Prefabricated underground stations (PUSs) are widely used for their low-carbon, high-efficiency benefits. The complex construction of large-scale PUS components can lead to stress concentration and cracking. This study analyzes the mechanical performance and crack identification of these components during the construction process. A high-precision monitoring system using optical frequency domain reflectometry (OFDR) was developed to measure the internal strain field of large-scale components. Additionally, a crack identification method based on principal component analysis (PCA) and hierarchical clustering (HC) was established to detect crack initiation and growth during assembly. The main findings of this study are that the monitored data demonstrated the ability of fiber optic sensors to identify areas of highest strain concentration with sub-millimeter spatial resolution. The topsoil backfilling process contributed the most to the bending moment in the roof, accounting for an average of 39.3%, while the foundation trench backfill process had the greatest impact on the base plate and sidewalls, with average contributions of 43.5% and 30.9%, respectively. Additionally, the PCA-HC method proved effective for identifying potential cracks in large-scale assembly components through strain field data measured by OFDR, with cracks observed at the mid-span thin-walled cavities of the roof during hoisting and positioning.
Post-flood recovery of underground metro stations requires coordinated restoration of interdependent power, communication, and pedestrian subsystems, yet existing models typically treat physical component repair as equivalent to effective service restoration, obscuring extended periods of zero service capacity behind apparent repair progress. This study develops a service-oriented decision-support framework coupling flood-depth-driven cascading-failure modeling, stochastic Monte Carlo recovery simulation, and a cost–time evaluation with an explicit service-level cascade-gating constraint. Six recovery strategies, spanning cost-prioritizing, time-prioritizing, hierarchy-following, balanced, and two resilience-informed greedy variants (RFVC-Greedy), are evaluated across 100-, 500-, and 1000-year flood scenarios. Three findings hold across scenarios: cost-prioritizing and strict hierarchy-following strategies produce prolonged zero-service intervals; cascade dependencies amplify functionality loss well beyond direct flood damage; and power restoration acts as a force multiplier dominating total resilience variance. The cost–time Pareto knee shifts with severity, with the Balanced strategy forming the knee under the 100-year scenario and the RFVC-Greedy variants under the 500- and 1000-year scenarios; the min cascade vulnerability (VC) variant additionally offers the greatest cross-scenario temporal predictability. The framework provides severity-adaptive operational guidance for emergency managers of flood-exposed metro systems and quantifies the structural decoupling between physical repair and service restoration.
Pre-existing underground structures would affect groundwater flow (i.e., water-blocking effect) and soil deformation (i.e., soil-blocking effect) during pit dewatering in dense urban environments, leading to ground responses different from those in scenarios without underground structures. In this study, a series of numerical models were developed to investigate water‑soil‑structure interaction, considering the barrier effect of an adjacent metro station embedded in multiple aquifers. The results demonstrate three key findings. First, the cut-off ratio of the station relative to the dewatering aquifer has a great effect on the water-blocking effect; a ratio exceeding 0.5 would significantly intensify the effect, resulting in considerable groundwater drawdown in front of the station and differential drawdown on both sides of the station. Second, the intensity of the two blocking effects is governed by the combination of the pit-station distance (D) and the station burial depth (H). A larger D and a larger H enhance the water-blocking effect, thereby increasing the ground settlement; however, a smaller D and a larger H intensify the soil-blocking effect, thereby reducing the settlement. Third, unified relationships of drawdown versus H and settlement versus H are proposed for preliminary analysis of dewatering-induced ground response under the barrier effect. The findings can help identify the dominant blocking effects and quantify key parameters, thus optimizing dewatering design in similar soft soil areas.
