Several typical stress threshold stages exist for rock materials during uniaxial compressive deformation failure process. It can help us to understand the mechanical behavior of internal crack propagation and rock failure mechanism in the progressive failure process. Two parameters，crack radial area strain and crack axial strain，are introduced based on the crack volumetric strain method in this paper. By conducting uniaxial compressive tests on five kinds of rocks，the proposed method and acoustic emission monitoring method are used to identify the stress thresholds of the five kinds of rock under uniaxial compression. According to the slope of the crack radial area strain curve when the rock breaks，the rock types are classified into type I and type II rocks. The testing results indicate that the stress threshold can be well identified by using the axial crack strain and radial area strain of the crack for type I and type II rocks respectively. It can accurately identify the stable crack development stress ，crack instability development stress ，and the crack connectivity stress of rock. Meanwhile，the results are similar to the stress threshold identified by the acoustic emission method. The mismatching phenomenon of stress thresholds identified by volumetric strain method，crack volumetric strain method and acoustic emission method is explained reasonably，the consistency of stress thresholds identification results using stress and strain data and acoustic emission data in the process of rock compression is realized.

Based on 3D printing and physical model test method，artificial irregular columnar jointed rock mass specimens with seven different dip angles were prepared by one-shot moulding method where brittle green wax was chosen as 3D printing material to print irregular columnar joint network in which white cement slurry was poured. Conventional triaxial compression tests were performed on these artificial irregular columnar jointed rock mass specimens，from which their strength and deformation anisotropy and typical failure modes were derived. From the experimental results，it showed that the strength，deformation modulus and lateral strain ratio of irregular columnar jointed rock mass exhibited moderate or high anisotropy and confining pressure had significant effect on its mechanical anisotropy. Furthermore，the typical failure modes of irregular columnar jointed rock mass under conventional compression condition had been classified into four kinds，that was，Y-shaped conjugate shear failure along columnar joints，splitting failure along both intact column and columnar joint，shear slip failure along columnar joint and axial splitting of intact columns. Finally，the influence of 3D printed green wax irregular columnar joint network on mechanical behaviors of irregular columnar jointed rock mass was analyzed.

The measured ground stress on the left bank of Shuangjiangkou Hydropower Station reached 37.82MPa. From the analysis of the spatial distribution law of ground stress and the spatial stress test results of rock mass in the technical construction stage，the ground stress has certain regional differences. From the macro qualitative analysis of the magnitude of ground stress in the cavern layout area and the angle between the tunnel axis and the direction of ground stress，the rock burst of the main and auxiliary power houses is the strongest and the destruction phenomenon is the most serious，and the rock burst damage on the upstream side is greater than that on the downstream side，Numerical simulation and quantitative analysis also get the same conclusion. After excavation，based on a large number of on-site investigations and statistics on the development of rock burst，the spatial characteristics of rock burst were analyzed，and the macro spatial development law of rock burst was summarized. Among the three caverns，the rock burst in the main and auxiliary power houses was the most serious，followed by the tail dispatching room，and the main transformer room was the weakest. The damage degree of rock burst in the main power house was greater in the pilot tunnel，the spandrels and side walls on the upstream side of the first and second layers than that on the downstream side，while the third layer The rock anchor beam and side wall at the downstream side of the fourth layer are larger than those at the upstream side，and the macro investigation law is consistent with the previous results. Combined with the spatial distribution law of microseismic monitoring events，it is found that Shuangjiangkou Hydropower Station has typical spatial characteristics of rockburst. By summarizing these characteristics，the characteristics of small-scale regional rockburst can be preliminarily estimated，which has certain guiding significance for the refinement of ground stress and support of water diversion system in the future.

Hydraulic fracturing is commonly applied into shale gas exploitation industry. However mechanical mechanism of permeability under the fracturing has been unclear so far. In view of this，a statistical analysis and discussion of the mechanical mechanism by the laboratory experiments on the hydraulic fracturing cracks propagation，and the activation of bedding plane of shale specimen are presented in the study. It is demonstrated that the microscopic observed micro-cracks could not be produced by fluid pressure as the latter generated compressive stresses on the places of micro-cracks. On the contrary，the micro-cracks are produced by tensile stress concentration in front of the propagating fracture. This implies that the bedding plane reactivation is caused by fracture propagation along the plane. An analysis of micro-crack lengths shows that shale exhibits anisotropy in fracture toughness with the resistance to fracture propagation parallel to bedding planes being as twice a small as compared with the resistance to fracture propagation in the directions normal to bedding planes.

