Karst-influenced deep excavations involve coupled geological, hydraulic, structural, and observational uncertainties that cannot be represented adequately by deterministic checks alone. This study develops a monitoring-updated reliability framework for the Baiyun Dongping Station excavation in Guangzhou, China. Project documents, karst investigation records, a documented water-inrush event, field monitoring data, APDL-derived geometry, and direct global–local coupled simulations are integrated through a source-path-response interpretation. Formal probability statements are restricted to two traceable criteria: a 24 mm wall-displacement warning threshold and a 270 m3/h severe-flow threshold, while support force, groundwater change, and adjacent-asset response are treated as diagnostic or validation evidence. Bayesian updating is used to reduce uncertainty in physically meaningful parameters, including rock stiffness, cavity-roof tensile strength, fracture conductance, and effective hydraulic head. Posterior standard deviations decrease by 38.5–81.8 %. Direct simulations give severe-flow exceedance probabilities of 0.0407, 0.1386, and 0.0631 at Stages 4, 5, and 6, respectively. Effective head and fracture conductance dominate post-connection flow, whereas roof strength controls pathway initiation and soil-rock stiffness controls stress transfer. External predictive coverage is 0.70 for wall response but only 0.20 for held-out settlement, indicating that the framework supports hydraulic reconstruction and wall-warning assessment, but not a general adjacent-asset reliability claims under the available field evidence.
Rapid urbanisation has led to extensive in-service metro lines adjacent to deep excavations, posing significant challenges to their safety and stability. This study investigates the deformation response of two orthogonally intersecting metro lines to three closely spaced deep excavations in Suzhou soft soils, along with the associated countermeasures. To accurately capture the complex soil-structure interactions, a high-fidelity three-dimensional numerical model was developed utilising the Hardening Soil model with Small-strain stiffness (HSSmall), with input parameters rigorously calibrated based on extensive in-situ testing data. The numerical model was thoroughly validated against field measurements of the displacements of both metro lines and the retaining structures. Comprehensive analyses were then conducted to assess the deformation response and effectiveness of various active and passive interventions, such as partition wall installation, excavation sequence optimisation, retaining system design, and tunnel reinforcement. Results show that temporary partition walls play an important role in controlling the deformation response of adjacent metro lines to deep excavation, with temporary partition walls installed perpendicular to the tunnel axis being particularly effective. Regarding excavation sequence, priority should be given to deep excavations farther from metro lines, with the construction sequence from small to large excavation sizes. Furthermore, provided the retaining wall embedded depth meets design requirements, increasing wall stiffness is more effective than increasing embedded depth. A more cost-effective approach is to enhance retaining walls locally near metro lines. While tunnel reinforcement significantly reduces the deformations, its effects on metro stations are limited, underscoring the necessity of integrating active and passive interventions in complex excavation scenarios. This study provides valuable insights into deformation control strategies for protecting existing metro lines adjacent to an excavation group.
With the increasing application of rectangular pipe jacking in underground space construction, clarifying the mechanisms and magnitudes of surrounding soil responses induced by jacking disturbance is critical to mitigating relevant environmental impacts. However, current understanding of subsurface soil responses (especially longitudinal soil movements), potential soil failure characteristics and the carrying-soil phenomenon induced by rectangular pipe jacking remains limited. Moreover, effective quantitative methods for evaluating the severity of the carrying-soil phenomenon are still lacking. To fill these research gaps, this paper develops a visual model test scheme based on Digital Image Correlation (DIC) technology to investigate the surrounding soil responses and failure characteristics induced by rectangular pipe jacking, as well as to quantitatively assess the corresponding environmental impacts, particularly the carrying-soil phenomenon. The reliability of the proposed scheme is verified by the good consistency between experimental results and theoretical solutions. Using this model test scheme, this study systematically reveals the stratum response characteristics under various key construction parameters, as well as the subsurface soil responses and soil failure characteristics when the overall carrying-soil phenomenon occurs. The results indicate that as the jacking length increases, the longitudinal stratum displacement along the jacking direction generally becomes more uniformly distributed. Meanwhile, the maximum longitudinal stratum displacement at the excavation face decreases and tends to reach a stable value. To effectively control stratum deformation and avoid severe carrying-soil hazards during rectangular pipe jacking, the burial depth of the jacking pipe should be maintained at no less than 1–1.4 times the excavation height, and the additional face thrust pressure should be optimally controlled within the range of −20 to 20 kPa. When the overall carrying-soil phenomenon occurs, the vast majority of soil above the pipe exhibits significant forward movement with strong integrity. The longitudinal ground surface displacement exceeds 41% of the longitudinal soil displacement near the pipe crown. In addition, tensile cracks perpendicular to the jacking direction may form on the ground surface, and shear cracks tend to emerge near the top of the excavation face. The proposed DIC-based model test scheme provides an effective approach to quantitatively evaluate the environmental effects of rectangular pipe jacking and assess the severity of the carrying-soil phenomenon in engineering practice.
