Based on the self-developed shear-seepage coupling test device for coal rock, a series of direct shear tests was performed on the unfilled joints and joints filled with gypsum, debris, and clay. The 3D optical scanning technology was also employed to examine the morphology and aperture evolution of joints during shearing. The results show that peak shear stress and normal displacement of joints exhibit same regularity, which decreases as filling material varies from gypsum, debris to clay, showing that the nature of filling materials contributes significantly to the shear properties of joints. Unfilled joint is evidently worn after shearing and its JRC decreases most significantly due to direct contacting between the upper and lower surfaces. The JRC of joints filled with gypsum and debris changes relatively little, while it remains nearly the same for joints filled with clay. The materials failure and surface abrasion are mainly controlled by the aperture evolution and local stress concentration. With shear displacement increasing, stress concentrates in the region aperture decreases, where materials failure and abrasion of unfilled joints occurred.

The current research on the periodic piles for vibration isolation based on the method of band gap mainly focused on theoretical and numerical analysis. In this paper, the method of principle experiment was adopted to validate the characteristics of attenuation zone of the periodic piles for vibration isolation. Four experimental cases were designed to consider two types of periodic piles (i.e., hollow steel piles and periodic hollow steel piles filled with soil) and two types of periodic layout (i.e., a hexagonal configuration and a square configuration). With the excitation of impulsive loads, the experimental cases with vibration isolation measures were compared to those without vibration isolation measures. Vibration attenuations of the periodic piles in various experimental cases were measured, analyzed and compared with theoretical band gaps. The measured vibration response in various experimental cases was greatly attenuated within the band gaps of theoretical calculations. The levels of attenuation were above 50% and the highest one was up to 98%. Both the attenuation effect and vibration isolation performance were greatly achieved, and thus the effectiveness of theoretical analysis is verified. The horizontal level of vibration attenuation of the periodic piles is better than the vertical. From this trend, with the same conditions, the vibration attenuation level of periodic piles in a hexagonal configuration was better than those in a square configuration. Moreover, soil-filling could effectively increase the width of the initial band gap and the attenuation zone of periodic steel pipe piles arranged in a hexagonal configuration, but reduce the lower bound frequency of the initial band gap and the attenuation zone of periodic steel pipe piles arranged in a square configuration. By analysing the band gap of periodic piles that were approximate to actual engineering dimensions corresponding to the periodic piles in the experimental cases, it is found that there are several band gaps appearing in the frequency range of 0-80 Hz which can be lower than 20 Hz. This study further illustrates that an objective frequency range can be designed to be isolated by periodic piles, and the periodic piles based on band gap theory can have a broad application prospect in the vibration isolation induced by metro train.

Reservoir rock is a type of heterogeneous material, which has pore structure with complex geometry, and microstructure affects the macro-mechanical properties and fracture characteristics of rocks. Therefore, understanding of reservoir rock micro-pore structure and its influence on rock mechanics is of great significance to oil and gas exploitation and reservoir transformation. For this reason, micro-pore structure of tight sandstone is quantitatively characterized by CT scanning experiment and digital image processing technology, and a discrete element model with pore structure is established. The micro-parameters of the model are calibrated by uniaxial compression experiment. The influence of pore structure on rock mechanical properties and crack propagation is analyzed by uniaxial compression simulation in PFC3D. The results show that the probability distribution of micro-pore structure of tight sandstone satisfies the lognormal distribution. The geometry and distribution of micro-pore structure are complex and have fractal features. Existence of micro-pore structure results in decrease of uniaxial compressive strength. The larger the fractal dimension is, the smaller the uniaxial compressive strength is, and the relationship is approximately linear. The micro-pore structure plays a decisive role in crack initiation position, propagation and penetration direction. Cracks initially appear at the tip of the pores. In the plane parallel to the loading direction, the cracks propagate along the 30° direction of the tip of the pore and parallel to the loading direction. In the plane perpendicular to the loading direction, the crack in the XY plane extends along the axis of the fracture or causes the hole diameter to increase. Cracks are likely to cause penetration between pores during loading.

Calcareous sand is a special medium with high calcium carbonate content or other insoluble carbonates, and has characteristics of irregular shape, porous and rich in inner pores. Capillary rise is a universal phenomenon in soil due to the pores. The tube inside rhizome is like a very fine plant capillary, which can absorb water from the soil. Therefore, the study on height of capillary rise in the calcareous sand has great significance to the ecological construction of the reef island. In this paper, the vertical tube tests are carried out to study the influence of gradation, particle size, dry density and salinity on the capillary rise in calcareous sand. Results show that the finer the particles and better gradation, the higher capillary water rise in the continuous graded calcareous sand. The height of capillary water rise decreases with increasing particle size in the single-sized calcareous sand. The capillary water rise decreases when the particle size is less than 0.075 mm with increasing dry density of the single-sized calcareous sand, while the height gradually increases with the particle size of 0.075-0.250 mm. Compared with fresh water, the NaCl solution inhibits the capillary water rise at low salinity, and the seawater promotes the capillary water rise but only weakly. The relationship between the height of capillary rise and time in calcareous sand fits well with the quadratic polynomial relationship in double logarithmic coordinates, and the physical properties of calcareous sand and the solution type are related to the parameters.

The Mohr-Coulomb (M-C) failure criterion has been widely used in engineering practice, which does not efficiently reflect the rock strength characteristics including the hydrostatic pressure effect, Lode angle effect and intermediate principal stress effect. To overcome this weakness, attempts have been made to develop the three-dimensional failure criteria, among which a recently proposed modified Hoke-Brown (H-B) failure criterion based on a two-parameter deviatoric function has an excellent performance. However, the M-C failure criterion is still widely used in practice. In order to avoid the difficulty resulting from the use of three-dimensional failure criterions in engineering, a new approach is proposed to determine the equivalent M-C strength parameters from the three-dimensional modified H-B failure criterion. The procedure for determining the equivalent M-C strength parameters from the modified H-B criterion is demonstrated in detail, and the equivalent M-C strength parameters influenced by the rock strength characteristics are investigated. Results show that dependence of M-C strength parameters on the hydrostatic pressure effect is nonlinear, and the effect of the intermediate principal stress and Lode angle is related to observation range. For a higher hydrostatic pressure, Lode angle has greater influence on the equivalent M-C strength parameters. As the minimum principal stress rises, the effect of intermediate principal stress on instantaneous cohesion is more significant. Verification of this determination method is conducted based on the true triaxial test data of different rock types, demonstrating its applicability. Findings in this paper have significant sense for practical application of the modified H-B failure criterion in the rock engineering.

