The construction of foundation pits in soil-rock composite strata presents significant challenges due to the mechanical incompatibility between overlying soils and underlying rock masses, often causing abnormal stress redistribution and local instability in conventional pile–anchor support systems. This study proposes a micropile and bored pile frame composite retaining system, wherein reinforced concrete bored piles constitute the primary bidirectional spatial frame and grouted steel micropiles serve as auxiliary members for inter-pile soil retention. The collaborative load-transfer mechanism is analytically characterized through Winkler elastic foundation theory, yielding a stiffness ratio of η = 16.2 between bored piles and micropiles. Field validation was performed at Hou Pantao Village Station on Xuzhou Metro Line 3, wherein full-field strain evolution was monitored using BOTDR distributed optical fiber monitoring throughout staged excavation to a final depth of 17.06 m, with subsequent comparative verification against three-dimensional finite element computations (MIDAS GTS NX). Results demonstrate that: (i) structural deformation concentrates predominantly within the upper 5-m soil layer, with negligible displacement below the rock surface; (ii) micropile peak bending moment migrates downward with excavation, stabilizing near the soil–rock interface; (iii) overburden thickness dominates system response, with a critical threshold at 5 m, beyond which micropile horizontal displacement increased from 5.32 mm to 19.80 mm, matching the field-monitored range(5.70–23.3 mm); (iv) reducing bored pile spacing from 9 m to 4 m decreases maximum micropile displacement by 42.1%, whereas increasing pile diameter yields only local stiffness enhancement without global deformation control.

Geological prospecting and the identification of adverse geological features are essential in tunnel construction, providing critical information to ensure safety and guide engineering decisions. As tunnel projects extend into deeper and more mountainous terrains, engineers face increasingly complex geological conditions, including high water pressure, intense geo-stress, elevated geothermal gradients, and active fault zones. These conditions pose substantial risks such as high-pressure water inrush, large-scale collapses, and tunnel boring machine (TBM) blockages. Addressing these challenges requires advanced detection technologies capable of long-distance, high-precision, and intelligent assessments of adverse geology. This paper presents a comprehensive review of recent advancements in tunnel geological ahead prospecting methods. It summarizes the fundamental principles, technical maturity, key challenges, development trends, and real-world applications of various detection techniques. Airborne and semi-airborne geophysical methods enable large-scale reconnaissance for initial surveys in complex terrain. Tunnel- and borehole-based approaches offer high-resolution detection during excavation, including seismic ahead prospecting (SAP), TBM rock-breaking source seismic methods, full-time-domain tunnel induced polarization (TIP), borehole electrical resistivity, and ground penetrating radar (GPR). To address scenarios involving multiple, coexisting adverse geologies, intelligent inversion and geological identification methods have been developed based on multi-source data fusion and artificial intelligence (AI) techniques. Overall, these advances significantly improve detection range, resolution, and geological characterization capabilities. The methods demonstrate strong adaptability to complex environments and provide reliable subsurface information, supporting safer and more efficient tunnel construction.

Insulators are crucial components for the safety and reliability of power systems. However, due to their small size and complex structure, precise segmentation of insulators is challenging. To address this issue, this paper proposes a directional attention-based PointNet++ model (PDA). The core module of PDA is the directional attention (DA) module, which consists of spatial self-attention (SSA) and channel self-attention (CSA). This module is designed to establish long-range relationships in both spatial and channel directions of the feature map, enabling global modelling. Additionally, to reduce computational costs, multi-scale pyramid pooling is embedded in both the SSA and CSA modules. Notably, by integrating DA into PointNet++, the model enhances the correlation between point cloud features and the long-range dependency of positional information without significantly increasing the computational burden. Experimental results demonstrate that the PDA model significantly outperforms existing models in segmenting insulator point clouds from multiple power transmission corridors.