The thermo-hydro-mechanical (THM) behavior of deep-buried rock tunnels in tectonic mélanges is governed by strong heterogeneity, seepage, and ground temperature, yet their coupled influence remains poorly understood. This study establishes an engineering-scale, two-dimensional heterogeneous continuum model for fully-coupled THM processes, based on the geological cross-section, lithological distribution, and in-situ conditions of a deep rock tunnel in southwestern China. A systematic numerical investigation is performed to investigate the effects of rock mass heterogeneity, seepage, ground temperature, seepage driving force, and virgin rock temperature on stress redistribution, deformation patterns, and seepage characteristics. The results demonstrate that heterogeneity dominates the mechanical and hydraulic response by triggering marked stress redistribution near lithological interfaces and localizing seepage concentration. Seepage fundamentally reshapes the vertical displacement profile, increasing settlement at the crown, shoulders, and sidewalls while suppressing heave at the invert and springlines. Elevated ground temperature substantially amplifies both deformation and Darcy velocity through thermally induced stiffness degradation, thermal expansion of the rock matrix, and reduced groundwater viscosity. Furthermore, Darcy velocity increases quasi-linearly with both the seepage driving force and virgin rock temperature, with the latter exerting a considerably stronger influence. These findings advance the mechanistic understanding of engineering-scale THM response patterns in tectonic mélange tunnels and provide practical guidance for identifying zones of elevated seepage risk.
Accurate estimation of fracture intensity in tunnel surrounding rock is essential for evaluating excavation stability. Traditional geological survey methods are limited by sparse sampling, restricted coverage, and low accuracy. To address these challenges, this study proposes a semi-supervised deep learning framework that combines an unsupervised constrained convolutional stacked autoencoder (CCSAE) with a supervised multi-branch extraction network (MBEN) to predict the volumetric fracture intensity (P32) using measurement while drilling (MWD) data and auxiliary information (including blasting and construction parameters). A novel group fusion strategy is proposed to integrate multi-source features extracted from MWD and auxiliary data. Through parameter tuning and ablation studies, the proposed CCSAE-MBEN framework demonstrates superior performance in P32 prediction compared to traditional approaches. The group fusion strategy outperforms conventional flattening and cascading fusion techniques in feature integration. Furthermore, extending the MWD data sequence length enhances model performance up to a certain point, beyond which excessive geological input may reduce accuracy. A complete dataset substantially improves the interpretability of MWD data, whereas omitting auxiliary parameters significantly reduces model interpretability.