Biased-pressure tunnel is a typical form of tunnel in engineering，and its cross-section shape can affect the mechanical performance of the lining structure and the stability of the surrounding rock-soil. Aiming at the shape optimization of biased-pressure tunnel，in this paper the investigation on optimal tunnel shape is performed based on the rational arch axis. According to its engineering characteristics，the FEM model of biased-pressure tunnel with initial circular shape is established，and the bending moment and axial force of tunnel lining are calculated to obtain the lining eccentricity. In optimization iterations，the lining axis is gradually approximated to the rational arch axis by repositioning the lining axis continuously，and the optimal cross-section of the biased-pressure tunnel is finally obtained when the eccentricity is approaching its minimum. On this basis，the mechanical performance of the lining and the stress distribution in rock-soil around the tunnel are studied. The results show that this tunnel with the optimal cross-sectional shape has a great improvement in mechanical performance of the lining，which is mainly subjected to axial compression and can fully utilize the compressive properties of the concrete material. Besides，the stress distribution in surrounding rock-soil is smooth and the stress concentration can be avoided effectively.

The understanding of coal fracture behavior and loading rate effect on the scale of laboratory test is also helpful for the analysis of the fracture propagation law and the regulation of the fracturing fracture formation in the process of underground reservoir fracturing. The fracture behavior and loading rate effect of lignite semi-circular samples taken from Erlian Basin in Inner Mongolia were studied by SHPB test in the direction of bedding parallel，bedding vertical and bedding oblique direction. The results show that the fracture process of lignite samples from Erlian Basin under impact includes elastic deformation，fracture and unloading stages；with the increase of loading rate，the fracture strength and fracture toughness of the samples increase linearly，and the effect of loading rate on the fracture behavior of lignite in the study area is very significant. With the increase of loading rate，the degree of sample breakage intensifies. The bedding parallel type specimens is mainly in the form of bedding shear fracture under impact，and the vertical bedding and oblique bedding specimens are mainly in the form of impact crushing and form a large number of fractures. Coal fines is unfavorable for coalbed methane drainage. The fracture orientation of lignite under impact is basically consistent with the loading direction. Meanwhile，with the increase of the loading rate，the energy dissipation of the lignite fracture process in the study area increases linearly. The above shows that the fault behavior of lignite in Erlian Basin is deeply controlled by the loading rate. Finally，this paper proposes a control program for coal reservoir fracturing mode based on loading rate effect. Among them，the deep stress-controlled lignite reservoir fracturing is mainly used to increase the loading rate step by step from near to far，while the shallow lignite reservoir controlled by natural fractures is fracturing，and the fracturing is from near to far. The pumping procedure of decreasing the loading rate step by step makes the fracturing fractures mainly controlled by stress in the initial stage，and turned into the main control by natural fractures in the later stage. The effect of“staged”fracturing is to form main fractures in the near-wellbore zone，complex fractures in the middle and complex fractures at the far end.

Based on the independently developed coal(rock) pressure water immersion test device，combined with Scanning Electron Microscope and XTDIC 3 D full-field strain measurement system，we carried out the Brazilian split test of four groups of coal samples under 0，3 and 5 MPa pressure water immersion(immersion time is 10 d)，and studied the tensile characteristics of coal samples under pressure water immersion. The results show that the uniaxial tensile strength and post-peak axial stress decline trend with the increase of water immersion pressure，and the plastic failure of coal samples increases at the same time. At the early stage of loading，the distribution of the maximum principal strain field of the coal sample is almost irregular. With the increase of the water immersion pressure，the strain isometric lines of the coal sample strain gradually become dense. The coal sample first produces stress concentration at the original macroscopic crack，forming the deformation localization zone. With the axial stress loads to the peak stress，the macro master control crack is generated in the middle of the coal sample，and the strain isometric line gradient increases，and gradually close to the control of the main macro tensile crack deformation localization zone area contraction. At the same time，the evolution of displacement mismomentum in the deformation localization zone of coal samples is related to the development，expansion and penetration of the localization zone，and the displacement mismomentum in the deformation localization zone at the stress peak increases with the increase of the water-immersion pressure. The increase of water immersion pressure intensifies the water-coal rock action，destroys the initial energy storage structure of coal samples，its pre-peak energy storage capacity gradually decreased，and the mechanical energy loaded to the coal samples by the test machine decreases. Compared with the dry coal samples，the average mechanical energy under 0，3 and 5 MPa pressure water immersion decreases by 29.93%，44.90% and 57.14%，respectively. As the immersion pressure increases，coal fracture morphology from relatively flat to "honeycomb" shape，measured its average porosity，pore maximum area and probability entropy，indicating that the internal damage of coal samples increased，which leads to its uniaxial tensile strength，and the tensile characteristic deterioration effect enhancement. The above research results are of great significance to the reasonable retention and stability guarantee of coal(rock) column in coal mine.

