The yield criterion of materials in geotechnical engineering is generally presented on the basis of the compression-shear failure type. However, brittle hard rock has three types of failure, i.e., tensile failure, tension-shear failure and compression-shear failure. With increasing the depth of rock engineering, the tension-shear failure turns into the main failure mechanism of surrounding rock. Tension-shear tests are conducted on brittle hard marble to investigate the characteristics of tension-shear failure. By analyzing tension-shear and compression-shear testing results, a modified Mohr-Coulomb criterion is established, which can consider the effect of tension-shear failure and stress state. It is noted that the failure of brittle hard marble is in tension-shear at the tension-shear stress state and at the compression-shear stress state with a low normal axial stress, but it shows compression-shear slip characteristic at the compression-shear state with a high normal stress. The failure of marble is dominated by obvious tension cracks at tension-shear stress state, while no obvious shear crack appears. The shear displacement increases with the increase of axial tensile stress when the shear stress is fixed. The cohesion and friction angle are influenced by the stress state. The cohesion decreases initially and then increases with increasing the normal stress, while it is opposite in the friction angle. The variations of the cohesion and friction angle with the normal stress can be divided into four stages: tension-shear, low compression stress, medium compression stress and high compression stress. The relationships between cohesion, friction angle of each stage and normal stress are linear, and the friction angle reaches up to 90 degrees while the cohesion approaches infinitely to tensile strength. By considering the effect of tensile shear failure and stress state, the Mohr-Coulomb criterion curve can be divided into two stages, the tension-shear stage which is fitted with a second-degree parabola and the compression-shear stage considering the development of cohesion and friction angle with the normal stress. It is demonstrated that the proposed Mohr-Coulomb criterion is comprehensive and precise.

Rheology has extremely significant influence on the long-term deformation and stability of engineering rock mass. Current experimental research on unloading rheology only considers the stress path under unloading confining pressure with the constant axial stress, however it is different from the process of stress adjustment in surrounding rock during tunnel excavating. Stepwise unloading rheological tests are conducted on sandy mudstone samples which are under unloading confining pressure with the axial compression or with the constant axial stress. The results show that the rheological deformation trends of rock samples under both unloading confining pressure are generally consistent. At the same initial confining pressure, the failure confining pressure under axial compression and the deviatoric stress are relatively higher than that under constant axial stress, but the ratio of long-term strength to failure stress is opposite. When the confining pressure unloads reach up to the point where the rock sample approaches failure, the growth rate of rheological strain under axial compression obviously becomes faster than that under constant axial stress, and the failure occurs more suddenly. The failure modes of rock samples under constant axial stress usually appear as one complete shear failure plane. By contrast, the failure modes of rock samples under axial compression are more complicated with a certain amount of secondary shear cracks and tension cracks, in addition to the controlled shear failure plane. Moreover, the secondary cracks increase with the increase of initial confining pressure. Therefore, the unloading confining pressure with the constant axial stress seems difficultly satisfy the requirement of engineering projects undergone a long-term deformation and stability of surrounding rock. The stress paths of tests should be conducted in more detail to simulate the actual stress change process of engineering rock mass. The results provide insights into the research on long-term deformation and stability of underground surrounding rock.

By the analytical solution of the average excess pore water pressures in a composite ground with a floating impervious pile, an analytical solution is obtained for the consolidation of surrounding and underlying soils under variable loading. The validity of the proposed solution is illustrated by comparing the present solutions to the FEM results. The consolidation behavior of the surrounding and underlying soils is investigated using the current solutions. The results indicate that the consolidation rate of the surrounding soil is considerably larger than that of the underlying soil, and it increases with the increase of area replacement ratio and constrained modulus ratio of a pile to surrounding soil. The consolidation rate of the underlying soil is hardly influenced by the constrained modulus ratio and area replacement ratio when the area replacement ratio is located within the range recommended in engineering design, and penetration ratio of a pile is not large enough. The consolidation rate increases with the increase of the constrained modulus and area replacement ratios. There exist some critical values of the constrained modulus ratio and area replacement ratio, beyond which consolidation rates of the surrounding and underlying soils are insensitive to the variations of these two parameters. The consolidation rates of the surrounding and the underlying soils always increase with the increase of loading rate.

