In tunnel engineering, construction activities usually traverse natural rock masses containing joints. The positional relationship between joints and holes is a key factor controlling the mechanical properties and damage evolution of surrounding rock. To investigate the effects of the joint-hole positional relationship on strain field evolution, crack propagation, and coalescence behavior, uniaxial compression tests were conducted on rock-like specimens containing a single circular hole and a 45° inclined joint. A multi-field analysis framework integrating acoustic emission (AE), digital image correlation (DIC), the strain quantitative calculation model (SQCM), and COMSOL Multiphysics simulation was established. The results indicate that the peak strength of specimens is governed by the coupling of the boundary constraint effect and the stress superposition effect. When joints and holes are arranged at the central position, the mechanical properties and stress distribution are optimized, whereas proximity to the specimen boundary weakens the boundary constraint and intensifies stress concentration, thereby promoting premature failure. The relative spacing between joints and holes dominates the damage evolution by adjusting the superposition of compressive stress fields: a small spacing aggravates early damage and accelerates strain localization, while an increase in spacing weakens the stress superposition and gradually improves the peak strength. Mechanistically, the top and bottom of circular holes are high-tensile-stress zones, the horizontal sides are high-compressive-stress zones, and significant shear stress concentration occurs at joint tips. These spatially distinct stress characteristics lead to four representative crack coalescence modes, including tensile, compression–shear composite, shear-dominated, and multi-mechanism composite, each corresponding to a unique stress-driven failure path. Among them, the compression–shear composite mode exhibits the highest structural stability and residual load-bearing capacity. Furthermore, the damage variable derived from cumulative AE counts is spatiotemporally synchronized with the evolution of localized strain bands, and the sharp exponential rise in the proportion of high-strain pixels quantified via the SQCM serves as a reliable precursor to macro-instability. In summary, this study provides a comprehensive mechanistic understanding from mesoscopic damage accumulation to macroscopic instability, and lays a solid foundation for the stability assessment of jointed rock masses in tunnel engineering.

Geothermal energy offers a reliable low-carbon pathway for decarbonization by providing continuous heat and baseload power. Enhanced geothermal systems (EGS) further expand this potential by enabling energy recovery from deep, otherwise inaccessible low-permeability crystalline rocks. However, wider deployment remains constrained by challenges in achieving effective reservoir stimulation while minimizing induced seismicity. These limitations largely stem from an incomplete understanding of fluid-driven fracturing mechanisms and fluid–rock interactions in crystalline reservoir rocks. Recently, CO2-based fracturing fluids and rock pre-conditioning techniques, which introduce distributed pre-existing damage into the rock mass, have emerged as promising stimulation strategies. Here, we numerically investigate fluid-driven fracturing of granite, a representative EGS reservoir rock, using a coupled hydro-grain-based model (Hydro-GBM) that explicitly captures grain-scale heterogeneity. Different stimulation scenarios were systematically examined by combining fracturing fluids (high-viscosity water versus low-viscosity CO2) with contrasting rock conditions (intact versus pre-conditioned). The underlying grain-scale mechanisms governing fracture development and induced micro-seismicity were also analyzed. The results show that low-viscosity CO2 reduces breakdown pressure and promotes more distributed fracture development, while rock pre-conditioning accelerates crack initiation, increases fracture complexity, and lowers the intensity of induced seismic responses. Among all scenarios, the hybrid strategy combining CO2 injection with rock pre-conditioning produced the most favorable outcome, generating the most extensive fracture network while yielding the lowest peak micro-seismic magnitude. Compared with conventional water fracturing in intact rock, this approach increased crack counts by 69

