Study on the crack propagation mechanism of single cracks in red sandstone based on stress field evolution characteristics Authors: Wenhua Zha, YiChao Cao, Tao Xu, BingWen Wang, XueJian Tan Published: 2026 ABSTRACT AbstractTo elucidate the influence of fracture dip angle on the mechanical response of red sandstone and the mechanism of crack evolution, this study employs the discrete element software PFC2D to construct a single-fracture sandstone model. Numerical uniaxial compression tests are conducted under varying fracture dip angles α. By integrating the ‘stress field–microcrack’ coupling criterion with fracture initiation displacement field characteristics, the study quantitatively derives the segmented enhancement patterns of peak strength and fracture initiation stress, systematically elucidating the mechanisms of microcrack initiation and propagation. Results indicate that as α increases, the normal clamping effect of the fracture surface intensifies, making it more difficult for the vicinity of the fracture surface to first reach the damage threshold. Microcrack activity gradually shifts from “preferential growth in the middle section of the fracture surface” to “easier triggering at the fracture tip,” driving the macrocrack path from local control at low inclination angles to synchronous evolution with high-stress zones at high inclination angles (This consistency becomes more pronounced at larger inclinations). The mechanical response manifests as an overall increase in both initiation stress and peak strength with inclination angle, exhibiting a more pronounced enhancement trend within the larger inclination range. Additionally, the strain value required for the first AE signal increases, the strain range where AE events occur narrows, and the peak ringing count rises.The crack initiation stress increased from 11.5 MPa (α = 0°) to 25.0 MPa (α = 90°), representing an approximately 117% increase; the peak stress rose from 24.2 MPa to 33.6 MPa, an increase of approximately 39%. Microcrack counts exhibited exponential growth with accumulated strain, dominated by tensile cracks, indicating a tensile-dominated brittle failure mechanism in the rock specimens.The research findings provide an interpretive framework based on stress thresholds for identifying failure modes under controlled joint dip angles and for engineering stability assessments.Keywords: Fracture dip angle, PFC2D, Single-fracture rock, Crack propagation, Stress field, Displacement fieldSubject terms: Engineering, Materials science, Solid Earth sciences Introduction During the excavation of underground structures, the surrounding rock often undergoes triaxial stress conditions1,2. Due to prolonged exposure to geological forces, numerous fractures have formed, varying in dip angle, location, length, and quantity. The presence of these fractures not only diminishes the rock mass’s mechanical properties but also significantly impacts the stability of underground works through their crack propagation behaviour3–6. As a widely occurring sedimentary rock, red sandstone plays a significant role in underground engineering. Among its fissures, single cracks constitute a fundamental fissure structure, and their crack propagation process under complex stress conditions has a substantial impact on the safety and stability of underground construction works. Numerous scholars have conducted extensive laboratory studies on the failure modes and crack evolution of various fractured rock masses. Gao Meiben et al.7 employed the axial strain stiffness method in conjunction with experimental data to determine characteristic strength values at different stages of rock failure, proposing that cracks undergo four phases during specimen failure: compression, initiation, propagation, and breakthrough. SU et al.8 employed Charge-Coupled Device Cameras (CCD) to track displacement data during uniaxial loading of rock specimens, investigating the combined effects of fracture and bedding planes on fracture evolution and mechanical properties in fracture rock specimens. They observed a novel phenomenon of tensile cracking. Wu Xu et al.9 conducted uniaxial loading tests on cross-orthogonal fracture specimens, revealing that the presence of fractures perpendicular to the loading direction exacerbates specimen failure. Ruide Lei et al.10–12 conducted uniaxial compression tests on defective sandstone supplemented with AE monitoring. Through analysis of AE statistical metrics, time-frequency energy distribution, and multifractal characteristics, they found that defect geometry significantly influences strength and failure modes. The combined use of multiple indicators enables more reliable identification of early warning points and critical fracture points, thereby achieving early instability warning. Chen Na et al.13 performed uniaxial loading tests on chlorite with varying foliation angles, employing AE detection to observe failure characteristics. They found that foliation angle correlates with mechanical properties, with distinct damage mechanisms corresponding to different angles. Guo Qifeng et al.14 employed AE and surface strain monitoring techniques alongside maximum distortion energy theory to test the Sanshan Island gold mine granite under uniaxial loading. They observed that rock strength did not exhibit a linear relationship with fracture dip angle, with specimen strength peaking when the loading direction aligned with fracture orientation. Tang15 analysed crack propagation patterns in marble under uniaxial compression using SEM imaging. This revealed factors accelerating crack development and established the fundamental relationship between damage and fracture parameter evolution at different scales and external loading. Li et al.16 conducted uniaxial loading tests on coal samples with varying fracture angles, analysing the influence of fracture inclination on mechanical properties and acoustic/thermal characteristics. They concluded that both the variation in mean infrared radiation temperature and peak emission counts first decreased, then increased with increasing