Final Failure Research Articles

Obtaining the longitudinal compressive response of unidirectional laminate composites from fiber misalignment micrographs through machine learning

The longitudinal compressive behavior of unidirectional composite laminates with fiber waviness is highly complex and plays a crucial role in determining final failure of composites. Evaluating this behavior, especially considering nonlinear failure mechanisms like kink-band formation, typically requires computationally expensive finite element analysis, which is impractical for large-scale quality inspection. To address computational challenges, this paper developed a new computational framework using Convolutional Neural Network (CNN) models, providing an ultra-efficient prediction of the entire stress–strain curve of composites with fiber waviness. The CNN models were trained on simulation data generated from an experimentally validated mesoscale finite element model. The microstructures of the composites with fiber waviness were taken from realistic micrographs, resulting in diverse stress–strain curves. The proposed CNN models showed high accuracy and efficiency for predicting the highly nonlinear stress–strain curves of the composites, which can be employed as a real-time evaluation method of the criticality of fiber waviness.

Investigating the Mechanical Characteristics and Fracture Morphologies of Basalt Fiber Concrete: Insights from Uniaxial Compression Tests and Meshless Numerical Simulations.

To explore the mechanical properties and fracture modes of basalt fiber-reinforced concrete, single-doped and hybrid-doped basalt fiber-reinforced concrete was prepared, and uniaxial failure tests under different basalt fiber-reinforced concrete contents were carried out. At the same time, the smooth kernel function in the traditional SPH method was improved, and the basalt fiber random generation algorithm was embedded in the SPH program to realize the simulation of the progressive failure of basalt fiber-reinforced concrete. The results show that under the circumstance with no basalt fiber, the specimen final failure mode is damage on the upper and lower surface, as well as the side edge, while the interior of the specimen center is basically intact, indicating that there is an obvious stress concentration phenomenon on the upper and lower surface when the specimen is compressed. Under the circumstance with basalt fiber, longitudinal cracks begin to appear inside the specimen. With the increase in the content, the crack location gradually develops from the edge to the middle, and the crack number gradually increases. This indicates that appropriately increasing the fiber content in concrete may improve the stress state of concrete, change the eccentric compression to axial compression, and indirectly increase the compressive strength of concrete. The numerical simulation results are consistent with the test results, verifying the rationality of the numerical simulation algorithm. For the concrete model without the basalt fiber, shear cracks are generated around the model. For the concrete model with basalt fiber, in addition to shear cracks, the tensile cracks generated at the basalt fiber inside the model eventually lead to the splitting failure of the model. The strength of concrete samples with basalt content of 0.1%, 0.2%, and 0.3% is increased by 1.69%, 5.10%, and 4.31%, respectively, compared to the concrete sample without basalt fiber. It can be seen that with the increase in the content of single-doped basalt fiber, the concrete strength is improved to a certain extent, but the improvement degree is not high; For hybrid-doped basalt fiber-reinforced concrete, the strength of concrete samples with basalt content of 0.1%, 0.2%, and 0.3% is increased by 14.51%, 15.02%, and 30.31%, respectively, compared to the concrete sample without basalt fiber. Therefore, compared with the single-doped basalt fiber process, hybrid doping is easier to improve the strength of concrete.

Study on the strength influence mechanism and cracking mechanism of stone powder-cement floor grouting materials

Underground grouting is a concealed project, and it is difficult to test the strength of grouting stones in engineering practice. Aiming at the problems of high confined water pressure and strong water-richness of the roadway floor, a pressure filtration test device was developed, and a PFC2D uniaxial compression model was established to study the variation law of stone strength and crack evolution mechanism of grouting materials under different grouting pressures. The results show that with the increase of pressure filtration value, the amount of slurry dehydration increases, the pore water pressure dissipates, the pores between particles are tightly compressed, the contact force between particle skeletons increases, and the strength of stone body increases. Under the condition of the same cement powder ratio (mass ratio of cement to stone powder) or stone powder particle size, the strength of stone body of stone powder-cement grouting material increases with the increase of pressure filtration value. The stress-strain curves of the stone body with different pressure filtration values all experienced three stages: continuous elasticity, fracture expansion and strength failure. Before the peak, the number of cracks increases slowly; after reaching the peak, the micro-cracks extend rapidly and the number increases rapidly, resulting in the final tensile failure of the specimen. This study can provide a basis for the selection of grouting engineering material ratio and grouting pressure parameters.

