In the seismic mountainous regions such as western China, it is usuallly inevitable to construct tunnels near active fault zones. Those fault-crossing tunnel structures can be extremely vulnerable during earthquakes. Extensive experimental studies have been conducted on the response of continuous mountain tunnels under reverse and normal fault movements, limited experimental investigations are available in the literatures on mountain tunnels with special structural measures crossing strike-slip faults. In this study, a new experimental facility for simulating the movement of strike-slip fault was developed, accounting for the spatial deformation characteristics of large active fault zones. Two groups of sandbox experiment were performed on the scaled tunnel models to investigate the evolution of ground deformation and surface rupture subjected to strike-slip fault motion and its impact on a water conveyance tunnel. The nonlinear response and damage mechanism of continuous tunnels and tunnels incorporated with specially designed articulated system were examined. The test results show that most of slip between stationary block and moving block occurred within the fault core, and significant surface ruptures are observed along the fault strike direction at the fault damage zone. The continuous tunnel undergoes significant shrinkage deformation and diagonal-shear failure near the slip surface and resulted in localized collapse of tunnel lining. The segments of articulated system tunnel suffer a significant horizontal deflection of about 5°, which results in opening and misalignment at the flexible joint. The width of the damaged zone of the articulated system tunnel is about 0.44 to 0.57 times that of the continuous tunnel. Compared to continuous tunnels, the articulated design significantly reduces the axial strain response of the tunnel lining, but increases the circumferential tensile strain at the tunnel crown and invert. It is concluded that articulated design provides an effective measure to reduce the extent of damage in mountain tunnel.

To improve the accuracy and efficiency of coal-rock interface recognition, this study proposes a model built on the real-time detection algorithm, you only look once (YOLO), and the lightweight bilateral segmentation network. Simultaneously, the regional similarity transformation function and dragonfly algorithm are introduced to enhance the quality of coal-rock images. The comparison with three other models demonstrates the superior edge inference performance of the proposed model, achieving a mean Average Precision (mAP) of 90.2 at the Intersection over Union (IoU) threshold of 0.50 (mAP₅₀) and 81.4 across a range of IoU thresholds from 0.50 to 0.95 (mAP_[50,95]). Furthermore, to maintain high accuracy and real-time recognition capabilities, the proposed model is optimized using the open visual inference and neural network optimization toolkit, resulting in a 144.97% increase in the mean frames per second. Experimental results on four actual coal faces confirm the efficacy of the proposed model, showing a better balance between accuracy and efficiency in coal-rock image recognition, which supports further advancements in coal mining intelligence.

High smoke extraction efficiency and a relatively stable smoke layer stratification are both expected in tunnel ventilation systems. The purpose of this paper is to explore the overall performance of mechanical board-coupled shaft under different ventilation strategies. A total of 57 simulations were conducted, and the effects of the distance between the shaft and board () and ventilation velocity on the overall performance were investigated. The results indicate that the performance of smoke extraction and control will be improved by the application of mechanical ventilation and board. Smoke movement patterns under different working conditions were different, for cases of the smoke could propagate through the whole tunnel without backflow, while for cases of, the backflow exists and the smoke movement can be separated into three periods (propagation, stagnation, and retraction). The critical criterion of backflow was investigated and a simple model was deduced to estimate the maximum propagation length. Moreover, the dimensionless time for the smoke flow to reach its maximum propagation length was established. Finally, a comprehensive index was proposed to evaluate the synergistic effects of smoke extraction and control performance. These studies may provide positive significance for the ventilation design.

This study investigated the potential use of microbially induced calcite precipitation (MICP) to prevent seepage in shield tunnels with the aim of decarbonizing tunnel engineering. An apparatus was developed to conduct scale model tests to evaluate the effectiveness of using MICP for shield tunnel seepage control. To understand the MICP process and its induced change in seepage flow rate, a series of 1-g physical model tests were conducted using the designed apparatus to investigate the effect of injection methods, grouting pressure, and calcium carbonate (CaCO₃) content produced as well as its distribution on the reduction of seepage flow rate for thephysical tunnel model with different backfills behind its linings. The variation law of the pore pressure near grouting hole of the tunnel segment was also revealed. Results indicated that when the amount of CaCO₃ precipitation in sand-grout mixtures was 10.53% and 10.12%, water seepage flow rate for thephysical tunnel modelwith Fujian- and coarse-sand-grout backfill respectively reduced by 94.3% and 73.8% of their respective initial values, and S-wave velocity increased by 89.6% and 84.9% for Fujian- and coarse-sand-grout mixture, respectively. The grouting pressure needed to be controlled within a certain range to prevent the unstable CaCO₃ precipitates from being washed away. The testing results also showed that the one-phase injection method was more effective in controlling seepage water into a shield tunnel. Based on the findings of the scale model tests, some vital considerations and suggestions were presented on the use of MICP approaches for shield tunnel seepage control.