In recent years，a large number of landslides blocking rivers induced by earthquake and extreme climate occurred frequently，seriously threatening the safety of people?s life and property and infrastructure in the basin. However，the current understanding of the formation mechanism of landslides blocking river is still not comprehensive，which is worth further research. Based on typical cases of landslide blocking river since the 21st century，the general process and main characteristics of landslide blocking river are deeply analyzed. In view of large scale of landslide blocking river，the necessary conditions for the formation of landslide river blocking and the meso-mechanism of fluid-soil coupling in the process of river blocking are revealed from three aspects of theoretical analysis，physical model test and numerical simulation respectively. Hazard assessment of landslide blocking river mainly include distinguishing of river blocking，evaluation of the landslide dam stability and assessment of dam failure hazard. However，there is no unified standard form for distinguishing of river blocking. The material and structure characteristics of the landslide dam have not been considered in the stability evaluation of the dam. The physical mechanism of the mathematical model for predicting peak discharge of dam-break flood still needs to be further revealed. Based on the limitations of the current research，four key issues that should be focused on in the study of landslide blocking river are prospected：(1) Study on the accumulation structure and prediction model of landslide damming induced by debris-avalanche considering the influence of fragmentation；(2) Physical and numerical simulation methods of evolution process of landslide blocking river；(3) Hazard assessment method of landslide blocking river based on the state of landslide material；(4) Early warning factors and key monitoring and early warning technologies for landslide dams. The results provide useful reference for the prediction of landslide blocking river disaster chain，emergency geological treatment of landslide dam and regional disaster prevention and mitigation planning in mountain and ravine region.

The near-surface surrounding rock of tunnels in southwestern China has been in a hypo-high temperature and high humidity environment for a long time，which is prone to thermal and humid diseases. The water vaporization of surrounding rock cracks produces crack air-vapor pressure，which affects the crack initiation strength and propagation path. In order to study the crack initiation mechanism of hypo-high temperature fractured air-vapor compression rock，based on the fracture gas state equation and linear elastic fracture mechanics theory，the rock crack initiation model and the calculation formula of stress intensity factor under the combined action of far-field compressive shear stress and fracture air-vapor compression are deduced. The crack initiation criterion and the equation of crack initiation strength and critical crack initiation angle under the combined stress state of compressive-shear are proposed. The circumferential stress distribution law and crack initiation characteristics at the crack tip are analyzed，and the influence of various parameters on the crack initiation strength and critical crack initiation angle are discussed. The results show that the fracture air-vapor pressure is jointly controlled by temperature and gas saturation，and temperature is the dominant factor. At low and high crack inclination angles， is much larger than ，and the crack initiation is mainly tensile，and the air-vapor pressure plays a promoting role. The crack inclination angle is between 15° and 75°，the value of and  are close. The crack initiation is mainly compression-shear，and the crack air-vapor pressure inhibits the crack initiation. The circumferential stress at the crack tip is sinusoidally distributed with the change of polar angle，and the maximum circumferential stress decreases with the increase of fracture air-vapor pressure，and the critical crack initiation angle increases with the increase of fracture air-vapor pressure. With the increase of temperature and gas saturation，the initiation strength decreases and the critical initiation angle increases. The theoretical calculation is in good agreement with the laboratory test results. The research results provide a theoretical basis for crack propagation analysis of tunnel near surface surrounding rock under hypo-high temperature and high humidity environment.