fracture inclination. Although the aforementioned scholars have achieved substantial results in rock macromechanical properties and crack evolution using techniques such as AE, laboratory test findings primarily focus on the macroscopic response of fractured rock. This makes it challenging to analyse the micro-mechanisms governing crack propagation patterns from the perspectives of displacement and stress fields. To overcome the aforementioned research limitations, numerous methodologies have emerged in recent years, particularly the Particle Flow Code (PFC) based on particle flow theory, which has become a crucial tool for investigating the evolution characteristics of microcracks within rock masses17. Xiao et al.18–20 conducted uniaxial and triaxial loading tests on single- and double-fractured rocks, combined with PFC numerical simulations, to analyse the effects of fracture length and confining pressure on fracture behaviour and crack propagation in fractured rock. Yang et al.21 employed PFC2D to numerically simulate uniaxial compression of red sandstone containing two non-parallel fractures, elucidating the stress field and fracture coalescence mechanism in uniaxially loaded fractured rock. Li Jianwang et al.22 combined DIC techniques with PFC2D simulations to investigate the failure process of pre-cracked rock specimens under uniaxial loading. They obtained full-field strain distributions across the observed surfaces and crack development patterns. Notably, PFC provided microstructural pathways for interpreting key parameters such as stress field and displacement field evolution—insights unattainable through conventional methods. In summary, the evolution of internal microcracks within fractured rock holds significant implications for the safety of rock mass engineering. However, its mechanical properties and crack propagation mechanisms warrant further investigation, particularly concerning the evolution of stress and displacement fields. This study employs the PFC2D software to establish a model of a rock specimen containing a single pre-fabricated fissure. Utilising a coupled stress field, displacement field, and microcrack criterion approach, combined with AE characteristics, it systematically analyses the crack propagation mechanisms and failure patterns of the fissured rock specimen under uniaxial loading conditions at different fissure inclinations. The aim is to provide a theoretical basis for the stability analysis and control of fissured rock masses during underground engineering construction. Laboratory testing and numerical modelling Laboratory testing To investigate the influence of fracture dip angle on the mechanical behaviour of rock specimens and crack propagation patterns, rock specimens with varying fracture dip angles were designed for uniaxial loading tests, providing a basis for subsequent numerical simulation studies. Considering the complexity of actual engineering geological environments and the difficulty in locating naturally fracture rock masses in the field, which cannot meet the substantial requirements for engineering testing, artificially prepared single-fracture rock specimens were employed as the research subjects. Natural red sandstone from a geological formation was selected. Following the testing methodology prescribed by the International Society for Rock Mechanics, specimens were prepared as cylindrical columns measuring 100 mm in height and 50 mm in diameter. Errors associated with waterjet cutting primarily include micro-damage caused by erosion of the cut edge walls and geometric deviations at the cut surface. These factors reduce the equivalent strength and stiffness in the cut zone, manifesting as lower crack initiation stresses and potentially altering the location and number of microcracks appearing during the pre-peak stage. However, since identical processing techniques and dimensional controls were applied to rock specimens with varying inclinations, this additional damage exhibits consistent systematic errors across all groups and can be disregarded. This approach enabled an investigation into the influence of fracture inclination on rock mechanical response and crack evolution mechanisms. The single pre-fabricated fractured specimens are illustrated in Fig. 1. The specific test setup is detailed in Table 1. Single pre-fabricated fractured rock sample. Table 1Uniaxial testing simulation protocol for rock samples.Rock sampleFracture dip α/(°)Fracture length (L/mm)A020B1520C3020D4520E6020F7520G9020 Uniaxial testing simulation protocol for rock samples. The experimental system employed in this study comprised two components: the loading system and the control system, as illustrated in Fig. 2. The loading system employs the uniaxial loading module from the TAW-20 soft rock creep triaxial testing system. The test process utilizes displacement-controlled loading to ensure the specimen axis is centered with the pressure plate, preventing eccentric loading. The loading rate is 0.006 mm/s.In indoor uniaxial loading tests, the axial deformation of the rock specimen is measured using a linear displacement sensor supplied with the testing machine. Uniaxial loading test system. Following uniaxial loading tests, the ultimate failure modes of rock specimens with varying fracture inclinations are illustrated in Fig. 3. Failure in specimens with different fracture inclinations commenced at the fracture terminus, subsequently propagating towards both ends of the specimen until complete failure occurred. When the pre-cut fissure inclination α was 0° and 15°, the fissured specimens exhibited ‘Y’-shaped failure. For pre-cut fissure inclinations of 30°, 45°, 60°, and 75°, the specimens displayed “X”-shaped failure. When α = 90°, the failure pattern resembled that of intact specimens, presenting as ‘H’-shaped failure. Fracture diagram of fractured rock sample. PFC2D single-fracture rock model and parameter calibration Construction of numerical models This paper employs the Fish language built into the particle flow numerical simulation software (PFC2D 5.0) for modelling. The rock specimen