Multiscale image-based modeling for failure prediction of sheet molding compound composite under uniaxial tension

Carbon fiber reinforced polymer (CFRP) composites are extensively utilized as primary load-bearing components in various engineering applications due to their superior strength-to-weight ratio and excellent mechanical properties. However, their intricate microstructural interactions within composite present a significant challenge for failure analysis of CFRP. Although finite element (FE) simulations have been proven feasible to conduct the failure analysis, the classical FE models are developed based on homogeneous fiber characteristics, ignoring the influence of internal structures on the damage evolution process. This paper presents a multiscale image-based modeling approach to predict the tensile failure procedure of chopped carbon fiber sheet molding compound (SMC) composite. To accurately reconstruct the representative volume element (RVE) model of the SMC composite, synchrotron micro-X-ray computed tomography (μXCT) was adapted to explore the SMC internal microstructure. Then microscale RVE models with different fiber volume fractions were constructed to predict the corresponding microcosmic mechanical properties, which were used as the inputs for mesoscale RVE models to determine the constitutive parameters of fiber chips having varied fiber volume and orientations. Finally, the YOLOv5_Seg algorithm was employed to extract the geometric feature parameters of the fiber chips for mesoscale RVE modeling and then the failure location and sequence under uniaxial tension were predicted. It is found that the final simulated failure behaviors were consistent with the experimental observations, confirming the feasibility of this approach for understanding the failure mechanisms of CFRP composites. Thus, once the internal microstructure is determined using experimental techniques or predicted by simulating the composite manufacturing process, this approach can also be utilized for design optimization and performance evaluation for CFRP composites.

Shear performance of perfobond leiste (PBL) connectors in engineered cementitious composites (ECC)

Substituting normal concrete with ductile Engineered Cementitious Composites (ECC) is a promising way to address the safety and durability issues caused by cracking in steel-concrete composite structures. To realize a synergistic deformation behavior between steel and ECC, Perfobond Leiste (PBL) shear connectors are preferred to be used owing to their superior shear resistance, shear stiffness, and excellent fatigue performance. This research comprehensively investigated the shear performance of PBL connectors in ECC through experimental, numerical, and analytical approaches. Firstly, the shear load-slip relationships of PBL connectors in ECC and concrete were studied through push-out tests of 36 H-shaped specimens. It indicated that the fracture of penetrating bars dominated the final failure of PBL connectors in ECC, whereas premature concrete spalling happened in concrete specimens. By establishing a refined finite element model, the parametric study was then conducted to clarify the influence of the compressive strength of ECC, hole diameter, penetrating bar diameter, and penetrated plate thickness on the shear characteristics of PBL connectors in ECC. Finally, an analytical model for predicting the shear carrying capacity of PBL connectors embedded in ECC was derived, matching well with the test results.

Dynamic strain ageing of austenitic Ni-based alloy during cyclic loading at 350 °C: Mechanism and its evolution

Dynamic strain ageing (DSA) in an austenitic Ni-based alloy was investigated by symmetric/asymmetric low cycle fatigue tests with different strain amplitudes at 350 °C. A quantitative analysis of DSA activity was provided in terms of the characteristic parameters (DSA transient and serration length), and associated microstructural features were explored. Particular attention was paid to the comparison in the DSA domain between Ni- and Fe-based austenite with different strain ratios. Results showed that DSA existed throughout life and led to the continuous cyclic hardening in austenitic Ni-based alloy. The DSA activity first decreased and remained almost constant under symmetric loadings. A gradual increase in DSA activity was observed for the asymmetric loadings before the final failure. In addition, a physical model was presented to explore the DSA mechanisms of austenitic Ni-based alloys in different temperature ranges. Specifically, at 350 °C, DSA is caused by C diffusion in austenite, while at higher temperatures, it may be caused by Cr diffusion. Moreover, for a given mechanism, the DSA occurred at a higher temperature range in Ni-based alloy than in Fe-based alloy.