With the expansion of international terrorism and the potential threat of attacks against civil infrastructure, the dynamic response and failure modes of underground tunnels under explosive loads have become a prominent research topic. The high cost and inherent danger associated with explosion experiments have limited current research on tunnel internal explosions, particularly in the context of scaled model tests of shield tunnels. This study presents a series of scaled model tests under 1g-condition simulating internal blast events within a shield tunnel based on the prototype of the Shantou Bay Tunnel, considering the influences of surrounding stratum and equivalent explosive yield. Three different TNT explosive yields are considered in the model tests, namely 0.2, 0.4, and 1.0 kg. The model tests focus on the damage behavior and collapse modes of the shield tunnel lining under internal explosive loads. The model tests reveal that the shield tunnel is prone to damage at the joints of the tunnel crown and shoulder when subjected to internal explosive loads, with the upper half of the tunnel lining experiencing segment collapse, while the lower half remains largely undamaged. As the TNT equivalent increases, the damage area at the tunnel joints expands, and the number of segment failures in the upper half of the tunnel rises, transitioning from a damaged state to a collapsed state. The influence of “stratum-structure” interaction is investigated by comparing two models, one with overburden soil and the other positioned at the ground surface. The model tests reveal that the presence of soil pressure and confinement can significantly enhance the tunnel resistance to internal blast loads. Based on the observation of the model tests, five different damage modes of segment joints under internal explosion are proposed in this study.

The present study develops a novel type of active joint node-bolt fasten wedge (BFW) active joints, aiming to investigate the load-bearing capacity of a BFW joint in a quantitative way and put forward precise formulas for its yield load and compression rigidity. To achieve this, indoor axial loading tests were conducted on two BFW joints, accompanied by a set of numerical simulations with the finite element approach implemented in ABAQUS. Parametric research was then conducted to assess the impact of various factors on the yield load and initial compression rigidity of BFW joints, leading to the derivation of precise calculation formulas for accurate prediction of these parameters. The key findings indicate that enhancing the bolt strength from 10.9 to 12.9 significantly improves mechanical performance. Under axial compression, the final bearing force, yield load, and initial compression rigidity increase by 0.86, 1.06, and 0.15 times, respectively. Numerical models accurately predict joint behavior under axial force, confirming their reliability. Parameter studies reveal that increasing web and eaves thickness, bolt strength, and diameter improves bearing capacity, while splint thickness has little effect. The fitting formulas introduced can precisely estimate yield load and rigidity, providing practical value for engineering applications.

The use of shield method in tunnel construction is limited by the engineering conditions of highwater pressure. This is mainly due to the uncertainty of the pressure-bearing capacity of the sealing chambers of the shield tail under different grades and conditions when subjected to different external water pressures. Therefore, it is crucial to determine the pressure-bearing capacity of the sealing chambers. However, there is a lack of studies on the calculating method of the pressure-bearing capacity, which requires more theoretical investigation. To explore the common patterns of multi-grade sealing-related parameters and quantify the pressure-bearing capacity of the sealing chambers, a breakdown and leakage model of the shield tail is proposed, targeting the basic sealing unit of the system. Based on non-Newtonian fluid dynamics and fractal theory of porous media, the model is used to calculate the breakdown pressure and grease seepage rate corresponding to tunneling and shutdown states. In addition, a hydraulic breakdown device of the sealing unit of the static shield tail is built to investigate the relationship between the shield tail clearance and the shield tail brush porous media area, which helps to verify the theoretical model. Finally, the analysis of sealing chamber geometry parameters, grease rheological parameters, and an environmental parameter using the proposed theoretical model shows that the pressure-bearing capacity of the shield tail can be improved by increasing the shield tail clearance and grease yield stress. It also shows that the length of the sealing chamber and the plastic viscosity of the grease do not have a significant effect on the breakdown pressure of the shield tail. The model proposed in this paper will provide ideas for the calculation of the pressure-bearing capacity of multi-grade sealing chambers in the future.