dimensions were set at 50 mm × 100 mm. A rock sample model comprising spherical particle elements of varying radii was generated using an automatic filling method, thereby reflecting the microstructural anisotropy of the rock sample. The ratio of the largest to smallest particle diameter was 1.6, with 13,388 spherical particles required to generate the complete rock specimen. A pre-crack was formed by removing a particle unit measuring 20 mm in length and 1 mm in width from the central portion of the rock sample, consistent with the laboratory test. Axial loading was simulated through relative movement between the upper and lower ‘walls’, conducting a uniaxial compression test. A parallel bonding model was employed between particles. This model generates parallel bonds between particles capable of transmitting forces and moments. Each contact point between two particles is treated as a spring with stiffness, undergoing tensile or shear failure when stress exceeds the bonding strength. The macroscopic strength of the rock specimen depends on the bonding strength between particles23. The mechanical structure of the parallel bonding model is illustrated in Fig. 4. The fractured specimen model is depicted in Fig. 5. Schematic diagram of parallel bonding model. Numerical simulation model of fractured rock samples. Microscopic parameter calibration Given that the micromechanical parameters governing particle-to-particle contacts within numerical models cannot be directly derived from macroscopic rock properties, a trial-and-error approach is required to iteratively determine these granular parameters24. Key micromechanical parameters include particle contact modulus, parallel cohesion modulus, inter-particle contact stiffness ratio, parallel cohesion stiffness ratio, tangential cohesion strength, and normal cohesion strength. Taking a specimen with a fracture dip angle of 60° as an example, this orientation lies within the intermediate transition zone of dip angle effects. It combines the joint effects of fracture surface clamping and stress concentration at the fracture tip, making its macroscopic response and crack propagation mechanism more representative. The numerical simulation schematic for uniaxial loading is shown in Fig. 6.The rock sample is loaded via upper and lower ‘wall’ elements at a loading rate of 0.006 mm/s, ceasing at 70% of peak strength. The stress-strain curve obtained from laboratory uniaxial loading tests on red sandstone was employed for calibration. This ensured the shape and trend of the stress-strain curve generated by the PFC2D numerical simulation closely matched that of the laboratory uniaxial compression test curve (as shown in Fig. 7). A trial-and-error approach was used to determine highly accurate micro-scale simulation parameters for the sandstone (as listed in Table 2). The initial concavity in the stress-strain curve from laboratory testing may stem from potential voids and minute fractures within the rock specimens. Numerical simulations, however, perform a ‘pre-compaction’ on fractured specimens, rendering them relatively denser. The peak intensity errors were 1.92% and 1.39%, respectively. Consequently, PFC2D simulations reliably capture the microcrack propagation patterns and failure mechanisms within rock specimens. PFC2D numerical simulation of uniaxial loading. Comparison of experimental and simulated results. Table 2Numerical simulation microscopic parameters.Density /kg·m− 3Radius/mmContact modulu/GPabond modulu/GPaInternal friction angle/°Contact stiffness ratiobond stiffness ratioTangential bond strength/MPaNormal bond strength/MPaDamping coefficient23900.25–0.49.79.726.71.371.3723.623.60.7 Numerical simulation microscopic parameters. Mechanical behaviour of fractured rock specimens and crack propagation mechanisms Analysis of mechanical behaviour of fractured rock samples The strength properties of rock serve as crucial indicators for evaluating its structural stability and load-bearing capacity, typically quantified through peak stress and its corresponding strain. During the deformation-fracture process, cracks undergo four distinct stages: compaction, initiation, propagation, and breakthrough7. Each stage corresponds to a specific stress threshold, with the associated characteristic stresses being the initiation stress, damage stress, and peak stress respectively. Therefore, to investigate the influence of fracture dip angle on the mechanical behaviour of rock specimens, the variation patterns of stress-strain curves under different fracture dip angles were obtained based on uniaxial compression test results (as shown in Fig. 8). The initiation stress, damage stress, and peak stress of fractured specimens under uniaxial compression were compared to reveal the influence mechanism of fracture dip angle on mechanical behaviour. Overall, the stress-strain curves for all fracture inclinations exhibit a broadly similar upward trend. As the fracture inclination increases, both the secant modulus and peak stress of the rock specimens rise, indicating enhanced strength. Fracture rock stress-strain relationship. By calculating the maximum and zero points of the lateral strain difference Δε across the rock specimen, the initiation stress and damage stress of the fractured specimen are determined19. Select one symmetrical measurement zone on each side of the specimen’s midpoint. Calculate the lateral average strain for each side and subtract the values. When microcracks first localise on one side, Δε exhibits its initial maximum value, thereby establishing the initiation stress. As the crack propagates and approaches full penetration, Δε recedes and crosses the zero value, thereby establishing the damage stress. Figure 9 illustrates the variation of peak stress, initiation stress, and damage stress with fracture dip angle. By comparing the initiation stress and damage stress of fractured specimens, the influence of the pre-fracture dip angle α on these stresses is analysed. As the fracture inclination angle α increases, the fracture initiation stress, damage stress, and peak stress