As one of the series large-scale shaking table tests on the failure mechanism of subway station structure, a large-scale shaking table test on frame type subway station structure model in soft silt clay under near- and far-field strong ground motions is conducted, in which the acceleration, pore pressure and earthquake-induced settlement of the model ground as well as the acceleration, strain and horizontal displacement of the model structure are measured and analyzed. The results show that as the seismic wave propagates in the model ground, the amplitude of the low frequency components increases while the amplitude of the high frequency components decreases from the bottom to the top of the ground. Under strong ground motions, the fundamental frequency of the model ground significantly decreases, and the ground exhibits a remarkable amplification effect for low frequency components and wave filtering effect for high frequency components. The pore pressure ratio of the model ground slightly increases. The development processes of pore pressure ratios under different ground motions are greatly different, and show an obvious spatial effect. Under near- and far-field strong ground motions, the acceleration responses of the model structure are obviously different. The model structure has a significant spatial effect on the acceleration response of the ground motion for the soft ground. The frame structure does not show an obvious floating phenomenon and its relative deformation is small. The frequency characteristics of the ground motion have an obvious influence on the deformation mode and magnitude of the side wall for the model structure. The most severely seismic damage part is located at the middle column of the model structure. The whole model structure is slightly seismic damaged and in nondestructive state.

The shape of grains is one of the main factors influencing the compaction, mechanical and hydraulic properties of breakstone material. Two sets of limestone gravel with diameters of 2-5 mm and 5-10 mm were examined using the image measuring instrument and special fixture, from which the grain appearance images were obtained at different rotation angles. The geometric sizes of grains were measured with graphic software. The common grain shape parameters were calculated from each image at different rotation angles. The mean values of each parameter at different rotation angles for each grain were analyzed, so that the influences of human factors on measured value at just one section were avoided on grain shape evaluation. All the grain shape parameters of limestone gravel show significant differences compared to spherical grains, and the differences of the set with larger size are more obvious. The grain shape parameters of the two set samples can be characterized by skew distribution. The parameters of length-width ratio, flatness and sphericity can more sensitively reveal the differences. Compared to other parameters, it is more straightforward to determine the length-width ratio and sphericity.

Considering the matrix suction and seepage force of unsaturated clay slope, the limit equilibrium state is discussed according to the sliding surface over or under the water line. Then, the force and moment balance equations are established in the limit equilibrium state of the sliding slope respectively. The formulation of the limit equilibrium slice method is obtained for calculating the unsaturated clay slope stability under rainfall infiltration. The related parameters of the equation are obtained through test, parameter transformation and the position relationship of the sliding force. Then, the interaction force coefficients and the safety factor of the unsaturated clay slope are determined in the limit equilibrium state with numerical calculation method. The case study results show that the safety factor of the unsaturated clay slope is lower about 13.8% considering seepage force than without considering the seepage force, and the change rate of coefficient between slices considering seepage force is obviously larger than that of coefficient between slices without considering the seepage force. The instability time of slope is approximately the same when the rainfall intensity exceeds a threshold with the formation of slope runoff soon.

To describe the time-dependent, over-consolidated and structural characteristics of natural soft clay, a simple elasto-viscoplastic constitutive model including the influence of structural behavior of clay is proposed. In the new model, the Asaoka's superloading and Hashiguchi's subloading yield surface is used as a reference yield surface at a specified strain rate. A relative overstress relation is used to obtain the dynamic loading surface at any strain rate based on the reference yield surface. A dynamic equilibrium condition, using the current stress, viscoplastic strain and viscoplastic strain rate as state variables, is adopted as a convergence criterion. Two material parameters and about the rate sensitivity are added into the new model, which can be determined by triaxial compression tests with different strain rates. The new model has been implemented into ABAQUS by a stress integration algorithm using Newton-Raphson iteration. The validity of the model and the reliability of the stress integration algorithm are verified by numerical calculation. The numerical results show that the model can describe the time-dependent behavior of clay such as rate sensitivity, structural characteristics and creep. The material parameters are easy to understand and measure with clear physical meaning. Furthermore, the prediction results of the model are in good agreement with experimental data. The model can be used in a finite element calculation under the complex boundary value problem.

Mechanical properties of heat-treated rock after the cooling treatment directly influence the exploration and development of underground space and resources, storage of nuclear waste and stability evaluation of underground engineering after the high temperature disaster. Uniaxial compression and acoustic wave tests were conducted on heat-treated marble samples under the conditions of natural cooling and water cooling. Then, the variations of peak strength, elastic modulus, attenuation coefficient, longitudinal wave velocity and dominant frequency under different conditions were analyzed. A decrease trend is noted in peak strength, elastic modulus and longitudinal wave velocity of heat-treated marble by water cooling with the increase of treated temperature. When the temperature is lower than 400 ℃, the attenuation coefficient gradually increases and the dominant frequency gradually declines with the increase of temperature. When the temperature is higher than 400 ℃, the attenuation coefficient and dominant frequency do not completely exhibit monotonically increasing or decreasing trend, though there is a turning point. The peak strength, elastic modulus and dominant frequency of heat-treated marble samples after water cooling are lower than those after natural cooling, but the longitudinal wave velocity and attenuation coefficient are opposite. These results can provide references for the engineering health monitoring and stability evaluation of cooling methods on heat-treated rocks.