The intactness index of rock mass is the main parameter to determine the rock mass basic quality index BQ. At present，the“national code”only considers the number of joints and does not reflect the influence of dig angel and filler on . In this paper the effects of joint number，dip angle and filler on P-wave velocity of rock mass are investigated through the P-wave velocity test of granite samples with different joints. Considering the influence of dip angle and filler，an evaluation method for quantitative calculation of is proposed. The is regarded as a random variable with normal distribution，and its relationship with is obtained through the estimated values of with a guarantee rate of 90%. The result is very similar to the recommended value of “national code”. From the aspect of rock mass mechanical strength evaluation，the applicability of this method in practical rock mass evaluation is verified. The proposed method can provide a reference for the quantitative evaluation of the rock mass basic quality BQ.

In view of the serious dynamic disaster caused by the hanging of the rock pillar and roof during the mining of steeply-inclined and extremely-thick coal seam group，by using comprehensive theoretical analysis，field monitoring and other research methods，a mechanical model of the rock pillar and roof was built and the expression representing the bending moment and energy evolution law of the rock pillar and roof was deduced，the law of the bending deformation and energy evolution of the rock pillar and roof with the increase of mining depth was studied，the impact risk of different coal seams was evaluated，the induced impact mechanism was analyzed，and the anti-impact strategy and scheme were put forward. The results show that：(1) with the increase of mining depth，the bending deformation and energy accumulation of rock pillar and roof in steeply-inclined and extremely-thick coal seam group show nonlinear acceleration changes，and the bending deformation is more serious and the energy accumulation is higher in the mining stage. After the mining depth of 300 m，the energy increased sharply，and it was determined that the mining depth was the critical value of impact disaster. (2) Rock pillar pry and energy accumulation in B1+2 coal seam mining are smaller than those in B3+6 coal seam mining，and the impact risk in B1+2 coal seam mining is smaller than that in B3+6 coal seam mining. Compared with the rock pillar，the bending deformation of B3+6 coal seam roof is more serious and the energy accumulation is higher. (3) Determine the impact prone area as the superposition position of each stress peak. The sources of impact force include static load and dynamic load，and static load is the basic condition of impact，which is the energy storage function. The dynamic load is the induced impact，and the dynamic and static loads act together to induce the impact. (4) The key to prevent scour is to weaken the static load. It is necessary to strengthen prevention and control in B3+6 coal seam mining. The key prevention and control objects are rock pillars and roof， and the key prevention and control areas are the superposition positions of stress peaks. After the rock pillar and roof blasting was carried out on site，the static load level was effectively weakened after the coal was injected with water，and the anti-scour control effect was good. The above research results provide a scientific basis for the prevention and control of rock burst disasters in steep and thick coal seam mining in Wudong Coal Mine.

To reveal the shear strength difference of direct shear and bi-direct shear of structural plane and better evaluate the stability of layered rock mass，four Barton standard profile lines were selected to establish the structural plane models，and the rock-like samples containing one or two structural planes were made for shear tests. Based on the experimental results，the differences of peak shear strength( ) between single and double structural planes under three normal stresses( ) as well as the reasons for them were analyzed quantitatively. Based on the particle flow simulation results，the empirical formula of the peak shear strength for the double structural plane with the same JRC value was derived firstly. Then，the combined joint roughness coefficient ( ) was introduced to characterize the roughness of double structural plane with different JRC values. After the relevance between low roughness coefficient( )，high roughness coefficient( ) and was found，an empirical formula of the double structural plane with different JRC values considering the effect of  and to peak shear strength was established. Eventually，an empirical formula of the peak shear strength for double structural plane with any combination of JRC values was established via a comprehensive analysis of the evolution of peak shear strength with the mixed joint roughness coefficient. During the process that the growth trend of  -JRC curve increases with the increase of  ，the peak shear strength of double structural plane increases with the sum of JRC value when   is constant. Importantly，when the sum of JRC value is small，the peak shear strength of double structural plane with different JRC values is controlled by the side of lower JRC. However，the control effect decreases with the increase of the sum of JRC value，and thus the peak shear strength is gradually controlled by the side of higher JRC.