Mineral veins are pervasive in the subsurface, and their interaction with hydraulic fractures is modulated by injection protocol, thereby influencing fracture network complexity and associated microseismic source mechanisms. To capture these processes, a hydro-mechanically coupled Discrete Element Method (DEM) framework incorporating full moment tensor decomposition is employed to simulate hydraulic stimulation in a veined granite model containing two heterogeneous, parallel shear veins. Hydraulic fractures initiate in a tensiledominated mode and propagate preferentially along the direction of the maximum principal stress. However, their subsequent evolution is governed by vein orientation. Low approach angles (30 degrees and 60 degrees) promote fracture deflection and interface reactivation, generating pronounced shear-dominated microseismic events. In contrast, vertical veins (90 degrees) are predominantly crossed by tensile fractures. Injection protocols exert additional control on both fracture evolution and source mechanism characteristics. Monotonic and stepwise injection promote a higher proportion of double-couple (DC)-dominated events, indicating enhanced shear reactivation, whereas cyclic injection produces a higher proportion of non-DC mechanisms. Cyclic injection also results in the lowest fracture breakdown pressure (approximately 2-7% lower than monotonic injection), enhances pore pressure diffusion, reduces main fracture aperture, and slows the accumulation of hydraulic energy and microseismic events. Although comparable moment magnitude distributions are observed across injection protocols, cyclic injection maintains the lowest seismicity rate over prolonged injection durations. These findings elucidate fracture-vein interaction mechanisms and injection strategy, providing mechanistic insights for optimizing hydraulic stimulation and mitigating injection-induced seismicity in fractured reservoirs.

Large-scale fracture simulation in rock engineering is vital in understanding the real geotechnical and geological behaviors. It has become increasingly important, as an enabling technique, when some emerging fields of underground exploration are being developed, e.g., geothermal energy, underground energy storage, faults and earthquakes, etc. Finite and Discrete Element Method (FDEM) based on cohesive crack model is a widely used simulation method for rock fracture propagation due to meaningful physical parameters and reliable algorithm implementation. However, in large-scale simulations, the cohesive element size often increases with the model scale and may become larger than the characteristic length of the fracture process zone (FPZ). It will cause significant errors or inaccuracy as the damage zone is insufficiently represented. This paper establishes a reduction law for fracture simulations when using relatively large element size. The size effect law of FPZ length is proposed by D-B model and Bazant size effect law. A series of three-point bending tests with Digital Image Correlation (DIC) technique and FDEM models are carried out to obtain FPZ length results and verify the size effect law. Worked examples on large-scale three-point bending tests and compressed air energy storage in rock caverns are demonstrated. This method can provide a useful and practical tool for low-cost simulation of large-scale rock fracture and failure, while not sacrificing the accuracy.

Understanding the micromechanical and creep behavior of shale under supercritical CO2 injection is essential for evaluating long-term reservoir integrity in carbon capture and storage (CCS). However, the microscale creep behavior of shale under different peak loads and holding times after exposure to natural-moisture, deionized-water-saturated, and brine-saturated conditions remains insufficiently understood. In this study, nanoindentation tests with varying peak loads and holding times were conducted to investigate the influence of fluid chemistry on the time-dependent microscale deformation of shale. Mineralogical composition and microstructural changes were characterized using X-ray diffraction, nuclear magnetic resonance, scanning electron microscopy, and automated mineral analysis with a TESCAN integrated mineral analyzer. Higher loads and longer holding times intensified particle sliding and pore rearrangement, resulting in greater creep depth and permanent deformation. This effect tended to plateau in deionized water but became more pronounced in brine because of sustained microslip. Deionized water produced the greatest reductions in indentation modulus and hardness but limited long-term creep, whereas brine maintained moderate stiffness and promoted greater creep displacement through ionic lubrication. A nanoscale creep constitutive model was developed by combining logarithmic and Newtonian components with load- and time-dependent parameters. The model accurately captured transient and steady-state creep responses in hard and soft phases and reproduced elastic recovery during creep, indicating that steady-state creep was reached after approximately 10 s. Parameter analysis showed that transient creep was primarily load-dependent, whereas steady-state creep was more sensitive to fluid chemistry. These findings support long-term stability assessments in CCS applications.

