Deep Underground Science and Engineering

The construction of tunnels inevitably leads to ground subsidence, while surface collapses caused by caverns during excavation can result in significant financial losses. However, previous research on tunnel excavation behavior in karst areas has predominantly assumed uniform soil strength, neglecting the geological uncertainty arising from soil spatial variability and rarely discussing the combined effects of soil variability and geological construction conditions. Therefore, this study conducted a series of random finite element analyses to investigate ground subsidence during tunnel excavation in karst cave areas, considering the combined effects of different karst cave distances and sizes as well as soil spatial variability. It was found that shorter distances between the cavern and the tunnel resulted in greater shifts in subsided trench toward the sides of the cavern. Furthermore, larger cavern diameters amplified the effect of geological variability on soils, leading to increased dispersion of maximum ground subsidence. Ground subsistence also showed sensitivity to changes in soil modulus with the coefficient of variation (COV). Neglecting cave influences could potentially underestimate failure probabilities. Additionally, three patterns for surface subsidence were presented based on different geological conditions. Finally, this study proposed a modified Peck formula to predict ground subsidence during tunnel excavation considering the combined effects of karst cave and soil spatial variability, which may enhance current design methods for tunnel engineering.

Cracks within the surrounding rock of roadways significantly affect their stability and failure characteristics. Investigating the failure modes of roadways under different crack distribution characteristics provides a theoretical basis for roadway support design. This study, based on the principle of superposition and the Mohr‐Coulomb criterion, derives calculation equations for the stress intensity factors of cracks and the radius of the plastic zone under the influence of stress concentration in roadways. Numerical simulations of roadway failures under varying crack distribution characteristics were conducted using PFC2D. The evolution patterns of crack fracture parameters under different lateral pressure coefficients and material properties were analyzed. The results show that the surrounding rock of the roadway develops tensile and shear cracks due to stress concentration. Under the influence of overburden load, the rock mass cracks evolve into wing cracks and coplanar cracks. The development and intersection of these four crack modes are the primary reasons for differences in roadway‐bearing capacity and failure modes. A larger distance between the roadway and cracks, shorter crack lengths, larger crack positional angles, and steeper crack dip angles effectively delay the time and distance of crack intersections, reduce the number of stress curve transitions, shift AE signals to a later stage, and enhance the strength and stability of the roadway. This study provides a foundation for evaluating roadway stability and selecting support parameters.

Shear tests were conducted at a shear rate of 1 mm/min and a shear displacement of 10 mm to investigate the effect of temperature on the shear behavior of the rock–concrete interface, with the aim of revealing the interaction between the surrounding rock and the lining during tunnel fires. The experimental results obtained several key findings: The mechanical properties of rock and concrete deteriorated with temperature. The mass loss rate of rock was 1.28%, with a decrease in compressive strength of 22.77%. The mass loss rate of concrete was 7.79%, with a decrease in strength of 33.27%. Concrete is more sensitive to temperature than rock under the experimental conditions. As the temperature rose, the shear fracture surface gradually shifted from within the concrete to the rock–concrete interface. Simultaneously, compaction shear displacement and peak shear displacement increased, while peak shear stress, shear stiffness and interface fracture energy decreased. In particular, the increase in compaction shear displacement and the decrease in shear stiffness were characteristic of the shear behavior of the rock–concrete interface at elevated temperatures. Subsequently, a temperature‐dependent shear constitutive model of the rock–concrete interface, based on damage mechanics and statistical theory, was developed to eliminate reliance on specific experimental variables and serve as a reference for similar situations beyond the experimental conditions. This study contributes to advancing our understanding of the shear behavior of the rock–concrete interface at elevated temperatures.

With underground engineering projects becoming deeper and more complex, the associated safety problems, especially rockburst, have increasingly increased. Despite decades of research, effective management of rockburst continues to be a formidable challenge in underground excavations. This study presents a scientometric visualization analysis of 2449 papers and conducts a comprehensive review of 336 key studies to explore the state‐of‐the‐art developments in rockburst research. With a primary focus on the prediction and prevention of rockburst, this review identifies existing research gaps and proposes a novel framework aimed at addressing these challenges in underground excavations. The results underscore a critical disconnect between advanced prediction methods and engineering practices, which limits the ability of engineers to carry out reliable assessments of rockburst potential. This disconnection prevents the prompt development of targeted prevention strategies, further aggravated by inadequate data sharing across large‐scale projects. The review also describes the limitations of relying solely on data‐driven methodologies to address the complex challenges in the lifecycle management of underground excavations. To overcome these challenges, this study proposes an innovative framework based on an ontological knowledge base. This framework is designed to integrate multisource data and diverse analysis techniques, exploring the means toward better decision‐making in future digital underground projects.