In recent years, climate change has led to an increased frequency of urban floods, which not only damage surface buildings but also submerge underground spaces. This review study compiled global flood cases in underground spaces from 2010 to 2024. Based on the analysis of statistical flood cases, the disaster mechanism of floods in underground spaces was explored. Moreover, risk analysis approaches for floods in underground spaces were summarized into three types: multi-criteria decision making (MCDM)-based quantitative evaluation, scenario-based qualitative prediction, and machine learning (ML)-based techniques. The advantages and disadvantages of these methods were discussed. Two typical subway flood incidents in Hong Kong, China and Zhengzhou, China were analyzed to provide further implications for flood mitigation in underground spaces. To effectively prevent floods, entrances and exits of underground spaces should be positioned at elevated locations and comply with established flood prevention standards. This review study puts forward new perspectives on strategies for flood mitigation in underground spaces, emphasizing the importance of data collection, risk assessment, and the establishment of early warning systems. It is hoped that this review will be helpful for flood mitigation in underground spaces and the safe operation of underground infrastructures.
Geological uncertainty is a major risk in shield tunnel construction, especially in upper-soft lower-hard mixed strata, where lithological differences can cause machine faults and construction hazards. Traditional drilling lacks spatial continuity, and single-modality methods based only on tunnelling parameters often show limited accuracy in mixed strata. To address this issue, this study proposes a multimodal fusion deep learning model that integrates numerical modality and image modality for mixed strata identification in shield tunnelling. In the proposed framework, a multilayer perceptron (MLP) is used to extract nonlinear features from tunnelling parameters and vibration root mean square (RMS) values in three orthogonal directions, while a convolutional neural network (CNN) is used to extract spatial features from vibration spectrograms. Middle-level feature concatenation is adopted to achieve deep fusion of heterogeneous data. To obtain realistic vibration responses, triaxial accelerometers were installed in the central chamber of a super-large-diameter shield machine. Under a chronologically constrained evaluation protocol, the proposed model achieved an average test accuracy of 88.71% ± 7.50% across the blocked temporal forward-chaining folds. On the independent later-stage test set, it achieved an overall accuracy of 97.36%. Ablation results show that multimodal fusion outperforms single-modality settings. Horizontal and axial vibration spectrograms contribute most to strata discrimination, and cutterhead thrust is the most sensitive numerical feature. These findings demonstrate the feasibility and effectiveness of multimodal fusion and vibration data for mixed strata identification.
Effective management and safe operation of tunnel infrastructure are critically challenged by the persistent vehicle tracking across non-overlapping camera domains and the lack of real-time situational awareness for emergency incidents. This work proposes a dynamic digital twin framework to create a high-fidelity virtual representation that synchronously updates with physical tunnel environments by leveraging existing Closed-Circuit Television (CCTV) networks. The core is a novel Tunnel Re-identification Model (TRIM) that can create discriminative feature embeddings for seamless vehicle tracking and re-identification across multi-camera domains. Comparative validations unequivocally demonstrate that the TRIM significantly outperforms Convolutional Neural Networks (CNNs) and standard Vision Transformer backbones, achieving a state-of-the-art mAP@10 of 0.684. Analysis of attention heatmaps reveals its ability to generate a comprehensive and granular feature representation, focusing on preserving high-quality characteristics for specific vehicle identity. Furthermore, the emergency response ability of the digital twin is augmented with a vision-based fire detection module to real-time localize and visualize emergent tunnel fire events in the digital twin. By consolidating and distilling vision data from CCTV cameras into an intuitive representation, the proposed framework enhances situational judgment for operators and significantly reinforces the safety resilience of tunnel infrastructure without requiring extensive hardware redevelopment.