SEM-DIC characterization of the damage mechanism of an AlSi10Fe0.7 casting alloy on the microstructure scale

Various brittle phases are present in commercial cast aluminum alloys, which strongly influence their mechanical behavior. Among these, silicon precipitates are nearly omnipresent, as Si is a common alloying element. In secondary alloys, usually Fe-containing phases cannot be avoided, and they tend to degrade the mechanical properties. The interaction between the silicon phase and the failure-critical intermetallic phase in the Al-Si-Fe phase system (β-Al5FeSi) is studied in this paper in high resolution. A model alloy AlSi10Fe0.7 was defined, which is composed of a large grain Al-matrix, Si-precipitates and the plate-like β-Al5FeSi phase. The goal of the study was to identify “hot spots” in the microstructure from which cracks may initiate under mechanical loading. The main tool was a deformation analysis via digital image correlation in the SEM (SEM-DIC). This allows the identification and tracking of developing strain localizations at different potential crack initiation sites with a high resolution as well as capturing an overview over the whole specimen. An adapted frame averaging script minimized measurement errors induced by drift. The SEM-DIC results show that the deformation field is governed by the elastic incompatibility of the microstructural constituents. Crack initiation occurs because of the detachment of the Si + β-Al5FeSi phase boundary. Cracks then cross the phase boundary and propagate along twin boundaries in the β-Al5FeSi phase. Final failure is caused by linking fractured brittle plate-like particles.

Axial behavior variation of 72-story concrete-filled steel tubes with different joint connections due to interface debonding considering long-term deformation of concrete core

Long-term concrete deformation including shrinkage and creep occurring seriously threats the longevity, safety and serviceability of concrete-filled steel tube (CFST) structures. In this study, numerical simulation on the variation of axial behavior of 72-story CFST structures with different joint connection details due to interface debonding considering shrinkage and creep of concrete core is carried out. Three connection details including connections with inner horizonal diaphragms only, two-direction through-beams only, and inner horizontal diaphragms combined with two-direction through-beam, are considered, respectively. Vertical loading from floors is treated as centralized vertical force applied on steel tube directly. The shrinkage of concrete core is simulated by decreasing concrete temperature. A cohesive constitutive law and a Coulomb friction model are employed to describe the interface behavior in tangential direction before and after debonding, respectively. The interaction between concrete core and steel tube in normal direction is treated as a hard contact. The concrete plastic-damage (CPD) constitutive model is used for concrete core and a plastic constitutive law is employed for the steel tube. The initiation of the interface debonding between steel tube, H-shaped steel beam flanges, inner diaphragms and concrete, the variation of the force transfer path, stress redistribution, steel tube yielding and the final failure are described in detail. Results show that the axial load-carrying capacity of each CFST structure has a decrease of 26-28% when compared with that determined by the current design code where the confinement effect is considered, and a decrease of 9-10% when compared with that determined by the current design code where the confinement effect of steel tube is not considered. Moreover, the axial stiffness decreases to 76-78% of their theoretical values when debonding is not considered. Numerical results show that the interface debonding negatively affects the long-term axial load-carrying capacity and stiffness obviously, which has not been considered in current design codes. Some comments on the design of CFST members are proposed based on the numerical simulation results.

Structural response of glass fiber-polymer composite bending-active elastica beam under long-term loading conditions

Bending-active elastica beams represent structural members which are initially installed as straight beams and then bent into arched shapes by applying horizontal displacements to one support. Designing such members for permanent structures made of fiber-polymer composites involves complex viscoelastic responses, which have not yet been thoroughly investigated. An experimental investigation of medium-scale bending-active elastica beams, consisting of pultruded glass fiber-polymer composite profiles, was thus conducted to investigate the long-term structural behavior of such members under imposed sustained bending and axial compression. The results revealed that viscoelastic responses are based on an interaction of stress relaxation and creep with their effects increased with increasing bending degree and time of exposure to sustained strains and stresses. The imposed horizontal displacement to one of the supports to maintain the bent beam shape induced sustained bending stresses in the beam. Beneficial relaxation of these stresses occurred with relaxation predicted to reach 12 % during a targeted 50-year design service life. Furthermore, the likelihood of the curved beam exhibiting in-plane deformations under sustained stresses enabled creep to occur simultaneously, with associated in-plane creep deformations and strength reduction. While creep deformations remained insignificant, progressive creep rupture occurred at highest bending degrees, exhibiting sequential creep rupture in the outer combined mat layers, delamination, crack opening and final fiber failure. Creep rupture can be prevented by postponing crack initiation in the combined mat layer beyond the targeted design service life. This can be achieved by limiting the bending degree to 50 % of the bending degree at which short-term crack initiation occurs.

Buckling analysis of PMMA hemispherical pressure shells with thickness variation

The buckling behaviour of polymethyl methacrylate (PMMA) hemispherical pressure shells under uniform external pressure was investigated experimentally and numerically. Six PMMA hemispherical pressure shells were prepared via the free blow-forming process. The geometry and wall thickness of each hemispherical shell were measured. The collapse loads and final failure modes of all shells were obtained via a hydrostatic pressure device. In addition, through optical 3D scanning, a numerical model of the hemispherical shell that reflects the actual geometric imperfection was established and used in the finite element buckling analysis. The numerical results were in agreement with the test results. These findings provide a reference for evaluating the buckling load of PMMA hemispherical shells prepared via the free blow-forming process.