This paper presents an investigation of well integrity during low‐temperature CO2 injection using a model of thermo‐poroelasticity with interface damage mechanics. The casing–cement and cement–formation interfaces are described using cohesive interface elements and a bilinear traction–separation law. Verification testing is performed to establish the correct implementation of the coupled thermal, hydraulic, and mechanical equations. Simulation scenarios are developed to determine well interface damage initiation and development for intact wells and wells with an initial defect in the form of a 45° $^\circ$ debonded azimuth. Each intact and defective well was simulated for 30 days of CO2 injection at selected temperatures. Under the conditions considered, tensile radial stress developed at both the casing–cement and cement–formation interfaces. Hoop stress in the cement sheath remained compressive after 30 days but with reduced magnitude at the lower injection temperature, indicating greater risk of tensile stress and radial cracking as the injection temperature was reduced. Damage occurred in two of four scenarios considered, namely, the intact and defective wells at an injection temperature of 10°C $^\circ {\rm{C}}$, and was limited to the casing–cement interface, with no damage to the cement–formation interface. Inclusion of the pre‐existing defect led to earlier damage initiation, at 2.75 days compared to 4 days, and produced a microannulus with over double the peak aperture at 0.077 mm compared to 0.037 mm. These findings emphasize the importance of accounting for initial defects and damage evolution when investigating the integrity of CO2 injection wells.

Tunnels are essential components of contemporary infrastructure, yet guaranteeing their safety, longevity, and efficiency remains a persistent challenge. Recent breakthroughs in artificial intelligence (AI) and digital twin (DT) technology provide innovative solutions for the real‐time monitoring of tunnel systems, suggesting proactive maintenance tactics and improved safety protocols. This review paper offers a comprehensive examination of the application of AI and DT methodologies in tunnel surveillance. We explore the core concepts of AI and DT and their applicability to structural monitoring, encompassing machine learning, computer vision, and sensor integration. Through the utilization of these AI‐powered technologies, engineers are equipped with unparalleled insights into the state and behavior of tunnels, facilitating the early identification of irregularities and the optimization of maintenance timelines. We discuss the array of AI techniques utilized for the immediate monitoring of tunnel systems, emphasizing their foundations, benefits, and practical uses. Numerous studies have showcased the effectiveness and adaptability of AI‐based monitoring systems in various tunnel settings. Moreover, we address the hurdles and constraints inherent in AI and DT methodologies and suggest strategies for overcoming them, such as data augmentation, interpretable AI, edge computing, and continuous monitoring. Ultimately, the incorporation of AI and DT technologies into tunnel surveillance signifies a paradigm shift, offering substantial advantages over conventional techniques. By adopting AI‐driven monitoring systems, tunnel operators can augment safety, prolong the lifespan of infrastructure, and decrease operational expenses, molding the future of subterranean infrastructure management.

Double‐wheel trench cutters display reduced efficiency while cutting through hard or extremely hard rock strata, leading to significant wear on their picks. A high‐pressure water jet‐pick combined rock‐breaking method (HPC) for double‐wheel trench cutters is proposed, addressing the challenges of low efficiency and high pick wear when cutting hard rock strata. The HPC mode integrates high‐pressure water jets to precut grooves on both sides of the picks, creating free surfaces that facilitate rock fragmentation. Experiments were conducted with the combination of high‐pressure water jet cutting and linear pick cutting under the HPC mode with granite. The HPC rock‐breaking efficiency was compared with that of conventional pick‐relieved cutting (PRC) and pick‐unrelieved cutting (PUC) modes. A numerical model based on the continuum‐discontinuum element method was developed to investigate the rock‐breaking mechanism with water jet assistance. The mechanism of rock breaking by pick with bilateral water jet assistance has been revealed. The main conclusions are as follows: (1) The HPC mode reduced the average horizontal and normal rock‐breaking forces by 36.02% and 48.78%, respectively, compared with PRC, and by 32.07% and 42.85%, respectively, compared with PUC. Furthermore, compared with the PRC and PUC modes, the HPC mode decreased specific energy consumption by 88.58% and 93.84% and increased the coarseness index of rock debris by 159.28% and 189.77%, respectively. These improvements indicate a transition from localized fragmentation to large‐chunk stripping during rock breaking, attributed to the free surfaces created by water jet grooving. (2) The free surfaces created by the water jet altered the rock mass displacement vector from radial to horizontal, promoting the formation of Λ‐shaped fractures and increasing tensile and tensile‐shear fractures. This mechanism reduced the intermediate principal stress and strain energy within the rock mass, reducing the difficulty of rock breaking. The HPC mode thus offers a promising solution for improving the efficiency of double‐wheel trench cutters in hard rock excavation, with the potential for broader application in underground diaphragm wall construction.