Hydraulic fracturing in high-temperature crystalline geothermal reservoirs requires reliable prediction of fracture initiation and growth under coupled thermal damage and microstructural heterogeneity. Most existing studies assume homogeneous rocks or ambient conditions, which limits mechanistic understanding of how thermal damage and grain-scale heterogeneity jointly govern fracture initiation, propagation and connectivity in high-temperature crystalline geothermal reservoirs. We develop a coupled fluid-solid discrete element framework (Hydra-Hete-GBM) that explicitly represents grains and grain boundaries. Using this framework, we systematically investigate the effects of temperature (150–600 ^∘C ), differential stress (2.5–10 MPa), and injection rate (2–5 × 10 ^-5 m^2/s ). Model predictions are validated against published laboratory hydraulic fracturing experiments reported in the literature, showing consistent trends in pressure evolution and fracture morphology. Results indicate that the maximum principal stress controls the macroscopic propagation path, whereas grain-scale heterogeneity regulates local deflection and branching. Uniform microstructures promote stable planar growth, while structurally heterogeneous domains trigger stress concentrations that increase branching and tortuosity. Fine-grained media are dominated by tensile failure and yield smoother paths; coarse-grained media exhibit stronger grain boundary guidance and more mixed-mode failure. Increasing injection rate and stress differential both enhance directionality and suppress heterogeneity-driven instability. Elevated temperature promotes distributed thermal cracking. Above 450 ^∘C , thermally induced microcracking increases total crack length, but reduces the coherence of the hydraulically effective fracture path, with network connectivity decreasing further as heterogeneity increases. These findings provide a mechanistic basis for tailoring injection strategy and stress management to improve fracture directionality, connectivity, and stimulation efficiency in high-temperature crystalline geothermal reservoirs.

Fault veins, mineral-filled fractures within fault zones, exhibit substantial grain-scale structural variation, yet, the mechanical consequences of this heterogeneity on fault deformation and seismicity remain poorly constrained. Here, we use the Discrete Element Method to investigate how grain-scale heterogeneity in quartz veins, which are common in crustal fault systems, affects shear rupture, frictional evolution, and microseismic activity. Two end-member vein models are constructed: a homogeneous vein with uniform mineral properties, and a heterogeneous vein containing coarse, irregular grains and weak grain-boundary contacts, each embedded within a representative host-rock domain. Direct-shearing simulations are performed to capture deformation from the initial cohesive phase through frictional sliding. Results show that the heterogeneous vein, relative to homogeneous veins, exhibits lower shear strength and friction during the cohesive-influenced phase, but develops slightly higher and more stable friction in the subsequent friction-influenced phase. Grain-scale heterogeneity promotes pervasive cracking not only within the vein but also in the adjacent host rock and along vein-host interfaces. This leads to more frequent, larger-magnitude microseismic events. These events form broad microseismicity clouds, splay-like off-fault fractures, and elevated off-fault b-values, whereas the homogeneous vein localizes deformation near the vein-host boundary. Our results show that grain-scale heterogeneity in fault veins strongly influences rupture style, distributed damage, and seismicity. Importantly, they show that shear failure of cohesive, mineralized veins can generate both near-field and far-field off-fault fracturing, a mechanism largely overlooked in laboratory and numerical studies focused on weak or granular gouges, with implications for rupture complexity and damage evolution in fault systems.