The multi-stage hydraulic-acid stimulation has become a key technology for permeability enhancement of geothermal reservoirs. In this study，the artificial fracture samples were artificially produced from the granite collected from Gonghe Basin，China，direct shear tests were performed on each fracture sample for six times after different acid immersion treatments. The results show that the shear behavior of rock fractures changes from brittle failure to ductile failure after during the previous five shear tests. The peak shear strength significantly reduces，but the residual shear strength does not change significantly during the previous five shear tests. After acid immersion treatment，the shear behavior of the rock fractures still exhibits ductile failure，but the peak and residual shear strengths significantly reduce. Based on wear theory，a formula of calculating shear-induced damage volume of rough fractures is developed，which is in good agreement with the experimental results. After acid immersion treatment，the differential dissolution rates of different minerals increase the fracture surface roughness，but the acid corrosion is further promoted by the large number of micro-cracks in the fracture surfaces induced during the previous five shear tests，which in turn degrades the mechanical properties of fracture asperities. Because the latter mechanism is dominant，rock fractures exhibit weakening effects after acid immersion treatment.

The spatial distribution and roughness characteristics of the structural planes have a significant impact on the anisotropic properties of jointed rock mass，which are important factors affecting the engineering stability. To study the influence of roughness on the mechanical properties of anisotropy of jointed rock mass，a rough joint network model was established based on the fractal theory. Numerical compression tests considering varied rotation angles and confining pressures were conducted. Comparison of failure modes and acoustic emission characteristics between the fractal fracture model and the linear fracture model were performed. The results show that significant differences exist consideing the joint roughness，comparing the mechanical behaviours，AE charateristics and size effect. Meanwhile，a quantitative analysis of the anisotropy of the compressive strength and shear strength parameters of the jointed rock mass was conducted. The anisotropic Mohr circles based on the least squares principle were discussed. The biaxial compressive strength and shear strength parameters of the jointed rock mass show slight anisotropy. Differences of peak stresses，Mohr circle shapes and shear properties were verified under different loading directions. Influence of the geometry of the discrete faractures on the strength and fracture patterns was verified. The results will provide a novel and scientific model for characterizing the fractured rock mass and give a good reference for analysing the anisotropic shear behaviours of jointed rock masses.

The new seepage channels were added to the original transparent triaxial compression servo-control test system to conduct the visual triaxial compression tests of rock under seepage-stress coupling conditions. Triaxial compression tests were conducted on sandstone under the conditions of different seepage states and different seepage pressure differences to verify the reliability and necessity of the test system. The experimental results show that the mechanical properties of rock were weakened under the water pressure at both ends of the rock，and the two seepage limit state tests cannot represent the influence of all the seepage-stress coupling conditions on the mechanical properties of rock. And the large strain data points in the radial strain field gradually gather near the fracture surface with the increase of axial stress，and the deformation localization occurs. Therefore，the mechanism of seepage field on rock mechanical properties can be better studied by the mechanical tests under complex seepage-stress coupling.

In this contribution，we developed a numerical model for describing a coupled effect of stress-strain-sorption-diffusion-permeability on evolution of deformation and internal stress in coal seams during coalbed methane(CBM) recovery，assuming that coal behaves as a poroelastic media. Three-dimensional discrete fracture network of a coal seam was first constructed using COMSOL，consisting of natural and hydraulic fractures. On this basis，we quantitatively determined the development of subsidence and internal stress of the constructed coal seam，using our proposed numerical model，given proper parameter values. The results indicate，that：(1) subsidence mainly occurred on the top of the coal seam，which was located near the both sides of fractures，resulting in a vertical dislocation along the natural fractures and the dislocation rate can be up to 2 mm per year. (2) Both the intermediate and minimum principal stresses in the area near the natural fractures decreased significantly，while the minimum principal stress exhibited a transition from compression to tensile，which can reach to 3 MPa. In addition，parameter sensitive analysis was performed to determine the main factors controlling CBM recovery induced dislocation and stress evolution. By the research we suggests that during CBM recovery，deformation and stress drops around the natural fractures should be paid more attentions as it may cause local damage and even earthquake，though the CBM recovery induced subsidence is relatively small.