Carbon capture and storage (CCS) remains one of the most feasible techniques for the control of Greenhouse gas emission levels. However, there will always be risks attached to the subsurface injection of CO2. These could be in the form of leakages from the injection wellbore due to completion failure; escape of the injected CO2 to neighboring aquifers due to the heterogeneous or fractured nature of the storage site; or seepage at the surface due to inadequacy of the sealing cap rock. While all these may occur, the most cost‐effective and timely way to reduce the risk of leakages is by plugging the pathways. This is done using either traditional Cementous materials or more augmented sealants like organic gels and resins. A lot of studies in the literature have described this collection of materials within the context of CO2 conformance control. So also, there are reviews on the classification and description of these chemicals. This review provides a more systemic evaluation of these classes of chemicals. This is by providing critical analyses of how external factors like CO2, pH, brine salinity and hardness, rock mineralogy, pressure, temperature, and injectivity could affect the performance of different sealants that can be utilized. Based on these assessments, best practices for the application of the sealants, both at the testing stage in the laboratory and the pilot stage and field deployment, are suggested.

Fluid flow through a single fracture is commonly described by the cubic law. However, deviations from this model are expected because natural fracture surfaces are rough and in contact with each other in discrete regions. In this study, the interactions between fracture closure, contact area, and hydraulic characterization of mesoscopic‐scale rough fractures were investigated. First, natural‐splitting fractures induced by Brazilian splitting were generated and digitally reproduced using a three‐dimensional scanner. The distribution characterization of the initial mechanical fracture apertures was described, and the variation in fracture contact area during fracture closure was evaluated. Second, numerical simulations of fluid flow between rough surfaces were conducted, and the Stokes equations were solved. The results reveal that the hydraulic aperture underwent three stages. When the contact area was less than 0.34%, changes in the hydraulic aperture resulted in flow rate changes that were consistent with the cubic law. When the contact area was between 0.34% and 14.41%, the hydraulic aperture gradually deviated from the cubic law and fell below the mean fracture aperture. At this stage, the hydraulic aperture should be fully considered rather than the mean fracture aperture. When the contact area exceeds 14.41%, although the mechanical fracture decreases with increasing strain, the influence of the contact area on the hydraulic aperture gradually diminishes. This is because the fractures contain less fluid, and the hydraulic aperture approaches the minimum. Finally, the relationship between mechanical and hydraulic aperture strains was established, improving geological engineering applications that account for fracture deformation. image

Acoustic vibrations applied in the range from 17 to 120 kHz restore the permeability of rocks during filtration of crude oil with heavy components due to a decrease in the injection pressure. The dominant frequency of 28 kHz is revealed, at which the pore space of rocks is unblocked from heavy components of the crude oil most effectively. A mechanism for blocking and unblocking the matrix of sedimentary clastic rock from plugs composed of heavy components present in crude oil is proposed. It is established that heavy components accumulate in narrow pore necks due to the occurrence of force, the nature of which is determined by the curvature of the inner surface of the pore channel. A mathematical model of rock permeability changes from the frequency and amplitude of acoustic vibrations reflecting the resonance effect is developed. A modernized technology for well treatment is proposed.

