Particle shape is one of the most important characteristics of particles and its correlation with mechanical properties has attracted great attentions. In order to systematize the influence of particle shape on the mechanical properties of aggregates, the research results of physical characterization, numerical simulation and laboratory test of particle shape were summarized and analyzed. The results show that the definition method of particle shape based on three scales (form, angularity and surface texture parameters) is the most ideal method to describe the particle shape characteristics. The numerical method can simulate the shape of a single convex particle and the arrangement structure of a convex particle assembly, but the characterization of concave particles by numerical methods is not accurate enough. There are correlations of particle shape with pore structure, natural angle of repose, limited void ratio and particle size. The permeability coefficient of the aggregate is largely controlled by the shape characteristics of the particles under the same gradation and porosity. Due to the difference in interparticle locking and the number of contact points, the particle shape in uniform particle size affects the small strain strength, peak strength, residual strength of the aggregate and the dilatancy. Finally, the problems in particle shape research are discussed and future research directions are proposed.

The visual physical model tesing technology based on transparent soils is playing an increasingly important role in the study of deformation and failure mechanism in geotechnical engineering. The preparation of transparent soil with high transparency, stable physical and mechanical properties, and suitable for simulating different rocks and soils is the basis for the development of this technology. Given the lack of transparent materials for clays and soft rocks, a transparent cemented soil has been formulated. The transparent cemented soil is synthesis using fused quartz as the skeleton particles, nano-level hydrophobic fumed silica powder as the cementation and mixed mineral oil of n-dodecane and #15 white oil as the pore fluid. By changing the content or proportion of the cementation, the size or gradation of the skeleton particles, a series of transparent cemented soils with similar refractive index and varying strength characteristics can be obtained. For the factors influencing the soil strength, such as cementation content and skeleton gradation, 11 groups of controlled experiments were designed and corresponding triaxial shear tests were conducted to study the strength characteristics. Results show that: 1) The shear strength varies complexly with the changes of the particle size and gradation of the fused quartz, the content and proportion of the silica powder, etc., but overall the poor quality of quartz gradation and the rich proportion of silica powder can reduce the shear strength; 2) By changing the preparation formulas, a series of transparent cemented soils with cohesions ranging from 5 kPa to 65 kPa and internal friction angles ranging from 25°to 44°can be prepared, which provides a basis for the selection and preparation of transparent materials in physical model experiments using natural clays and soft rocks.

The true triaxial loading and unloading tests were conducted to obtain the characteristic stress, failure mode and energy evolution characteristics of the granite in an underground cavern. Results show that under the true triaxial loading and unloading stress paths, the failure modes of granite are both tensile and shear composite failure, and characteristics of high damage stress and brittleness are obviously observed. A new brittleness index is proposed to evaluate rock brittleness by using volumetric strain curve. The brittleness of granite under unloading condition is higher than loading condition. In the true triaxial loading test, the change trend of total energy with axial strain goes through three stages: slow increase, rapid increase, and steady increase. In the true triaxial unloading test, the dissipated energy increases rapidly at the moment of unloading, and its proportion in the energy distribution increases, which becomes the main energy consumption. The energy dissipation value of the granite in the loading test is obviously greater than that in the unloading test. It indicates that more energy is required for samples under the loading path to cause damage. More elastic strain energy can be released under the unloading path, which is more dangerous than the loading path.

In order to study the long-term creep law of extremely soft coal rock, the uniaxial and triaxial creep tests of extremely soft coal rock subjected to single-stage loads were carried out by using the self-developed triaxial creep test system. The following results and conclusions are obtained. 1) In the uniaxial long-term creep test, decay creep stage, steady-state creep stage, and accelerated creep stage occur successively during 232 hours, and the cumulative creep strain is as high as 3.45%, which is 10.5 times of the instantaneous deformation. The strain rate for the total steady-state creep stage is as high as 8 ?10 /h, the maximum accelerated creep rate reaches up to 0.043/h, and the strain rate of total creep process is distributed in a U-shape. 2) In the triaxial long-term creep tests subjected to the same axial pressure (0.96 MPa), the ability of resisting long-term deformation of extremely soft coal rock increases continuously with the increase of confining pressure from 0 to 0.6 MPa, which is shown as follows: the creep strain decreases significantly, the strain rate of steady-state creep stage decreases by order of magnitudes, the duration before creep failure increases obviously, the ratio of creep to instantaneous strain decreases dramatically, the intensity of accelerated creep failure decreases markedly. The creep strain and deformation rate are especially sensitive to the confining pressure change in the range of 0?0.2 MPa. 3) The accelerated creep stage of extremely soft coal rock is characterized by “gradual” time dependent instability, which is significantly different from the “abrupt” accelerated fracture instability of ordinary rock. 4) By connecting a new nonlinear viscoelastic element considering the concept of accelerated creep start-up time with Burgers model in series, a nonlinear viscoelastic plastic creep model is established to describe the three creep stages of extremely soft coal rock subjected to single-stage load. Then, the creep parameters were identified by employing the Levenberg-Marquardt optimization algorithm. The fitting curves are highly consistent with the experimental curves, verifing the validity of the proposed model. These findings can provide reference for theoretical analysis of nonlinear large creep deformation and support design of extremely soft rock.