At present，there are few studies on the horizontal load behavior of large-diameter rock-socketed single pile foundations for offshore wind turbines, and it is difficult to consider the impact of rock foundation breakage on the bearing performance of foundations in traditional numerical simulations, so the failure mechanism is still unclear. In order to solve the difficulties in accurate numerical simulation and analysis of horizontal bearing characteristics of large rock-socketed monopile foundations for offshore wind turbines. Firstly，according to the geological survey data of the offshore wind field，the finite-discrete element method(FDEM) was used to study the mechanical properties of the weathered granite in the rock seabed. Then we improved the algorithm of mesh remeshing and contact detection in the FDEM program，and realized the embedding of cohesion elements in the pile-rock interface. Based on the self-developed FDEM preprocessing program GiD–Y2D，we established the FDEM plane strain model of the pile-rock interaction of offshore wind turbines. Finally，we analyzed the failure mode of the pile-rock contact interface of the offshore wind turbine and the horizontal bearing characteristics of the rock-socketed pile foundation. The results show that the failure modes of rock seabed can be divided into tension failure stage and compression failure stage. When the tension zone is damaged，the horizontal displacement of the single pile is about 1.37×10－4 times the diameter of the pile. After local failure，the lateral capacity per unit length decreases by about 30%，and the lateral capacity stiffness decreases to 2/5 of the original. When the compression zone is damaged，the horizontal displacement of a single pile is about 2.17×10－3 times the diameter of the pile. After the failure，the foundation reaches the ultimate lateral capacity. Due to the brittle failure of the rock mass around the pile，the lateral capacity per unit length of the pile decreases rapidly by about 62.7%.

Long-term strength is crucial for studying the time-dependent mechanical properties of deep engineering rock masses. Based on the anisotropic time-dependent fracture characteristics of hard rock induced by true triaxial stresses，the steady-state creep rate method for determining the long-term strength of Jinping marble was proposed in this paper，in which the stress when the steady-state creep rates in the directions of and    are equal is determined as the long-term strength of the rock. In addition，the long-term strength of Jinping marble was calculated by the isochronous curve method，the transition creep method，and the steady-state creep rate method，respectively. The results show that transition creep method is not suitable for solving the long-term strength of Jinping marble under true triaxial stress. It also shows that the isochronous curve method can only obtain the long-term strength range of Jinping marble under true triaxial stress and its accuracy relies on the stress increment selected by multistage creep loading. And the long-term strength of Jinping marble under true triaxial stresses can be accurately obtained by steady-state creep rate method，which proves the correctness and reliability of this method.

To explore how the strength，deformation and failure characteristics of layered rock mass change with bedding angle and specimen scale，3D printing technology is adopted to produce bedding specimens of different angles and scale，and the uniaxial compression test and Brazilian splitting test are conducted. Besides，the combined finite-discrete element method(FDEM) is employed to testify the laboratory results. Three findings are reported as below. Firstly，compressive strength，elastic modulus and deformation modulus have a“U-shaped” curve relationship with bedding angle，respectively. When bedding angle is 90°，the values of these three factors are obviously higher than the results under the condition of other bedding angles. Tensile strength has a negative correlation with bedding angle. When bedding angle is 0°，the tensile strength is apparently lower than that of homogeneous specimens. Secondly，the specimens subjected to uniaxial compression load feature tension- dominated failure(with the bedding angles of 0°and 90°) and shear-dominated failure(with the bedding angles of 45°and 67.5°). As bedding angle increases，failure mode gradually transits from ductile failure to brittle failure. The specimens suffering from Brazilian splitting test show tension-dominated failure and the fractures assume linear-type(with the bedding angles of 0° and 90°) and arc-type(with the bedding angle of 45°). Lastly，as specimen scale gradually increases，compressive strength tends to be stable and peak strain decreases gradually and tends to stabilize around 0.005. Tensile strength and specimen scale have an“inverted U-shaped”curve relationship，elastic modulus and deformation modulus fluctuates around 2.1 GPa，and these three factors don?t show stable tendency.