The depth of coal mining is expected to increase continuously owing to the exhaustion of shallow coal resources. However, with the continuous increase in mining depth, transportation and lifting difficulties in deep mines have significantly increased, and traditional wire rope lifting methods can no longer meet the needs of deep transportation. Based on the principle of pipeline hydraulic lifting, a deep coal fluidization pipeline lifting system has been proposed. To address the problem of particle settlement in the horizontal connection section of large particles, a scheme involving the installation of guide vane‐type swirlers in the conveying pipeline is proposed. First, the impact of the guide vane parameters on the liquid flow field and solid particles within the pipeline was studied, and a mathematical model of the characteristic parameters of the swirler was established. Suitable guide vane parameters for the swirler were determined by considering factors such as the alleviation of particle settling, energy utilization efficiency, and the structural strength of the swirler. On this basis, the movement patterns of particles of different sizes in pipelines with and without swirlers were investigated. The study found that under conditions of high velocity and large particle size, the swirler was more effective in improving the slurry flow state within the pipeline. Subsequently, a quantitative method for determining the slurry flow state in the spiral flow pipeline was established, using the particle proportion within the pipeline section as an evaluation index, while considering flow velocity and particle size. Finally, the bench test results show that adding a swirler can reduce the critical nonsilting velocity and resistance loss of the slurry conveying pipeline by 9% and 42.9%, respectively, and we elucidate the internal mechanisms behind these reductions.

The Great Dyke of Zimbabwe is a major geological formation renowned for its rich deposits of platinum group metals. This study addresses the geological and geotechnical challenges faced during mining on the Great Dyke, focusing on the implications for hardrock pillar design. The Great Dyke's geological complexity includes diverse rock types—dunites, harzburgites, pyroxenites, and norites—and notable structural features like joints, faults, and shear zones. These factors complicate the stability of underground workings. Traditional empirical methods and numerical modeling are used in pillar design but fall short in capturing the full complexity of the Great Dyke. The study highlights the absence of advanced methods such as machine learning (ML), artificial intelligence (AI), and geostatistical techniques in current pillar design practices. Incorporating these methods could significantly enhance pillar stability. Geostatistical techniques like kriging offer detailed estimates of rock quality and quantify prediction uncertainty, while ML and AI can analyze extensive data sets to uncover patterns and improve predictions. Integration of real‐time data from Industrial Internet of Things sensors into these models allows for dynamic updates and better risk management. Continuous monitoring and adaptive design are essential for maintaining stability in this challenging geological environment. The study's findings aim to guide future mining practices, ensuring enhanced safety and efficiency on the Great Dyke.

This study focuses on hydraulic fracturing experiments conducted under triaxial conditions on tight sandstone specimens from Shivpuri district, Madhya Pradesh, India. The experiments assessed the effect of confining pressure on specimens with 10 and 12 mm borehole diameters. Breakdown pressure observations indicates that increasing confining pressure results in higher breakdown pressures for both borehole sizes, with slightly higher breakdown pressures for the 10 specimens. Using BSE‐SEM (backscattered electron‐scanning electron microscope) imaging, the fracture was obtained, and four distributions that are Weibull, Lognormal, normal, and exponential, Anderson Darling coefficient were calculated to determine their goodness of fit corresponding to measured fracture width data from BSE‐SEM images. On the basis of the Anderson Darling coefficient, lognormal distribution fitted best for both 10 and 12 mm fracture widths. Further, the Probability density function corresponding to fracture widths for both 10 and 12 mm borehole samples was determined by the maximum likelihood approach through Minitab software to reflect the variability and standard deviation of fractures. The research findings demonstrate that increasing confining pressure leads to decreased values of maximum fracture width, eventual fracture width, process zone width, and fractal dimension. In contrast, closure fracture width and tortuosity values were increased with increased confining pressures. Additionally, photomicrographs were utilized to study micro‐crack propagation through quartz and feldspar minerals, highlighting the influence of confining pressures and quartz cementation and the absence of clay inclusions on micro‐crack development. These findings contribute to the current understanding of hydraulic fracturing in tight sandstone and open up potential future research directions, such as investigating the influence of sandstone minerals and their fracture characteristics or studying the long‐term effects of hydraulic fracturing on tight sandstone specimens.

The geothermal resources in hot dry rock (HDR) are considered the future trend in geothermal energy extraction due to their abundant reserves. However, exploitation of the resources is fraught with complexity and technical challenges arising from their unique characteristics of high temperature, high strength, and low permeability. With the continuous advancement of artificial intelligence (AI) technology, intelligent algorithms such as machine learning and evolutionary algorithms are gradually replacing or assisting traditional research methods, providing new solutions for HDR geothermal resource exploitation. This study first provides an overview of HDR geothermal resource exploitation technologies and AI methods. Then, the latest research progress is systematically reviewed in AI applications in HDR geothermal reservoir characterization, deep drilling, heat production, and operational parameter optimization. Additionally, this study discusses the potential limitations of AI methods in HDR geothermal resource exploitation and highlights the corresponding opportunities for AI's application. Notably, the study proposes the framework of an intelligent HDR exploitation system, offering a valuable reference for future research and practice.