Structural loads, internal forces, deformations, and surrounding environmental monitoring of shield tunnels are critical for understanding structural response and construction risk. However, the complex construction environments and confined spaces in shield tunneling projects often hinder the effective implementation of monitoring equipment and methodologies. This paper proposes an intelligent perception monitoring system based on IoT technology. By constructing perception, transmission, and application layers, the system achieves intelligent and automated real-time monitoring of structural loads, internal forces, deformations, and surrounding environmental conditions in shield tunnels. Its reliability and stability are validated through an engineering case study. The results demonstrate that the monitoring system enables comprehensive tunnel surveillance, with advantages including high reliability, scalability, compact size, and ease of maintenance, making it suitable for both short-term and long-term monitoring. Implementation in practical project shows that measured data for external loads, segment ring misalignment, and longitudinal joint opening accurately reflect tunnel load and deformation states, confirming the system’s robustness. However, exposed sensors in tunnel layouts are susceptible to interference from construction machinery or personnel; thus, developing miniaturized sensors with rapid installation/removal capabilities can mitigate this issue. The proposed intelligent perception monitoring system provides foundational data support for intelligent shield tunnel construction and structural health assessment.
Buried steel pipelines are critical lifeline systems that are highly vulnerable to permanent ground deformations (PGDs) induced by geohazards such as fault rupture and landslides. Although the vulnerability of pipelines under seismic wave propagation has been widely investigated, research on landslide-induced deformation remains relatively limited, particularly studies that consider multiple failure mechanisms using advanced three-dimensional continuum models. This study presents a comprehensive framework for the structural integrity (SI) assessment of buried steel pipelines by developing both individual failure-mechanism-based and system-level fragility surfaces. A total of 216 advanced three-dimensional finite element continuum models were developed considering different steel grades, diameter-to-thickness (D/t) ratios, and 18 landslide-width scenarios. Four potential failure mechanisms, namely local buckling, beam buckling, tensile rupture, and cross-sectional ovalization, were investigated within SI limit state. Lognormal fragility surfaces were first developed for individual failure mechanisms using probabilistic demand models based on landslide displacement and width as intensity measures. The resulting fragility surfaces were subsequently integrated using the first-order reliability bounds approach to derive system-level fragility surfaces. The results indicate that local buckling and beam buckling are the dominant failure mechanisms governing pipeline performance. The D/t ratio was identified as the primary parameter governing vulnerability, with the probability of exceedance increasing significantly as the pipe wall thickness decreased. In contrast, steel grade exhibited a secondary and comparatively limited influence. Landslide width generally showed an inverse relationship with failure probability, and the magnitude of this effect varied depending on pipeline geometry and material properties. The proposed system-level fragility surfaces provide a practical and computationally efficient tool for predicting the likelihood of damage in buried pipelines subjected to landslide-induced deformations.
Local stress/strain concentration in the sealing steel liner of lined rock caverns (LRCs) can govern serviceability even when the global cavern response remains moderate. This paper develops a two-scale mechanical framework to link cavern-scale load sharing to crack-induced local response in the liner. A closed-form crack-bridging model is established for a liner strip spanning a concrete crack, accounting for membrane action, local bending and friction-controlled transfer at the steel–concrete interface. A global plane-strain FE model of a demonstration LRC simulates excavation, lining installation and pressurization, whereas refined local sub-models with explicit cracks and contact are used to resolve the near-crack hot spots. The global model indicates that the rock mass and the 650-mm concrete lining carry most of the internal pressure, so the mean hoop strain in the 10-mm steel liner remains about 0.44 ‰ at 10 MPa. Once the concrete lining cracks, however, local bending and interfacial slip produce pronounced stress concentration near the crack. Parametric analyses show that the maximum von Mises stress increases with internal pressure, interface friction coefficient and crack width, but decreases with liner thickness; in the studied case, the local peak reaches the steel yield strength when crack width approaches 0.37 mm. The framework quantifies how pressure, crack width, interface friction, and liner thickness affect local liner demand, and may support crack control and liner design in LRCs.