The main reason for the deterioration of rock mechanical properties under water-rock chemical reaction is that the rock cement is dissolved in an aqueous solution to varying degrees. This paper proposes an improved discrete element three-dimensional cementation contact model based on reasonable assumptions. In the size-dependent discrete element contact model，the soluble bond proportion and the bond width coefficient related to chemical damage degree are introduced to consider the dissolution of water solution to randomly distributed soluble cement. The model is implanted into the discrete element to calibrate sandstone’s conventional and micro parameters after chemical damage. The uniaxial and triaxial compression tests of sandstone with different chemical damage degrees are numerically simulated. The stress-strain curves and micro information are obtained，and the influence of chemical damage on mechanical properties is analyzed from the macro and micro points of view. The simulation results show that the compression test of chemical damage sandstone is dominated by tensile shear and tensile bending failure，while the water-rock chemical reaction has a great influence on the bond number of tensile failure and little influence on compressive bending failure. With the increase in chemical damage degree，the number of macro cracks on the rock surface increases，and there is a changing trend from one through a crack to multiple cracks. The energy storage limit of rock is greatly reduced，and the energy storage limit and growth rate of elastic energy are reduced to varying degrees. The energy storage limit of bonded elastic energy tends to be stable after falling to a certain extent. With the increase of soluble bond proportion，the sensitivity of peak strength and elastic modulus of chemical damage rock to bond width increases. When the proportion is 50%，the rock strength is less than 1/3 of intact rock，while the mechanical properties decrease significantly.

As a natural porous medium，coral reef limestone owns a complex pore structure，which lays great influences on its macroscopic mechanical and seepage properties. In order to accurately characterize the pore structure of coral reef limestone and explore the structural features of its pore network，this paper constructed digital 3D models and equivalent pore network models by combining  CT scanning and 3D reconstruction techniques for coral reef limestone samples collected from Hainan. The spatial features of pore structure distribution of the two coral reef limestones were analyzed，and statistical analyzes were performed on pore structural parameters including porosity，pore throat size and coordination number. A topological network model of the internal pore structure was constructed to carry out complex network analyses，whereby the overall and individual property indicators of the core reef limestone pore network were obtained. The key nodes affecting the network characteristics were identified. The study shows that the coral reef limestone has high porosity and non-homogeneity，and the distributions of both pore throat size and coordination number conform to the log-normal distribution rule. The internal pore network is well connected with a high degree of compactness，and there exist some key pore nodes with high importance，which have a greater influence on the overall network characteristics.

Aiming at the long-term deformation of the surrounding rock in the underground rock engineering under the high-stress condition，the triaxial creep tests and real-time acoustic emission monitoring were performed on the marble samples that taken from a high-stress underground engineering by the MTS815 Flex Text GT and PCI–II Acoustic Emission Monitoring System. And the creep mechanical properties and acoustic emission characteristics of the marble were investigated. The results demonstrate that the stress-strain curve comprises five stages throughout the entire process of multi-level creep of marble showing the brittle failure property. Meanwhile，it is noticed that the lateral deformation has more significant creep features than the axial deformation. The acoustic emission shows a clear stress response characteristics and time-dependent effect，which matches the typical three stages of creep. The acoustic emission ringing count rate and energy rate linearly increase in the initial and steady creep stages，while exponentially increase during the accelerated creep stage. Furthermore，the time point corresponding to the change of acoustic emission characteristics is earlier than the strain acceleration point，which can be used to predict the creep damage in marble. Based on the characteristics of internal microcrack development in marble defined by AF and RA values，the creep damage of marble is the result of a combination of tensile and shear damage.

Due to the difficulty of obtaining large size real cores in the well and the limited number，the analog method is usually used to make rock-like material specimens in the laboratory to simulate the actual coal rock，so as to carry out the hydraulic fracturing test. In order to accurately explore the mechanical properties of rock-like material specimens，cement，quartz sand，plaster，and coal powder were used as raw materials to cast specimens with different combinations and ratios. Through uniaxial compression test to explore the effects of cement-plaster ratio，sand-binder ratio and coal powder on the mechanical properties and destruction form of the specimens. The results show that the compressive strength of cement-quartz sand-plaster specimen increases with the increase of cement-plaster ratio，and decreases with the increase of sand-binder ratio；The sand-binder ratio has a more significant effect on the compressive strength of the specimen. When the sand-binder ratio and the cement-plaster ratio are 0.33，0.75，1 and 0.5，1，and 2，respectively，the specimen compressive strength decreases from 23.19 MPa to 19.14 MPa；The peak failure time of the same type of rock-like material specimens is advanced with the increase of the cement-plaster ratio，and the internal damage degree of the specimen intensifies. The cement-plaster ratio increases from 0.5 to 2，and the acoustic emission ringing count increases by about 5.1 times，macroscopically，it is manifested as increased cracks in the specimen，Compared with quartz sand and plaster，Coal powder will aggravate the development of cracks in the specimen and weaken its mechanical properties，so as to ensure that the ratio remains unchanged，and replacing quartz sand and plaster in cement-quartz sand-plaster specimens with coal powder，the maximum acoustic emission ringing count of the specimen at peak failure is 1.5 times and 2 times that before replacement，respectively，and the surface changes from more regular cracks to more developed through cracks；After the specimen is damaged，the cracks are distributed in reverse arches on both sides of the specimen surface，which is the X-shaped conjugate inclined plane shear failure，and the maximum principal strain is located at the crack of the specimen，which the cracks on the surface of the specimen are basically consistent with the phenomenon described by the acoustic emission monitoring data.