Experimental study on seismic performance of seismic-damaged precast steel reinforced (recycled) concrete frame structure after strengthened and repaired

In order to understand the seismic performance and failure mechanism of the seismic-damaged prefabricated steel (recycled) concrete frame after strengthening and repairing, this paper uses four different strengthened methods, including carbon fiber reinforced polymer (CFRP), high-strength steel wire cloth (HSSWC), oblique support and enveloped steel jacket. Four seismic-damaged prefabricated steel (recycled) concrete frames were strengthened by single or composite strengthening, and low-cycle reciprocating failure tests were carried out on all frame specimens. The final failure mode of the reinforced specimens under simulated seismic load was observed. The seismic performance indexes such as load-displacement (P-Δ) hysteresis curve, skeleton curve, and stiffness degradation of each specimen before and after strengthening were compared and analyzed. The test results show that the strengthening methods used in this test can effectively restore the seismic performance of the seismic-damaged prefabricated steel (recycled) concrete frame. The specimens show a typical ductile failure mode, and the P-Δ hysteresis curve of the specimens is relatively full. For a single strengthened form, CFRP strengthened can more significantly restore and improve the ductility of the specimen than HSSWC strengthened, but the cumulative energy dissipation capacity of the specimen after HSSWC strengthened is stronger. The bearing capacity of the specimens after composite-strengthened is significantly improved, and the damage degree of the structure is reduced. The addition of oblique support and an enveloped steel jacket can provide greater lateral stiffness for the specimens and improve the failure mode of the specimens. The results of this paper can provide a scientific basis for the selection and design of strengthened methods for prefabricated steel-reinforced concrete structures after earthquake damage.

Percolation versus depinning transition: The inherent role of damage hardening during quasibrittle failure.

The intermittent damage evolution preceding the failure of heterogeneous brittle solids is well described by scaling laws. In deciphering its origins, failure is routinely interpreted as a critical transition. However at odds with expectations of universality, a large scatter in the value of the scaling exponents is reported during acoustic emission experiments. Here we numerically examine the precursory damage activity to reconcile the experimental observations with critical phenomena framework. Along with the strength of disorder, we consider an additional parameter that describes the progressive damageability of material elements at mesoscopic scale. This hardening behavior encapsulates the microfracturing processes taking place at lower length scales. We find that damage hardening can not only delay the final failure but also affect the preceding damage accumulation. When hardening is low, the precursory activity is strongly influenced by the strength of the disorder and is reminiscent of damage percolation. On the contrary, for large hardening, long-range elastic interactions prevail over disorder, ensuring a rather homogeneous evolution of the damage field in the material. The power-law statistics of the damage bursts is robust to the strength of the disorder and is reminiscent of the collective avalanche dynamics of elastic interfaces near the depinning transition. The existence of these two distinct universality classes also manifests as different values of the scaling exponent characterizing the divergence of the precursor size on approaching failure. Our finding sheds new light on the connection between the level of quasibrittleness of materials and the statistical features of the failure precursors. Finally, it also provides a more complete description of the acoustic precursors and thus paves the way for quantitative techniques of damage monitoring of structures-in-service.

Comparison of the flexural and flexural-bond behaviour of beams reinforced with CFRP rods with different surfaces

Carbon-fibre-reinforced polymer (CFRP) bars are available with diverse surfaces and with a wide range of tensile strengths compared with traditional rebars. However, CFRP rods in beams require a larger cross-section than steel rebars in order to meet the required anchorage length in current design standards. The goal of this work was to compare the flexural and flexural-bond behaviours of CFRP-reinforced beams and steel-reinforced beams using four-point flexural tests. CFRP rods with thread-wrapped and sand-coated surfaces, each with distinct characteristics, were used. Additionally, two shear span/depth ratios were considered (2.5 and 3.5). The results indicated pseudo-plastic behaviour in the load–deflection curves of the CFRP-reinforced beams. This behaviour was attributed to the slip due to the reduced bond strength between the concrete and reinforcement, leading to the final failure. Based on the surface type of the CFRP rods, a cross-sectional design with the maximum load exceeding the design load was realised, even when the CFRP-reinforced beam failed to satisfy the required anchorage length. These findings suggest that the current anchorage length equations may be overly conservative, indicating the need for appropriate adjustments to the design approach.