of the fractured specimens all exhibit an upward trend, indicating that the specimens exhibit enhanced mechanical properties with increasing fracture inclination. When the fracture angle α ≤ 45°, the increase in initiation stress and peak strength is relatively stable and gradual. When α > 45°, the rise in initiation stress and peak strength becomes markedly steeper. Damage stress, however, increases steadily throughout. As the fracture angle increases, the initiation stress, damage stress, and peak strength of the fractured specimens all increase with nearly identical trends. This is due to the following: Initiation stress, damage stress and peak stress in fractured rock. Under uniaxial loading, the presence of pre-existing fractures induces stress concentration near the fracture tip. The normal stress (σₙ) and shear stress (τ) on the fracture plane are calculated as follows: According to the Mohr-Coulomb criterion, failure requires that: Therefore, the peak intensity σ₁ is: When α = 0°, σ₁ is unaffected by the internal friction angle ϕ, yielding the lowest peak strength. For 15° ≤ α ≤ 60°, σ₁ increases with greater fracture dip angles, as the friction force within the fractured rock specimen exceeds the shear force, thereby enhancing its peak strength. At α = 90°, where the fracture orientation aligns with the loading direction, maximum strength is attained, exhibiting mechanical behaviour approximating that of intact rock specimens. Analysis of microcrack quantity and propagation patterns Microcrack counting in PFC is defined as follows: Each irreversible bond break event at a bonded contact is counted as one microcrack. When σmax > σt (Normal bond strength), tensile failure is determined, and the bond is removed; when τmax > τs (Tangential bond strength), shear failure is determined, and the bond is removed. The stress-strain curves of rock specimens with different pre-existing fracture inclinations during uniaxial compression can be divided into the fracture compaction stage, fracture initiation stage, fracture unstable development stage, and failure stage25. During the crack compaction stage, the number of microcracks is nearly zero. In the crack initiation stage, the number of microcracks begins to rise slowly, with the curve remaining relatively smooth but showing initial signs of damage. During the unstable crack growth stage, the number of microcracks shifts from slow to rapid growth. In the failure stage, the number of microcracks increases exponentially, ultimately leading to complete failure. Fracture propagation often causes fluctuations in the stress-strain curve. Due to variations in fracture inclination, the stress-strain curves exhibit distinct characteristics at each stage. As illustrated in Fig. 10, which plots stress-strain against microcrack count, the initial loading phase (oa segment) corresponds to crack consolidation, characterised by a smooth stress-strain curve with no discernible microcracks. Upon entering the crack initiation phase (ab segment), tensile cracks gradually emerge and grow steadily. As loading continues, the rock specimen undergoes a phase of unstable crack propagation (bc segment) prior to peak stress, characterised by rapid tensile crack growth accompanied by minor shear crack formation. Beyond peak stress, the failure stage (cd segment) commences, where both tensile and shear microcracks increase exponentially. Tensile cracks dominate, while shear microcrack propagation arises from intergranular slip and residual bearing capacity, collectively indicating tensile failure. In summary, the number of cracks in all fractured rock specimens exhibited exponential growth, with the greatest increase occurring during the failure stage. As the fracture dip angle increased, both the strain required for crack initiation and the number of cracks formed showed a marked increase. Evolutionary characteristics of microcracks within rock specimens with varying fracture dip angles. Cracks provide a direct manifestation of internal damage within materials. Their evolution process reflects stress redistribution and localised failure characteristics in rock under loading, whilst also revealing failure pathways and structural instability mechanisms26. During loading, mechanical properties also differ among rocks with varying fissure characteristics. As illustrated in Fig. 11, tip-end propagation failure in single-fissure rock specimens primarily generates two crack types: wing cracks (propagating perpendicular to the pre-existing fissure direction) and secondary coplanar cracks (rapidly propagating along the main structural plane from the tip towards the upper and lower ends). A small number of secondary inclined cracks may also form (running in the opposite direction to the wing cracks). Distribution of primary crack types at the pre-fracture tip27. To investigate the influence of fracture dip angle on microcrack propagation within rock specimens, numerical tests were conducted using uniaxial loading on specimens with varying fracture dip angles. Four stress-strain points—a, b, c, and d (as shown in Fig. 10)—were selected from the stress-strain curve for crack evolution analysis. Black regions denote tensile cracks, while red regions indicate shear cracks. When the fracture dip angle α < 30°, no significant microcracks were observed during the initial loading phase. As loading progressed, tensile cracks first formed near the fissure centre and propagated towards both ends of the specimen. Near point c, microcracks also emerged at the fissure ends: some converged with central tensile cracks to form wing-shaped cracks, while others propagated from the fissure tip towards the specimen ends as secondary coplanar cracks, locally accompanied by minor shear microcracks. Ultimately, through-fracture occurs along the specimen’s primary structural plane in the form of combined tensile-shear cracks, with the macroscopic failure pattern tending towards a ‘Y’ shape. When the fracture dip angle α is 30° and 45°, microcracks first form at the fracture tip and propagate towards both ends of the specimen