The accurate and timely identification of lithology and adverse geology is crucial for the safe and efficient construction of tunnels. However, traditional methods for lithology and adverse geology identification rely excessively on the experience and accumulated knowledge of geologists, making them highly subjective and prone to misjudgement and omission. This study aims to introduce the latest advancements in lithology and adverse geology identification. First, we present an innovative high-precision method for the intelligent identification of lithology based on “pure image,” “infrared spectral,” and “image and spectral fusion” analyses. Second, we propose methods of adverse geology identification, including “element and mineral anomaly analysis,” “geological and geophysical joint inversion,” and “multi-source data fusion of borehole information,” which realize comprehensive identification of the location, shape, scale, property, and type of adverse geology ahead of a tunnel working face. Finally, we present new theories and methods for the quantitative testing and inversion of elements and minerals, multi-source data fusion for intelligent lithology identification, and adverse geology identification dual-driven by knowledge and data. Integrating and analyzing multi-source data on geology, geophysical prospecting, and advanced drilling is conducive to overcoming the limitations of single-source data and is the future development direction of accurate and intelligent lithology and adverse geology identification.

With the improvement of multisource information sensing and data acquisition capabilities inside tunnels, the availability of multimodal data in tunnel engineering has significantly increased. However, due to structural differences in multimodal data, traditional intelligent advanced geological prediction models have limited capacity for data fusion. Furthermore, the lack of pre-trained models makes it difficult for neural networks trained from scratch to deeply explore the features of multimodal data. To address these challenges, we leverage the fusion capability of knowledge graph for multimodal data and the pre-trained knowledge of large language models (LLMs) to establish an intelligent advanced geological prediction model (GeoPredict-LLM). First, we developed advanced geological prediction ontology model, forming a knowledge graph database. By using knowledge graph embeddings, multisource and multimodal data are transformed into low-dimensional vectors with a unified structure. Secondly, pre-trained LLMs, through reprogramming, reconstructs these low-dimensional vectors, imparting linguistic characteristics to the data. This transformation effectively reframes the complex task of advanced geological prediction as a "language-based" problem, allowing the model to approach the task from a linguistic perspective. Moreover, we propose the prompt-as-prefix method, which enables output generation, while freezing the core of the LLM, and significantly reduces the number of training parameters. Finally, evaluations show that compared to neural network models without pre-trained models, GeoPredict-LLM significantly improves prediction accuracy. It is worth noting that as long as a knowledge graph database can be established, GeoPredict-LLM can be adapted to multimodal data mining tasks with minimal modifications.

In order to improve and optimize the advance classification and prediction method of tunnel surrounding rock, a prediction method based on Tunnel Seismic Prediction (TSP) and Probabilistic Neural Network (PNN) is proposed. Based on the characteristics of science, maneuverability and representativeness, several factors that greatly affect rock mass classification are selected as evaluation indices based on analysis of numerous TSP data, establishing an advance classification index system for surrounding rock, and designing the "Advance classification and prediction system for surrounding rock" to predict the classification. Engineering application of Jinpingyan Tunnel of Chenglan Railway in high altitude and high intensity area of China is taken as a case study, and proved that the evaluation indices are easy to obtain and the evaluation results are accurate and reliable, and compared with Back Propagation (BP) neural network prediction results, the results show that PNN has some advantages in predicting the calculation speed of surrounding rock classification, the ability to add samples and the classification accuracy in practical engineering applications. The PNN-TSP method can be further used for other tunnel engineering.

As more and more TBM tunnels are constructed under complex conditions such as high stress and fault fracture zones, large deformation disasters of squeezing soft surrounding rocks occur frequently. However, TBM method shows poor adaptability to major engineering challenges. Equipment damage, track deviation and support structures failure are widely observed. It is important to propose an effective support scheme and to fully clarifying the mechanical properties of the support structures in TBM tunnel. In this case, this paper aims to conduct laboratory and numerical tests considering steel–concrete interface bond-slip behavior so as to verify the effectiveness of the proposed support scheme. Thus, the quantitative relationship between arch type, arch spacing, longitudinal connection strength, concrete strength, steel fiber content, and tunnel surrounding rock deformation can be obtained. The research results show that the support scheme proposed in this paper shows good applicability for deformation control of TBM tunnels under squeezing soft rock conditions, which provides a reference for TBM tunnel support under squeezing soft rock conditions.