To solve the problem of carbon emission and high energy consumption of traditional binder, the coal-bearing metakaolin (CMK) geopolymer was used to stabilize sludge. The unconfined compressive strength (UCS) test was conducted to determine the proportion of alkali activator firstly. On this basis, the effects of alkali-precursor ratio, precursor content, NaOH molarity, curing time, curing temperature and slag content on the mechanical properties of stabilized sludge were investigated. Finally, the SEM and XRD tests were conducted to analyse the microstructure of stabilized sludge. The results showed that the UCS of stabilized sludge first increased and then decreased with increasing Na2SiO3:NaOH ratio and the rational proportion of alkali activator of geopolymer was Na2SiO3:NaOH=75:25. The UCS increased as the alkali-precursor ratio, precursor content and NaOH molarity increased. As the curing time, curing temperature and slag content increased, the UCS of stabilized sludge improved more obviously. The microstructural analysis showed that no new minerals were formed in geopolymer stabilized sludge. With the increase of alkali-precursor ratio, precursor content, NaOH molarity, curing time, curing temperature and slag content, the amount of amorphous N-A-S-H and C-(A)-S-H gels increased, which made the soil structure denser through bonding and filling effects, thus increasing the strength of stabilized sludge.

To study the effectiveness of filled cracks as a means of anti-blasting and wave-elimination for underground protective structures, 36 one-dimensional impact tests were carried out on the intact rock specimens and rock specimens with filled cracks using the SHPB apparatus. The transmittance, stress wave velocity and peak stress value of the specimen under different working conditions were obtained. The impacts of confining pressure level, filling thickness and filling material types on stress wave attenuation were analyzed. The results show that, with the increase of the thickness of the filled cracks, the transmittance of the specimen, stress wave velocity, and stress wave peak stress attenuate clearly. The greater the thickness of the filled crack, the smaller the transmission energy of the stress wave during the propagation process. With the increase of confining pressure, the wave- eliminating effect of filled cracks is reduced due to the closure of pores, but still demonstrate stress wave attenuation effect to some extent. Therefore, using particles with higher compressive strength as crack-filling material and increasing filling thickness can effectively improve the wave-eliminating effect in engineering protection.

Cyclic loading with the upper limit loading of 80% uniaxial compressive strength (UCS) and monotonic loading are consequently applied on argillaceous quartz siltstone. The evolution law of peak strength, dilatancy point stress, elastic modulus, Poisson’s ratio, hysteretic loop, strain, compliance and mechanical properties of the studied siltstone under cyclic loading near fatigue strength are studied. The test results show that: 1) With the increase of cycle numbers, the rock peak strength decreases slightly at first, then increases continuously, and finally tends to be stable. The minimum and maximum peak strengths are respectively 7.35% lower and 16.57% higher than the UCS. 2) There is a negative linear correlation between stress and the volumetric strain at dilatancy point. The ratio of stress corrseponding to dilatancy point at the last cycle to peak strength of different specimens follows a logarithmic function distribution with the number of cycles, and a formula is proposed to predict the uniaxial peak strength of rock after cyclic loading near fatigue strength. 3) The monotonic loading strain ratio and the peak strain ratio of the rock are stable around 0.58 and 1.02 after several cycles, respectively. It indicates that the deformation resistibility capacity of argillaceous quartz siltstone is significantly improved after cyclic loading near fatigue strength. The peak strain of uniaxial compression test can be used as the reference value of rock failure under cyclic loading. 4) In loading stage, the loading compliance corresponding to axial strain increases first and then decreases linearly, and the loading compliance corresponding to volumetric strain shows a hook shape. In unloading stage, the loading compliance corresponding to axial strain presents a trend of accelerating growth, and the loading compliance corresponding to volumetric strain shows a V-shape. The loading and unloading compliance corresponding to volumetric strain at the lower limit load exhibits positive linear correlation with the residual volumetric strain, which can reflect the accumulated damage of rock under cyclic loading. 5) Based on the evolution law of mechanical parameters and deformation, and the microstructure characteristics of argillaceous quartz siltstone, the evolution mechanism of peak strength and dilatancy point stress under cyclic loading near fatigue strength is revealed.