perpendicular to the fracture plane, forming wing-shaped cracks (the wing-shaped cracks are shorter at 45° compared to 30°). As loading progressed, secondary coplanar cracks rapidly developed at both ends of the fissure along the main structural plane in the form of mixed tensile-shear cracks, dominating the through-fracture and resulting in an overall ‘X’-shaped failure pattern. When the fracture dip angle α = 60°, the specimen initiates failure at the fracture tip. Wing-shaped cracks and secondary coplanar cracks emerge almost simultaneously, with the latter propagating rapidly along the main structural plane and controlling the X-shaped through-fracture. At a crack inclination angle α = 75°, fracturing again initiated at the crack tip. However, only one distinct secondary coplanar crack formed, propagating along the main structural plane to cause through-fracturing. Wing cracks appeared sporadically, locally coupling with the secondary coplanar crack. At a crack inclination angle α = 90°, microcracks first developed at the crack tip; As loading progressed, a crack propagated from the lower left corner of the specimen towards the lower tip of the fissure, converging with microcracks extending downward from the lower tip. Simultaneously, microcracks developed from both the upper and lower tips of the fissure towards the top of the specimen, rapidly forming and extending secondary coplanar cracks along the principal structural plane. A small number of wing cracks initiated near the tips but did not dominate. These fractures developed in an upper-lower symmetrical pattern to achieve full penetration. In summary, under uniaxial compression, the fractured specimen first produced wing-shaped fractures perpendicular to the fissure extension. Secondary coplanar fractures then rapidly propagated along the main structural plane to dominate full penetration. When the fracture dip angle α < 30°, cracking initiates at the fracture centre. As α increases, initiation progressively shifts towards the fracture tip. Variations in fracture dip angle influence the initiation location, propagation pattern, and macroscopic crack type of microcracks. The presence of fractures also increases microcrack density, significantly degrading the mechanical properties of the rock specimen. Analysis of crack propagation mechanisms in fractured rock specimens based on displacement field, stress field and AE evolution. To investigate in depth the influence of fracture dip angle on the evolution behaviour of microcracks in rock specimens, the following section introduces displacement field and stress field visualisation techniques. Combined with AE evolution characteristics, this analysis examines the microstructural causes of mechanical behaviour changes and crack evolution mechanisms in fractured specimens during uniaxial loading. Relationship between instantaneous displacement of surface particles in cracks and dip angle To further elucidate the influence of fracture dip angle on the mechanical properties of rock specimens and microcrack initiation characteristics, PFC2D simulations were employed to obtain the instantaneous displacement field distribution during fracture initiation under varying fracture dip angles (as illustrated in Fig. 12). From a micromechanical perspective, particles within fractured rock specimens undergo relative motion under applied forces. When external forces exceed the cohesive strength between particles, cohesive bonds fail, leading to microcrack initiation28. Displacement field during fracture initiation in rock specimens with different fracture dip angles (unit: m). When α ≤ 30°, the displacement of particles on both sides of the fissure is significant, indicating that the fissure surface markedly weakens external loading and readily induces microcracks to initiate primarily along the fissure surface. When α ≥ 45°, the fissure-induced effect diminishes substantially, with crack initiation governed chiefly by the high stress concentration zone at the tip rather than the weakened fissure surface, demonstrating an enhancement in the rock specimen’s overall load-bearing capacity. As the fracture dip angle increases, the normal stress component on the fracture surface intensifies, enhancing the local confinement effect. This inhibits relative slip between particles on either side of the fracture, causing the initiation location of microcracks to gradually shift from the fracture surface towards the fracture tip. In summary, as the fracture dip angle increases, the overall displacement of particles on both sides of the fracture shows a decreasing trend. This phenomenon indicates that the influence of fractures on the initial microcrack initiation gradually diminishes, with microcracks tending to be generated more by the overall stress field. It also provides a microstructural basis for the pattern where the mechanical properties of rock specimens strengthen with increasing fracture dip angle. Stress field analysis of fractured rock To reveal the micromechanism of crack propagation in fractured rock, stress field evolution maps at various locations within rock specimens with different joint inclinations were analyzed during uniaxial loading. As shown in Fig. 13, the local stress tensor was obtained from contact force inversion at each computational step. Its eigenvalues were then computed to derive the maximum principal stress of each granular element, characterizing the correspondence between stress zone evolution and crack propagation. The results indicate that stress field distribution significantly influences both the initiation location and propagation path of cracks, serving as a crucial basis for analysing the failure mechanism of fractured rock specimens. Stress contour plots at various locations during uniaxial loading for rock specimens with different fracture dip angles (unit: MPa). The essence of microcrack initiation lies in damage threshold control rather than the principle that ‘the highest equivalent stress causes failure first’. Within the particle flow model, microcrack nucleation is governed by local damage