Underwater tunnels beneath rivers and lakes are frequently subjected to leakage issues. However, the instability mechanism caused by particle loss remains insufficiently understood, particularly in pebble formations. This study combines experimental and numerical approaches to investigate the influence of particle loss on formation instability. The results indicate that the process of particle loss can be divided into intense erosion, erosion mitigation and stable seepage, and the anti-erosion capacity of gap-gradation pebble soils is the worst. When leakage occurs at the tunnel vault or sidewall, a continuous seepage erosion channel eventually propagates from the leakage location toward the ground surface. For leakage at the tunnel bottom, the particle disturbance zone is confined within the tunnel contour. As the leakage width increases, the disturbance zone continuously expands, and the instability evolution accelerates. The particle loss illustrates an initial rapid increase, followed by a trend toward stabilization after a certain step. Moreover, the calculation step required to reach the steady state decreases with increasing hydraulic pressure. The microscopic contact force above the vault deflects horizontally to form temporary arching effect near the leakage, determining the evolution of particle loss and formation deformation. These finds can provide references for the early warning and risk assessment of tunnel leakage disasters in pebble formations.
This study investigates that the flooding induced vulnerability of Asia metropolitan area (MA) with the impact on urban underground space (UUS). Three exploratory questions guide the analysis: (1) how socioeconomic, population, and infrastructure indicators are associated with national flood economic losses and fatalities; (2) what trends and spatial patterns characterize flood vulnerability in Asian MA, and which regions show the strongest vulnerability; and (3) how metro-system indicators modify economic-loss and casualty-vulnerability patterns in large cities and urban agglomerations. The national flood deaths and direct economic losses from 2014 to 2024 are used to derive empirical weights for the relative importance of indicator families. This design is consistent with Analytical Hierarchical Process (AHP) logic, which ranks factors through pairwise importance comparison before producing a weighted overlay. The calibrated weights are then applied to 1 km gridded exposure layers in an Asia equal-area projection. The following new findings are obtained: (i) economic-loss vulnerability is shaped by a relatively balanced combination of GDP, population, built-up area, road length, and metro-system length, whereas fatality vulnerability is mainly population-led. (ii) high vulnerability is spatially concentrated in coastal, deltaic, river-basin, and major metropolitan corridors, especially the Seoul MA, Greater Tokyo MA, the Pearl River Delta MA, the Yangtze River Delta MA, the Ganges-Brahmaputra Delta MA, western coastal India, and major Southeast Asian urban corridors. (iii), metro-system length makes hotspots more prominent in high rail-oriented metropolitan regions, while high population and high built-up area make hotspots more severe.
Cutterhead clogging is a common problem in Earth Pressure Balance (EPB) tunnelling through clayey and mudstone-dominant ground, where adhesion and accumulation of excavated muck may reduce penetration capability, increase energy consumption, and disturb normal excavation. This study develops an interpretable data-driven warning framework for cutterhead clogging by integrating structural dependency analysis, hybrid deep learning prediction, warning-oriented evaluation, and field validation. A Peter–Clark (PC)-based analysis is first used to examine the structural relevance of geological, operational, and soil-conditioning variables. A three-branch Hybrid Temporal Deep Neural Network (HTDNN) is then developed to predict three clogging-related excavation-response indicators, namely the Field Penetration Index (FPI), Torque Penetration Index (TPI), and Specific Energy (SE). Based on the coupled evolution of these indicators, an entropy-weighted rank-sum ratio (RSR) method is adopted to construct a warning score for stage-wise identification of clogging development, while Monte Carlo-based analysis is used to supplement the interpretation of engineering-risk evolution. The results indicate that cutterhead clogging in the studied section is a progressive and cumulative process rather than a sudden event. The proposed framework captures the transition from the early warning stage to the proximity warning stage and finally to the clogging stage, and the warning results agree with the chamber-opening observation of mud cake adhesion on the cutterhead. In practical application, the resulting warning level and risk tendency can support timely tunnelling-parameter adjustment, enhanced soil conditioning, muck-discharge inspection, and chamber-opening preparation.