A series of splitting tests was carried out on the intact and remolded loess specimens by particle image velocimetry (PIV) test system, and the microstructure of the intact and remolded loess specimens were quantitatively analyzed by mercury intrusion porosimetry (MIP), so as to discuss the influence of the structure and initial dry density on the tensile strength. The test results show that the peak load of the intact specimens is bigger than that of the remolded specimens when the specimens splitting failure. The peak load of the remolded specimens increases with increasing initial dry density. The displacement vector field shows that the primary cracks incline and the secondary cracks do not develop during the split failure of the loess specimens. When the remolded specimens split, the primary cracks are radially vertical, the secondary cracks more develop, and the crack morphology of the remolded specimens with different initial dry densities is basically the same. For the equal initial dry density, the cumulative intruded porosimetry volume curve and pore diameter distribution density curve of the intact specimen are both higher than those of the remolded specimen, and the inter-aggregate pores of the intact specimen are more than those of the remolded specimen. However, due to the obvious structure, the tensile strength of the intact specimen is higher than that of the remolded specimen. As the initial dry density increases, the cumulative intruded porosimetry volume curve of the remolded specimen moves downward, and the peak of the pore distribution density curve moves to the left. The inter-aggregate pores gradually decrease or even disappear. As a result, the tensile strength increases with increasing initial dry density.

To reveal the mechanical characteristics and failure modes of stratified rock under combined static-dynamic loads, a split Hopkinson pressure bar device was used to apply one-dimensional load on stratified sandstone specimens. Meanwhile, digital image correlation technique was used to monitor the fracturing process in real time. Fracture evolution and energy consumption characteristics of different dip angles were further summarized. The results show that: under the combined static-dynamic loads, the dynamic strength of stratified sandstone specimens with holes initially increases and then decreases with the increase of bedding angle. The failure modes of the specimens can be divided into four types. The energy consumption characteristics of different dip angles under combined loads show that when the bedding dip angle is 45°, the strength of sandstone specimens and the proportion of absorbed energy reach the maximum values. In underground engineering, reasonable arrangement of breaking position and the angle between the direction of impact load and bedding can significantly improve the efficiency of rock burst, reduce the adverse impact of dynamic disturbance on surrounding rock, and enhance the stability of deep rock engineering.

The frost heaving pressure generated by water-ice phase change and volume expansion of saturated fractured rock mass in cold regions promotes the initiation and expansion of new fractures, and leads to further damage and deterioration of fractured rock mass. To reveal the frost heaving pressure degradation mechanism of fractured rock mass subjected to multiple freeze-thaw cycles, the repeated frost heaving pressure monitoring test was carried out on saturated fractured granite with different macroscopic fractures under different freezing temperatures. The evolution characteristics of cyclic frost heaving pressure and the effect of crack size, freezing temperature and the number of freeze-thaw cycle on the frost heaving pressure of saturated fractured granite were analyzed. The results show that the evolution characteristics of frost heaving pressure in single freeze-thaw cycle is similar, which can be roughly divided into five stages. As the number of freeze-thaw cycles increases, the peak frost heaving pressure decreases exponentially, and the amplitude of peak frost heaving pressure drop increases. The peak frost heaving pressure increases linearly with the increase of crack length. The smaller the crack width, the earlier the frost heaving pressure appears in the freezing and thawing stages. The lower the freezing temperature, the earlier the frost heaving pressure appears. The peak frost heaving pressure increases linearly with the decrease of temperature, and the effect of freezing temperature on the frost heaving pressure weakens with the increase of freeze-thaw cycles. As the freezing temperature decreases, the effect of crack size on the peak frost heaving pressure becomes more significant. The research results can provide reference for the theoretical calculation and numerical analysis on frost heaving pressure of fractured rock masses in cold regions.