thresholds rather than the maximum equivalent stress principle. Local fracture occurs when either the normal or tangential stress component within parallel interparticle bonds first exceeds the bond strength. As illustrated in Fig. 10, during loading progression from a→d across rock specimens with varying fracture inclinations, a high stress concentration zone consistently emerges and intensifies at the fracture tip. This high-stress band progressively elongates with loading, gradually aligning with the actual crack propagation path, ultimately leading to complete failure along this zone. Under uniaxial compression conditions, the axial stress is σ and α is the crack inclination angle. The normal stress (σn) and shear stress (τ) on the crack surface can be decomposed as follows: Shear slip along a fracture surface requires satisfying the Mohr–Coulomb strength threshold: Where c is the equivalent cohesion and ϕ is the equivalent internal friction angle. When the fracture dip angle α ≤ 30°, during the initial stage of uniaxial compression, although significant stress concentration exists at the fracture tip, the high normal stress σn serves a dual purpose: it inhibits the initiation of tensile-type cracks while simultaneously raising the shear failure threshold according to the Mohr–Coulomb criterion ()29. Conversely, near the free surface adjacent to the crack, constraints are weaker and local σₙ is lower, resulting in a lower failure threshold and greater susceptibility to microcrack initiation. As loading progresses, the stress field redistributes gradually. When the crack tip region reaches the failure threshold, crack propagation follows the direction of maximum principal stress, exhibiting an evolutionary characteristic from ‘low-constraint-area initiation’ to ‘tip-dominated failure’. When the crack inclination angle satisfies 45° ≤ α ≤ 75°, the shear stress () decomposed onto the crack plane from the external load reaches its maximum near 45°, while the normal stress () continuously decreases with increasing angle. This results in enhanced shear stress and weakened normal constraint, causing the crack tip to exceed the failure threshold first. Consequently, crack propagation initiates and progresses along high-stress zones. From initiation to through-cracking, the crack path closely aligns with these high-stress regions. When the fracture dip angle α = 90°, the fracture orientation coincides with the loading direction, resulting in uniform stress distribution throughout the specimen. This condition approximates that of intact rock, leading to relatively uniform stress distribution within the specimen during the early loading stage. Ultimately, the specimen fails via through-cracking along its primary structural plane. In summary: When α ≤ 30°, high normal clamping strength elevates the failure threshold, causing early cracks to deviate from high-stress zones before converging later; When 75° ≥ α ≥ 45°, the high-stress zone at the fracture tip first exceeds the failure threshold, and the crack propagates along the high-stress region, showing a high degree of correspondence from initiation to through-fracture; when α = 90°, the stress distribution becomes more uniform, and the rock specimen fractures through along the direction of secondary coplanar fractures. The evolution path in the high-stress region closely matches the actual microcrack propagation path shown in Fig. 10 of Sect. 3.2. Consequently, crack propagation is essentially the synergistic result of stress field redistribution and microcrack evolution, reflecting the micromechanical mechanism whereby the stress field drives crack propagation. Analysis of AE evolution characteristics in fractured rock samples During uniaxial loading of rock specimens, the initiation and propagation of microcracks are accompanied by the release of AE signals. Both microcracks and macroscopic fissures generate acoustic signals, enabling AE techniques to monitor the formation and propagation of microcracks within fractured rock through acoustic means. Consequently, AE technology is frequently employed in the damage analysis of fractured rock30. Within the PFC framework, each intergranular bond failure corresponds to a single AE event. The fish programme can be employed to statistically quantify the number of AE per computational cycle. The ring count, defined as the number of oscillations exceeding the threshold signal, serves as a typical indicator of signal activity. It is utilised for evaluating AE activity and provides a rough reflection of signal intensity and frequency. Given the established correlation between AE signals from rock specimens and crack evolution31, an AE detection system for fractured rock specimens was subsequently developed based on PFC2D. Stress-strain-AE ring count variation curves during loading were obtained for specimens with varying fracture inclinations, as illustrated in Fig. 14. These curves provide further insight into the initiation and propagation of microcracks within rock specimens featuring different fracture inclinations. Stress-strain-ringing count variation curves for rock specimens under different fracture dip angles. Based on the evolution characteristics of the AE signals, the failure process of the fractured rock specimen can be divided into four distinct phases: the initial phase, the intense phase, the decline phase, and the quiescent phase32. As illustrated in Fig. 14, distinct characteristics emerge in the ring count during each phase. The initial stage of sonic emission yields a minimal number of signals. This occurs because, during the early loading phase, the fissured rock specimen undergoes continuous compaction of its fissures alongside elastic deformation. This induces relative displacement between particles, resulting in a negligible ring count. During this period, the applied load exceeds the fracture initiation stress of the fractured rock specimen. However, due to the limited number and stable propagation of microcracks, the resulting ring count remains low. The intense phase of AE occurs around the peak of the stress-strain