The free energy and dissipation function for over consolidated clays are established, which are based on the generalized thermodynamic method and combined with the thermodynamic functions form of the modified Cam model. The energy generated by the plastic shear deformation is considered a part of free energy in the free energy function, and this partial free energy is related to the degree of over consolidation. When the degree of over consolidation decreases, the free energy generated by the plastic shear deformation also decreases. The dissipation function is independent of the current stress state, so the associated flow rule is still employed in this paper. Based on the free energy function and the dissipation function, all of the yield function, the flow law, the hardening law and the elastic law can be obtained with strict mathematical derivation from the two scalar functions. The final elastic-plastic incremental relation can then be obtained. Finally, the rationality of the model is verified by comparing the calculation and experimental results in stress and strain, volumetric strain of four different over consolidation clays.

Basalt fiber (BFRP) anchor has many advantages, such as high tensile strength, good corrosion resistance, etc. It is a good substitute for reinforcement in geotechnical anchorage structure, and has attracted much attention in recent years. Four groups of field pull-out tests were carried out in the loess stratum to study the failure modes and anchorage performance difference between the 25 mm diameter BFRP bolts and steel bolts. The results show that the failure signs on inner and outer surface of grouting coexist in the process of drawing for large diameter soil anchor. However, the ultimate failure mode is controlled by the interface between grouting and soil (the 2nd interface), and the damage degree of the interface between BFRP bolt and grouting (the 1st interface) is evidently higher than that of steel bolt. The ultimate bearing capacity of two bolts is similar, and the interfacial bond strength decreases with increasing anchorage length. The bond performance between the steel bolt and grouting is better than that of the BFRP bolt, which is mainly caused by the different processing technology and material mechanical properties between the two bolts. At the same load level and position, the axial force of BFRP bolt is larger than that of steel bolt, and the attenuation rate of axial force is slightly smaller than that of steel bolt; the peak shear stress of BFRP bolt is also smaller than that of steel bolt.

In order to study the diffusion mechanism of slurry in soil-rock mixture and its influencing factors, a reusable indoor grouting model test system was developed with the Chongqing soil-rock mixture as the research object. Experimental study of grouting model of soil-rock mixture under conditions of different stone contents, porosities, slurry viscosities and grouting pressures was carried out. The results show that slurry diffuses mainly in the soil-rock mixture by means of osmotic and splitting. When the stone content increases, the main diffusion mode of the slurry shifted from splitting to osmotic. When the medium contains stone, the splitting diffusion is equivalent to the osmotic. On the basis of experiments, the grouting diffusion radius prediction model of soil-rock mixture is established based on the theory of support vector machine. Based on the calculation results of the model, the order of influencing factors for the infiltration diffusion radius and the splitting radius are obtained respectively. Which are: stone content>slurry viscosity>porosity> grouting pressure; stone content> grouting pressure>slurry viscosity>porosity. The stone content has the greatest influence on the diffusion form, and it domains the diffusion radius value. It is suggested that the engineering should design the grouting parameters according to the stone content.

To explore the energy dissipation of coal-rock combinations failure, uniaxial compression experiments were conducted to investigate the failure process characteristics in terms of the energy. Meanwhile, the relationships of failure and acoustic emission signal were discussed based on fractal dimension. Method for evaluating the risk of rock burst was proposed, based on the investigation of the different combination structures influence on energy dissipation of coal-rock combination. The results show that when the rock height increases, the strength and elastic modulus increase, and the softening process decreases. The raise of rock height could increase the impact risk, as well as the reduction of difference between the elastic modulus of rock and coal. In addition, a new decision method was proposed based on the analyses of height and mechanical properties of coal-rock combination, compared with the “four index” of traditional fuzzy evaluation. The result of assessing the rock burst index by the decision method has higher accuracy and practicability. These results provide the data and theoretical support for further development study of rock burst risk in coal-rock combinations.

The laterite is prone to shrinkage and crack due to dehydration, which may result numerous engineering disasters. Lime (5%) and meta-kaolin (4%) were added into laterite for mitigating their shrinkage behavior through improving moisture sensitivity. The samples were compacted at optimal water content and cured for 180 days. And then, samples were vacuumed and saturated immediately. After that, the saturated samples were dehydrated to predetermined moisture content for conducting a series of tests such as shrinkage, unconfined compressive strength, suction and pore analysis. The results showed that compressive strength of laterite increased at first and then decreased with moisture content reducing, which may be caused by micro-cracks derived ascribing to dehumidification. However, the strength of laterite adding with lime, especially lime and meat-kaolinite mixture, would increase with approaching to dry status, even if which has a slight reduction during drying process. This proved that meta-kaolin and lime were able to inhibit the shrinkage effect of laterite, and increase their overall strength. This may be ascribed to the following reasons, meta-kaolin contains lots of amorphous silicon and aluminum oxides and has an edge-surface contacted structure, both enable meta-kaolin to quickly capture calcium ions in calcium hydroxide solution and form cementation hydrates between grains or particles of laterite.

To investigate the dynamic response of reinforced earth abutment structure under dynamic action of the top strip footing, dynamic load failure test of the indoor reinforced abutment retaining wall was carried out by MTS servo loading system with cyclic dynamic load. Comparative analysis of the reinforced earth retaining wall parameters with three grid lengths and three grid types, such as settlement of reinforced soil retaining wall, horizontal displacement of the panel, earth pressure and strain of the reinforcement, etc. hence revealing the dynamic bearing capacity of reinforced abutment retaining wall. The test results show that there are differences in failure modes of different geogrid lengths and types of reinforced abutment retaining walls under cyclic dynamic loading. When 1.0H, the failure modes of retaining wall of M-A-B-type geogrids are punching shear failure. And when 0.7H and 0.4H, the failure mode of retaining wall of A-type and B-type geogrids are local shear failure. The lateral displacement of the reinforced abutment retaining wall panel decreases with increasing reinforcement length, and the lateral displacement coefficient of the A-type grid reinforced retaining wall is generally smaller than that of the B-type. Due to the length and type of the reinforced geogrid, the abutment retaining wall has significant difference in the attenuation law of the dynamic earth pressure. When 1.0H, the vertical dynamic earth pressure attenuation coefficient of M-type and A-type geogrids are parabolic function along the wall height. When 0.7H, the vertical dynamic earth pressure attenuation coefficient of A-type and B-type geogrids are exponential function along the wall height.