Model experiments are conducted on polymethyl methacrylate (PMMA) materials to investigate the influence of empty holes with different shapes on directional fracture controlled blasting of rock. A new digital laser dynamic caustics is applied to obtain caustics patterns around empty holes. Experimental results show that the preset rhombic hole can achieve fine directional fracture controlled blasting and effectively ensure the forming effect of blasting around the roadway. By comparing the propagation velocity of blast-induced cracks in specimens containing empty holes with three different shapes, it is found that the velocity of blast-induced main crack in a specimen with a rhombic hole is the greatest, then the velocity in a specimen with a circular hole is followed, and the velocity in a specimen with a circular hole with groove is the lowest. The values of dynamic stress intensity factor (DSIF) at the tips of blast-induced main cracks in specimens with a circular hole are generally higher than that in specimens with a rhombic hole and a circular hole with groove. The values of DSIF in specimens with rhombic holes are relatively low during the middle and late stages of blast-induced main cracks propagation.

Large-scale cyclic triaxial tests are conducted on remoulded kaolin specimens with 300 mm diameter and 600 mm height. To achieve radial drainage during and after the cyclic tests, a single prefabricated vertical drain (PVD) is installed in the centre of the soil cylinder. The test results verify that radial drainage can effectively dissipate the pore water pressure induced by cyclic loads. A radial consolidation model under cyclic loading is proposed to capture the behavior of soft clays subjected to cyclic loading when radial drainage is allowed during the loading and rest periods. It is achieved by combining radial consolidation theory with an undrained cyclic loading model. The effects of the stress history and dissipation of pore water pressure on the generation of pore water pressure are considered in the proposed model, which is verified by large-scale cyclic triaxial tests. It is found that for a high cyclic stress load, the radial drainage decreases the cumulative rate of pore water pressure up to the critical value, so the soil could undergo more loading cycles prior to failure. With a moderate cyclic stress load, the radial drainage prevents the pore water pressure from accumulating to the critical value. Furthermore, the effect of rest period on the generation of pore water pressure is investigated. It is indicated that the cumulative pore water pressure begins to decrease after three sets of cyclic loading, showing that no substantial pore water pressure will be observed if more sets of cyclic loading are applied.

Generally, rock strength criteria are obtained by the triaxial compression tests at the low confining pressure. However, the strength in these criteria tends to be higher than experimental data at the high confining pressure. To overcome this deviation, we propose a concept of rock critical state confining pressure. It can be deduced that deviator stress is constant and critical rupture angle is 45 degrees at critical state using this concept. It is then introduced to modify Mohr-Coulomb and Hoek-Brown strength criteria. The accuracy and applicability of these two modified criteria are verified by a series of triaxial experimental tests. It is noted that the parameters in modified criteria have lower sensitivity to the confining pressure than those in the original criteria. The modified criteria can not only maintain the consistency and accuracy in strength assessment at the low confining pressure, but also assess the consistency of experimental strength at high confining pressure. Thus, the modified criteria can be used to accurately estimate the strength of rock at the high confining pressure.

The stress-strain, resistivity and temperature variations of backfill materials in the whole process of uniaxial compression are studied using a self-designed synchronous monitoring system, which is composed of the loading control system, strain measurement system, the resistivity collection system and the thermal infrared (TIR) observation unit. The precursory anomaly characteristics of the stress-strain, resistivity and TIR are observed before backfill body failure. The sensitivity and difference of different monitoring systems to the same failure event are also compared. The results show that the time-space evolution process of the resistivity and TIR is quite close to the whole process from compression deformation to failure of backfill body, showing apparent staged characteristics. The precursory information point of resistivity occurs prior to the TIR and the stress points. The resistivity variation has an apparently anti-symmetry feature in the process of compression, compared to the stress-strain relation and temperature variation. The variation of resistivity can give a detailed description of the internal structure change at every phase during compression. However, the TIR information mainly represents the surface temperature evolution during the process of compression before plastic yielding of the backfill. The multiparameter methodology used to evaluate the precursory information during loading process can overcome the disadvantages of early-warning approach, say low reliability and inaccuracy. The proposed method is an effective one in determining the stable state of backfill body.