Aerosol injection technology (AIT) is a new type of active drainage technology. Aerosol injection is carried out in an injection hole through a rotating pipe with a high-pressure pump, which plays three major roles including deep three-dimensional drainage, deep pressurization and air-lift precipitation to improve drainage efficiency. Aiming at the effect of air-lift precipitation, a large-scale air-lift model test device has been independently developed, and a series of experimental studies has been carried out on variables such as the lower medium, surrounding medium, air injection pressure, air-lift drainage channel size and air injection depth. The research results show that: 1) The surrounding medium has little effect on the air-lift precipitation depth but has a great impact on the water level drop speed; 2) The air injection pressure is not directly proportional to the maximum air-lift precipitation depth. The maximum air-lift precipitation depth can only be reached when the air injection pressure is well matched with the diameter of the vertical drainage body; 3) When the diameter of the vertical drainage body is constant, the air injection depth determines the maximum precipitation depth and precipitation rate. The research results can guide the implementation of AIT and provide a basis for further study on the air-lift precipitation mechanism of AIT.

In order to study the deterioration effect of swelling pressure of Gaomiaozi sodium bentonite under long-term strong alkali- heat action, 24 groups of orthogonal optimization-tests were designed for Gaomiaozi sodium bentonite from Xinghe County, Inner Mongolia. The test conditions were as follows: one NaOH concentration (0.5 mol/L), four alkali-heat temperatures (25, 90, 150 and 210 ℃), six alkali-heat interaction durations (2, 4, 6, 8, 10 and 12 months). X-ray diffraction (XRD) tests were performed on the bentonite after alkali-heat treatment, and constant volume swelling (CVS) experiments were carried out on the bentonite sample with the initial dry density of 1.7 g/cm3 using self-made 2D swelling-apparatus, with the distilled water as an infiltration medium. The results showed that: 1) With the increase of the time and temperature under alkali-heat interaction, the content of montmorillonite, the maximum swelling pressure and the hydration time decreased; 2) The characteristic curve of swelling pressure-hydration time tended to be stable and was independent of the strong alkali-heat condition when the ratio of the maximum radial and vertical swelling pressures was 0.5; 3) A transferring-function of swelling pressure with hydration-time was established on the basis of damaging-factor of swelling pressure and accelerating-factor of aging duration, the model can better predict the experimental data; 4) A total damage increment was mostly attributed to the coupling effect of damage rate of temperature and time under the alkali-heat interaction, additionally, there was an accelerating effect of the degradation characteristic of swelling pressure facilitated by an increasing temperature.

At present, the quantitative evaluation models of mechanical reinforcement of soil by roots lack consideration of root-soil interface bonding effect. In order to improve the accuracy and applicability of the quantitative model, this study was premised on the widely used Wu model, combined with the failure mode of root system in root-soil, and quantified the bonding effect of interface between root and soil interface by interface bonding strength parameters. Thus, an estimation model of limit values of shear strength of root-bearing soil based on interface bonding was established, which considered the influence of soil stress, root diameter, root length, and initial angle between root and shear direction in root-bearing soil. Furthermore, the comparison of shear strength of root-bearing soil between the measured values of direct shear test and the calculated values of the model showed that the measured values were within the range of the calculated values of the model established in this article. When the angle between the root and the shear direction was 45°, the measured value was close to the maximum value of the model calculation, and when the angle was 90°, the measured value was close to the average value of the limit values of the model calculation. Compared with the calculated value of Wu model, the accuracy of the estimation model established in this article was increased by five times on average. Moreover, the sensitivity analysis of model parameters showed that the additional soil strength provided by roots was greatly affected by the root number parameters, root size parameters and root-soil interface bond strength parameters.