Natural gas hydrate is a new type of clean energy that has broad application prospects. However, the change of temperature and pressure conditions in the process of hydrate mining will cause the dissociation of hydrate, which lead to the loss of cementation strength of hydrate-bearing sediments. Meanwhile, micro-cracks and defects in the sediments expand gradually, the hydrates between the soil particles are damaged and broken in the loading process. However, in the previous investigations of the multi-field coupling model on hydrate dissociation process, the damage evolution process of sediment structure and its influence on the coupling process are ignored. Therefore, the Weibull distribution of three parameters and residual strength correction coefficient are introduced into the damage statistical constitutive model in this paper based on the continuous damage theory, and the damage statistical constitutive model of hydrate-bearing sediment considering the influence of damage threshold and residual strength is established. And then, the constitutive model is embedded in the multi-field coupling model of the hydrate dissociation process, and the thermo-hydro- mechanical-chemical (THMC) multi-field coupled mathematical model considering damage of hydrate-bearing sediments is proposed. Finally, based on the model, the influence of sediment structural damage on the deformation, pressure and temperature of sediment reservoirs during hydrate dissociation are discussed. It can be shown from the computational results that the structural damage of hydrate-bearing sediments has a significant effect on the multi-field coupling process about the dissociation of hydrates, and its influence increases gradually with the increase of dissociation time.

A large-scale shaking table experiment of loess slope supported by frame anchors with a 1:10 similarity ratio was designed and completed. By loading seismic waves of different types, acceleration peaks and durations, the dynamic response laws of model slope under earthquake action were discussed, and the seismic performance of the supporting structure with frame anchors was evaluated. The test results show that the anti-seismic effect of the supporting structure with frame anchors on the loess slope is significant. Different seismic wave inputs are used to verify its stability in the seismic design of the supporting structure with frame anchors. Under the action of the supporting structure, the loess slope has evident filtering effect on seismic waves and amplification effect on frequency spectrum, and the slope acceleration response has the amplification effect on the near-empty surface and vertical amplification effect. The acceleration response exhibits ‘attenuation’ phenomenon when the peak earthquake input acceleration is small. Under the support of frame anchors, the displacement of loess slope is effectively restrained, its dynamic earth pressure distribution curve shows ‘hyperbolic’ change, and the dynamic earth pressure is distributed in an inverted triangle from the foot of the slope to the top along the slope height direction. The test results can provide some reference for the seismic design of the supporting structures with frame anchors of loess slope.

Existence of rock microcracks has great influence on rock shear fracture under compressive loadings. However, the effects of microcrack geometries and confining pressures on shear fracture angle are rarely studied in rocks. In this study, an improved expression of stress intensity factor containing crack angle effect is derived, which is based on the proposed expression at crack tips in Ashby’s microcrack model. Using this improved stress intensity factor, the constitutive relationship between crack growth, strain and stress is obtained, and it predicts the peak strength in rocks. Coupling this constitutive relationship at state of peak strength and Mohr-Coulomb failure criterion, the correlations of internal friction angle, cohesion, shear strength and damage are obtained. Effects of confining pressure, initial crack size, crack angle and friction coefficient on shear fracture angle are discussed. The rationality of theoretical model is verified with experimental data. The results show that internal friction angle, cohesion and shear strength decrease with increment of damage; shear fracture angle increases with the increment of confining pressure and friction coefficient, or the decrement of initial crack size; shear fracture angle decreases firstly, and then increases with increment of crack angle.

A novel visualizing constant pressure penetration grouting diffusion and reinforcement simulation test device was designed. The ordinary Portland cement 42.5 (OPC), microfine Portland cement (MC), microfine sulphoaluminate cement (MSAC), and self-developed effective microfine cement-based grout (EMCG) were selected, and the penetration grouting diffusion and reinforcement test has been performed. The variation rules of diffusion distance and volume under different suspension and grouting pressures were studied, as well as the effects of grouting material, pressure and sand gradation on the reinforcement. The ANOVA method was adopted to determine the dominant factor for grouting effectiveness, and the macroscopic failure modes of reinforcement specimens were obtained. The microscopic reinforcement model of rock-slurry interface and mineral characteristic were analyzed through SEM method, and the core reasons of reinforcement differences caused by different grouts were revealed. The results show that the grouting material and particle percent in slurry were the dominant factors for the groutability of fine sandy stratum, the increase of groutability was not evident with the increase of grouting pressure. The groutability of EMCG slurry was the best, followed by MC and MSAC, and the groutability of OPC slurry was the worst. The strength improvement of reinforcement specimen was better when the sand gradation was finer when suspensions penetrated fully. The 7 d and 28 d strengths of EMCG reinforcement specimen are higher than those of MSAC, MC and OPC, the 7 d of strength of EMCG reinforcement specimen is greater than 70% value of its 28 d strength. The grout is the dominant factor for reinforcement effectiveness, and EMCG is the best while OPC is the worst. After grouting, the dense C-S-H gels produced in EMCG rock-slurry interface can enhance its bonding strength effectively. Finally, improvement suggestions for practical penetration grouting design have been put forward based on three aspects, which are grouts selection, grouting dynamic control and drilling holes layout.

To accurately predict the dynamic permeability of low-rank coal, a new model was established on the basis of coal cube model, with consideration of the slippage effect and effect of matrix porosity on the permeability. Sensitivity analysis of factors affecting absolute permeability and slippage coefficient was carried out, and the effects of methane and nitrogen on matrix shrinkage and slippage effect were discussed. The results show that the permeability prediction models based on match stick hypothesis are special cases when the matrix porosity is not considered in the new model. Compared with the new model, P-M and S-D models have more evident effect on matrix shrinkage, the new model considering matrix porosity and slippage effect is more realistic. Langmuir strain is the key to matrix shrinkage, whether the absolute permeability of coal can rebound or not is the result of joint action by cleat compression, matrix pore expansion, matrix elastic deformation and matrix shrinkage. Under the same conditions, the matrix shrinkage of methane is stronger than nitrogen, and the opposite on the slippage effect. The factors affecting slippage coefficient include internal and external cause, variation law of slippage coefficient with pore pressure is opposite to that of cleat width. The slippage effect and matrix shrinkage effect together enhance the gas permeability. The lower the pore pressure is, the more significant the effect of slippage effect and matrix shrinkage is.