Based on true triaxial simulation experimental system, servo control system of hydraulic fracturing equipment and tracer technique, a large-scale true triaxial physical simulation test on hydraulic fracturing of shale horizontal well is conducted. The extension of the hydraulic fractures is analyzed by dynamically monitoring of fractures and sectioning the fractured specimen. The formation mechanism of fracture network is discussed. The results indicate that: (1) A relatively complex fracture morphology is formed due to the hydraulic fractures extending along and perpendicular to the bedding planes after initiation. (2) The fracture network which contains longitudinal and transverse fractures is formed in hydraulic fracturing of shale specimens. Main hydraulic fractures perpendicular to the bedding plane and secondary fractures extending along the week bedding plane all exist in the fracture network. (3) Hydraulic fractures will gradually turn to the direction perpendicular to the minimum principal stress in extension. (4) Arresting, branching, penetrating and re-orientation of hydraulic fractures in bedding planes are the main reason for the formation of complex fracture network of shale gas reservoir. The existance of the weak structural planes is the foundation of the forming of complex fracture morphology. The results can provide technical supports for staged exploitation of horizontal well of shale gas reservoir.

The subgrade bed layer is the core part of railway subgrade composed of coarse-grained soil (CGS), which is generally about 2.5-3.0 m thick. It is this layer that is directly subjected to long-term effect of repeated traffic loading. The deformation of the subgrade bed layer under cyclic dynamic loading is one of the key factors to evaluate the operating performance of subgrade. To study the stress-strain characteristics of CGS under cyclic loading, a series of large-scale dynamic triaxial tests is carried out at different confining pressures (simulating different depths under the effect of lateral pressure) and dynamic stress amplitudes (simulating different axle loads), to simulate dynamic loading of train and actual situation of CGS filling by one-way cyclic stress-controlled loading. The results show that, under the cyclic loading, the change for soil stiffness is closely related to the vibration number and confining pressure. According to different dynamic stress amplitudes, the dynamic strain of saturated CGS under cyclic loading can be divided into three types according to development trends of vibration times: stable, destructive and critical. Based on the experimental results, a backbone curve model is developed including confining pressure and vibration times. Compared with the traditional backbone curve model, the proposed model can reflect the variation of soil stiffness with vibration times and the actual situation of cyclic effect of train. In addition, the model can also be used to determine the dynamic strength of subgrade soil, providing reference for evaluating dynamic stress deformation stability for subgrade bed core layer of railway and subgrade bed design based on dynamic deformation control.

To understand the deformation and mechanical behavior of soil-rock aggregate mixture, a series of large-scale triaxial tests is conducted under different stress paths (conventional, constant stress ratio and constant p paths). The experimental results indicate that the strain softening of soil-rock aggregate mixture is insignificant under the low-confining pressure condition; the soil cohesion is generally low and the internal friction angle is relatively high. The internal friction angle and deformation modulus increase as the rock content increases from 25% to 70%. The stress-strain curves of soil-rock aggregate mixture are similar under the conventional stress path and the constant p path. Their shear strength indices are also similar. However, the form and development process of stress-strain curves under constant stress ratio path are significantly different. The specimen presents dilatation under low confining pressure and shrinkage at high confining pressure under conventional stress path. The higher the rock content, the more significant the dilatation under low confining pressure is. Conversely, the more significant the shear shrinkage under high confining pressure is. The relationship of volumetric strain and axial strain is linear under constant stress ratio path, and its slope decreases with the increase of stress ratio (q/p). The soil-rock aggregate mixture generally exhibits dilatation in constant p path tests, which is especially significant under low average principal stress.

To further understand the failure mechanism of water inrush from mining face floor, the water inrush failure mechanism of mining floor is analyzed using the Mohr diagram and fracture mechanics theory. By considering the effect of saturated pressure on the branching crack tip, an analytical formula for the plastic fractured region and a damage threshold of rock mass failure are established with damage fracture mechanics and the unified strength theory. An equation of damage fracture energy is determined by coupling crack propagation and rock damage, and the factors influencing the damage fracture energy are analyzed in depth. The results show that the stress intensity factor (SIF) of the crack tip at the complete unloading state of the maximum principal stress is higher than that at the biaxial stress state with a confining pressure coefficient 0.5, in which the mining floor has been found to be more prone to failure. By taking into account the influence of the plastic fractured region of the branching crack tip, it is noted that the crack damage fracture energy tends to be higher and water inrush risk of mining floor is more obvious. There is a positive correlation between and the crack half-length a, the ratio of crack connected area to the crack total area , crack seepage pressure and the minimum principal stress . Whereas it shows a negative correlation with the friction coefficient of crack surface and modulus of rock mass . The results provide some insights into the prediction of water inrush failure mechanism of mining floor.

Disastrous failures of shallow residual soil slopes, which result in great loss of lives and properties, are mostly attributed to heavy rainfall. To evaluate the stability of colluvial landslide during rainfall, a new analytical infiltration model for slope is derived from the traditional Green-Ampt (GA) model. The new model assumes the pore water flow parallel to slope surface in saturated layer above the wetting front, and overcomes the deficiency of GA model which is only applicable to the infinite slopes. In addition, a new method is proposed to calculate the safety factor for soil slice based on the shear strength criterion related to the depth of saturated layer, and in this method the safety factor is deduced by taking the seepage force into account. The results show that the size effect of rainfall infiltration is significant, and the movement rate of wetting front increases with the increase of slope length. The proposed infiltration model is identical with GA model in analyzing infinite slope, demonstrating that the GA model is a special case of the proposed model. It is shown that the stability of slope decreases sharply at the initial time of rainfall, but the rate of decrease slows as rainfall continues. The predicted results of the new infiltration model show a high consistency with that of slope model test, showing the reliability of the proposed method.

