During tunnel excavation, the accumulated wall displacement and the tunnel support load result from both the tunnel advance and the time-dependent behavior of the surrounding rock mass. One approach to analyze the interactions between tunnel wall displacement and support load is the Convergence-Confinement Method (CCM) using analytical closed-form solutions or empirical Longitudinal Displacement Profiles (LDP). This approach neglects the influence of the time-dependency of ground response resulting in delayed deformation increasing significantly within time after the excavation stage. This time-dependency is particularly crucial in tunnels in squeezing ground, which remains one of the most difficult problems in tunneling. Predicting large tunnel convergence and the influence of time on tunnel deformation in tunnel squeezing, which leads to very high loads on tunnel support, remains a major challenge in tunneling. Failure to consider the added delayed displacements in the preliminary design can result in a false selection of the installation time and the support system type, causing safety issues, cost overruns, and project delays. This paper discusses a revised CCM to estimate the tunnel's support system loads in squeezing ground conditions considering the time-dependent ground response. The proposed approach combines laboratory-scale physical model test results and observations from squeezing tunnels worldwide. The paper explains the new time-dependent CCM and the improvement offered to extend the methodology in the analysis and design of squeezing tunnels. The proposed methodology is validated against Venezuela's Yacambu´-Quibor water conveyance tunnel, which has experienced extreme ground squeezing.

In this study, the authors compare analytical voussoir-analogue based solutions with numerical ones, implemented through 2D distinct-element-method based simulations, in the context of the application of these analyses to the design of underground rooms in bedded rock masses, based on the conditions found in a particular underground room and pillar carbonate mine. Some guidelines are given on how to derive key parameters required for both approaches based on laboratory rock testing, rock mass characterization, and in-situ observations. Particular attention is devoted to quantifying the influence of the roof span, bedding dip, and the occurrence of an overload (associated with highly-fractured beds over the already-detached roof). The performance of the analytical solution in calculating the maximum deflection and compressive stresses within a beam is verified through several 2D numerical models, which, in turn, are capable of capturing the instability mechanisms occurring in the mine.

Modulus is an essential input parameter for the design of excavations and can be referred to as the loading modulus due to its determination under loading conditions. The loading modulus is usually calculated from results of uniaxial compressive strength tests, without considering confining pressure. In this study, a series of uniaxial and triaxial compressive tests on synthetic isotropic specimens has investigated the impact of confining pressure on the resulting loading modulus. The results demonstrate that confining pressure significantly influences the determined loading modulus, highlighting its importance as a key parameter. A Tangent technique calculated at 50% peak strength with a 10% range is concluded as the most reliable method for determining the loading modulus. In addition, contrary to deviatoric stress, total axial stress is identified as the most competent parameter for evaluating deformation and strength, due to its inclusion of confining pressure in the axial direction.

Rockburst, a phenomenon occurring in highly stressed grounds, poses a significant risk due to the safety of underground mines due to a sudden and violent rock failure. Integrated ground support systems are generally utilised to enhance stability and minimise rockburst hazards. This study employs ABAQUS software to assess the performance of integrated support system in potential burst-prone areas. The effects of bolt spacing, and fibre-reinforced shotcrete were examined based on deformation in burst-prone areas (PBPA) and energy absorption of the support system. The results indicate that rockbolt spacing of 1 m to 1.3 m and 2 m leads to a decreased volume of PBPA by 47%, 41% and 34%, respectively. Stress and displacement are highest in the bolts within these areas, as indicated by the energy absorption results. The result shows that energy absorption is 6.90 kJ/m^2 when applying the integrated support system, which comprises rockbolts with 1 m spacing and 75 mm thick fibre-reinforced shotcrete.

Blasting is widely used in tunneling when mechanical excavation methods cannot be applied due to rock conditions or cost constraints. The blast design of the contour holes defines the damage to the remaining rock, which might change the rock support requirements. This study investigates the crack behavior in sequential boreholes through small-scale experiments on rock-like specimens. Cylindrical samples, prepared with speckles for Digital Image Correlation (DIC), varied in decoupling ratio, and the detonation cord was detonated simultaneously in the blast holes. The data was collected with an ultra-high-speed camera (UHSC) for DIC. The results indicated the development of the cracks between the boreholes and their behavior towards the boundary of the samples. The results showed that in this experimental configuration, there is no significant difference between the different decoupling ratios. This study shows the importance of an optimum blast design to minimize the damage to the remaining rock.

In the early stages of a deep tunnel design, it is essential to identify and assess the potential hazards that could occur during construction. One of these hazards is the development of significant deformations around the tunnel after excavation, which is often referred to as "squeezing” phenomenon. The most widely used tool for estimating the squeezing potential is an abacus introduced by Hoek and Marinos (H&M) in 2000. It allows to simply estimate the percentage of tunnel convergence by knowing only the uniaxial compressive strength of the rock mass (σcm) and the in-situ stress. The chart has several important advantages, such as ease of use and the relevance of the information obtained, making it a very useful tool. However, there are also some limitations, which are poorly known in the engineering community. Consequently, this abacus may be used outside of its scope. Firstly, the chart is based on analytical models that aims to catch the behavior of a tunnel excavation in a perfectly plastic rock mass characterized by a Mohr-Coulomb behavior (Duncan Fama 1993) and Hoek&Brown behaviour (Carranza-Torres and Fairhurst 1999) which is not always realistic in the area near the excavation. The other problem lies in estimating the uniaxial compressive strength of the rock mass. While easy to comprehend, it is very hard to estimate in practice. Several formulations can be found in the literature that lead to very different values of σcm. Consequently, important discrepancies exist in the estimated convergence values. This paper aims to clarify the underlying hypothesis and to update the H&M chart by replacing σcm with the uniaxial compressive strength of the intact rock (σci) and by introducing curves of GSI (Geological Strength Index) isovalues. To achieve this, a Monte Carlo simulation is conducted on the input parameters (in-situ stress, GSI, σci, the Hoek&Brown parameter mi, dilation angle), like the original approach. However, unlike Hoek and Marinos, who relied on analytical expressions, the new chart is based on numerical modelling, allowing the rock mass to be characterized using a generalized Hoek and Brown failure criterion. The variation intervals of the input parameters for the Monte Carlo simulation and the expressions linking these parameters to those used in the numerical modelling have been updated. Finally, results give an improved assessment of the expected convergence and allow to consider uncertainties about the in-situ stress, σci and mi to get a confidence interval of the estimation.