To investigate the expansion and shrinkage characteristics of the Lushi expansion rock, experimental study on the expansive rock expansion and contraction characteristics under dry-wet cycle was carried out. After experiencing dry and wet cycles, the expansive rock was analyzed by scanning electron microscopy (SEM) and nitrogen adsorption experiment (NA). The phenomenon of swelling and water shrinkage of expansive rock was analyzed from the microscopic point of view, and the reason for the change of expansion and contraction characteristics was explained. The results show that the expansion rate of expansive rock increases with the increase of dry and wet circulation, and the absolute expansion rate increases by 25%. The contraction curve shows an evident contraction inflection point, which usually occurs at 20% of the total shrinkage time. At this time, the water loss state of the expansion rock changes from the free water loss to the combined water loss. Cracks appear during the first expansion and contraction of the expansive rock, which are transfixion. In the later expansion and contraction process, the cracks appeared shallower gradually stabilized with the increase of dry and wet circulation. After the number of cycles reached 6-8, the expansion and shrinkage rate of Lushi expansion rock reached a stable value, with the absolute expansion rate stable at 17% and the absolute contraction rate stable at 9%. As the number of dry and wet cycles increased, the aggregation of clay particles in the microstructure of expansive rock changed from compact state to loosen. In addition, the pore characteristics of the sample show that total pore volume gradually increases, the pore diameter gradually decreases, and the specific surface area gradually increases with the increase of the number of dry and wet cycles.

To study the influence of mining disturbance and temperature on coal permeability in front of the working face, the fracture deformation factors caused by mining stress and temperature were divided into three parts: effective stress, adsorption and thermal expansion. The fracture strain expression was derived from damage mechanics, adsorption theory and thermal stress theory. The expression included the coupling damage effect of disturbance stress and temperature. Furthermore, the permeability model of coal under the influence of mining disturbance and temperature was established. The permeability model was verified and analyzed according to two permeability tests. One was changing temperature and the other was changing the disturbance stress and temperature simultaneously. In addition, the parameter sensitivity was also analyzed. The results show that the model can well represent the evolution process of coal permeability with the mining disturbance and temperature effect. Under the same temperature, permeability increases with the increase of internal swelling ratio. Permeability decreases with the increase of thermal internal swelling ratio. The research results will provide useful reference for the development of coal mining and gas extraction technology.

The roadway surrounding rock often produces single or multiple bed separation due to deformation incoordination. Based on the mechanism of bolt pullout force during the relative movement to the strata, an elastic-plastic mechanics model of additional stress occurred by bed separation was established. The influence of bolt loading caused by bed separation in elastic state was firstly discussed. The interface of anchorage body will enter elastic-plastic stage under further expansion of bed separation. The sliding ranges on both sides of the bed separation were determined through the two-stage linear function shear slide model. Loading process of the anchorage body under the bed separation function was divided into four general stages, which included elastic stage, one side elastic-plastic stage, two sides elastic-plastic stage, one side sliding further with one side sliding all stage, and both sides sliding all stage. The axial force distribution during bed separation was verified by experimental results. The results show that calculated value has good agreement with the measured value. The bolt stress distribution caused by the multiple bed separation was summarized according to the superposition principle. The multiple peak curves of stress distribution were proved. Finally, the parameter analysis of bed separation value location was made through an example. The influence of bed separation on bolt loading should be considered in future anchorage design.

Prestressed anchor bars are widely used since it can effectively control the displacement between rock and soil layers and structures. Due to the complexity of geological conditions, prestressed anchor bars are corroded and damaged continuously during long service, and even failures. In this paper, the accelerated corrosion test was conducted in lab, and the evolution characteristics of apparent corrosion of prestressed anchor bars were qualitatively described. The corrosion damage mechanism of prestressed anchor bars under the synergistic action of pH and O2 was then analyzed by means of electrochemical test system. A time-varying model of corrosion current density based on acidic oxygen environment was developed. The corrosion degree was characterized by the corrosion amount per unit length. Furthermore, the relationship between the oxygen flux rate and corrosion amount per unit length was established. The results show that corrosion degree reached a threshold value with the increase of the oxygen flux rate. Based on the test data analysis, the polarization curves of typical prestressed anchor bars were obtained, and the corrosion behavior of prestressed anchor bars in different periods was investigated. It is found that there is no apparent passivation zone in the working part of the prestressed anchor bars, and the anode is always in the active state. Positive shift of corrosion potential in oxygen environment has been revealed, while the corrosion potential of the prestressed anchor bars in the non-oxygenated acid corrosion solution is negative shift, thus the corrosion resistance is poor. This study can provide analytical means and data support for damage evolution behaviors analysis of prestressed anchor bars in aerobic rock and soil environment.

Based on the Biot’s theory of wave propagation, a saturated virtual soil pile model is proposed. Considering the three-dimensional wave effect and saturation characteristics of surrounding soil and beneath the pile toe, a three-dimensional system including saturated viscoelastic soil, virtual soil pile and solid pile is established. The analytical solution for displacement of saturated soil is also derived by potential function method. The vertical dynamic impedance at the pile head is obtained using the pile-soil compatibility conditions. Finally, the obtained solution is reduced to verify its validity with existing solutions. Furthermore, extensive parametric analysis is performed to investigate the effects of saturated soil parameters on the vibration characteristics at the pile head. The computational results show that the amplitude and resonance frequency of dynamic impedance at the pile head in saturated soil decrease with increase of saturated soil thickness beneath the pile toe. When the thickness of saturated soil beneath the pile toe increases to a certain extent, the pattern of staggering peak become evident. The porosity of saturated soil beneath the pile toe has significant effects on both the resonance amplitude and resonance frequency of the dynamic impedance at the pile head, while the porosity of saturated surrounding soil only influence the resonance amplitude. With increasing shear modulus of surrounding soil and soil beneath the pile toe, the amplitude of the dynamic impedance at the pile head is significantly reduced, and it is more affected by the shear modulus of surrounding soil.

Aging effect on thermal conductivity of Gaomiaozi (GMZ07) and MX80 bentonites with different water contents and dry densities were investigated in this paper. Water contents of compacted bentonite specimens were kept under constant volume condition during curing periods of 1, 5, 30, 60 and 100 days, and the measurements of thermal conductivity were conducted on two bentonites using a thermal probe method. The mercury intrusion porosimetry (MIP) tests were also performed to observe the pore-size distributions of compacted specimens. The test results show that the thermal conductivities of GMZ07 and MX80 bentonites decrease with increasing aging time, the thermal conductivities decrease significantly at early curing periods and then tend to be constant when the aging time exceeds a given day. The aging effect on thermal conductivity increases with increasing water content for the same dry density. According to the change in microstructure of compacted bentonite specimens with aging time, it is considered that the aging effect on thermal conductivity may be attributed to the smectite hydration during the aging process. With the hydration, part of soil water turned into “inert water” with poor heat conduction property, which leads to a decrease in thermal conductivity of bentonites.