The changes of permeability coefficient and pore structure of clays with high liquid limit and low liquid limit after drying-wetting cycles are compared. For each clay three densities are examined. The results show that after three drying-wetting cycles a larger increase in permeability coefficient is observed in clay with high degree of compaction than that with low degree of compaction when the liquid limits are the same. At the same dry density, the clay with high liquid limit exhibits larger increase in permeability coefficient than low liquid limit clay after drying-wetting circles. During the drying-wetting process, the influence of pore structural damage on permeability coefficient increases as compactness and liquid limit increase. The influence of cracks on permeability coefficient, however, decreases when the compactness increases and the liquid limit decreases. In contrast to low liquid limit clay which does not crack but shrinks after drying-wetting cycles, high liquid limit clay forms significant cracks. The change in permeability coefficient of the samples of the small size is mainly caused by the pore structure, and cannot reflect the influence of cracks. It is suggested that further research on the difference between permeability coefficients of clays with different liquid limits under drying-wetting cycles be carried out through permeability test on site or indoor large-size permeability test.

The time-effect characteristics of shear displacement of soil under different shear stress levels are analyzed. Four states of soil deformation are identified as the shear stress level increases, namely, fast steady, slowly steady, slowly breaking and fast breaking states. Based on the creep direct shear test of silty clay with different degrees of compaction, these four states are investigated. The test data are fitted by a power function which serves as the classification criterion of deformation state. Based on the shapes of shear stress level-power curve, three shear stress level thresholds corresponding to these four states are obtained. An expression between the state strength parameter and the ultimate strength parameter is developed based on the Coulomb strength equation. The post-construction settlement of subgrade in high-speed railway is of a magnitude of millimetre. It is found that the choice of embankment fill material and the compaction standard must meet the control principle and technology condition of fast steady state. The research results provide theoretical and experimental bases for controlling deformation states of soil structures in high-speed railway subgrade.

In order to investigate the law of water and salt migration, understand the deformation characteristics of saline frozen soil, and explore the mechanism of soil salinization in cold and arid regions, a model test of silty clay with different salinities is carried out. Results indicate that the temperature, water, salt and deformation of the saline frozen soil are coupling with each other. The temperature reduction is advantageous to precipitate the salt crystals and freeze the unfrozen water, while the temperature rise is liable to dissolve the salt crystals and thaw the ice. Heat release or absorption during the phase change of salt and water influences the soil temperature. Salt influences the water content significantly by changing the dynamic viscosity of the fluid and freezing temperature the soil, and generating a great suction when salt crystals precipitate from the supersaturated solution. Salt migrating with unfrozen water in ionic form means that water is the medium for salt migration. Salt and water are moving upward to the surface during cooling period, while the opposite direction is observed during warming period. The migration rate is related to the soil suction, and the greatest soil suction in the freezing fringe results in the fastest migration velocity. The deformation of saline frozen soil is the result of the combined effect of salt and water. Frost heave and thaw settlement caused by water freezing and thawing are the mainly deformation of saline frozen soil with low salinity, while salt heave and dissolve settlement caused by salt crystallization and dissolution play the dominant role with high salinity.

The strength development of solidified soil would be limited by the incompletely hydrated cement particles for the gradually reduced Ca2+ concentrations as the hydration reaction proceeds. Theoretically the unhydrated cement particles would be catalyzed by the active substances in cementitious capillary crystalline waterproofing (CCCW) materials. Using sulfur aluminate cement (SAC) as cementitious material, CCCW as additive, a series of experiments was conducted in a single-mixed or admixed way. The properties of solidified soil including unconfined compressive strength (UCS), water stability, the ability resisting wetting-drying cycles and microstructure were analyzed with X-ray diffraction (XRD) and scanning electron microscope (SEM) characterization. The results show that the UCS of soil solidified by 16% mixture (4% CCCW+12% SAC) is 1.5 times the soil solidified by the same content of SAC and 1.41 MPa higher than the soil solidified by 20% SAC. The average softening coefficient of the soil solidified by 16% mixture (4% CCCW+12% SAC) reaches 0.97 after 2-8 days soaking in water while only 0.73 of the soil solidified by SAC. The UCS of soil solidified with single-mixed gradually decreases with the increase of wetting-drying cycles while wavy development of solidified soil with admixed. The generation amount of AFt increases and the microfractures are repaired in solidified soil by the active substances in CCCW. Two or more soil particles are connected by AFt and a three-dimensional net structure is formed with a significantly increased aspect ratio. It is shown that the properties of soil solidified by SAC including the UCS, water stability, ability resisting wetting and drying cycles are improved due to the application of CCCW.