Mineralized faults-fractures that have accommodated displacement and were subsequently filled or coated with mineral precipitates-are commonly found in hydrothermal regions. Precipitation from circulating fluids or magmatic intrusions can substantially alter fault structure, reduce permeability, and influence frictional stability. Despite their geological importance, the slip behavior and frictional evolution of mineralized faults during fluid injection remain poorly constrained. To address this knowledge gap, we conduct triaxial shear-flow experiments on critically stressed mineralized faults in granite to examine their hydro-mechanical response to fluid pressurization. Deionized water is injected at a constant rate of 0.6 mL & sdot;min-1 under different confining pressures of 20 and 30 MPa. Fault slip occurs in two distinct stages: an initial unstable stick-slip phase followed by stable slip, with higher confining pressure promoting the transition. The effective friction coefficient fluctuates during stick-slip but increases progressively during stable slip. Greater hydraulic energy input does not necessarily induce earlier fault reactivation, but it results in greater deformation moment accumulation and higher energy release. Microstructural analysis reveals distinct deformation mechanisms: foliation-like microfractures develop near the fault surface at 20 MPa, while 30 MPa conditions lead to grain crushing and the formation of interconnected fracture networks. These findings advance our understanding of injection-induced slip in mineralized faults and provide insights relevant to geothermal energy extraction, energy storage, and waste disposal.

Temperature variations strongly influence the fracture behaviour of sedimentary rocks, with direct implications for the safety and performance of thermally affected subsurface systems. This study investigates the Mode I fracture behaviour of mudstone over a temperature range of 25–350°C using semi-circular bend (SCB) tests combined with Digital Image Correlation (DIC). The results show a systematic increase in fracture toughness with temperature. Full-field measurements reveal that deformation localisation consistently initiates at the notch tip and propagates towards the specimen boundary, but with varying full-field features with increasing temperature. Strain development follows three distinct stages: a quasi-linear stage, a nonlinear acceleration stage, and a rapid escalation stage immediately prior to failure. Normal and shear strains evolve concurrently, indicating coupled deformation during crack propagation. Crack growth behaviour exhibits strong temperature dependence. Higher temperatures promote increasingly unstable crack propagation, characterised by sharper acceleration near peak load and more irregular crack paths, reflecting enhanced microstructural heterogeneity and thermal damage. These findings provide new insights into temperature-dependent fracture mechanisms in mudstone and inform improved design, risk assessment, and long-term stability evaluation of underground infrastructure and underground engineering applications.

ABSTRACT: Natural fractures can be reactivated during subsurface fluid injection operations such as enhanced geothermal system development and wastewater disposal. Under hydrothermal conditions, these fractures are commonly mineral-filled, and such mineralization can alter their frictional properties and dynamic response to fluid injection, particularly under different stress states. However, the combined effects of stress state and mineralization on injection-induced fracture slip behavior and seismicity remain poorly understood. To address this gap, we performed laboratory fluid injection experiments on saw-cut fractures in veined granite samples, injecting water at a constant rate under varying confining pressures to examine how stress state governs fracture slip mode and seismic response. The results show that increasing confining pressure favors a transition from unstable stick-slip to stable sliding. When stick-slip occurs under higher stress, the associated seismic events display larger slip displacements and higher peak slip rates. The friction coefficient fluctuates during stick-slip but gradually increases during stable sliding. For similar injection durations, greater hydraulic energy input is observed at higher confining pressures. Microscopic analyses further reveal intensified grain crushing and the development of interconnected fracture networks near the fracture surface under elevated confining pressures.

Under repeated loading and unloading, the mechanical response of crystalline rock exhibits hysteresis. To elucidate the linkage between macroscale hysteresis, microscale deformation, and microstructural evolution in crystalline rock, we perform in-situ time-integrated X-ray Computed Tomography (X-CT) test to image the microstructural evolution along a complete loading-unloading hysteresis loop followed by reloading until failure of Blue Hone granite. The microstructural evolution is quantified by the development of geometries and orientations of segmented voids. The evolved accumulative strain fields are calculated by Digital Volume Correlation (DVC) to characterize distribution of micro-scale deformation and reveal the interaction between strain localization and microstructural evolution and build relationship across length scales. The orientation distribution of first principal strain is analysed to characterize the influence of microstructural evolution on overall deformation. The results show that the macroscale hysteresis comes from the microscale hysteresis in dilation zones, which are highly correlated with high shear strain zones. Crack-like voids (perpendicular, inclined and parallel to axial load) in dilation volumes exhibit prominent hysteresis compared to those in contraction volumes, resulting in the delayed strain releasing of dilation volumes. The evolution of inclined crack-like voids confirms that the hysteresis mainly results from the newly developed inclined crack-like voids (shear cracking). After reloading to the same stress at the onset of unload, a further evolution of damage and strain localization is observed, while the sample deforms in a higher efficiency way to accommodate applied stress. This observation is discussed with the insight of rock fatigue.