Effect of geometry and adhesion on the performance of fiber reinforced thermoplastic composite joints with metal inserts

Joints for laminated composite structures using mechanical fasteners or adhesives often face issues such as reduced strength from fiber breakage in mechanically fastened joints and complex preparation for adhesive joints. Additionally, adhesive joints require compatibility between the adhesive and the composite matrix, limiting their suitability for many thermoplastics. To address these challenges, this study explores a technique that uses metal inserts embedded during the consolidation of laminated thermoplastic composites. It investigates how the geometry of metal inserts and their adhesion to the fiber-reinforced thermoplastic composite affect the joint’s performance under out-of-plane loads. Results indicate that while mechanical locking assists in load transfer, the adhesion between the metal insert and the composite is crucial for the joint’s failure mechanism and overall performance. Strong adhesion delays final failure by requiring fiber breakage through the composite’s entire thickness, while weak adhesion leads to failure through fiber breakage on the bearing side.

Analysis of the progressive collapse of a parking garage concrete structure due to punching shear

This work aims to clarify the causes leading to the structural collapse of the garage of a residential development located in Santander, Spain, in 2020. Visual inspections of the collapsed area were carried out to establish the first hypotheses about the collapse. Subsequently, the materials, geometry and acting loads were studied to compare them with the original project. The main conclusion was that the slab was subjected, in the area where the collapse began, to gravity loads of 34 kN/m2 when it was designed to support 15 kN/m2. Then, structural verifications were performed in accordance with EHE-08 and Eurocode 2, in force during project design, to identify the failure mechanism(s) that caused the progressive collapse of the structure. Unacceptable safety factors were observed for punching shear, which was the main failure mechanism. Punching was aggravated by cavities detected in the concrete, which caused both a decrease in the slab strength (lower effective depth of the slab) and an increase in the stresses in the compression struts of about 20 % to 30 %, which considerably reduced the punching safety levels. This calculation was carried out by means of a finite element analysis with SAP2000. The assessment of this type of defect, very common during the execution phase, is not well developed either from a theoretical or regulatory point of view. This paper proposes, as a corrective measure, the inclusion both in the European and National Standards of a mandatory inspection procedure based on Ground Penetration Radar or ultrasounds, during the execution phase, to detect and correct this type of defect, not detectable by visual inspection. Finally, a bending and cracking analysis were also conducted, giving values of the corresponding safety factors of less than unity. Therefore, they also contributed to both the progressive loss of strength of the slabs and the final punching failure. These last two causes are also crucial to understand why the collapse occurred 15 years after its commissioning.

Influences of Pre-Existing Fissure Angles and Bridge Angles on Concrete Tensile Failure Characteristics: Insights from Meshless Numerical Simulations.

The existence of cracks is a key factor affecting the strength of concrete. However, traditional numerical methods still have some limitations in the simulation of crack growth in fissured concrete structures. Based on this background, the numerical treatment method of particle failure in smoothed particle hydrodynamics (SPH) is proposed, and the generation method for concrete meso-structures under the smoothed particle hydrodynamics (SPH) framework is developed. The concrete meso-models under different pre-existing micro-fissure inclinations and bridge angles (the inner tip line of the double pre-existing micro-fissure is defined as a bridge, and the angle between the bridge and the horizontal direction is defined as the bridge angle) were established, and numerical simulations of the crack propagation processes of concrete structures under tensile stress were carried out. The main findings were as follows: The concrete meso-structures and the pre-existing micro-fissures all have great impacts on the final failure modes of concrete. The stress-strain curve of the concrete model presents four typical stages. Finally, the crack initiation and propagation mechanisms of fissured concrete are discussed, and the application of smoothed particle hydrodynamics (SPH) in crack simulations of fissured concrete is prospected.