During the construction of slurry shield, a large amount of slurry is easy to permeate and filtrate, which leads to the instability of the excavation face. It is of great significance to maintain the stability of the excavation face that the slurry permeates and forms a membrane. Fully understanding the law of slurry permeation is the key to study the stability of the excavation face of slurry shield. Considering the filtration effect of soil skeleton on permeable slurry particles, a slurry concentration diffusion model is established, and the relationship between slurry concentration and stratum porosity is obtained. Based on the theory of porous media fluid mechanics, the slurry permeability parameters are solved, and the analytical model of slurry penetration-diffusion is established. The accuracy of this model is verified by indoor model test and comparison with existing theories. Finally, based on an engineering example, the temporal and spatial evolution law of slurry penetration-diffusion is analyzed, and the analytical results are consistent with the engineering understanding. The stratum porosity increases with the increase of slurry penetration distance, and then remains unchanged, and decreases gradually with the increase of slurry penetration time; the slurry penetration distance increases gradually with the increase of penetration time, and then tends to be stable. With the increase of penetration time and penetration distance, the slurry concentration, penetration velocity and seepage discharge decrease. The results show that the analytical model of slurry penetration-diffusion can reveal the law of slurry penetration-diffusion under different stratum parameters, and quantify the slurry permeability parameters, which has a good guiding role for the actual slurry shield construction.

The analytical solution for the nonlinear large strain consolidation model is significant for the actual engineering. However, since the large-strain consolidation theory of saturated soft soil involves complex material nonlinearity and geometrical nonlinearity problems, it is usually challenging to obtain an analytical solution for the large strain consolidation models. At present, the nonlinear compression and permeability characteristics of soils are usually described by e- and e- (e is the void ratio, is the effective stress, and is the permeability coefficient). However, large numbers of laboratory consolidation tests show that those nonlinear relationships are no longer suitable for the high compressibility soft soil with large strain. In contrast, the nonlinear relationships between the void ratio and effective stress as well as the void ratio and permeability coefficient of soft soil with high compressibility can be described by - and - more effectively. Therefore, in this study the one-dimensional nonlinear large strain consolidation model of the high compressible soft soil under a ramp loading is developed, and the exact analytical solution with the specific parameters and the approximate solutions under the general conditions are developed. The reliability of the proposed analytical solutions is validated by comparing the calculated settlement curve with the experimental settlement curve under a ramp loading in the laboratory. On this basis, the influences of parameters ( and ), external load, and loading rate on the consolidation behavior are investigated by numerous calculations. The results show that at a certain compression index , the smaller the , the faster the consolidation rate of the soil layer, and the faster the dissipation of the excess pore-water pressure at the bottom of the clay layer. Conversely, as is constant, the smaller the , the faster the consolidation rate of the soil layer, and the faster the dissipation of the excess pore-water pressure at the bottom of the clay layer. The average degree of consolidation in terms of settlement is generally larger than that of the consolidation accounting for the excess pore-water pressure , which means the development rate of settlement is faster than pore pressure dissipation. In addition, when ( 2) 1, the average consolidation rate decreases with the final external load increasing. When ( 2) 1, the average consolidation rate increases with the final external load increasing. When ( 2) 1, the average consolidation rate is independent of the final external load. Moreover, when the final external load is constant and the loading rate increases, the average consolidation rate increases while the excess pore-water pressure at the same depth decreases.

The key to solving the safety factor of the vector sum method lies in determining the real stress state and the main sliding trend direction of the slope. The safety factor obtained by the vector sum method is closely related to the real stress state of the slope and the final sliding trend direction. The existing vector sum method is mostly based on a single anti-slide force or sliding force to determine the main sliding trend direction of the landslide. However, since slope sliding is actually the result of the combined action of the anti-slide force and the sliding force, the combined action of the two needs to be considered in the definition of the main sliding trend direction. In this study, we take the whole landslide body as an analysis object and the main sliding trend direction is first determined based on the overall vector resultant direction of the external load of the isolation body, the normal force, and the ultimate anti-sliding force at the bottom of the sliding body. After that, the safety factor of the landslide body is determined by the definition of the safety factor of the vector sum method. Finally, with the theoretical derivation and verification of standard test questions, the most dangerous sliding surface position, safety factor, and main sliding trend direction derived based on the new main sliding trend direction are comparable with those of the conventional slope stability method. What’s more, the slope safety evaluation based on the new main sliding trend direction are more safe, and more convenient for engineers to take more targeted slope support strategies.