Tunnel excavation in the southwestern area of China encounters many difficulties caused by the large deformation of extremely fractured surrounding rock, which is the product of high plate tectonic stress. Improving the initial cracking strength and ductility of shotcrete layer is one of the most important means to reduce the frequency of support structure repair, which will improve tunnel construction efficiency and ensure personnel safety. In this case, this paper introduced fiber reinforced polymer (FRP) mesh to replace the steel mesh as reinforcement to optimize the support effect of shotcrete lining. Both large-scale model test and mechanical test are conducted to compare the reinforcement effect between FRP and steel mesh. Research results indicate that replacing steel mesh with FRP can increase the ultimate deformation capacity of tunnel lining by more than four times. The parametric analysis results indicate that increasing the lining thickness, the diameter of the FRP bars and the mesh density have different degrees of effect on improving the stiffness and strength of tunnel lining. The research results can provide reference for relevant projects.

This study aims to prepare an anti-washout grout using marine soft clay to stabilize the wind power pile foundations. The cement, hydroxypropyl-methyl cellulose ether (HPMC), and sodium silicate were chosen as the main materials to modify the marine soft clay, and the fluidity, bleeding rate, setting time, rheology, anti-washout characteristics, and compressive strength were investigated. The results showed that high cement content, HPMC, and sodium silicate decreased the workability of grout while enhancing the washout resistance. The addition of HPMC and sodium silicate would increase the yield stress and plastic viscosity of the grout. The sodium silicate content was the key factor influencing the early strength, while the cement content and HPMC affect the later strength. The microstructures including the morphology of C-S-H gels and the content of ettringite and calcium hydroxide significantly changed with the addition of sodium silicate, and Friedel's salt was observed through the microstructure analysis. This study can provide effective guidance for the utilization of marine soft clay, the preparation of anti-washout grout, and the control of pile foundation stability.

Cavities under roads are one of the main reasons for early structural damage to pavements. It is necessary to conduct a structural analysis of road sections with cavities and evaluate the possibility of pavement cracking caused by different cavity sizes. In this study, an analysis method for evaluating the possibility of pavement cracking based on the load-mechanical response is proposed. An example library of the mechanical response of asphalt concrete (AC) pavements was established by numerical simulation. Based on the tensile cracking characteristics of pavements in the mechanical response research, the tensile strain at the bottom of the AC layer was selected as the key analysis parameter. Sensitivity analysis of the tensile strain was conducted, and the main factors controlling pavement cracking were determined. A tensile strain response prediction model was established using multiple linear regression, and its reliability was verified. The cavity influence coefficient ( CIC ) and pavement cracking factor ( PCF ) were constructed to analyze the cracking possibility. The variation in PCF with the cavity size and pavement structure parameters was studied. A quantitative relationship between the depth and length of the cavity for a given PCF was obtained. This law conforms to a power function. The possibility of pavement cracking can be determined by measuring the cavity size. Compared to the existing cavity management system, the proposed method provided analysis results of the cracking possibility that were more consistent when the cavity depth was small and the length was long. The findings of this study provide new insights for evaluating the possibility of pavement cracking.

The anomalous enrichment of clay minerals in fault zones is considered a key factor affecting the seismic behavior of faults, fault permeability, and mineralization. This process also weakens the rock strength and triggers deformation, instability, and geological disasters. Nevertheless, the geological factors affecting the growth and anomalous enrichment patterns of clay minerals in fault zones are not systematically understood. In this study, the anomalous enrichment patterns of clay minerals within 220 natural faults are investigated (detailed information can be found in Appendix A). The main geological factors and processes affecting the anomalous enrichment of clay minerals in fault zones are revealed. The main formation modes and occurrence forms of the clay minerals in fault zones are investigated. The anomalous patterns of clay minerals in fault zones are classified into four categories and eight types, i.e., the felsic anomaly category (illite type and kaolinite type), mafic anomaly category (chlorite type), ultramafic anomaly category (smectite type, serpentine + talc type, vermiculite type, and palygorskite + sepiolite type), and argillaceous anomaly category (mixed-layer type). The 220 case studies include different types of faults and fault rocks. Thus, the results are promising for a wide range of engineering geological applications, e.g., analysis of fault activity and permeability changes and studying the genetic mechanisms of geological disasters near faults.

Tunnel route selection is the most important component of tunnel engineering design, and it is also a complex system engineering. It needs to integrate topography, geomorphology, engineering and hydrogeological conditions, with construction and operation needs in consideration, and then adopts reasonable engineering technical plans and measures, and thus ensures the safety of tunnel construction and operation.