The structural behavior of a 640 m long stretch of the segmental lining in the north tube of the Koralm tunnel’s construction lot KAT3 is analyzed. Over a period of 3.5 years, circumferential normal strains were measured in nine measurement rings, each consisting of seven tubbings equipped with a centrally located strain sensor pair. A hybrid analysis combines the monitoring data with (visco)elastic modeling of concrete and steel as well as kinematics of slender circular arch theory. This allows for computing circumferential normal forces and axial bending moments at longitudinal sections through the measurement tubbings containing the strain sensor pairs. Cubic splines are used for circumferential interpolation between sensor positions and longitudinal interpolation between measurement rings. The obtained fields of internal forces are translated into fields of utilization degrees of both the reinforced concrete tubbings and the longitudinal joints consisting of plain concrete. The utilization degrees are illustrated in interaction diagrams. All computed pairs of normal force and bending moment – including those transmitted across reinforced tubbing sections – fall within the ultimate capacity boundary of the longitudinal joints. This is consistent with the interpretation that the longitudinal joints play a role in controlling lining stresses. This is particularly the case in highly utilized joints, where concrete exhibits nonlinear creep. This amplifies the overall deformability of the segmental tunnel ring, and thus provides a possibility for the lining to follow the deformation of the surrounding ground mass without activating overly high internal forces.
The planning of metro-led underground spaces often neglects public perception, while traditional methods struggle to translate unstructured social media data into actionable insights. This study introduces a novel framework that integrates large language models (LLMs) and machine learning to quantify public experience systematically. Using 25,349 Google Maps reviews for 91 Hong Kong metro stations, an LLM engine performs fine-grained perception extraction across 14 indicators. We propose a “perception value” (PV) metric that synthesizes perception frequency and preference, significantly improving modeling performance. Machine learning analysis identifies “transfer” as the most critical contribution to public satisfaction. Notably, “barrier-free design” emerges as a high-contribution but low-preference factor, signaling an urgent need for improvement. Spatially, our analysis reveals a distinct “center-periphery” pattern in perception quality, enabling targeted renewal strategies. Our framework provides a scalable, interpretable paradigm for transforming public perception into planning intelligence, advancing human-centered governance for urban underground spaces.
Geological tectonic processes lead to the widespread development of joints and fractures in deep rock masses. During service, such rock masses are affected not only by rainfall infiltration and groundwater seepage but also by dynamic disturbances such as blasting and earthquakes. The combined effects of seepage, static stress, and dynamic loading alter the stress and deformation states of fracture surfaces, promoting crack initiation, propagation, and coalescence and thereby inducing complex failure responses. As structural weaknesses in rock masses, joints and fractures substantially reduce the overall load-bearing capacity and constitute key factors governing the mechanical response and failure modes of rock masses subjected to seepage and dynamic disturbances. To investigate the dynamic mechanical response and energy-absorption characteristics of bolted fractured rock under a simulated deep hydro-mechanical environment, bolts were fabricated from high-strength, high-toughness, and high-ductility negative Poisson’s ratio (NPR) steel, 45# steel, and Q235 steel. Hydro-mechanically coupled split Hopkinson pressure bar (SHPB) tests were then conducted to comparatively examine the dynamic strength, secant modulus, failure modes, and energy-dissipation characteristics of fractured sandstone reinforced with the three bolt types at different fracture inclination angles. The tests were performed under a fixed hydro-mechanical environment comprising an axial static stress of 27 MPa, a confining pressure of 27 MPa, and a seepage water pressure of 6 MPa, with fracture inclination angles of 0°, 15°, 30°, 45°, and 60°. The results show that NPR bolts markedly improve the dynamic load-bearing and energy-absorption capacities of fractured sandstone. The dynamic peak stresses of the NPR-bolted specimens were 10.22%–39.44% and 19.85%–48.12% higher than those of the specimens reinforced with 45# steel and Q235 steel bolts, respectively. For all three bolted specimen groups, the peak stress first increased and then decreased with increasing fracture inclination, while the NPR-bolted specimen reached a maximum peak stress of 80.00 MPa at an inclination of 15°. Failure-mode analysis indicates that the bolts restrained relative fracture displacement by carrying axial and transverse loads and enhancing resistance to sliding along the fracture surface. At 15°, the NPR-bolted specimen was characterized primarily by tensile cracking and slight particle spalling, without evident block instability. Based on the experimental analysis, a theoretical model incorporating dynamic loading, axial static stress, confining pressure, and seepage water pressure was established. Parameter sensitivity analysis was further conducted to examine the potential effects of environmental parameters on the distribution of axial and shear forces in the bolts, thereby revealing the governing role of the angle between the bolt and fracture plane in the mechanical response stages and energy evolution. These findings provide a theoretical basis and parameter support for the design of rock-bolt reinforcement in deep engineering rock masses subjected to dynamic disturbances.