To investigate the long-term stability of surrounding rock in cold region tunnels, the gneiss in the Huibai tunnel in Jilin Province was selected for triaxial creep test under freeze-thaw cycles. The mechanism of the freeze-thaw cycle influence on the creep properties of gneiss was analyzed. The experimental results show that with increasing number of freeze-thaw cycles, the creep deformation of gneiss gradually increases, while the creep failure stress, creep duration and long-term strength have a significant decrease trend. Simultaneously, the freeze-thaw cycle temperature has a certain influence on the creep parameters of gneiss. For the same cycle number, the lower the temperature, the larger the initial creep of the rock sample, and the smaller the creep duration and the long-term strength. In addition, based on the experimental results, the parameters of the Burgers model were identified by the 1stOpt optimization software. It performs that the theoretical curve of the model and the experimental curve agree well in the deceleration and stable creep stages. Finally, the influence mechanism of freeze-thaw cycles on the gneiss microscopic structure was analyzed by SEM scanning and energy spectrum EDS. The failure modes of rock samples under long-term load are shear failure. As the number of freeze-thaw cycles increases, the degree of rock rupture becomes more and more serious, and the rock failure block becomes smaller. The research results can provide reference and basis for the support and antifreeze design of rock engineering in cold area.

The impact mechanical model changes of granular flow caused by the changes in slope angle and rigid wall slope are neglected problem presently. The discrete element method (DEM) is utilized to investigate the influences of slope angles on the impact properties of dry granular flow impacting on the rigid wall, which is based on the laboratory experiment of granular flow impacting on low-angle retaining wall. Two impact mechanical models are proposed according to the accumulation characteristics of dead zone, the impact characteristics of flowing layer and their interaction characteristics. The results show that the slope and the angle of the retaining wall change the accumulation characteristics of the dead zone and change the impact direction and impact force of the flowing layer. The maximum normal impact resultant force (NIRF) can be estimated by the formula of static earth pressure when the slope is less than 40 degrees, since the flow layer impact the retaining wall indirectly at the moment of maximum impact force. With the increasing of slope, the kinetic energy of the flowing layer increases. At the moment of the maximum normal impact force, the flowing layer impacts the retaining wall directly. However, the dead zone has the buffer and deceleration effects on the flowing layer. This leads to decrease of direct impact force on the retaining wall. The load of dead zone on retaining wall mainly includes the direct impact force of the flowing layer on the dead zone along the slope, the shear friction force of the flowing layer on the dead zone and the static earth pressure of dead zone. The ratio of impact force of the dead zone on the retaining wall increases to 90% of the maximum NIRF. When the slope angles increase from 40 to 50 degrees, the ratio of shear friction force increases from 15% to 49% of the maximum NIRF. The friction coefficient between dead zone and flowing layer also changes from rolling friction coefficient to static friction coefficient. The shear friction force by the flowing layer onto the dead zone provides a new research idea for the estimating model of granular flow impacting on rigid retaining wall.

Desaturation is a mitigation method for sand liquefaction. Reducing saturation has been proposed to improve liquefaction resistance of saturated sandy foundation. Comparison of the centrifuge test numerical simulation was carried out, based on the modified single-phase fluid numerical method with simultaneous updating fluid modulus and the simplified method with fixed fluid modulus. The results show that the modified method is more reasonable because it considers the change of fluid modulus due to pore pressure change. The simplified method will underestimate the accumulation of pore pressure in the desaturated sand. Numerical simulation of liquefaction deformation on gently tilting desaturated sandy ground with different saturations and angles is carried out based on the modified single-phase fluid numerical method. The results show that the increasing velocity and peak value of excess pore pressure increase with the increase of saturation. The peak value of the excess pore pressure decreases by 20%-65%, and the peak value of acceleration response decreases significantly at the same depth with the saturation decreased from 100% to 96.4%. Along the foundation depth from 0.75 m to 9.00 m, lateral deformation decreases by about 20%-50%. It can conclude that the saturation reduction plays a positive role in reducing lateral deformation of liquefiable sandy foundation. The influence of saturation on lateral deformation is increasing evident with decrease of foundation depth.

Due to rainfall and water-level fluctuation, deformation increases on a large number of landslides , which create severe geologic hazards problems in Three Gorges reservoir area. The study of influence factors, deformation evolution rules and instability conditions of reservoir bank landslide is conducted by physical modeling test. Taking Huangtupo riverside slump-mass No.I as an example, a typical reservoir landslide in the Three Gorges Reservoir area, , a series of model tests of landslide was carried out by taking into account the induced factors such as water level fluctuation, rainfall and their combination. Test results show that the deformation mainly locates at the front edge of the model slope during water level fluctuation. The deformation acceleration stage occurs during the decline of water level, and the deformation rate is directly proportional to the water level decline rate, while deformation is relatively small with rising water level. Therefore, the landslide of Huangtupo riverside slump-mass No.I is a dynamic-water-pressure type. The model deformation shows obvious spatial and temporal zoning under rainfall. The deformation mainly occurs when the shallow part of the model is saturated. Spatially, large deformation is generated at the front and trailing edge of the model slope. The model slope fails from local slip failure to the leading edge failure under heavy rainfall and decline of water level, showing a progressive retrogressive mode. Finally, the experiment reveals that abnormal fluctuations of pore water pressure and earth pressure occur at the near-sliding stage, which can provide some reference for landslide warning and prediction.

Packing characteristics are the comprehensive embodiment of particle shape, size and other factors. It also reflects the mechanical properties of particle aggregates to an extent. Round gravels of river alluvial gravel soil with size range of 10-50 mm is used as research objects. By using the digital image technology and analyzing the morphology of large quantities of round gravels, aspect ratios of round gravels are taken as the shape parameters and the method to build the ellipsoid models is formed. Based on the above, the ellipsoid model database is established with the same distribution in shape and size as the actual round gravels, and the models are used to numerically simulate the packing tests of single set round gravels. By comparing with the physical packing experiments, the following conclusions are made: The morphological features of the round gravels are mainly reflected in the macroscopic contour shape, and the angular characteristics are not obvious. The numerical simulation results show that the void ratios of packing test in different size groups vary little. The coefficient of interparticle friction has significant effect on the void ratios. The void ratio increases rapidly when the friction coefficient is 0.0-0.4 and then becomes flat. The physical experiments results verify the validity of the ellipsoid model simulation, and the coefficient of interparticle friction under gravity is calibrated to be approximately 0.3.