The peril posed by time-delayed rockburst significantly undermines the safety of deep-buried tunnel construction and operation, with dynamic disturbance recognized as a pivotal triggering factor. Vibration monitoring of critical Tunnel Boring Machine (TBM) components and surrounding rock surfaces within TBM-constructed tunnels shows that low-frequency dynamic disturbance affects the failure behavior of excavated surrounding rock. However, the mechanism underpinning instability of hard rock under dynamic disturbance during stress adjustment in deep excavations remains elusive. To bridge this gap, multilevel dynamic disturbance experiments (A =1 MPa, f = 20 Hz) under true triaxial loading and unloading paths were conducted. The results indicate that dynamic disturbances can accelerate the failure of granite, with its strength reduced by 10 %-12 % under the combined effects of high-stress unloading damage. These effects diminish the energy storage capacity of rock mass, thereby lowering the threshold for rockburst occurrence. The deformation rates of granite increase with each phase of 61 loading and 63 unloading. Based on real-time deformation rate observations at the final stage, critical instability characteristics are categorized into two types: stress adjustment and dynamic disturbance. With prolonged exposure to dynamic disturbance, a distinct acceleration-stabilization-acceleration (V-shaped) pattern appears in the strain rate of dynamic disturbance-induced failure. Acoustic emission monitoring demonstrates the degradation mechanism in hard rock subjected to dynamic disturbance, leading to the initiation of tensile cracks and accelerating the propagation of cracks across fracture surfaces. Ultimately, the effect of supporting stress was evaluated through comparative testing. A regulatory strategy was proposed for controlling the evolution process of dynamic disturbance-triggered rockbursts, entailing the construction of a three-dimensional wave-absorbed support system for the excavated rock surrounding deep-buried tunnels. These findings provide a valuable reference for understanding, early warning, and controlling the mechanisms underlying time-delayed rockbursts triggered by dynamic disturbances.

Traditional methodologies primarily depend on macroscopic observations and post-experimental fracture surface analysis to identify transversely isotropic rock’s failure modes and mechanisms. This study employed discrete element method (DEM) simulations to analyze Brazilian splitting and uniaxial compression tests on transversely isotropic rocks. By applying moment tensor inversion, seismic information was extracted from acoustic emission (AE) events. The key findings are as follows: (1) The failure modes of the specimens can be categorized into three types: failure that cuts through the bedding planes, failure along the bedding planes, and composite failure, where fractures penetrate the bedding planes while partially following them. (2) In uniaxial compression tests, the sample with a bedding angle of 30 ^∘ exhibits frequent small-scale AE activity along the slip zone, resulting in progressive shear slip failure. In contrast, those at 45 ^∘ and 60 ^∘ angles show more high-magnitude events along the slip zone, with a shorter time from crack initiation to final failure, indicating catastrophic shear slip failure. (3) In the Brazilian splitting tests, specimens with a bedding angle of 60 ^∘ experienced progressive shear sliding failure, while those with a bedding angle of 75 ^∘ experienced catastrophic shear sliding failure. These findings refine the failure mode classification of transversely isotropic rocks across different bedding angles, enhancing understanding of their mechanical behavior.