A data preprocessing workflow is proposed to address key challenges in rockburst data analysis, including dimensionality differences among various sample features, variations in data values within the same feature, missing data, poor data consistency, and sample class imbalance. The workflow is divided into four steps. Each step introduces multiple algorithms, which are sequentially combined according to the order of the four steps. Then, these coupled algorithms are utilized to preprocess the rockburst data set. The rockburst data set contains 459 samples, and the maximum tangential stress (MTS), the uniaxial compressive strength (UCS), the uniaxial tensile strength (UTS), the elastic strain energy index (WET), the rock stress factor (SCF), and the rock brittleness coefficient (B) are selected as the feature parameters. Subsequently, three architectures, Deep Neural Network (DNN), Convolutional Neural Network (CNN), and Recurrent Neural Network (RNN), are used to evaluate the data sets processed by different coupled algorithms. The hyperband algorithm is introduced to optimize the hyperparameters of the RNN model, and the prediction accuracy of different architectures is compared between the RNN model with dense layers and without dense layers. Finally, a rockburst prediction model based on data preprocessing and the Hyperband‐DNN model is developed. The prediction results show that data preprocessing can significantly improve the model prediction accuracy; the model architecture with the highest prediction accuracy can be found quickly using the hyperband algorithm; and adding the dense layer can improve the stability and prediction accuracy of the model.

Excavation unloading damages rock masses, with preferential failure along geological defects in rock engineering, which may induce catastrophe geological hazards. It is important to study the failure of jointed rock under true‐triaxial unloading conditions. 3D DEM true‐triaxial unloading modeling tests on rock with through‐going joint considering the contributing factors, that is, the joint inclination, the initial confining pressure, and the unloading point, were conducted to study the rock mechanical properties, cracking behaviors, and failure characteristics. Six typical rock failure modes were summarized based on the development of main cracks, and the associated cracking mechanisms were studied by analysis of the ratio of tensile crack to shear crack. An energy criterion determining whether unloading‐induced rock failure occurs instantaneously was proposed by comparing the accumulated elastic strain energy at the unloading point in true‐triaxial unloading simulations with the limit elastic strain energy stored in rocks in biaxial compression modeling tests. Furthermore, the strength characteristics and applicability of the Mogi–Coulomb failure criterion in describing the failure of specimens with through‐going joint under true‐triaxial unloading conditions were studied. This study provides some new insights into true‐triaxial unloading‐induced failure of jointed rock, which may be valuable for revealing the mechanism of rock instabilities influenced by geological defects in deep excavations.

Aiming at the creep characteristics of the surrounding rock of the Haidong water conveyance tunnel of the Central Yunnan water diversion project and taking the diabase at the tunnel fault zone as the research object, compression creep tests were conducted under a range of confining pressures, revealing the axial and lateral creep characteristics of diabase. Based on the theory of fractional derivative, a nonconstant coefficient Abel dashpot considering damage was established, and the simplified equation form was used to facilitate parameter identification and calculation. By combining a nonlinear Kelvin model, a classical element model, and a simplified Abel dashpot that accounts for damage effects, a constitutive model was established to describe the diabase creep process. The constitutive relations for one‐dimensional and three‐dimensional creep were provided. The results show that the sample exhibits both a decay creep stage and a steady‐state creep stage when the stress level is below the diabase's stress threshold; when the stress level surpasses the threshold, the sample goes through an accelerated creep stage with a quickly growing creep rate and ultimately fails. Diabase mostly experiences axial creep deformation at low‐stress conditions, followed by more significant lateral creep deformation under high‐stress conditions. Comparing the triaxial compression creep test curve with the theoretical curve revealed that the model more accurately characterized the nonlinear creep properties of diabase.