Shield tunnel is a type of linear underground structure assembled by lining segments, characterized with long joint, weak stiffness, and strict deformation control requirement. The situation of the long-term deformation and defect of the shield tunnel in soft ground in coastal area of China is severe, mainly attributed to the tunneling-induced ground consolidation, frozen cross passage, groundwater pumping, cyclic train load, and nearby construction. Shield tunnel is buried in ground, and the above factors could result in underlying ground settlement, overlying ground loading/unloading, and at-side ground unloading. As a result, the tunnel could suffer from different types of structural deformation and defect. Based upon the aforementioned different reasons, this study investigates the characteristics of the tunnel deformation and defect corresponding to the different types of ground stress change and deformation. It is found that tunneling-induced ground consolidation, frozen cross passage, groundwater pumping, and cyclic train load mainly contribute to the longitudinal differential settlement but negligible transverse convergence, associated with water leakages at circumferential joints. In comparison, surface surcharge and at-side unloading not only cause significant longitudinal differential deformation but also increase transverse lining internal forces, resulting in water leakages at circumferential joints, longitudinal lining concrete cracks and water leakages. Finally, nearby construction could strongly disturb the ground and cause the generation of excess pore-water pressure, making the shield tunnel deformation develops continuously after the nearby construction is completed.

Relating the ground motion intensity measure (IM) and the structural engineering demand parameter is a crucial step in the performance-based earthquake engineering framework. This study investigates the selection of IM for development of probabilistic seismic demand model of urban shield tunnels subjected to earthquake ground motions in liquefiable and non-liquefiable soils. Nonlinear dynamic effective stress analyses are conducted to develop a database of the intensity measures and structural seismic responses exposed to ground shaking and soil liquefaction. Two advanced soil constitutive models (i.e., Pressure DependMultiYield03 and PressureIndependMultiYield for liquefiable and non-liquefiable soils, respectively) are employed to capture the nonlinear behavior. A suite of 23 ground motion intensity measures is selected and assessed based on the evaluation criteria of correlation, efficiency, practicality and proficiency. Eventually, the multi-level fuzzy comprehensive evaluation method is employed to comprehensively consider the four evaluation criteria and establish the optimal ground motion IM suitable for probabilistic seismic demand analysis of shield tunnel structures. The obtained results show that the sustained maximum acceleration is the optimal IM for evaluating the structural seismic response, followed by the peak ground acceleration in both liquefiable and non-liquefiable soils. Peak pseudo velocity spectrum, displacement square integral and Housner spectral intensity are found to be not suitable for the probabilistic seismic demand analysis of shield tunnel structures.

The excavation of deep foundation pits can cause variations in the displacement and stress fields of surrounding soils, which hence induces adverse effects on adjacent structures. This study presents a two-stage method to quantify the impact of the excavation of a deep foundation pit on the adjacent double-curved arch bridge in the historical city of Nanjing, Southeastern China. The entire process of the foundation pit excavation was simulated and the induced deformation of the arch foot was obtained in the first stage by hardening soil model with small-strain stiffness. Then, the obtained deformation of the arch foot was applied to the bridge structure as a displacement boundary in the second stage to calculate the internal forces and deformations of the double-curved arch bridge structure. The tensile strength of concrete is taken as the limit value of the tensile stress of the double-curved arch bridge. The limit values of arch foot displacement under four evaluation conditions are obtained by step loading calculation. The present results provide construction guidance and safety warning for the process of foundation pit excavation adjacent to double-curved arch bridges for historical preservation.

Group excavations are composed of several individual excavations adjacent to each other with simultaneous or successive construction sequences (CS), which are distinctive from individual excavation in terms of the performance of excavation. In this study, a hyper-scale 3D finite element model was established to investigate the deformation behavior of a diaphragm wall system retaining a deep and oversized group excavation (DOGE) in Shanghai soft clay deposits. The numerical model simulated the practical construction stages and sequences, and it was verified by a series of comparisons with field measurements. Based on the numerical model, the spatial effect of the performance of DOGE in the process of excavation stages was investigated in this study, which cannot be addressed by limited field measurements. Furthermore, the effects of partition walls and CS on the deformation control were discussed to provide practical suggestions for oversized and deep excavations. The results indicate that the employment of bi-partition walls to divide the oversized excavation into several small pits and mono-partition walls and cross walls to further divide the pits near the metro lines into smaller ones, was proved to have significant effectiveness in controlling the wall deflection and protecting the adjacent metro line. For the partition wall, the magnitude and direction of the wall deflection primarily depended on the initial excavation, while the influence of subsequent excavation activities proved insignificant. Thus, it should be noted that the effect of the initial excavation should be especially concentrated. The findings can help optimize similar DOGE engineering.