In order to study the long-term loss law of anchorage force and the calculation method of pre-stress loss of anchorage force, a calculation method of the pre-stress loss of anchorage force was proposed to account for three factors. The proposed method is based on the generalized Hooke’s law, relaxation rate time-history response equation and improved coupling constitutive model. It was adopted to calculate the theoretical loss value of the anchor cable pre-stress coupled with the internal sliding lock of the anchor plate, the stress relaxation of the cable body steel strand, and the rock mass creep, respectively. In order to verify the accuracy of the long-term loss calculation method, the anchor cable pull-out failure test was also carried out. By analyzing the transfer mechanism of the pull-out load, it was determined that the pull-out stage of the free section of the load-incremental displacement curve was the working load interval of the anchor cable. The median value of the working load interval was taken as the effectiveness force of the anchor cable, and the measured value of the long-term loss of prestress was also obtained. Results show that the pre-stress loss of the anchor cable is about 17.5%?27.5% during 20 years. The rock mass creep is the most significant factor affecting the long-term loss of the anchor cable. Besides, there is an error of about 4.2% between the calculated value of the three-factor loss calculation method and the measured value of pre-stress loss at site. This verifies the accuracy of the three-factor loss calculation method, and also shows that the three-factor calculation method of long-term pre-stress loss mentioned here meets accuracy requirements and is feasible.

Aiming at the disaster of rock burst in graben structural area of coal mine, based on the particularity of graben structure and the occurrence characteristics of overburden rock, the movement characteristics of overburden and stress evolution law of coal and rock mass in the graben structural region are analyzed and the corresponding mechanical model is established by means of field investigation, theoretical analysis and numerical simulation. The mechanism of rock burst in the graben structural region is consequently studied. At the same time, the evolution characteristics of microseismic energy and frequency before and after rock burst are analyzed, and the time sequence of microseismic precursory information of rock burst in graben structural region is obtained. The results show that: the sliding subsidence of wedge is closely related to the width of goaf. The existence of goaf in graben structural area will break the stable state of wedge and lead to more severe stability conditions. The local slip and dislocation of FD6 and FD8 faults lead to wedge sliding subsidence, and the hanging roof structure of graben structural area also has corresponding extrusion effect on the undeveloped coal mass. The same action provides high static stress conditions for the occurrence of rock burst, and the breaking of suspended roof structure provides dynamic load disturbance conditions for the occurrence of rock burst. The joint action of static load and dynamic load results in the upper drift and triple entry in the graben structure. The energy and frequency of microseismic show obvious deviation before the impact, and the daily average energy of microseismic is measured. There are obvious sudden drop and sudden rise. The anti-scour technology of “chain-brokening and consumption-increasing” is proposed based on the analysis results and field practice. It provides some theoretical reference value for occurrence mechanism, early warning and prevention and control of rock burst on working face in graben structural area.

In China's existing blasting-related standards, the control parameters (K and a) of blasting vibration in the Sadovsky formula are selected only based on the softness and hardness of the rock, which is subjective and arbitrary. In order to solve this problem, the blasting vibration attenuation law and the rock mass quality classification of some engineering examples were statistically analyzed, and the control parameters K and a were quantitatively expressed. In addition, in view of the uncertainty of blasting vibration caused by the discrete characteristics of rock mass parameters, a probability analysis of the peak particle velocity (PPV) was also carried out. The results showed that: 1) By comparing with the field blasting test, the established relationships between K, a and BQ (the basic quality classification index of rock mass) can be applied to the selection of control parameters in case of the absence of field blasting test data; 2) The proposed method for probabilistic analysis of blasting vibration can take into account the discrete properties of rock mass and the attenuation law of blasting vibration measured on site. The obtained probability distribution of PPV uncertainty factor can quantitatively evaluate the reliability of blasting vibration, which is a necessary supplement to reasonably predict PPV.

Due to the influence of fracture undulation, there is an obvious difference in the grout flow rule between the rock mass with rough fractures and the rock mass with existing results. Based on the parallel plate model of fluid flow in a single crack, a flow rule model of viscous fluid in a rough fracture was developed using a corrected cubic law. The derived results were agreed well with that from a single crack grouting test. Based on a discrete fracture network (DFN) model, the grouting process in a fractured rock mass was simulated and studied by changing the fracture network characteristics and the grouting pressure. The results show that the grouting diffusion volume is controlled by the permeability of rock mass, but the connectivity plays a key factor affecting the fluid diffusion process, which can affect the fluid diffusion in local areas. The increase of grouting pressure changes the slurry diffusion velocity, which can make the slurry easier to diffuse to the remote areas that far from the grouting point and those areas with poor local connectivity, however, excessive grouting pressure may cause the fracture opening and even lead to the rock mass failure.