curve. As loading continues, new cracks nucleate and propagate within the specimen, achieving local penetration. Near the peak stress, AE signals surge sharply, with ringing counts reaching a peak. Due to the active unloading employed in the uniaxial loading numerical simulation tests, the decay and silent phases of AE were not observed. Comparing AE events across different fracture inclinations reveals: as fracture inclination increases, the strain value corresponding to the first AE signal in the rock specimen rises, and the concentration of AE events intensifies with increasing fracture inclination, narrowing the strain range within which they occur. The peak ring count occurs at peak stress and increases significantly with greater fracture inclination, indicating more vigorous energy release from the fractures. In summary, fracture dip angle exerts a pronounced regulatory effect on AE activity, manifesting as delayed signal initiation coupled with heightened signal intensity. Results and discussion Results This study combines indoor uniaxial compression tests with PFC2D numerical simulations to investigate the influence of fracture dip angle on the mechanical response and crack propagation behavior of single-fractured red sandstone. By analyzing the variation characteristics of different stress types and integrating microcrack propagation, stress field, displacement field, and AE ringing counts, a connection is established between macroscopic stress-strain response and crack evolution at the microscale. Results indicate that fracture dip angle not only governs the characteristic stress level of rock specimens but also determines the mechanisms governing crack initiation, propagation, and ultimately through-fracture failure. During the initial stage of uniaxial loading, high-stress zones first appear at both ends of the crack and their adjacent regions. Subsequently, as strain increases, these zones expand into the interior of the rock specimen and gradually connect to form high-stress bands. The orientation of these high-stress bands generally aligns with the primary crack propagation direction, indicating that cracks primarily develop along high-stress regions controlled by stress transfer pathways. As α increases, stress concentration zones transition from localized tip-focused areas to directional accumulation along primary structural planes. High-stress zones form earlier with enhanced continuity, and the alignment between crack paths and these zones becomes markedly stronger. At low inclination angles (α ≤ 30°), the early connectivity of high-stress zones is weaker, with larger low-constraint regions existing in the fracture neighborhood. Consequently, microcracks may initially nucleate in low-strain areas before transitioning to propagation dominated by the tip and high-stress zones. This demonstrates that fracture inclination shifts the crack propagation mechanism from being primarily guided by the fracture surface to being primarily governed by the overall stress distribution. At the macro level, as α increases, the overall strain levels corresponding to crack initiation and peak values rise. Microcrack counts and AE ringing counts exhibit a pattern of gradual accumulation before the peak, accelerated growth near the peak, and explosive release after the peak, with tensile cracks dominating. Regarding crack morphology, specimens generally first develop wing cracks, followed by rapid propagation and dominance of secondary coplanar cracks at the crack tip. When α < 30°, initiation is more likely to occur at the mid-crack region. As α increases, the initiation location progressively migrates toward the crack tip. Correspondingly, the macroscopic propagation pattern shifts from locally controlled propagation at low inclination angles to propagation controlled along high-stress zones at medium to high inclination angles. Discussion Existing research generally indicates that the fracture dip angle significantly alters the initiation stress, damage stress, and peak strength of rock specimens33. This aligns with the pattern observed in this study—that initiation stress, damage stress, and peak stress increase collectively as the fracture dip angle increases—further validating the reliability of the discrete element method. The classical understanding of single-crack specimens emphasizes stress concentration at the crack tip and dominant propagation via wing cracks34, which holds true under many inclination conditions. However, this study observed that initiation can occur first in low-strain regions near the crack, followed by a shift to tip-dominated growth when α ≤ 30, indicating that the location of maximum equivalent stress is not always the first point of failure. To address this, Sect. 4.2 introduces the decomposition of normal and tangential stresses on the crack surface along with the MC threshold. Combined with stress field evolution analysis, it demonstrates that the crack initiation sequence is determined by whether local conditions first satisfy the threshold criteria, rather than solely by the maximum equivalent stress at any given moment. Although this study elucidates the crack propagation mechanism in single-fracture rock under varying fracture dip angles, certain limitations remain. First, the numerical model is two-dimensional and simplifies rock into a homogeneous aggregate of spherical particles, failing to adequately reflect the three-dimensional structure and complex fracture networks of actual rock masses. Second, the research considers only uniaxial compression conditions, whereas in-situ stress conditions in underground engineering often involve confining pressure effects and more complex stress paths. In subsequent research, this study can be expanded into both numerical simulation and experimental aspects. Regarding numerical simulation, models incorporating multiple fractures, varying fracture lengths and spacings, as well as non-coplanar fractures, should be developed to investigate the influence of fracture geometry and dip angle on crack propagation behavior. Concurrently, three-dimensional