To address the limitation of traditional machine learning models in explaining the rockburst intensity prediction process, this study proposes an interpretable rockburst intensity prediction model. The model was developed using 350 sets of actual rockburst sample data to explore the impact of input metrics on the final rockburst intensity level. The collected data underwent pre-processing using the isolation forest algorithm and synthetic minority oversampling technique. The random forest model was optimized through 5-fold cross-validation and the Optuna framework, resulting in the establishment of an Optuna-random forest (Op-RF) model that generates decision rules through its internal decision tree, utilizing the properties of the random forest model. The model was further interpreted using the Shapley additive explanations algorithm, both locally and globally. The results demonstrate that the proposed model achieved an area under curve score of 0.984. In comparison to eight other machine learning models, the proposed Op-RF model demonstrated superior accuracy, precision, recall, and F1 score. The model provides a transparent explanation of the prediction process, linking impact characteristics to the final output. Additionally, a cloud deployment method for the rockburst intensity prediction model is provided and its effectiveness is demonstrated through engineering verification. The proposed model offers a new approach to the application of machine learning in rockburst intensity prediction.

Accurate geomechanical parameters are key factors for stability evaluation, disaster forecasting, structural design, and supporting optimization. The intelligent back analysis method based on the monitored information is widely recognized as the most efficient and cost-effective technique for inverting parameters. To address the low accuracy of measured data, and the scarcity of comprehensive datasets, this study proposes an innovative back analysis framework tailored for small sample sizes. We introduce a multi-faceted back analysis approach that combines data augmentation with advanced optimization and machine learning techniques. The auxiliary classifier generative adversarial network (ACGAN)-based data augmentation algorithm is first employed to generate synthetic yet realistic samples that adhere to the underlying probability distribution of the original data, thereby expanding the dataset and mitigating the impact of small sample sizes. Subsequently, we harness the power of optimized particle swarm optimization (OPSO) integrated with support vector machine (SVM) to mine the intricate nonlinear relationships between input and output variables. Then, relying on a case study, the validity of the augmented data and the performance of the developed OPSO-SVM algorithms based on two different sample sizes are studied. Results show that the new datasets generated by ACGAN almost coincide with the actual monitored convergences, exhibiting a correlation coefficient exceeding 0.86. Furthermore, the superiority of the OPSO-SVM algorithm is also demonstrated by comparing the displacement prediction capability of various algorithms through four indices. It is also indicated that the relative error of the predicted displacement values reduces from almost 20% to 5% for the OPSO-SVM model trained with 25 samples and that trained with 625 samples. Finally, the inversed parameters and corresponding convergences predicted by the two OPSO-SVM models trained with different samples are discussed, indicating the feasibility of the combination application of ACGAN and OPSO-SVM in back analysis of geomechanical parameters. This endeavor not only facilitates the progression of underground engineering analysis in scenarios with limited data, but also serves as a pivotal reference for both researchers and practitioners alike.

Due to the planning of the subway route, it is difficult to avoid crossing soft soil site conditions at subway stations. The seismic response of subway station structures is closely related to the surrounding soil site. In this paper, centrifuge shaking table tests were designed and carried out for subway station structures at three typical soft soil sites (all-clay site, liquefiable interlayer site, and fully liquefiable site). The test results are as follows. The structure is most severely damaged in all-clay site, while the damage is low in liquefiable interlayer site and fully liquefiable site. For liquefiable sites, site liquefaction results in a lower soil-structure stiffness ratio. Thus liquefiable interlayer site and fully liquefiable site provide a natural seismic isolation system for structures compared to all-clay site. The limits of the inter-story drift ratio of the structure were used to evaluate the post-earthquake performance stages of the model structure in the three sites. In all-clay site, the structure is in the “immediately operational” stage after the loading condition of 0.1g and 0.32g, and in the “reparable operational” stage after the loading condition of 0.52g and 0.72g. In the liquefiable interlayer site and full liquefiable site, the underground structure is in the “normal operational” stage after the loading condition of 0.1g and in the “immediately operational” stage after the loading condition of 0.32g-0.72g.