Underground CO2 storage is a key strategy to achieving net-zero targets by 2050, which requires gigatonne-scale containment of gaseous CO2 in geological formations. Shale rocks play a role in both trapping CO2 and preventing its escape, thus ensuring containment security. However, shale integrity can be compromised upon interaction with CO2, which should be carefully evaluated. This study explores the dynamic behaviour of CO2-shale interaction at the pore scale, focusing on the physiochemical interactions between CO2 and shale, including the impact of shale swelling, where CO2 adsorption causes matrix deformation and alters fracture sizes. Here, we utilise image-based analyses to develop a triple-porosity pore network model (PNM), reflecting the complex nano- to micro-scale structure of shale, to examine CO2 injection into methane-saturated environments. The study particularly focuses on the impact of matrix deformation caused by gas sorption (swelling), competing with mechanical stress effects. Findings indicate that CO2 injection leads to a reduction in fracture permeability by up to 17 % and 10 % in low- and high-density fractured shales, respectively, under high confining pressure (50 MPa), and by 15.5 % and 8 % under lower confining pressure (25 MPa). Although fracture permeability versus CO2 injection pressure reduces monotonically at the lower confining pressure, that of the higher confining pressure is non-monotonic, where the fracture permeability shows an increase due to effective stress change. Additionally, the average fracture aperture size decreases by 50 nm in low-density and 25 nm in high-density fractured shales, highlighting the critical balance between swelling effects and mechanical stresses in the geological sequestration of CO2.

The natural faults show a spectrum of slip modes due to different geological conditions. How the principal slip zones (PSZs) evolve in these faults remains a puzzle. Here, a series of numerically simulated sheared granular faults that incorporate breakable grains under different loading conditions (normal stress, shear velocity, and load stiffness) are adopted to explore the detailed evolution of PSZs. With the increased normal stress, the faults show more distinct stick-slip cycles, and the slip modes gradually switch from stable sliding to stick-slip. The faults produce larger slip events when the shear velocity increases to the scale of slip velocity of the slip events, and higher load stiffness tends to yield stable sliding states. Additionally, when subjected to higher normal stresses, tenser shear localization in the granular gouge can be observed, and PSZs prefer to initiate from places further away from the plates and can expand more widely before entering a stable state. Also, when subjected to higher normal stress or lower shear velocity, the PSZs expand faster. However, the load stiffness has a minor influence on the development of PSZs. This paper illustrates the progression of PSZs under various loading conditions, providing a more comprehensive understanding of the formation and evolution mechanism of PSZs in sheared granular gouges.

Fault cores have layered shear structures that evolve with time. Probing details of the accompanied gouge grain breakage and granular flow can unveil the complex evolving mechanism of fault. Here, irregular breakable grains are incorporated in a numerically simulated sheared granular gouge to show the evolution of shear structures. Granular quantities (e.g. non-affine displacement and local granular temperature) are adopted to depict the microstates of gouge in different macro shear stages and during stick-slips. Granular plasticity (defined as plastic granular flow here) is scattered in the linear stage, while it starts to cluster near the plate in the yield stage. Then, the shear localization becomes the strongest in the unstable shearing stage and weakens in the later stick-slip stage. In the stick phase during the stick-slip stage, some particles are more prone to unlocking due to the geometry of the fault and the arrangement of particles. Small local slips occur between these particles. The accumulation of small local slips between these particles can cause large dynamic failures. Finer grains in the principal slip zone decrease the transmitted shear force and render it easier for the granular gouges to rearrange themselves, causing the shear to localize in these zones. The long-term stick-slip cycles of natural faults result in the layered shear structures in the fault cores. Although the granular gouge between fault planes is composed of discrete particles or grains, the granular mass ruptures like a solid when slip events happen.

