the impacts and specific uncertainty caused by fracture network structures on solute transport in discrete fracture-matrix systems have yet not been fully understood. In this article, we aim to investigate the influence of fracture network structure on solute transport in stochastic discrete fracture-matrix systems. The fluid flow and solute transport are simulated using a three-dimensional discrete fracture matrix model with considering various values of fracture density and size (i.e., radius). The obtained results reveal that as the fracture density or minimum fracture radius increases, the corresponding fluid flow and solute transport channels increase, and the solute concentration distribution range expands in the matrix. This phenomenon, attributed to the enhanced connectivity of the fracture network, leads to a rise in the effluent solute concentration mean value from 0.422 to 0.704, or from 0.496 to 0.689. Furthermore, when solute transport reached a steady state, the coefficient of variation of effluent concentration decreases with the increasing fracture density or minimum fracture radius in different scenarios, indicating an improvement in the homogeneity of solute transport results. The presented analysis results of solute transport in stochastic discrete fracture-matrix systems can be helpful for uncertainty management in the geological disposal of high-level radioactive waste.

When a fire occurs in an underground shield tunnel, it can result in substantial property damage and cause permanent harm to the tunnel lining structure. This is especially true for large-diameter shield tunnels that have numerous segments and joints, and are exposed to specific fire conditions in certain areas. This paper constructs a full-scale shield tunnel fire test platform and conducts a non-uniform fire test using the lining system of a three-ring large-diameter shield tunnel with an inner diameter of 10.5 m. Based on the tests, the temperature field distribution, high-temperature bursting, cracking phenomena, and deformation under fire conditions are observed. Furthermore, the post-fire damage forms of tunnel lining structures are obtained through the post-fire ultimate loading test, and the corresponding mechanism is explained. The test results illustrate that the radial and circumferential distribution of internal temperature within the tunnel lining, as well as the radial temperature gradient distribution on the inner surface of the lining, have non-uniform distribution characteristics. As a result, the macroscopic phenomena of lining concrete bursting and crack development during the fire test mainly occur near the fire source, where the temperature rise gradient is the highest. In addition, the lining structure has a deformation characteristic of local outward expansion and cannot recover after the fire load is removed. The ultimate form of damage after the fire is dominated by crush damage from the inside out of the lining joints in the fire-exposed area. The above results serve as a foundation for future tunnel fire safety design and evaluation.

The ability to predict tunnel deformation holds great significance for ensuring the reliability, safety, and sustainability of tunnel structures. However, existing deformation prediction models often simplify or overlook the impact of spatial characteristics on deformation by treating it as a time series prediction issue. This study utilizes monitoring data from the Grand Canyon Tunnel and introduces an effective data-driven method for predicting tunnel deformation based on the spatio-temporal characteristics of the historical deformation of adjacent sections. The proposed model, a combination of graph attention network (GAT) and bidirectional long and short-term memory network (Bi-LSTM), is equipped with robust spatio-temporal predictive capabilities. Additionally, the study explores other possible spatial connections and the scalability of the model. The results indicate that the proposed model outperforms other deep learning models, achieving favorable root mean square error ($\mathrm{RMSE}$), mean absolute error ($\mathrm{MAE}$), and coefficient of determination ($R^2$) values of 0.34 mm, 0.23 mm, and 0.94, respectively. The graph structure based on intuitive spatial connections proves more suitable for meeting the challenges of predicting deformation. Integrating GAT-LSTM with transfer learning technology, remains stable performance when extended to other tunnels with limited data.

During the operation of a deep geological repository in crystalline rocks for disposal of high-level radioactive waste, understanding the seepage behaviors of fractured crystalline rocks under coupled thermo-hydro-mechanical conditions is essential for the performance assessment of deep geological repositories. In this study, radial flow tests on cylindrical Beishan granite specimens with a single artificial fracture were conducted using the MTS 815 rock mechanics testing system to investigate the influence of normal stress and temperature on radial flow behaviors of rough rock fractures. Steady state method was used to measure fracture permeability, and an axial extensometer was used to measure fracture deformation during compression. A three-dimensional blue light scanner was used to characterize fracture surface morphology. Experimental results indicate that fracture permeability decreases nonlinearly with the increase of normal stress or temperature, and normal stress has a more significant influence on fracture permeability than temperature. The evolution of three-dimensional non-uniform distribution of voids under compression was numerically obtained, and the variogram was employed to quantify the non-uniform distribution characteristics of mechanical apertures. In addition, a radial flow model considering non-uniform distribution of apertures is proposed to predict the normal stress- and temperature-dependent seepage behaviors of rock fractures, and the predictions were found to be in good agreement with experimental data.