For the construction safety of underground engineering and prediction & warning of disasters, the fracture mechanical behavior and acoustic emission (AE) characteristics of fractured rock masses under hydro-mechanical coupling condition are key concerning problems. This study carries out compression failure experiments with the cement mortar material and AE technique, and investigates the failure modes, mechanical properties and AE characteristics of specimens containing 3-D crack under dry and hydro-mechanical coupling condition. The influences of crack dip angle and water pressure on the specimen failure and mechanical behavior, as well as the relation of dominant frequency characteristics of AE signals with the specimen failure status are analyzed. The experimental results indicate that the failure mode of specimen changes from tensile-dominant mode to tensile-shear mixed mode and finally shear-dominant mode as the crack dip angle increases. High water pressure enhances the tensile failure mode of specimens, while the shear failure mode is weakened to some extent. With the increase of water pressure, the crack initiation stress, damage stress and the peak strength of specimens decrease continuously, and the influence of water pressure on the initial stress and peak strength exhibits a threshold effect (4 MPa). The AE hit rate during the failure process of specimens presents an evident stage characteristic, which provides a reference base for the adjudgment of crack initiation and damage stress. Additionally, the AE signals exhibit different dominant frequency characteristics at different failure statuses of specimens, and the dominant frequency ranges of relatively high, intermediate and low frequency signals are discontinuous. The proportions of relatively-low-frequency and relatively-high-frequency signals correspond to different failure statuses of specimens, which can provide references for the rock mass failure prediction and warning.

To investigate the hydration evolution process of sodium montmorillonite, the water vapor isothermal adsorption-desorption experiment was conducted for samples with relative humidity ( ) ranging from 0 to 0.98. The water adsorption rate curves and BET curves were proposed to define the hydration evolution process of sodium montmorillonite and the corresponding control factors. By measuring the variation of value, the influence of adsorption water on the thickness of clay mineral layer was studied for sodium montmorillonite hydration characteristics. Based on the theory of Fourier transform infrared spectroscopy, the rationality of the proposed method is also quantitatively validated with water molecular structure vibration information. The relationship between adsorption water characteristics of sodium montmorillonite and its hydration mechanism is explained in terms of the energy required by the phase transition of adsorption water and the mass loss of the adsorption water by thermal analysis. The test results indicate that sodium montmorillonite adsorbsion is mainly on the external surface to form surface adsorption water at 0 0.15, the stage of 0.15 0.40 is the cationic hydration stage between sodium montmorillonite layers, and the stage of 0.40 0.98 is the hydration stage on the inner and outer surface of the crystal layer. The water molecules gradually cover the montmorillonite completely, forming a multi-layer adsorption layer. The hydration process of sodium montmorillonite is controlled by interlayer sodium ions and crystal energy, and the hydration of interlayer sodium ions affects the starting sequence of hydration evolution of sodium montmorillonite.

Based on a large number of experimental data, the evolution law of strength parameters and dilation angle in hard rock elasto-plastic deformation and failure process is studied to establish their evolving models. Firstly, by analyzing the relationship between critical plastic strain and confining pressure of twenty-six types of hard rocks, a widely suitable three-parameter power-type confining pressure function is proposed and introduced into the definition of the plastic internal variable. The evolution law of strength parameters and dilation angle with the plastic internal variable in the plastic strain hardening-softening process is studied with respect to the Mohr-Coulomb criterion. The Gaussian model for describing the nonlinear evolution of strength parameters is then proposed, and the evolving model of dilation angle considering the influence of both confining stress and plastic internal variable is established. And then, considering that the pre-peak stress-strain relationship of hard rocks is usually simplified as linear in practical engineering, the strength parameters and dilation angle evolving models specifically in the post-peak process are studied. Finally, by comparing the simulated and experimental results of multiple sets of hard rocks, it is found that the models of strength parameters and dilation angle in the plastic strain hardening-softening process can correctly reproduce the axial and circumferential deformation under different confining pressures. The proposed evolving models can also reasonably describe the post-peak deformation in the plastic strain-softening process under different confining pressures. The proposed models are applicable and showing good prospect for different types of hard rocks.

Based on field prototype tests, the anti-sliding performance of single pile with steel-tube grouted for multiple times is studied. The steel-tube is analyzed in aspects of landslide thrust, soil pressure and bending moment of pile body, and the grouting effect and failure mode. The test results show that the root shaped anti-sliding bodies composited by cement column are formed by multiple sectional control grouting technology. This structure can effectively improve the shear strength of the soil around the pile, and enhance the anti-sliding performance of the steel-tube single pile structure effectively. Compared with the conventional steel pipe, the horizontal anti-sliding force of single pile is increased by 35 kN, and the horizontal bearing capacity is increased by about 50.72% compared with the conventional steel-tube, and better prevents landslide. The bending failure of the steel-tube occurs under the action of landslide thrust, and the maximum bending moment appears near the slip surface. It is suggested that the reinforcement ratio of the piles near the sliding surface should be properly enhanced to improve the anti-sliding ability.

To overcome the difficulty of deep buried tunnel microseismic source positioning and low positioning accuracy problems, a heuristic algorithm, gravitational search algorithm (GSA), is used to in this paper to search the location. The algorithm is compared with results of the particle swarm optimization and the simplex algorithm. Research shows that, under the two and three velocity models, the gravitational search algorithm has the advantages of fast convergence and high precision over the particle swarm optimization and the simplex algorithm. The distance from the source location can be controlled within 10 m, which meets the need of tunnel construction better. For the two velocity model, the accuracy of gravitational search algorithm is 83.71 percent higher than the simplex, and 7.77 percent higher than the particle swarm optimization. For the three velocity model, the accuracy of gravitational search algorithm is 70.67 percent higher than the simplex, and 39.36 percent higher than the particle swarm optimization. Through the analysis, gravitational search algorithm (GSA) can provide a new idea for the position of micro-seismic source in deep buried tunnels.

The penetration effect of the rectangular pile is different from the conventional circular cross-section pile. However, conventional theoretical models and experimental techniques are not suitable for investigating the rectangular cross-section pile. Based on deformation visualization technique of transparent soil, this paper investigated the displacement field variation law induced by the rectangular pile. Experimental results show that after driving of rectangular pile, the soil displacement of the surrounding soil after installing a rectangular pile can be divided into two typical areas. A transition area is produced around the rectangular pile, where soil displacement is non-columnar symmetry characteristics, and its radius is around 4deq-5deq. The expansion area of the circular hole is located away from the pile, and its displacement presents the characteristics of cylindrical symmetry. According to pile driving compaction tests of the rectangular cross-section pile, this study proposed a modified cavity expansion mode for rectangular pile and verified the rationality of the modified model by comparing theoretical values with experimental data.