The red beds are widely distributed in Yunnan, Guizhou, Sichuan of China and Southeast Asia. This material is not suitable as subgrade filling due to seriously weathering and its softening characteristic after absorbing water. In the project of the section between Chengdu and Ya'an of Sichuan-Tibet Railway, the graded weakly weathered red-mudstone gravel was added to red beds subgrade fillers with specific proportion to improve size distribution and compactness of the material. The stress-strain characterization of improved red beds subgrade fillings has been obtained from laboratory test. A method of randomly generating gravel geometry model with concave convex polyhedron shape has been proposed. Based on the numerical compression model established by PFC3D, the proportion of different strong force chains during loading and probability density distribution at the peak intensity have been studied. The results show that the improved red beds subgrade fillings with the stone content of 50% performs better mechanically with good deformation characteristics, and its unconfined compressive strength is 421.9 kPa. The gravels form super chains easily by friction under bond strength of red clay and that strengthens the system. The mechanical properties degree of improved red beds subgrade fillers follows the order of the stone content 50%> 40%> 60%. Therefore, it is recommended that weakly weathered red-mudstone gravel with stone content of 50% should be added to the red beds as the subgrade fillings of Sichuan-Tibet Railway.

There are columnar collapse in many reservoir gorges and coastal zones, these collapse may be exhibited by situated collapse due to the failure of weak base rock mass. The debris will impact into the water and form impulse waves. It is a challenge to predict and analyze the large area and long-distance impulse wave quickly and accurately. The coupling process of collapse body and water body is black boxed, a high-precision initial surge generation model is formed by using the physical test formula, and the Boussinesq model is used to control the propagation of initial surge event wave. A simplified surge numerical model 3D-LWBM with similar calculation accuracy to the physical test is constructed. The results show that: 1) The timing sequence and accuracy of 3D-LWBM simulation of pseudo-two-dimensional numerical liquid level and measured point water level are in good agreement with the physical experiment results, and the accuracy is high. 2) The generation and propagation of impulse waves are predicted by 3D-LWBM in the complicated topographic channel of Longmenzhai dangerous rock mass in the Three Gorges Reservoir area. Compared with the results of fluid-landslide full coupled simulation, 3D-LWBM has better calculatiion accuracy and faster calculation speed. 3) When the mesh size is 1/6?1/4 of the initial wave length, 3D-LWBM can obtain a more accurate and resonable result. At the same time, 3D-LWBM has its own scope of application. This study can provide technical support for disaster prevention and mitigation of relevant impulse waves.

The critical state model for sand is capable of describing the stress-strain relationships over a range of densities and stress levels, as well as the feature of dilatancy. Considering that the soil is generally under true triaxial stress condition in practical engineering problems, the extension of the state-dependent model in the multi-axial stress space will be taken into account in this paper. The iteration scheme for the implicit integration algorithm is formulated in the three-invariant plasticity model platform. With this algorithm programed into the user subroutines, Vumat, the constitutive model is implemented into the finite element analysis based on ABAQUS. Through the Explicit procedure, the sand element tests with different relative densities are simulated first. Then the bearing capacity of circle and strip foundation is analyzed subsequently. The calculated results show that this algorithm is converged and robust, as well as being potential to deal with large-scale boundary value problems.