During the construction of deep-buried tunnels, dynamic disturbances may adversely affect surrounding rock stability, while research on vibration control measures and their influence on wave attenuation remains limited. In this study, a wave-absorbing anchorage composed of wave-absorbing anchoring mortar and rock bolts was employed to conduct field tests on wave attenuation in a deep-buried tunnel. The attenuation characteristics of blasting vibration intensity were analyzed for a single wave-absorbing anchorage (single-anchorage) and for the region between two adjacent wave-absorbing anchorages (inter-anchorage). Both the single-anchorage and inter-anchorage configurations reduced vibration intensity in the axial direction by 28.7% and 21.8%, respectively, and in the radial direction by 10.8% and 18.9%, respectively, whereas a slight increase was observed in the tangential direction. Subsequently, numerical simulations were performed to investigate the variations in vibration velocity and stress before and after stress waves passed through the wave-absorbing anchorage. The reflection and refraction of stress waves upon encountering the anchorage are important factors in the attenuation and the redistribution of three-directional vibration components. Based on these findings, the wave-absorbing performance of group-anchorage support was further evaluated through numerical simulations. The offset-densified group-anchorage exhibited stronger attenuation capability than the conventional rectangular group-anchorage. This study provides a reference for the prevention and control of hazards induced by dynamic disturbances in deep-buried underground engineering.
Tunnel construction safety is influenced by complex factors such as geological conditions and environmental uncertainties. To enable comprehensive monitoring and improve construction safety, this study proposes a multi-dimensional monitoring system spanning space-air-ground-underground (SAGU) domains. The multi-dimensional monitoring technologies encompassed by the SAGU system are summarized based on their key technical features, research methods, and development status, with particular emphasis on our contributions and innovations. An intelligent monitoring system based on the cloud-edge-terminal architecture, together with the corresponding GPR-AI Master platform, is proposed to manage diverse monitoring technologies and enable the analysis and dynamic feedback of multi-source and multi-fidelity monitoring data. This study not only introduces innovations in the application system but also analyzes the development, characteristics, and current status of these technologies.
The water pressure withstood by the gasketed joints in shield tunnels is continuously increasing, but the influence mechanism of the gasket configuration remains unclear under ultra-high water pressure. This study presents a sensitivity analysis of the waterproof performance of the gaskets to identify the key driving factors and clarify the interaction patterns among configuration factors, providing suggestions for design optimization. The influence mechanisms of the gasket configuration on waterproof performance were revealed. Four gasket configurations designed for tunnel joints under ultra-high water pressure were innovatively developed. The compression and waterproof tests were employed to verify their waterproof performance, identifying the optimal configuration. The results indicate that the waterproof performance of the gasket is most influenced by the joint opening, followed by hardness. The significant and antagonistic interactions exist among configuration factors. At large joint openings, the assembly force is critical to the waterproof performance of the gasket, whereas at small joint openings, deformation behavior becomes more influential. The stress transfer between the close holes is determined by the stiffness of the gasket legs. The gasket configuration with triple-row close holes and flat-bottomed structure improves waterproof performance and overall deformation stability. Four novel gasket configurations designed using the response surface method achieved waterproof performance above 2.5 MPa under 6 mm joint opening and 15 mm offset, with the optimal configuration exceeding 2.7 MPa.
Prefabricated underground stations (PUSs) are widely used for their low-carbon, high-efficiency benefits. The complex construction of large-scale PUS components can lead to stress concentration and cracking. This study analyzes the mechanical performance and crack identification of these components during the construction process. A high-precision monitoring system using optical frequency domain reflectometry (OFDR) was developed to measure the internal strain field of large-scale components. Additionally, a crack identification method based on principal component analysis (PCA) and hierarchical clustering (HC) was established to detect crack initiation and growth during assembly. The main findings of this study are that the monitored data demonstrated the ability of fiber optic sensors to identify areas of highest strain concentration with sub-millimeter spatial resolution. The topsoil backfilling process contributed the most to the bending moment in the roof, accounting for an average of 39.3%, while the foundation trench backfill process had the greatest impact on the base plate and sidewalls, with average contributions of 43.5% and 30.9%, respectively. Additionally, the PCA-HC method proved effective for identifying potential cracks in large-scale assembly components through strain field data measured by OFDR, with cracks observed at the mid-span thin-walled cavities of the roof during hoisting and positioning.