Municipal sludge production in China is huge, and will seriously affect the comprehensive management of the environment if not handled properly. To investigate an efficient, resourceful and stable municipal sludge treatment technology, in this study, the municipal sludge is first subjected to quicklime digestion treatment, and then orthogonal test. Taking the unconfined compressive strength as the index, the optimum ratio of raw lime, soil, municipal sludge and curing agent are selected. Using this ratio, a new type of municipal sludge solidified soil is prepared and its mechanical properties are discussed. The test results show that the new municipal sludge curing agent has a good effect on the treatment of municipal sludge and the heavy metals leaching amount meets the national standard requirements. The strength of sludge solidified soil increases with the increase of curing age, and it tends to be relatively stable after 28 days. In addition, alkali has a promoting effect on the strength, and it is suggested that the initial formulation of the solidified soil is controlled to be between 45% and 50%. Furthermore, in the triaxial shear test, the stress peak and structural yield stress increase with the increase of confining pressure and curing age. In the ring shear test, the residual strength is proportional to the curing age and effective normal stress, and inversely proportional to the shear displacement.

Several investigations after earthquake indicate that areas where tunnels cross through faults are damaged severely. For a tunnel running through fault, an anti-seismic design concept is established and a joint of segmental lining is proposed in this paper, based on characteristics of ground motion energy transmission and release. A shaking table test was carried out to study the response of lining structures with joints under single fault movement loading mode, and combined action of normal fault rupture and subsequent seismic shaking based on the Longxi tunnel engineering. The test results show that effect of earthquake wave on the tunnel can not to be ignored, and the loading mode of combined normal fault rupture and subsequent seismic shaking was reasonable. The new joint designed segmental linings could alleviate the tunnel structure damage by self-adaptive deformation, and adjust the longitudinal deformation mode of the tunnel structure to improve the aseismic capability of the whole tunnel lining. Simultaneously the joints could mitigate the circumferential failure of lining, and weaken the seismic force transfer between segments to realize the local damage for tunnel lining. The damage influence length of the tunnel lining located in the hanging wall is 1.8 times of the tunnel span, while the damage influence length of the tunnel lining at the footwall is 1.2 times of the tunnel span. The tunnel damage of the hanging wall is mainly caused by the combined action of the normal fault movement and seismic shaking, while the damage of the lining is mainly affected by the ground motion at footwall.

To explore the influence of different mining sequences on the fault stability, this paper put forward the hypothesis of stress rotation based on pressure arch theory. Firstly, a numerical model was set up to simulate the process of working face passing through the fault from the hanging wall and footwall to the fault. And then it analyzed the stress state and evolution law of contact surface, and verified the stress rotation induced by mining. Finally, comparative analysis of the fault damage variables and their growth rates was carried in this paper. The results show that during the propulsion process of working face from different directions of hanging wall and footwall of fault, the distance between start position of principal stress rotation in roof and fault is 120 m and 40 m, respectively. The maximum value of principal stress deflection angle when the working face located at footwall is 1.68 times than the value when located at hanging wall. The starting slip points of fault damage variables are 130 m and 40 m from fault, respectively. The fine correspondence between them indicates that stress rotation is significantly related to fault slip instability. The caption of additional stress induced by stress rotation can well explain the difference of fault stability due to different mining sequences, and provide new ideas for the design of fault coal pillar and disaster prevention measures.

To extract accurately rock discontinuities from the borehole images and measure automatically the corresponding geometric information, an intelligent recognition method based on the edge detection algorithm was proposed in this study. Firstly, the feature parameters related to gray level co-occurrence matrix were calculated to locate the rock discontinuities areas in the long borehole images, and Canny edge detector was employed to identify the rock discontinuities edge in these located rock discontinuities areas. Secondly the upper and lower edges of rock discontinuities were extracted through the edge-edge connection and appropriate threshold selection. Thirdly, the rock discontinuities edge trigonometric function was fitted to obtain the sinusoidal curve of the upper and lower edges of the rock discontinuities, and several geometric parameters of the rock discontinuities, such as the dip direction, dip angle and aperture were computed through spatial geometric theory. In the case study, images of borehole wall were recorded in two boreholes in the dam site of Rumei hydropower station, and the geometric data of rock discontinuities derived from the proposed method matched the data obtained from other existing approaches, verifying the feasibility of this method.

The mechanical properties and design-calculation methods of pile-net composite foundation were analyzed, with the condition of pile’s end into and not into the hard soil layer. A mathematical model for the optimal design of pile-net composite foundation was established by taking pile diameter, pile length, pile cap length, pile cap thickness and pile spacing as design variables. The foundation bearing capacity, settlement, tensile strength of reinforced geotextile, bending capacity and shearing capacity of pile cap were constraint conditions, and engineering cost was the objective function to be optimized. By introducing the perturbation weighting and reducing parasitic operations, the perturbation-weighted symbiotic organisms search (PWSOS) algorithm was proposed to increase the convergence speed and accuracy of the symbiotic organisms search algorithm. Based on the perturbation-weighted symbiotic organisms search, the global optimization of the mathematical model of the pile-net composite foundation was carried out, and the optimal reinforcement scheme that meet the requirements was the output. The results show that the optimal design scheme has better economic benefit and the perturbation-weighted symbiotic organisms search algorithm has better performance.

The determination of deep rock stresses is a focused and difficult issue in the study of regional stress field. Most estimating methods for deep rock stresses depend on the extrapolation of fitting empirical formula, which are short of theoretical basis and unreliable. This research proposed the lateral pressure coefficient polygon with reference to stress polygon theory, combined with stress factor R to better constrain stress distribution polygon. The Zijingguan fault belt in Yi county, Hebei province was chosen as the application area. Based on focal mechanism and in-situ stress measurement data in this area, , the maximum lateral pressure coefficient, and , the minimum lateral pressure coefficient, were estimated for the deep rock masses of 6 km, 11 km and 19 km respectively. The results showed that at three depths are 1.07 0.07, 1.14 0.14, 1.09 0.09 and are 0.85 0.15, 0.88 0.12, 0.86 0.14 respectively. The relative deviation of results is between 6% and 17%. The orientation of the maximum horizontal principal stress, jointly determined by hydraulic fracturing and focal mechanism methods, is N44.4°E. The method proposed and established in this research provides a new estimate means for deep rock stresses.