The two‐phase flow in porous media is affected by multiple factors. In the present study, a two‐dimensional numerical model of porous media was developed using the actual pore structure of the core sample. The phase field method was utilized to simulate the impact of displacement velocity, the water–gas viscosity ratio, and the density ratio on the flow behavior of two‐phase fluids in porous media. The effectiveness of displacement was evaluated by analyzing CO2 saturation levels. The results indicate that the saturation of CO2 in porous media increased as the displacement velocity increased. When the displacement velocity exceeded 0.01 m/s, there was a corresponding increase in CO2 saturation. Conversely, when the displacement velocity was below this threshold, the impact on CO2 saturation was minimal. An “inflection point,” M3, was present in the viscosity ratio. When the viscosity of CO2 is less than 8.937 × 10−5 Pa·s (viscosity ratio below M3), variations in the viscosity of CO2 had little impact on its saturation. Conversely, when the viscosity of CO2 exceeded 8.937 × 10−5 Pa·s (viscosity ratio greater than M3), saturation increased with an increase in the viscosity ratio. In terms of the density ratio, the saturation of CO2 increased monotonically with an increase in the density ratio. Similarly, increasing density ratios resulted in a monotonic increase in CO2 saturation, though this trend was less pronounced in numerical simulations. Analysis results of displacement within dead‐end pores using pressure and velocity diagrams reveal eddy currents as contributing factors. Finally, the impact of pore throat structure on the formation of dominant channels was examined.

One of the significant challenges faced in shield tunnel construction is the risk of collapse resulting from abrupt changes in soil conditions in advance of the working face, particularly within the sandy cobble stratum. This study aims to effectively manage these sudden variations and to further investigate the failure mechanisms associated with shield tunneling. The limit support pressure, instability patterns, and soil arch ranges at the excavation face under varying burial depths were analyzed using PFC3D discrete element software, specifically in relation to the shield section of the Beijing Subway New Airport Line 06 cigezhuang~1# wind shaft project. The findings of this research indicate the following: (1) The limit support pressure at the shield excavation face in sandy cobble strata increases with the burial depth ratio; however, the ratio of the limit support pressure decreases as the burial depth ratio increases. (2) A burial depth ratio threshold of 1.0–1.5 suggests that no soil arch forms in the stratum ahead of the excavation face when the burial depth ratio is below this range. Conversely, a soil arch develops when the burial depth ratio exceeds this threshold. (3) When the burial depth ratio is less than the threshold range of 1.0–1.5, the longitudinal instability zone transitions from a wedge shape to a barrel shape, extending to the surface. Conversely, when the burial depth ratio surpasses this range, the longitudinal instability zone changes from a wedge shape to a bulb shape and does not extend to the surface. (4) This paper proposes a calculation model for determining the limit support pressure of the excavation face under both shallow and deep burial conditions in the sandy cobble stratum, providing a calculation formula for the limit support pressure and establishing a reference range for the calculation parameters.

In deep underground engineering design, the true‐triaxial compressive strength of intact rocks is a critical evaluation index. Traditional methods for acquiring true‐triaxial strength data are hampered by labor‐intensive manual operations. To mitigate the time‐consuming nature of true‐triaxial experiments, this study leverages the unique capabilities of the relevance vector machine (RVM) to develop machine learning prediction models. These models aim to streamline the process and enhance predictive accuracy, thereby offering a more efficient alternative to conventional experimental approaches. The proposed models establish a correlation between the major principal stress (σ1) and the material constants, alongside other Hoek–Brown (H–B) strength parameters. A comprehensive data set, encompassing 408 sets of true‐triaxial experimental data from 12 different rock types, was collated from previous studies. This true‐triaxial strength data set was systematically divided into three groups based on the intact rock material content (mi), facilitating subsequent validation efforts. To enhance prediction accuracy and generalization capability, particle swarm optimization (PSO) is employed to optimize the hybrid kernel function parameters of the RVM. This study introduces a dynamic inertia weight decreasing method, demonstrating superior prediction accuracy compared to conventional PSO improvement techniques. In comparison with five three‐dimensional H–B type criteria and two other machine learning models, the improved PSO‐RVM model demonstrated superior performance across three distinct mi groups. Additionally, the proposed model is capable of generating probabilistic predictions, thereby effectively capturing the inherent uncertainty associated with rock strength. The probability distribution of model prediction errors closely aligns with that indicated by the generalized Zhang–Zhu criterion, underscoring the improved PSO‐RVM model's ability to capture the uncertainty in true‐triaxial compressive strength. Furthermore, this study explores sample selection for combined tests integrating true‐triaxial experiments and the proposed improved PSO‐RVM model, providing a tentative optimal ratio for predicting the true‐triaxial compressive strength of intact rocks.