Fault rocks vary in mineralogy but understanding the role of mineralogy in injection‐induced fault slip behavior remains challenging. We conduct laboratory fluid injection experiments on representative faults with distinct surface mineralogy, injecting fluid at a constant rate of 0.6 mL/min and under a confining pressure of 20 MPa, to examine how mineralogy affects fluid‐driven slip and seismicity. We find that the fault surface with a higher content of tectosilicates (quartz, microcline, albite, totaling 96%) exhibit higher shear strength. The tectosilicate‐rich fault takes a longer time to reactivate but accumulates seismic moment more rapidly, indicating more intense seismicity. In contrast, the fault containing abundant phyllosilicates (50% biotite) promotes a transition from unsteady stick‐slip to steady slip during fluid injection. Stick‐slip events in the tectosilicate‐rich fault exhibit higher peak slip rates and greater slip displacements. We propose a microscopic mechanism to elaborate how fluids may influence mineral friction and thereby affect injection‐induced fault slip.

Fault zones can be highly lithified, often manifesting as veins or cemented faults. Understanding vein structures and their dynamic response to reservoir stimulation is essential for managing anthropogenic earthquakes; however, this remains poorly constrained. We implement a novel algorithm for full moment tensor inversion and decomposition within the Discrete Element Method (DEM) framework and enhance the pipe network flow model by dynamically updating the domain volume and fluid saturation at each time step. We inject fluids into a simplified fault zone model containing heterogeneous veins to capture their dynamic response to reservoir stimulation. We find that shear stress on the veins decreases upon fracture initiation around the borehole. Normal stress decreases until the hydraulic fractures reach the veins, after which it begins to increase. We also observe relatively high double-couple (DC) components during interactions between hydraulic fractures and shear veins, whereas non-DC components are more prominent during fracture propagation within the rock matrix. Effective friction decreases as fractures interact with the veins. We argue that pore pressure diffusion and poroelastic stress transfer play critical and synergistic roles in reservoir stimulation.

During earth pressure balance shield tunnelling in water-rich sandy ground, both foam and other conditioning agents, such as bentonite slurry, are injected to prevent water spewing. Permeability tests were conducted to investigate how water pressure affects the permeability of sand conditioned with foam and bentonite slurry. Experimental results demonstrate that increasing water pressure at the top and bottom of the specimen extends the initial stable period of the permeability coefficient, significantly slowing down its growth rate during the fast growth period. Soil grain migration was observed in specimens exposed to sufficiently high water pressure. During the slow growth period, the permeability coefficient decreased as water pressure increased, and this decrease rate correspondingly decreased. Under a consistent hydraulic gradient, increased water pressure led to enhanced stability of foam bubbles and extended the time-dependent curves for the permeability coefficient. Furthermore, the relationship between chamber pressure dissipation and foam stability was discussed during the standstill period of shield machines. To prevent water spewing, it is recommended to use the permeability coefficient of the muck at the outlet of the screw conveyor with the lowest water pressure as the evaluation index during permeability testing.

Fluids are often injected into tight geothermal reservoirs to create conductive fractures and enhance reservoir productivity. Mineral veins are ubiquitous in reservoir rocks because of the hydrothermal mineralisation process. We hypothesize that vein strength anisotropy and geometrical complexity could introduce complex hydraulic fractures through vein–fracture interactions. To test this hypothesis, we develop an improved hydro-mechanical coupled scheme using the Discrete Element Method to examine hydraulic fracture propagation in veined rocks. The coupled scheme considers incremental updates of domain volume and fluid saturation. We follow topology in root architecture to develop a benchmark model with simplified yet representative vein networks. We find that vein strength significantly influences the fracture pattern of the tested veined models during hydraulic fracturing. Cracks are preferentially reactivated in the calcite (soft) veins, leading to much more complex fracture networks compared with those created in quartz (hard) veins. An increase in stress anisotropy (high-stress contrast) leads to a reduction of the maximum aperture of dominant fractures. Under a high-stress contrast, fluid lag is not evident; the number of reactivated cracks in soft-veined models is also reduced. It is interesting to note that the bifurcation angle of veins has negligible influence on fracture patterns and propagation in the model. We also find that high-stress contrast leads to a higher level of the maximum moment magnitude and a lower b value. This research could provide some insights into reservoir stimulation to enhance production and mitigate seismic hazards.