The natural barrier function of the host rock in a geological repository for high-level radioactive waste (HLW) disposal is essential for long-term safety, as excavation-induced disturbance may adversely affect its isolation capacity. To investigate the microcracking behavior of the surrounding rock during tunnel boring machine (TBM) excavation, an in situ acoustic emission (AE) monitoring test was conducted at the −280 m experimental level of the Beishan underground research laboratory (URL), China. A monitoring tunnel was pre-excavated, and eight boreholes were drilled toward the TBM-excavated ramp. Sixteen AE sensors were installed to monitor two representative zones: Zone I (moderately fractured) and Zone II (sparsely fractured). Monitoring results indicate that a total of 12 954 AE events were recorded during TBM advance, with approximately 85% confined within 1.5 m of the excavation boundary. Zone I exhibited 75% more events and a higher maximum energy than Zone II, indicating stronger fracture-controlled disturbance. Low-energy events (<10−13 J) dominated the dataset (97.9%), whereas high-energy events (>10−11 J) were rare (0.2%) and concentrated near the excavation boundary and fracture intersections. Fracture mechanism analyses show that shear failure along fracture surfaces is the dominant mode. However, tensile cracking occurred more frequently in Zone I, where a fracture network promoted crack opening under localized stress redistribution. In contrast, Zone II, characterized by fewer fractures and more consistent orientations, was dominated primarily by shear failure. Comparative analyses with microseismic (MS) and stress monitoring provided consistent validation of the AE results. These findings confirm that TBM excavation induces limited disturbance at Beishan URL, with microcracking strongly controlled by pre-existing fractures, providing important insights for long-term safety assessment of HLW repositories.
Traditional passive support design methods are increasingly inadequate for complex geological conditions and high-stress environments, posing significant threats to tunnel construction and operational safety. To ensure the overall stability of the support-surrounding rock system, this study systematically proposes a stiffness design theory and a synergetic control method. First, the concept of the structural ring is introduced to quantify the self-bearing capacity of the surrounding rock, accompanied by its determination and stability analysis protocols. Based on this, deformation control standards and rock load effects are established as the foundation for structural design. Subsequently, a dynamic analysis model for the structural ring is developed, establishing a characterization method that links tunnel deformation to support stiffness. By accounting for the stiffness growth property of the support system, a design method based on the active distribution of deformation and load is proposed. This approach transcends the limitations of traditional strength-based designs that focus solely on the final structural state. Furthermore, a stiffness synergetic control model is established, integrated with a multi-objective optimization method to achieve both global safety and economic efficiency. This methodology was successfully applied to the New Badaling Tunnel, which is a super-large cross-section project, forming an active control technology framework centered on the synergy between the structural ring and the anchorage system. Finally, the technology was extended to address large-deformation disasters in tunnels under extremely high geo-stress. These research findings provide a robust theoretical basis and systematic technical support for the scientific design and safe construction of tunnels in complex environments.
Amidst the accelerating pace of global urbanization and increasing climate urgency, the decarbonization of energy systems requires strategic utilization of underground space in general and underground networks (UNs), in particular, where tunnels represent critical nodes of pollutants and greenhouse gas emissions, as well as consumption of energy, raw materials, and final resources. Their high impact is due to their resource intensive construction, operational energy, and resource demands. Although lifecycle assessment (LCA) is commonly used to quantify carbon footprints, existing research on low-carbon tunnels (LCTs) still faces three persistent challenges: fragmentation across isolated lifecycle stages, lack of synergy between different phases, and insufficient consideration of scenario-specific adaptation such as geography and traffic load. To address these gaps, this study conducts a bibliometric analysis to map the current research landscape and identify technological hotspots, systematically integrates innovative low-carbon technologies across the entire lifecycle of both the tunnel itself and its installations, from design to decommissioning, and evaluates their effectiveness through case studies in different scenarios. Finally, based on the integrated analysis of low-carbon technologies across the tunnel lifecycle, this study discusses their potential for scaling up to support the decarbonization of broader underground networks, offering scenario-adaptive strategies and decision-support tools to promote sustainable and resilient underground network development under carbon neutrality goals.