A well‐designed cut‐hole layout is crucial for improving drilling and blasting efficiency in deep underground engineering. For this purpose, this study investigates the patterns of blast‐induced rock damage under various cut‐hole layouts and ground stress conditions through numerical simulations and field tests. The key cut‐hole parameters include uncoupling coefficients (K), the empty hole diameter (De ${D}_{{\rm{e}}}$), and hole spacing (Dc−e ${D}_{{\rm{c}}-{\rm{e}}}$). In simulations, when K exceeded 1.15, indicating the use of uncoupling charge, both the rock damage scope and the blast shock wave transmission were significantly suppressed. In contrast, the coupled charge (K = 1.0) demonstrated better performance in promoting damage. Additionally, large‐diameter empty holes (De ${D}_{{\rm{e}}}$ = 8 or 10 cm) showed pronounced free surface effects on the facing‐blasting side, completely breaking the rock mass between the charge hole and the empty holes and effectively guiding crack propagation on the back‐blasting side. Hole spacing had an inhibitory effect on effective plastic strain near empty holes; the optimal Dc−e ${D}_{{\rm{c}}-{\rm{e}}}$ value is 14 cm. Finally, field tests were conducted, and the results revealed that the suggested cut‐hole parameters enhanced cycle advancement and borehole utilization by 9.0% and 12.3%, respectively, while reducing the powder factor by 27.4%. These findings provide valuable insights for similar underground engineering applications.

The development of deep reservoirs is an emerging topic in the energy industry. This paper analyzes the challenges in simulations of flow and transport in deep reservoirs and introduces models and algorithms aimed at resolving these challenges. A fast, accurate, and robust phase equilibrium model is developed with the aid of deep learning algorithms to accelerate the thermodynamic analysis of deep reservoir fluids. A pixel‐free search algorithm is developed to generate a pore‐network model that describes pore connectivity and porous media fluidity. A fully conservative Implicit Pressure Explicit Saturation algorithm is developed to simulate the Darcy‐scale two‐phase flow while achieving a reliable result for production evaluation. Numerical examples are presented to validate the performance of the developed models and algorithms. This paper also presents suggestions for future studies on deep reservoirs to achieve both scientific and engineering progress.

Considering the expansion of mining operations into increasingly deeper areas, it is imperative to assess the influence of dynamic disturbance loads on the security of deep tunnels. Here, via AUTODYN finite difference software, a numerical analysis of the fracture characteristics of a fractured tunnel was employed under the coupled action of in‐situ stress and dynamic disturbance loads. The experimental setup comprised a tunnel model with “I‐shaped” cracks, and a drop impact device (DID) was employed to generate dynamic wave loads. A crack fracture test (CFT) was utilized to gather information on the fracture process, including initiation time and average propagation rate. A series of combined scenarios were subsequently simulated to replicate various in situ stress levels (ranging from 0.5 to 2.5 MPa) and dynamic loads. The results indicate that with increasing in situ stress, the crack propagation rate, crack propagation length, and crack break time (CBT) decrease; moreover, the circumferential tensile stress concentration factor in the tunnel also decreases, enhancing tunnel stability. Finally, changes in ground stress influence the propagation path of cracks.

During geothermal resource exploitation, the potential deterioration of mechanical properties in high‐temperature granite subjected to cooling poses a significant safety concern. To address this, the present study investigates the coupled thermo‐mechanical behavior of granite during heating and cooling through a combination of laboratory tests and finite difference method analysis. Initial investigations involve X‐ray diffraction, thermal expansion test, thermogravimetric analysis, and uniaxial compression test. Results show the significant variations of granite properties under different thermal conditions, attributed to temperature gradients, water evaporation, and mineral phase transitions. Subsequently, a model considering temperature‐dependent parameters and real‐time cooling rates was employed to simulate linear heating and nonlinear cooling processes. Simulation results indicate that the thermal cracking predominantly occurs during the heating stage, with tensile failure as the primary mode. Additionally, a faster real‐time cooling rate at higher temperatures intensifies the thermal cracking behavior in granite. This study effectively elucidates the thermo‐mechanical coupling behavior of granite during heating and cooling processes, providing insights into the mechanisms of mechanical property changes with rising or decreasing temperatures.