In order to investigate the behavior of geogrid-reinforced soil retaining wall with concrete-block panel under local loads and the related interaction mechanism, laboratory model tests are conducted to determine the distribution of the panel lateral displacement, the vertical settlement of the top of wall, the vertical and horizontal soil pressure, the diffusion angle of additional stress, the coefficient of lateral soil pressure and the tension strain of geogrids. The results show that the cumulative lateral displacement of the panel increases gradually from the bottom to the top. The distribution of vertical and horizontal soil pressures along the geogrid length shows a decreasing trend from the middle portion to both ends. The measured diffusion angle of additional stress, which ranges from 40° to 75°, is greater than that of soil mass without geogrids. The distribution of the lateral soil pressure coefficient is different in different sections. The distribution of geogrid strain along its length can be represented by a curve of single-peak, where the horizontal distance between the position of peak and the wall foot decreases gradually from the top to the bottom.

Temperature, water content and pressure strongly influence the deformation and strength behaviors of frozen soil. In order to study the effect of initial water state on deformation and strength behaviors of frozen silty clay, a series of triaxial compression tests has been conducted under initial water contents varying from 12.5% to 20% at ?6 ℃. The results show that the frozen silty clay presents strain softening and strain hardening successively under low initial water content with increasing the confining pressure, but the frozen silty clay mainly exhibits strain softening when the initial water content is more than 16%. With increasing the initial water content, the relationship between initial tangent modulus and confining pressure changes from linear increase to parabola form. Meanwhile, in order to describe the nonlinearity between strength and confining pressure of frozen silty clay under the initial water contents of 12.5%, 14% and 16%, a nonlinear Mohr-Coulomb strength criterion is built based on the envelope theory. The linear Mohr-Coulomb strength criterion can describe the strength development with confining pressure of frozen silty clay under the initial water contents of 18% and 20%.

Laboratory tests on the liquefaction properties of saturated sand with different contents of non-plastic fines were performed by a hollow cylinder apparatus. Results show that: (1) The maximum void ratio emax and minimum void ratio emin decrease and then increase, reaching minimum at 20% and 40% respectively. (2) As fines content increases from 0% to 20%, volumetric strain increases. When fines content increases from 20% to 40%, volumetric strain decreases. Afterwards, volumetric strain increases again when fines contents are between 40% and 60%. Finally, volumetric strain decreases when fines content is more than 60%. (3) The sand with larger fines content has smaller peak strength. Stress-strain curves transform from strain-hardening behavior to a perfect elastic- plastic stress-strain behavior. The phase transformation angle reaches minimum at fines content of 30%. (4) The larger fines content is, the smaller number of cycles and strain to liquefaction are. (5) Liquefaction resistance curve and cyclic resistance ratio have the same trend. They decrease when fines content increases from 0% to the threshold fines content of 30%, and increase when fines content increases from 30% to 50%, and decrease sharply at fines content of 60%, and then increase with increasing fines content. The threshold fines content is about 40%.

The complex mechanical properties of soft broken rock mass lead to a series of challenging issues in deep mining engineering. In-depth study of the post peak mechanical behavior of broken rock mass is significant for analyzing the stability of surrounding rock and designing support structures in deep roadway. By considering the effect of confining pressure, a mechanical model of the post peak strain softening is established on the basis of a quantitative geological stremth index (GSI) evaluation system of surrounding rock and the continuous media theory. The reliability of the proposed model is verified by comparing numerical results with an application of a field engineering project. By calculating the relationships between surrounding rock and support structure, it is indicated that calculated results by the ideal elastic-plastic model and by the strain softening model are obviously different. The use of ideal elastic-plastic model may lead to a low safety and even instability of the support structure. Field application results are in good agreement with numerical results. The proposed model can reflect the nonlinear failure behavior of broken rock mass and provide new insights into the stability analysis of surrounding rock and the design of support structure.

The interaction between earth berm and retaining structure is analyzed with finite element method. A simplified method is proposed to consider the effect of earth berm on the retaining pile. The calculated results show that the existence of earth berm has a significant influence on reducing the internal force and displacement of retaining pile. The internal force and displacement of retaining pile decrease with the increase of the width or height of earth berm. The pile deformation gradually tends to steady when the width of earth berm increases to the excavation depth. The deformation and internal force of the retaining piles considering the effect of earth berm can be calculated by the following method. Firstly, the deformation and internal force of the retaining piles can be calculated by elastic foundation beam method without considering the effect of earth berm. Then the area reduction factor of the earth berm is calculated. Finally, the deformation and internal force can be calculated. Through actual engineering verification, the force and deformation of retaining structure can be more easily calculated by the simplified calculation method.