Tensile properties and damage behaviors of the prepreg-RTM co-cured composite bolted T-joint with a novel configuration

The tensile properties and damage behaviors of prepreg-resin transfer molding (RTM) co-cured composite bolted T-joints, consisting of an internal skeleton and an external skin, were investigated through experimental and numerical methods. With the validated finite element model, the effects of adding load-carrying beams, the corner radius and the stacking sequence were investigated. Results indicate that the novel T-joint, weighing only 55.7% of an equivalent sized 2A12 aluminum T-joint, can achieve 82% of the aluminum T-joint’s tensile stiffness and 107% of its proportional limit load. Damage in the joint initiates near the bolt holes and leads to delamination failure within the skeleton. With load-carrying beams including bolt holes, the tensile stiffness and ultimate load increase to 1.11 and 1.51 times that of the original configuration, respectively, with the final failure mode shifting to the fracture of the base panel. With the final failure mode unchanged, the ultimate load remains essentially unchanged with increasing the corner radius. Under tensile loads, the preferred stacking sequence for the skeleton is [0/+45/90/−45]nS. For the skin, the greater the number of ±45° plies, the higher the ultimate load, while the greater the number of 90° plies, the higher the tensile stiffness.

Adhesive strength improvement by providing steps in joints and differentiating initial and final debonding stresses

Step adhesive joints have a special characteristic quite different from other joints. When initial delamination occurs in other joints such as butt, lap and scarf joints, final failure always occurs. In these cases, the external stress causing initial delamination σcInitial is equal to the final failure stress as σcFinal = σcInitial. However, in step joints, the final failure stress σcFinal can be greater than the initial delamination stress σcInitial < σcFinal. To clarify the adhesive improvement mechanism, first, this paper discusses the ISSF (Intensity of Singular Stress Fields) for fully bonded step joint by varying the number of steps NS. Second, the singular stress field causing 2nd debonding is discussed by analyzing partially delaminated step joint. The results show that 2nd debonding requires larger external load than the initial debonding as σcInitial < σcFinal. This is because under the same external load the singular stress causing the 2nd deboning is smaller than the one causing the initial debonding. When NS≥6 and suitable overlap length, the final bond strength σcFinal can be more than 3.6 ∼ 4.4 times larger than the initial delamination stress σcInitial≪σcFinal resulting in much larger bond strength.

Study on seismic failure mechanism and damage model of unconstrained adobe block walls

Through the quasi-static cyclic in-plane loading test of three kinds of unconstrained adobe block walls, namely plain adobe wall, modified mud adobe wall and modified adobe wall, the failure mode and seismic performance of each specimen were clarified. Using residual deformation and cumulative energy dissipation as failure parameters, a two-parameter seismic damage model of unconstrained adobe block walls was established. and a calculation method for the adjustment coefficients of energy consumption items was provided. The results showed that the failure mode of the typical unconstrained adobe block wall under earthquake action was not affected by the modification of adobe materials, and the final failure mode was shear failure along the mud crack and the inclined crack extending through the block. Furthermore, the proposed seismic damage model can well reflect the evolution law of earthquake damage of adobe block walls, and the recommended damage index boundary values for five performance levels were given, providing a reference for performance-based seismic design and damage assessment of adobe block walls.

Does double pre-notches have a greater impact than single pre-notch on the mechanical and fracture behavior of rock? Insights from three-point bending tests using the numerical approach of grain-based model

The mechanical performance of rock masses is significantly impacted by the existence of cracks. Distribution of the cracks may result in different mechanical or fracture behaviors of rock. To gain a preliminary understanding of such problems, a series of numerical tests of three-point bending (TPB) are designed and carried out with the consideration of single pre-notch and double pre-notches. The numerical method of grain-based model (GBM) approach is employed in this paper, with a validation implemented by comparing the results with previous experiments to demonstrate the reliability of this model. Afterward, a comprehensive investigation, including the mechanical behavior, strength characteristics, crack evolution and progressive failure mechanism of the above mentioned TPB tests are conducted. The conclusions are as follows: (1) During TPB tests, as the offset distance increases, the fracture mode of the single pre-notched specimens gradually transitions from Mode I to Mode II, with the peak fracture toughness occurring at a mixed Mode I-II fracture condition (2p/s = 0.4). Moreover, the failure angle, notch open degree and expansion area all tend to rise regardless of specimens of single or double pre-notches. (2) A thorough exploration of crack evolution shows that the specimens primarily fracture during the crack propagation and failure stages. Shear failure predominates among these stages, with microcracks mostly occurring inside the rock mineral grains. (3) When offset distance increases to a certain degree (2p/s = 0.8), the fracture of specimen ceases to be governed by the pre-notch; instead, it occurs directly from the middle of the beam specimen. This phenomenon can be attributed to the reduction of stress concentration at the notch tip, as inferred from analysis of the contact force chain. (4) In double pre-notched specimen, microcracks initiate at both pre-notches, but the final failure results from crack propagation at one notch tip. Compared to single pre-notched specimen, the presence of double pre-notches exhibits superior fracture resistance performance.