In this paper, we develop a two-dimensional (2D) numerical model based on the finite element method to analyse the impact of fracture networks on the behaviour of pressurised lined rock caverns (LRCs). We use the discrete fracture network approach to represent the fracture system in rock obeying a power law length distribution. The LRC consisting of an inner steel lining and an outer reinforced concrete is situated within the rock mass characterised by spatially distributed and intersected fractures. An elasto-brittle constitutive relationship is adopted to characterise the deformation/failure of intact rocks, while the classical Mazars damage model is used to simulate the cracking of concrete linings. For pre-existing fractures in rock, a non-linear stress-displacement formulation is implemented to capture their normal and shear deformations. The 2D model, representing the horizontal cross-section of an LRC with its surrounding rock mass, is subject to a prescribed in situ stress condition. We explore various fracture network scenarios associated with different values of power law length exponent and fracture intensity. We analyse the damage evolution in rock/concrete and tangential strain in the concrete/steel linings. It is found that the damage within the rock mass mainly evolves in the form of wing cracks that emanate from the tips of pre-existing fractures. For damage development in the concrete lining, it is primarily induced by tensile cracking under cavern pressurisation. The damage emerges in the lining sections where pre-existing fractures are located in the tensile region around the cavern and either intersect with the cavern wall or could reach the cavern wall by promoting wing crack propagation. The results and insights obtained from our study have significant implications for the design optimisation and performance assessment of LRCs for sustainable hydrogen storage.

Small-section hydraulic tunnels are characterized by small spaces and various section forms, under complex environments, which makes it difficult to carry out an inspection by the mobile acquisition equipment. To resolve these problems, an arbitrarily adjustable camera module deployment method and the corresponding automatic image acquisition equipment with multi-area array cameras are proposed and developed. Such method enables the acquisition of full-length surface images of the hydraulic tunnels with different cross-section forms and diameters by a one-way travel, and the overlap rate and accuracy of the acquired image sets meet the requirements of three-dimensional reconstruction and panoramic image generation. In addition, to improve the speed and accuracy of traditional algorithms for tunnel surface defects detection, this paper proposes an improved YOLOv5s-DECA model. The algorithm introduces DenseNet to optimize the backbone feature extraction network and incorporates an efficient channel attention ECA module to make a better extraction of features of defects. The experimental results show that mAP, and F1-score of YOLOv5-DECA are 73.4% and 74.6%, respectively, which are better than the common model in terms of accuracy and robustness. The proposed YOLOv5-DECA has great detection performance for targets with variable shapes and can solve the problem of classification imbalance in surface defects. Then, by combining YOLOv5-DECA with the direction search algorithm, a “point-ring-section” method is established to allow rapid identification of common surface defects by detecting them layer by layer with the bottom image of the stitched panorama as the seed. The presented method in this paper effectively solves the problem that a single image fails to show the overall distribution of the defects and their accurate positioning in a whole large tunnel section and the effective features of defects in an excessively large panoramic image size are difficult to be captured by the neural network. Field applications demonstrated that the presented method is adequate for high-precision and intelligent surface defect detection and positioning for different small-section hydraulic tunnels such as circular, arch-wall, and box-shaped hydraulic tunnels.

Assembled monolithic subway station partly synthesizes the advantages of cast-in-place and precast subway stations. However, the related seismic response analysis considering the influences of vertical ground motion and aboveground structure is still scant. In this study, we firstly performed the statistical analysis on bidirectional bedrock ground motion parameters (response spectrum, duration and envelope function) using KiK-net data, and obtained some suggested values of the above parameters. Then, four sets of artificial bedrock ground motions with statistical meanings were generated and a three-dimensional finite element analysis of the seismic response of an existing two-story three-span subway station was conducted. The main results are summarized below. (1) The significant damage to assembled monolithic station under far-field strong motion firstly occurred at side middle slab; middle slab, upper column and related grouting sleeve joints were more damage-prone. (2) When horizontal peak ground acceleration stayed constant, overall the damage of far-field motion was stronger than that of near-fault motion. (3) Vertical ground motion obviously accelerated the damage progresses of various structural members at various positions, then aboveground structure further enhanced the damages and vertical displacement responses of parts of top slab. (4) For the axial force time-history of upper column during far-field strong motion, aboveground structure uplifted the baseline, and vertical ground motion increased the amplitude and advanced the obvious drop of the baseline, among which the latter effect of vertical ground motion on assembled monolithic station was stronger than that on cast-in-place station. (5) Vertical ground motion enhanced inter-story displacement during far-field strong motion, among which the influence on the upper story of assembled monolithic station could be obviously amplified by aboveground structure, and the amplification effect lagged behind the influence of vertical ground motion. Based on the results of this study, some suggestions for the seismic design of subway station are also provided.