The subgrade defect of expansive soil is of main concern in high-speed railway construction. Based on the monitored data of the cutting subgrade of Yun-Gui high-speed railway and using software FLAC3D, a three-dimensional dynamic analysis is performed to study the dynamic behavior of the cutting subgrade under the moving train load. It is shown that the dynamic response can be described by numerical model and the predicted variation is consistent with the observations. The vertical dynamic stress of subgrade is exponentially attenuated and the variation of vertical dynamic displacement with depth follows a power law. The dynamic displacement value of railway surface is 0.95 mm which satisfies the requirement of high-speed railway specification. The effect of working conditions on dynamic responses of subgrade is significant. The dynamic stress and vibration velocity of subgrade surface increase due to immersion condition. Laying waterproof and drainage structure layer will improve dynamic stress distribution, reduce dynamic displacement of railway surface and enhance anti-vibration performance of subgrade. The conclusions can provide references for engineering practice and theoretical research of high-speed railway in expansive soil area.

The blasting vibration control is one of the most significant factors during deep tunnel excavation. In this study, the characteristics and influencing factors of vibration induced by ground stress transient unloading were investigated by an integrated approach of field monitoring and numerical simulation of the diversion tunnel No.2 of Pubugou hydropower station. It is shown that the induced vibration by in-situ stress transient unloading of excavation surface plays an important role in the whole blasting excavation vibration under the medium and high in-situ stress conditions. When the wave impedance of rock is fixed, the induced vibration is controlled by the in-situ stress level, initial stress of excavation surface, excavation surface area and vibration attenuation index. The intensity of induced vibration is the highest at the station when the comprehensive effect of excavation surface in-situ stress and excavation surface area reaches the greatest. The blasting vibration intensity of deep tunnel is governed by the combined effect of explosion and in-situ stress transient unloading. The induced vibration can be effectively reduced by decreasing the row spacing of blast-holes and excavation footage, and adopting small holes diameter and partial excavation methods during deep tunnel excavation.

Initial anisotropy caused by granular arrangement effects the mechanical behaviors of pure sand evidently. In order to study such effects, pure sand specimens consisting of elliptical particles at four different sedimentary angles (?) are prepared by the self-developed software NS2D firstly, and then biaxial compression tests are conducted on those DEM specimens. Secondly, the simulated results are compared with those of laboratory tests to validate this method. Finally, the influence of ? is studied by analyzing the simulated results. The results show that all of these four specimens show strain softening and shear dilation phenomenon. With the increase of ?, the strain softening phenomenon of stress-strain curve tends to become strain hardening, and shear dilation phenomenon becomes more and more unapparent. The peak and residual internal friction angles achieve the maximum when ? =0°. With the increase of ?, the peak internal friction angle decreases and then keeps constant, while the residual internal friction angle decreases. The initial principal direction of particle arrangement is parallel to the sedimentary direction. With the increase of axial strain, the principal direction of particle arrangement rotates to the direction parallel to the loading plane gradually, and the principal direction of contact rotates from perpendicular to sedimentary direction to perpendicular to the loading plane. And the increments of both rotations aforementioned increase with the increase of ?. The principal direction of strong contact force is perpendicular to the loading plane all the time.

A particle system has a complex force chain network, and its balance and stability depend on the geometrical structure and mechanical properties of the internal force-chain framework. To depict the micro-mesoscopic mechanical behavior of the particle system, we start with specifying the local arrangement morphology of particles, the contact force between particles and the physical parameters of particles and external load, and then conduct the deformation tests on the particle system subjected to a concentrated force based on the two dimensional-digital image correlation (2D-DIC). With combining with the DIC test method, the Newtonian mechanics theory and the particle linear momentum balance are used to obtain the contact force and the direction of contact force between the particles, and then determine the kinetics of the force chain. Once the distribution and variation of contact force between the particles are obtained, the force chain is demonstrated. It is suggested that there are two states in particle system, namely, the transverse force transmission state and the longitudinal force transmission state. The parameters such as the angle ? between the normal of the contact surface of particles and the horizontal axis, the force transmission angle ? are analyzed. The influences of ? and ? on the force chain damage are discussed. The parameters, the ratio R of the particle diameter d to the external load diameter D and the maximum number N of particles contacted with the loading rod, are analyzed. All of the above can provide a guideline for describing the evolution, destruction and reconstruction mechanism of the force chain.