Determining reasonable spacing between two adjacent piles is one of the important contents in the anti-slide pile design. In this study, the anti-sliding ability of pile is mainly combined with two parts: the anti-sliding ability from the pile back face and ability from the pile side. Assuming the anti-sliding ability of the pile is putting into full play, a new spacing calculation formula between two adjacent anti-slide piles has been presented based on Mohr-Coulomb strength criterion. The landslide thrust of the present method is balanced by the pile-soil shear friction on pile side surface, and the soil-soil shear friction within the ultimate shearing thickness of soil arching on the pile back face. Both the pile-soil and the soil-soil shear frictions are simply assumed to have linear distributions along its shearing thickness. Based on the proposed method, a parameter study has been conducted to investigate the influence of multiply factors, such as cohesion of landslide soil, slide force and cross section size of plie on the pile spacing. The results show that the anti-sliding ability from the pile side, which mainly controlled by the pile side width, plays an important proportion in the whole anti-sliding ability of the pile. Moreover, the maximum spacing between adjacent anti-slide piles is significantly influenced by the cohesion and internal friction angle of landslide soil and the width of piles.

The statistical uncertainty characteristics of random parameters in geotechnical engineering makes the reliability design have certain risks. The robust geotechnical design (RGD) method can minimize the effects of variation in the geotechnical parameters on the system response, by explicitly considering robustness in the design process along with safety and economic requirements. Aiming at the influence of the statistical uncertainty characteristics of random parameters on the independent foundation design under the column, the foundation geometry is regarded as controllable design parameters to conducting design analysis. Analysis considers the influence of the statistical uncertainty of the mechanical parameters of rock and soil, concrete and steel material and based on reliability theory and robust design method of geotechnical engineering. In this paper, the robustness design of the structural system under multiple failure modes is carried out, and the relationship between structural geometric parameters and structural system reliability under multiple failure modes is analyzed by viewing four failure modes: independent foundation bearing capacity, foundation deformation, infrastructure punching shear failure and foundation bending failure as series system. Finally, the optimal solution of the design of independent foundation under column is determined combining with robustness and economic efficiency.

Discontinuous deformation analysis (DDA) is a new numerical method parallel to the finite element method. Based on the principle of minimum potential energy, the method unifies the deformation, motion and contact of each discrete block into the equilibrium equation for implicit solution. However, the traditional DDA method requires assembling the whole stiffness matrix to solve the equations in parallel, which occupies a large amount of memory, takes a long time and has a very low computational efficiency in the three-dimensional numerical simulation of large geotechnical engineering problems. Therefore, a three-dimensional spherical DDA method based on explicit time integration is proposed. This method does not require assembling the whole stiffness matrix. Because of the mass matrix is diagonal matrix, it can be stored as a one-dimensional vector with less memory, and can be solved block by block with higher efficiency in the acceleration solving process. The maximum displacement criterion is used to simplify the contact algorithm in the contact judgment, and the smaller time step ensures the accuracy of calculation. Finally, the accuracy and efficiency of the method are verified by several classical examples.

Fracture grouting is an effective method for soil reinforcement, yet the theoretical studies are far behind engineering practice. Using secondary-developed ABAQUS, a finite element model of fracture grouting is proposed based on the smeared crack model and VOF (volume of fluid) method. Results of FEM analysis are in good agreement with the laboratory tests. The effects of grouting depth and injection rate on the grout vein shapes are then investigated. The results show that fracture process can be divided into two stages: fracture initiation and extension. When the grout veins reach model boundary, higher grouting pressure is needed to increase the width of grout veins. With the grouting depth increases, the grout veins have less branches, reducing length and increasing width. The main factors affecting the shape of grout veins change from randomness of soil properties to the difference between major and minor principle stress. For given injection volume, higher injection rate results in shorter and thicker grout veins and higher grouting pressure. The research provides theoretical support for engineering applications of fracture grouting.

Using the newly developed cement mortar materials, horizontal natural fissures and inclined precast flaws were arranged in the testing specimens, and different tests were carried out under uniaxial and biaxial compression conditions. Crack propagation and penetration process under different loading conditions were analyzed in detail. The test results show that: the propagation path and direction of pre-cracks change significantly after the pre-cracks meet the natural fissures during the coalescence process, and the phenomenon of ‘displacement jump’ appears. The increase of lateral confining stress inhibits the initiation of wing cracks and secondary cracks on the inclined crack tips. Under high lateral confining stress, these cracks do not penetrate through natural fissures, and the failure mode changes from splitting to shear failure. Under the action of hydraulic coupling, the maximum principal stress distribution of the wing crack tip in uniaxial compression gradually decreases with increase of internal water pressure, and the initiation stress, initiation angle and peak strength are also gradually reduced. The crack propagation and penetration law obtained by numerical simulation is in good agreement with test results. The numerical analysis of inclined flaws with different internal water pressures extending through the waterless pressure horizontal fissures is carried out. With the increase of the water pressure, the inclined flaws firstly initiate the coplanar cracks and then generates the wing crack. When the wing crack propagates through the horizontal fissures, the expansion path also changes. Under the action of hydraulic coupling, the lateral pressure also suppresses the initiation of the wing crack. The increase of water pressure weakens the inhibition of lateral pressure on wing cracks.

The maximum tangential strain criterion is selected to determine the fracture initiation and propagation, and the incremental crack growth method is proposed to simulate the hydraulic fracturing propagation process. Numerical results show that the proposed method can effectively simulate the path of fracture propagation during hydraulic fracturing. By using the proposed crack growth method, we discussed the effects of the perforation length, the perforation angle, the stress anisotropy coefficient, Poisson’s ratio and the water injection pressure on the initial critical water pressure, the fracture path and the deflection angle. According to the initial critical water pressure and the fracture propagation path, the suitable perforation geometry can be optimized. It leads to a decrease in the initial critical water pressure but an increase in the contact area with the formation by decreasing the stress anisotropy coefficient. Poisson’s ratio has significant influence on fracture path under the critical water pressure conditions but does not have an effect on the fracture initiation angle under high water injection pressure. The sensitivity analysis indicates that the perforation angle is a critical factor to affect the fracturing. The numerical model provides a comprehensive understanding of the characteristics of hydraulic fracturing under complex loading conditions. The results also provide a basis for quantitative investigations of the engineering design of hydraulic fracturing treatments.