The development towards deeper underground is an inevitable trend in rock mechanics research and rock engineering construction. During deep underground construction, disasters such as extensive rock spalling, deep fracturing, large-volume collapse, large extensive deformation of hard rocks, and rockbursts of various types frequently occur. They are characteristic of high-stress-induced internal rock fracturing and disintegration of geological structures usually accompanied by energy release, which are completely different from the instability issues in shallow rock engineering. Therefore, a research approach for stress controlled and stress-structure controlled failure processes in deep engineering rockmasses has been proposed. Starting from recognition of geological characteristics of deep underground engineering and the characteristics of different types and levels of disasters, this paper elucidates the effects of stress and geological structure on the process of rockmass failure, and presents the ideas of developing laboratory and in-situ testing techniques, establishing predictive analysis theory and developing warning methods with some typical research achievements. The principles, methods, and technical measures for controlling the stress controlled and stress-structure controlled disasters mentioned above are proposed. Typical results and applications were used to illustrate the proposed research approach.
The increasing proximity of deep excavations to existing tunnels in urban underground development necessitates accurate safety assessment and early warning mechanisms during design and construction. Traditional approaches suffer from inefficiencies in multi-source data integration, low interoperability between modeling and analysis platforms, and limited capability in capturing complex soil-structure interactions. This paper presents an intelligent building information modeling (BIM)-finite element method (FEM) collaborative workflow that enables intelligent modeling, automated data conversion, and high-fidelity mechanical analysis for the digital design and safety assessment of foundation pit excavation adjacent to sensitive tunnels. A parametric modeling framework combining Kriging interpolation and Dynamo visual programming is developed to construct high-precision 3D geological, support, and tunnel models despite sparse borehole data. A seamless BIM-to-FEM interface based on the advanced computer-aided software integrated system (ACIS) kernel and HyperMesh is implemented to ensure lossless data transfer, automated mesh optimization, and boundary condition assignment. The proposed workflow is validated through a real-world case study, showing existing tunnel deformation errors of 1.59%–4.80% and horizontal displacement increments of over 58%–72% occur during the excavation phase in weak soil layers. The accuracy of diaphragm wall deformation capture in weak soil layers reaches 97.45%. A deformation-sensitive area is defined based on a 3 mm displacement threshold. Parametric analysis further reveals that tunnel deformation decays exponentially with horizontal distance and embedment depth, and identifies critical design thresholds for deformation-sensitive zones. This framework offers an efficient digital twin solution for risk control and safety assessment in complex urban underground projects.
Gas explosion in an urban utility tunnel is a crucial safety issue which may cause heavy casualties and economic losses. In this study, a series of gas explosion experiments were conducted on a large-scale concrete utility tunnel structure (10 m in length) to evaluate overpressure, structural responses, and impact on the surrounding environment under varying methane volume concentrations and venting conditions. Time history characteristics of explosion overpressure, structural displacement, acceleration of structure, and adjacent ground motion were recorded. Results indicated that the maximum overpressure was observed at stoichiometric volume concentration (9.5%), which similarly corresponded to structural responses and ground motions intensity. Comparative experiment between unconfined and confined gas explosions revealed that while the latter exhibited longer duration, the maximum overpressure magnitudes remained identical between both scenarios. Furthermore, a significant pressure increase was observed near boundary regions, attributed to the accumulation of ignited gas mixture and energy, which can be mitigated through the implementation of a top vent. Additionally, the presence of soil and vibration isolation walls adjacent to the utility tunnel demonstrated a significant reduction in acceleration magnitudes resulting from gas explosion events.