Gas permeability is a significant parameter influencing gas extraction and coal-gas outburst prediction in coal seams. To investigate the effects of coal damage and shear deformation on fracture permeability, a damage variable is introduced to reflect the failure state of coal damage. A constitutive model is established with the increment of volumetric strain. The evolution model of fracture permeability influenced by shear and compression is obtained, and Hurst coefficient is applied to represent the roughness of fracture surface. Based on the dual permeability media of matrix pore and fracture, a gas-solid coupling software named TOUGH2(CH4)-FLAC is built by the secondary development of the codes of TOUGH2 and a DLL file for FLAC3D. A case study is performed to investigate the influences of damage and shear dilation on gas flow in the process of uniaxial compression using the proposed model. Numerical results show that the accumulation and coalescence of damaged elements largely control the failure of coal, and the damage zone is the main factor resulting in the failure of coal. The location of permeability change is highly related to the damage zone. The fracture permeability can be increased by two orders of magnitude in most of damage zones. The variety of fracture permeability in the shear damage zone increases drastically with the increase of shear dilation. The proposed method provides a theoretical basis for the development of gas control.

The stability of rock engineering is strongly dependent on the shear strength of jointed rock mass. Based on the particle flow code (PFC2D), the reasonable mesoscopic parameters are selected in combination with experimental results to analyze the meso-properties of crack propagation, energy transmission, and acoustic emission phenomenon of jointed rock. The strength models and failure patterns of jointed rock are numerically simulated. The main research results are summarized as follows. Abrasive and shear failure patterns heavily exist in jointed rock, and different failure patterns are corresponding to different strength models. Rock mass is damaged along joint plane with the increase of shear deformation. Normal cracks prevail within elastic stage, whereas shear cracks dominate along the rough surface within plastic stage. The joint plane slides owing to appearance of crushed zone induced by the coalescence of R and P cracks. Boundary energy is mainly converted into strain energy and more normal cracks are generated prior to the peak shear strength. With the increase of shear stress, the friction energy grows rapidly and a large amount of shear cracks are produced at the same time. Compared with experiments, PFC2D can be used to simulate the shear properties of jointed rock mass well, which remedies the challenge of simulating behaviors of jointed rock at meso-scale in the laboratory test and provides a useful reference for further research on direct shear tests of jointed rock mass.

Based on the whole slope stress field obtained from block element and the joint stress field calculated by distinct element method, combining with the definition of vector sum method (VSM), a new slope stability analysis method is proposed, namely VSM-universal distinct element code (VSM-UDEC) method. Firstly, the accuracy of contact stress and element stress calculated by UDEC under various conditions is verified by the slider test. It is shown that the contact stress and element stress fields have high precision (i.e., the relative error of contact stress is less than 0.2%), which satisfies the condition of safety factor calculation. Numerical simulations of a linear sliding surface and a circular sliding surface are analyzed by the VSM-UDEC method. Numerical results are compared with the solutions by the corresponding theory and the rigorous limit equilibrium method (Morgenstern-Price method). It is indicated that the safety factor obtained by the VSM-UDEC method is almost identical with that by the theoretical solution in the case of linear sliding surface, and is consistent with that by the limit equilibrium method in the case of circular sliding surface. Finally, the VSM-UDEC method is applied to analyze the stability of the left bank slope of the Jinping-I Hydropower Station. The results demonstrate that the safety factor by the VSM-UDEC method is good agreement with that by the limit equilibrium methods (Sarma method, Morgenstern-Price method) when blocks are regarded as rigid. Moreover, the VSM-UDEC method can be considered as the effect of the structural surface in the slope body on the slope stability, and thus it has more advantages than the limit equilibrium method to reflect the actual stability and deformation distribution situation of the slope. UDEC software has good applications of simulating slope failure process and VSM has the clear physical significance of safety factor, and therefore, the VSM-UDEC method may have good prospects of applications in the slope stability analysis.

This paper describes the applications, principle and common packaging method of existing fiber Bragg grating (FBG) testing technology. The applicability of FBG to strain detection of seasonally frozen subgrade which is a concealed work is explored. The main works are as follows: (1) Based on the method of embedded-beam for the testing technology of FBG, two ways for the packaging and layout of sensors are introduced taking account of the difference of testing objects, namely the shallow-setting semi-flexible feeler rod for short distance and the deep-setting semi-rigid probe for long distance. (2) To meet the need of tests in field, the systems of power supply, fiber optic sensors, data collection and analysis are compared and selected. (3) Two sites of seasonally frozen subgrade are chosen and a one year observation is conducted. The results indicate that FBG has a high survival rate within the temperature range from -20 ℃ to 30 ℃. The system of strain detection for subgrade in seasonally frozen soil regions presented and achieved here has a high testing sensitivity and reliability. The system provides a new technical means for testing physical indices of concealed works.