To study the time-space effect of the dynamic response of a long tunnel under non-uniform ground motions, firstly, a random ground motion synthesis method was used to generate different ground motion time histories, which can reflect local site effects such as traveling wave effects and attenuation effects of long-distance tunnels. Through vertical input, the dynamic responses of the surrounding rock and liner of the long tunnel under the action of uniform and non-uniform ground motions were analyzed. On this basis, through the Fourier spectrum, transfer function and coherence function, the spectral correlation between the input earthquake and the dynamic response of the surrounding rock was revealed. Secondly, the spatial inconsistency of the dynamic response of surrounding rock in different parts was studied in terms of acceleration, displacement, and relative displacement time history. Finally, the coordination of seismic dynamic response of surrounding rock and liner was studied by analyzing their dynamic characteristics under the action of ground motion. The results show that the transmission coefficient of the surrounding rock monitoring points at the same position in the axial direction of the tunnel is inconsistent in the main frequency band under non-uniform ground motion. During the upward propagation of ground motions, the maximum coherence coefficient between the response of each monitoring point along the elevation and the input ground motion has a basically quadratic polynomial correlation with the elevation, but the maximum coherence coefficient of the non-uniform ground motion input is significantly greater than that of the uniform ground motion input. Under non-uniform ground motion input, although there is a certain delay in the displacement time history of the surrounding rock monitoring point at the same location along the tunnel axial direction, the maximum displacement of the surrounding rock at the exit of the tunnel occurs earlier than that at the entrance of the tunnel. Also, the peak acceleration and displacement of the surrounding rock monitoring points near the tunnel and the maximum relative displacement of the top and bottom monitoring points show the same distribution trend that all of them are larger at both ends and smaller at the middle along the tunnel axis. Due to the traveling wave effect, the shear action will be generated at the left and right ends of the tunnel under non-uniform ground motion input. This study can provide a comprehensive understanding of the destruction mechanism of long tunnels under the action of earthquakes.

Based on the Hertz-Mindlin with bonding (HMB) contact model in EDEM, the modeling method of the rock discrete element model was investigated. First of all, all of the bonds in HMB rock model must be “weak bond”, i.e., the tensile strength of the bond is less than its critical normal stress. Then, the regression relationships between the elastic modulus , Poisson’s ratio , tensile strength of rock model and the particle and bond parameters were established, respectively. According to the three regression relationships, a mathematical model that can fit the real-rock’s E, , and through the HMB model was proposed. For the compressive strength of rock, when the critical stress ratio ( = / , is critical normal stress, and is critical tangential stress) ranges from 1 to 2, is influenced by , and the ratio of critical normal to tangential stress. Especially, when the is closed to 1.5, the failure mode of the rock model presents a better agreement with the experiment result. In addition, when the is constant, a linear relationship between the and can be clearly observed. In conclusion, a HMB model based modeling method that can fit the elastic modulus, Poisson’s ratio, tensile strength, compressive strength and the failure form of rock sample was proposed, and the detail steps were also given. Finally, an example was conducted to verify the proposed modeling method. The results showed that the rock models established through the proposed method can fit both the static and dynamic mechanical properties of rock.

To better explore the effect of different hydrate habits on the mechanical properties of hydrate-bearing soils, a combined injection synthesis and direct shear testing system has been developed. The system includes a hydrate injection synthesis setup and a hydrate-bearing soil direct shear apparatus under low temperature and high pressure conditions. The sprayed water mists can widely contact with hydrate-forming gas using the spraying water equipment, and thus gas hydrate can be synthesized quickly and efficiently under the high-pressure and low-temperature environment in the container. The system can be used to prepare hydrate-bearing samples by combining the unsaturated method and the hydrate powder mixing method. Different sample preparation methods correspond to different hydrate habits. Hence, the new direct shear tests can compare the shear strength of hydrate-bearing soil under different hydrate habits. Using silty sand and carbon dioxide hydrate as the tested materials, the direct shear tests of cemented and pore-filling hydrates were carried out. The results show that the shear strength of hydrate-bearing soil is higher for the cementation habit than for the pore-filling habit, which verifies the reliability of the instrument.

Tilt test is a common method to obtain the basic friction angle of rock. However, the existing tilt test technology has not formed a recognized standard, leading to an obvious difference between the test results of the same kind of rock materials. To solve this problem, an inclinometer was developed to measure the basic friction angle of rock based on the existing tester structure and testing mechanism. A series of multi-scale specimens with dimensions ranging from 50 mm to 150 mm in length and 25 mm to 75 mm in height can be tested. The high-precision rotation control of angular speed ranging from 6(°)/min to 60(°)/min can be realized through the servo motor and the high-speed ratio reducer. The corresponding test parameters, such as specimen’s sliding distance and inclination angle, can be monitored and recorded in real time. Then, the operation flow of testing basic friction angle was put forward. The main factors that have obvious effects on the experimental results were analyzed, such as specimen size and processing technology, angular velocity, test results criterion and the dryness of the structural plane. In this paper, the basic friction angles of 11 specimens, including granite and diabase, with discontinuity size of 100 mm×100 mm were tested, and it provided a reference for the test specification of inclinometer.