To date, the transport and discharging of rock chips, considering single-particle breakage behavior in secondary breaking system, such as jaw crusher, remains under-explored. This paper uses the Qingdao Jiaozhou Bay Second Undersea Tunnel in China, as an engineering background and case study. A three-dimensional coupled model considering the rock particles breaking and slurry flushing behavior within the crushing box of jaw crusher was established using coupled computational fluid dynamics-discrete element method (CFD-DEM) approach. The breakage behavior of rock chips was described based on Tavares Universidade Federal do Rio de Janeiro (UFRJ) breakage model. The results revealed two key findings: first, the jaw crusher employs a hybrid rock-breaking mechanism that combines extrusion and impact crushing. Second, the slurry flushing inlet generates concentrated flow pathlines in the lower region of the crushing box. These flow pathlines direct rock particles toward the jaw crusher and intake screen, partially preventing particle aggregation and clogging. Meanwhile, the locations of reinforcing plates or abrasion-resistant layers are identified based on the severe wear and stress concentration areas of the equipment. Besides, a series of parametric investigations were performed to investigate the effects of operating parameters of jaw crusher. Lastly, the optimal operating parameters of jaw crusher were suggested. The findings of this study could provide references for addressing slurry chamber clogging during slurry tunnel boring machine (TBM) advancement in hard rock in similar future projects.
To our knowledge, the “Yangtze Estuary No.2” ancient shipwreck was salvaged using the world’s first curved pipe jacking curtain method. For underground construction in the underwater soft soil environment, this study employs the Material Point Method (MPM) to analyze the resulting large geotechnical deformations. Vertically variable depth construction, a characteristic feature of the novel underwater shipwreck integral salvage method (which uses tight-radius rectangular curved pipe jacking to form a bottom tray), presents difficulties in determining the corresponding excavation jacking force. Therefore, the frictional resistance variation along the entire vertical curved pipe jacking was first derived, obtaining the required jacking force at different positions during construction as key input parameters for the MPM simulation. Then, the MPM numerical solutions were compared with the results of a 1:10 scaled experiment for validation. A full-scale MPM numerical model of shipwreck salvage construction was built to study the environmental impact of the curved pipe jacking process and its soil deformation mechanisms. The results show that during the curved pipe jacking construction process, the soil near the jacking end experienced significant disturbance, with vertical displacements of −17 cm to −16 cm observed on the initial jacking side after construction. Validated against on–site measured data analysis, the jacking force calculation method established in this paper yields higher accuracy (an average improvement of 36.79%). Compared with conventional numerical simulation methods, the displacement distribution obtained by MPM is closer to the actual observations. Under the influence of the jacking load, the ship hull experienced approximately 16 cm of displacement on the jacking side, and the maximum displacement difference between both sides of the hull was about 9 cm. The research findings provide practical guidance for thrust calculation and refined evaluation of construction disturbance in underwater curved pipe jacking excavation with variable overburden.
Shotcrete is critical for providing rapid support and deformation control in drill-and-blast tunnel construction; however, the suppression of its rebound from the rock surfaces remains a persistent challenge in practice. This study developed a real-time shotcreting simulation engine based on the position-based dynamics algorithm. A staged physical–virtual calibration strategy is proposed to identify key parameters across velocity regimes: low-velocity fluid and collision parameters governing deformation, rheological, and adhesion behaviors are calibrated using dynamic slump and inclined-plate tests, whereas high-velocity jet-flow parameters governing transport, impact, and rebound are calibrated using jet mass-field and velocity-field measurements. This calibration workflow links observable responses to the dominant behaviors in different velocity regimes and enables the engine to reproduce both low-velocity deformation/adhesion and high-velocity jetting/rebound processes. Using the calibrated model, the spatial pattern of rebound can be reproduced and assessed quantitatively. Field validation in the Yangjiawopu Tunnel demonstrated excellent agreement between simulation and measurement results, with a maximum local error of 13.7% at the near-wall location where rebound was the highest. The proposed approach maintained stable real-time computation, achieving 38 fps for a 2 m × 2 m wall section. Overall, this simulation engine provides a physically grounded and computationally efficient basis for smart shotcreting in tunnel excavation, supporting future integration with sensing, automation, and digital-twin-based construction control.