The laboratory test and numerical calculation were carried to explore the durability and service life of the grout-reinforced geological foundation. In this paper, the simulation test for determining penetration properties of grouting was carried out with the self-designed experimental device, and the mechanical properties and permeability coefficient were obtained. The results show that the seawater ions corroded the grout-reinforced body, and the properties had deteriorated. Meanwhile, the service life cycle prediction model was obtained according to the weakening of the deformation modulus. The calculated results demonstrate that the grout-reinforced body will fail in 29–63 years, which was far less than the concrete service life (100–120 years). In addition, the compressive strength obtained after the end of the service life was obtained as 9–27 MPa, less than the concrete strength at the same curing age. In conclusion, the failure of marine engineering is not only caused by concrete but more by the failure of the weak geological foundation strengthened by the grouting reinforcement.

Multi-layer composite support structure is widely applied in tunnel excavation but shows complicated failure behavior under rock pressure. Local joint broken, out-of-plane instability and arch-shotcrete interface slip widely exist and highly weaken the overall strength of support structure, which is inconsistent with the ideal assumption, leading to sever engineering disaster. Based on above situation, this paper adopts laboratory and numerical methods to investigate the bearing mechanism and failure behavior of arch joint, arch frame and arch-shotcrete composite lining. Then, a modified safety evaluation method of arch is proposed based on convergence-confinement method. The research results proved that, component strength and arch critical instability strength check are necessary in arch design. On the other hand, compared with thin-walled steel arch, the concrete-filled steel tube arch shows higher critical instability strength and better ductility, effective for controlling large deformation of soft rock. The research conclusions can provide reference for related tunnel design and construction.

Cement–sodium silicate grout (CSG) is widely used to control water inrush disasters, and its apparent viscosity considerably impacts its water-plugging effect. However, traditional grouting materials and methods are inappropriate for high-temperature environments as high temperatures can affect the grout's viscosity. To thicken grout and increase its apparent viscosity, viscosity-modifying admixtures can be used. In this work, two types of bentonite (calcium bentonite (Ca-B) and sodium bentonite (Na-B)) and hydroxyethyl methyl cellulose (HEMC) were used to modify CSG and laboratory tests were performed to evaluate the fluidity, gelation time and rheology. The results showed that both bentonites and the HEMC decreased fluidity and prolonged the gelation time. HEMC, Ca-B and Na-B decreased the fluidity by 46.8–60.4%, 12.5–31.5% and 17.7–39.1%, respectively, at different temperatures. HEMC, Ca-B and Na-B increased the gelation time by 23.8–50.1%, 23.3–71.4% and 20%–57.1%, respectively. Bentonite can partially resist high temperatures and improve the apparent viscosity of grout owing to its water-absorption capacity. Conversely, HEMC has a negative effect on apparent viscosity, which is attributed to the formation of a complex microstructure resulting from intermolecular cross-linking between the cement particles and HEMC, preventing the connection of sodium silicate.

Accurate and effective identification of adverse geology is crucial for safe and efficient tunnel construction. Current methods of identifying adverse geology depend on the experience of geologists and are prone to misjudgment and omissions. Here, we propose a method for adverse geology identification in tunnels based on mineral anomaly analysis. The method is based on the theory of geoanomaly, and the mineral anomalies are geological markers of the presence of adverse geology. The method uses exploration data analysis (EDA) to calculate mineral anomaly thresholds, then evaluates the mineral anomalies based on the thresholds and identifies adverse geology based on the characteristics of the mineral anomalies. We have established a dynamic expansion process for background samples to achieve the dynamic evaluation of mineral anomalies by adjusting anomaly thresholds. This method has been validated and applied in a tunnel excavated in granite. As shown herein, in the tunnel range of 142 + 800–142 + 860, the fault F37 was successfully identified based on an anomalous decrease in the diagenetic minerals plagioclase and hornblende, as well as an anomalous increase in the content of the alteration minerals chlorite, laumonite, and epidote. The proposed method provides a timely warning when a tunnel enters areas affected by adverse geology and identifies whether the tunnel is gradually approaching or moving away from the fault. In addition, the applicability, accuracy, and further improvement of the method are discussed. This method improves our ability to identify adverse geology, from qualitative to quantitative, and can provide reference and guidance for the identification of adverse geology in mining and underground engineering.