Loess is a special type of soil whose properties are significantly affected by water. However, the grout diffusion law for backfill grouting in loess shield tunnels remains unknown. Based on a visual model experimental device, three experiments were conducted with 10%, 20%, and 30% loess moisture. A finite discrete element method was used to verify the grout diffusion mode, and parameters such as the tunnel buried depth, grout viscosity, and elastic modulus were considered to analyse the grout diffusion law. Experiments and numerical simulations show that the screening diffusion of grout occurs at low loess moisture, whereas splitting diffusion occurs at high loess moisture. The farthest splitting diffusion distance decreases as the tunnel buried depth, grout viscosity, and elastic modulus increase. In addition, based on capillary theory and geotechnical strength criteria, screening diffusion and splitting diffusion models were established. This study investigated the grout diffusion law and grout diffusion model, providing a reference for the design and construction of loess shield tunnels.

The dynamic response and failure characteristics of tunnels vary significantly under various dynamic disturbances. These characteristics are crucial for assessing structural stability and designing effective support for surrounding rock. In this study, the theoretical solution for the dynamic stress concentration factor (DSCF) of a circular tunnel subjected to cylindrical and plane P-waves was derived using the wave function expansion method. The existing equivalent blast stress wave was optimized and the Ricker wavelet was introduced to represent the seismic stress waves. By combining Fourier transform and Duhamel’s integral, the transient response of the underground tunnel under near-field blasts and far-field earthquakes was determined in both the frequency and time domains. The theoretical results were validated by comparing them with those obtained from numerical simulations using ANSYS LS-DYNA software. Numerical simulations were conducted to further investigate the damage characteristics of the underground tunnel and evaluate the effect of initial stress on structural failure under both types of disturbances. The theoretical and numerical simulation results indicated that the differences in the dynamic response and damage characteristics of the underground tunnel were primarily due to the curvature of the stress waves and transient load waveform. The locations of the maximum DSCF values differed between near-field blasts and far-field earthquakes, whereas the minimum DSCF values occurred at the same positions. Without initial stress, the blast stress waves caused spalling damage to the rock mass on the wave-facing side. Shear failure occurred near the areas with maximum DSCF values, and tensile failure occurred near the areas with minimum DSCF values. In contrast, damage occurred only near the areas with maximum DSCF values under seismic stress waves. Furthermore, the initial stress exacerbated spalling and shear damage while suppressing tensile failure. Hence, the blast stress waves no longer induced tensile failure on the tunnel sidewalls under initial stress.

A thermal-hydraulic-mechanical (THM) field coupling three-dimensional (3D) finite element (FE) program is developed for complex THM coupled problems in engineering practice. This 3D program incorporates a thermo-mechanical coupled constitutive model known as Tsinghua-Thermo-Soil. The program solves the hydraulic and mechanical fields together and the thermal field separately (i.e., the T-HM scheme). Validation is done against the analytical solutions of one-dimensional (1D) steady-state forced convection-conduction and 1D thermo-elastic consolidation processes. Additionally, effects of the dynamic viscosity coefficient and thermal expansion coefficient of water are analyzed for 1D thermo-elastic consolidation coupled problem. It is revealed that for soils in long-term consolidation and under high loading levels, convective effect is significant and the temperature distribution differs from that obtained by considering only heat conduction. A coupled THM problem of foundations involving an actual engineering energy raft is analyzed. The response of a linear elastic foundation under the combined effect of assumed long-term cyclic thermal loading and mechanical loading process is studied. The results demonstrate that heating leads to the locally accumulation of excess pore pressure and reduces settlement and differential settlement, while cooling has the opposite effects. Due to the heat injected into the foundation exceeding the heat extracted, the ground temperature within several meters of burial depth gradually increases over time, meanwhile the average differential settlement decreases.

This paper proposed a framework for muck types identification based on data augmentation-assisted image recognition during shield tunnelling. The muck pictures were collected from the shield monitoring system above the conveyor belt. The data augmentation operations were then used to increase the quality of the original images. Furthermore, the Bayesian optimisation algorithm was employed to adjust the parameters of augmenters and highlight the features of the photos. The deep image recognition algorithms (AlexNet and GoogLeNet) were trained and enhanced by the augmentation images, which were used to establish the muck types identification models and assessed by the evaluation indices. Model efficiency was analysed through the performance and time cost of training and validation processes to select the optimal model for muck types identification. Results showed that the performance of identification models could be highly increased by data augmentation with Bayesian optimisation, and the enhanced GoogLeNet performed the highest efficiency for muck types identification.