numerical simulation techniques should be introduced to capture more realistic crack distribution characteristics. Experimentally, tests under triaxial and other complex stress conditions can validate and refine the crack initiation and propagation mechanisms identified in this study. Such investigations will deepen understanding of fracture behavior in single-fractured rock masses, providing more reliable foundations for stability assessment and construction safety in underground engineering projects involving fractured rock formations. Conclusion This study employed the discrete element software PFC2D to construct numerical models of red sandstone with varying fracture inclinations. A series of uniaxial compression simulations was conducted to systematically analyse, from a micromechanical perspective, the initiation, propagation, and failure characteristics of microcracks in fractured specimens. By integrating analyses of the particle stress field and displacement field, the influence of fracture inclination on the micromechanical response and failure mechanisms of single-precast fractured specimens was investigated. Key conclusions are as follows: Fracture dip angle is a key parameter governing rock mechanical response. As fracture dip angle increases, initiation stress, damage stress, and peak stress all rise collectively, with more pronounced increases occurring when α > 45°. During loading, the number of microcracks exhibits exponential growth, peaking most significantly in the post-peak stage. Tensile cracks dominate, leading to tensile failure of the specimens. Fracture dip angle is a key parameter governing rock mechanical response. As fracture dip angle increases, initiation stress, damage stress, and peak stress all rise collectively, with more pronounced increases occurring when α > 45°. During loading, the number of microcracks exhibits exponential growth, peaking most significantly in the post-peak stage. Tensile cracks dominate, leading to tensile failure of the specimens. Variations in fracture dip markedly influence microcrack propagation paths and macroscopic failure patterns. Wing-shaped cracks initiate first, propagating perpendicular to the fracture at a slower rate. Secondary coplanar cracks subsequently develop, rapidly extending from the fracture tip towards both ends of the specimen and dominating through-crack formation. When α < 30°, microcracks initially nucleate near the midpoint of the fracture; when α ≥ 30°, nucleation shifts to the fracture tip. Variations in fracture dip markedly influence microcrack propagation paths and macroscopic failure patterns. Wing-shaped cracks initiate first, propagating perpendicular to the fracture at a slower rate. Secondary coplanar cracks subsequently develop, rapidly extending from the fracture tip towards both ends of the specimen and dominating through-crack formation. When α < 30°, microcracks initially nucleate near the midpoint of the fracture; when α ≥ 30°, nucleation shifts to the fracture tip. As the fracture dip angle increases, the instantaneous displacement of particles on the upper and lower surfaces of the pre-fractured rock specimen generally decreases during fracture initiation. This indicates a diminished relationship between the emergence of initial microcracks and the fracture itself, with the fracture’s inductive effect on the initiation location gradually weakening. As the fracture dip angle increases, the instantaneous displacement of particles on the upper and lower surfaces of the pre-fractured rock specimen generally decreases during fracture initiation. This indicates a diminished relationship between the emergence of initial microcracks and the fracture itself, with the fracture’s inductive effect on the initiation location gradually weakening. The coupling between the stress field and crack field exhibits phased characteristics: at early low inclinations (α ≤ 30°), although stress concentration occurs at the tip, enhanced normal loading facilitates cracking initiation in the low-constraint zone near the fracture surface. As loading progresses, high-stress regions gradually coincide with crack paths, manifesting as ‘high-stress zone-driven’ propagation when α ≥ 45°. The coupling between the stress field and crack field exhibits phased characteristics: at early low inclinations (α ≤ 30°), although stress concentration occurs at the tip, enhanced normal loading facilitates cracking initiation in the low-constraint zone near the fracture surface. As loading progresses, high-stress regions gradually coincide with crack paths, manifesting as ‘high-stress zone-driven’ propagation when α ≥ 45°. As the fracture inclination angle increases, the strain value corresponding to the first AE signal from the rock specimen rises. Furthermore, the concentration of AE events intensifies with increasing fracture inclination angle, while the strain range within which events occur narrows. The ringing count increases significantly with greater fracture inclination angle, indicating more vigorous energy release from the crack. As the fracture inclination angle increases, the strain value corresponding to the first AE signal from the rock specimen rises. Furthermore, the concentration of AE events intensifies with increasing fracture inclination angle, while the strain range within which events occur narrows. The ringing count increases significantly with greater fracture inclination angle, indicating more vigorous energy release from the crack. Acknowledgements The project supported by the National Natural Science Foundation of China (Grant No. 52264003). Author contributions Wenhua Zha: Writing - review & editing, Project administration, Funding acquisition. Yichao Cao: Writing - original draft, Data curation, Conceptualization. Tao Xu: Writing - review & editing, Supervision. Bingwen Wang: Writing - review & editing.Xuejian Tan: Writing - review & editing. Funding The project supported by the National Natural Science Foundation of China (Grant No. 52264003). Data availability The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.