The master planning of urban underground space (UUS) is essential for guiding orderly development, spatial coordination, and sustainable resource allocation. However, existing UUS planning remains largely static and blueprint-oriented, with insufficient attention to the spatial equilibrium among development demand, supply capacity, and existing utilization performance. Taking the central urban area of Jinan, China, as a case study, this research develops a data-driven spatial equilibrium framework based on multi-source urban data. A multi-dimensional spatial autocorrelation analysis matrix (MSAAM) is constructed by integrating bivariate spatial autocorrelation analysis (BSAA) results to identify local–neighbor relationships and differentiated UUS development strategies. The identified combinations are classified into five types: Internal Development, External Development, Balanced Development, Limited Development, and Self-contained Development. Among the 869 classified spatial units, Limited Development accounts for the largest proportion (57.5 %), followed by Balanced Development (23.2 %), Internal Development (7.8 %), Self-contained Development (7.5 %), and External Development (3.9 %). Balanced Development areas are concentrated in mature built-up areas, whereas Internal and External Development areas occur mainly in urban expansion and transitional zones. Limited and Self-contained Development areas are primarily located in urban–rural fringe areas, ecological corridors, and low-density peripheral zones. These findings show that UUS development is shaped by spatial differences in the alignment of demand, supply, and benefit. The framework supports differentiated master planning, strategic zoning, development prioritization, and more efficient allocation of underground space resources.
To address the critical challenges of constructing deep shafts in urban central areas, where water‑rich soft soils, strict disturbance control requirements, construction safety, and operational efficiency impose significant constraints, a novel Actively Controlled Prefabricated Press‑in Shaft (ACPP) method has been developed. The core of this method lies in an integrated system that unifies active sinking control, automated underwater excavation, and digital monitoring within a human‑in‑the‑loop control framework. The system is built upon a dedicated boring machine featuring four key hardware innovations: (1) a synchronized hydraulic jack system enabling active force and attitude control; (2) a multi‑degree‑of‑freedom subaqueous excavation robot with automated kinematic control for precise soil removal; (3) a specialized friction‑reducing grouting system to minimize sinking resistance; and (4) an auxiliary support system designed to maintain hydraulic equilibrium and eliminate dewatering risks. Serving as the central platform that orchestrates these hardware components, a real‑time data‑driven visualization and control system integrates multi‑source information to enable remote precision decision‑making, state transparency, and closed‑loop active regulation. The ACPP system performs active sinking regulation through force‑capacity modulation and attitude correction, achieved via differential jacking and zoned excavation management. The method’s effectiveness was demonstrated through a successful application in a 29.5‑m‑deep emergency shaft in Shanghai, where environmental disturbance was minimized (settlement ≤ 5 mm, compared to approximately 30 mm in conventional methods), construction efficiency doubled, and structural quality exceeded standards (verticality 0.62‰, ovality 1.2‰). The ACPP method shows strong potential for widespread application in deep shaft excavation within highly congested urban environments, and its applicability in hard or heterogeneous strata remains to be further validated.
Time-delayed rockbursts have become a major and fatal technical disaster as deep plateau tunnel construction advances. Using in-situ microseismic monitoring and heterogeneous numerical modelling, this work investigates the disaster-inducing process of rockburst in locally degraded surrounding rock during a typical time-delayed rockburst occurrence at the crown of a plateau railroad double-line tunnel. The spatiotemporal evolution of rockburst precursors is investigated using seismological metrics (b-value) and spatial fractal dimensions (D-value). We describe a numerical equivalent approach that uses MS energy inversion for initial damage. The Weibull distribution and dynamic simulation tools are used to reconstruct the physical contact process between the softened crown zone and the lining. The results show that the main fundamental cause of delayed collapse is the localised deterioration of nearby rock brought on by previous blasting. Both the b-value and D-value show an abrupt and simultaneous fall (b < 0.7, D < 1.7) before to the rockburst, which can serve as a reference for identifying such latent disasters. Additionally, the lining’s uniform radial support is weakened by the softened region in the surrounding rock. The subsequent blasting in the next tunnel quickly releases the elastic energy contained in the deep load-bearing arch. As a result, the crown lining experiences uneven loading, which is characterised by “outer compression and inner tension,” which eventually results in macroscopic spalling failure. In order to clarify the energy-driven disaster mechanism in potential hazard zones subjected to combined dynamic and static loads, this study combines MS precursor data with a damage evolution model for surrounding rock. This creates a solid theoretical foundation for the early warning and prevention of rockbursts in deep underground engineering.
Axial Chain Rockbursts (ACRs) represent a critical hazard in the construction of deep-buried tunnels under high-geothermal conditions, yet their triggering mechanisms under such environments remain inadequately understood. This study systematically investigates ACRs through integrated field surveys, temperature monitoring, microseismic analysis, and true triaxial testing in a typical high-geothermal tunnel in Southwest China. Key findings reveal that ACRs preferentially develop in high-temperature dry rock masses with steeply inclined gneissic foliation aligned at small angles to the tunnel axis. Microseismic activity and energy release are significantly more pronounced in moderate-to-high temperature zones compared to low-to-high temperature zones. While the strength of gneiss remains relatively temperature-insensitive, its energy storage capacity increases markedly with temperature, exhibiting a 40.44 % rise in elastic strain energy at 91.3 °C compared to 33.4 °C. The nucleation of ACRs evolves through four distinct stages: energy accumulation with initial tensile fracturing, progressive fracturing and event clustering, sequential rockburst occurrence accompanied by a transition from tensile to mixed-mode and shear failures, and eventual energy depletion leading to stabilization. These insights provide a mechanistic basis for predicting and mitigating ACR risks in future deep geothermal tunnel projects.
Compaction grouting is widely used to mitigate excessive and differential settlement of operational metro tunnels in soft clay. This study proposes a new analytical model to predict the uplift force, grouting-induced additional loads, and uplift displacement of metro tunnels. The grouted zone is represented by multiple discrete spherical grout bulbs. The uplift force is derived using spherical cavity expansion theory, while the additional loads is determined by combining the superposition principle with Mindlin’s solution. The tunnel is modelled as a Timoshenko beam on a Winkler foundation to obtain its uplift response. The proposed model is validated against field monitoring data from a compaction-grouting project on Nanjing Metro Line 10, showing good agreement with the measured tunnel deformation. Parametric analyses indicate that the additional loads and tunnel uplift increase with grout volume and grouting pressure, but decrease with tunnel burial depth and the distance between the tunnel and grouting zone. Among all influencing parameters, grouting pressure exhibits the highest sensitivity to both additional loads and uplift displacement, whereas tunnel burial depth shows the weakest effect. When the grouting pressure increases by 0.2 MPa, the additional loads and uplift displacement increase by about 70 kN/m and 5 mm, respectively, corresponding to a growth rate of about 50%. In contrast, when the tunnel burial depth increases by 2 m, the additional loads and uplift displacement decrease by approximately 4.2 kN/m and 0.45 mm, respectively, corresponding to a reduction rate of about 2.1%. The proposed model can provide theoretical support for the uplift design of operational metro tunnels induced by compaction grouting in soft clay.