Taking the rock anchor beam on the third floor of the underground main power house on the left bank  of Shuangjiangkou hydropower station，the world?s highest dam，as the research object，the rock anchor beam surrounding rock failure is characterized by fan shape，plate crack and uplift deformation to the free face. A typical rock burst failure type of super high stress and brittle granite-“sector plate crack uplift type”concept is proposed. It is found that the number and angle of new cracks increase with the increase of the groove width when the groove depth remains unchanged；When the width of the groove is constant，the number and angle of the new cracks increase with the depth of the groove；When the depth and width of the groove remain unchanged，the overbreak increases the number and angle of the new cracks. The concepts and calculation methods of ultra-high stress，brittle rock fracture angle and maximum rock fracture angle are put forward. By using the concept of maximum fracture angle，the concept design of natural formation of rock anchor beam excavation and the refined excavation method of rock anchor beam can be used for reference to similar underground hydropower plants and Sichuan Tibet railway projects in the future.

With the deep underground excavation of hard brittle surrounding rock，time-lagged rockburst frequently happens in high geostress environment. Considering the influence effect of unloading stress path，triaxial time-lagged deformation failure test and scanning electron microscopy(SEM) have been performed under different unloading confining pressure rates and different unloading confining pressure degree. A series of tests is to study the time-lagged deformation and failure characteristics of sandstone under different influence factors. The results indicated that the total time of time-lagged deformation stage exhibited increasing and then decreasing trends with the confining pressure unloading rates，and it exhibited decreasing trends with the confining pressure unloading levels. However，compared with unloading confining pressure with constant axial stress tests，the total time of time-lagged deformation stage under unloading confining pressure with loading axial stress tests was longer. Based on this and combining macroscopic observations，the time-lagged deformation failure was more significant under unloading confining pressure with loading axial stress tests. Additionally，through SEM image observation and EDS analysis，it indicated that mineral debris on fracture surface were closely related to the total time of time-lagged deformation stage under different influencing factors conditions. Finally，according to renormalization group theory，the staged damage constitutive model based on logistics function was established，and the effectiveness and rationality of model was proved. These research results can be used as a theoretical reference to study the development process of time-lagged rockburst.

In order to investigate the influence of different roughness and filling gap width on the seepage characteristics of rough single fracture of limestone. Combined with three-dimensional scanning and engraving technologies，samples of limestone fractures with different joint roughness coefficient(JRC) and undulation( ) were prepared. Seepage characteristics experiments of rough single fracture of limestone under triaxial stress condition were conducted to study the influence of different JRC and filling gap width( ) on fracture seepage flow，equivalent hydraulic aperture and permeability. The results show that under different JRC and Be conditions，the evolution process of seepage flow，equivalent hydraulic aperture and permeability of fractures with seepage time has three stages: rapid decline，slow decline and stabilization. With the increase of JRC or decrease of ，all seepage parameters gradually decrease. The seepage flow and permeability of fracture has quadratic function and logarithmic function relationship with JRC respectively，and the minimum value of square root( ) is 0.985 5. There is a quadratic function between steady seepage flow and ，and the minimum value of is 0.974 5. The equivalent hydraulic gap width of the fractures with different filling gap width decreases with the increase of seepage time. According to the quantitative relationship between the initial seepage flow and various parameters，the seepage behaviour of the fracture after stabilization can be predicted.

It is an effective method for soft and fractured rock mass to obtain the deformation parameters of surrounding rock by in-situ conventional bearing plate tests. However，it is be restricted by the site condition and the results is significantly size effective due to the structural surrounding rock mass. Relying on the diversion tunnel of a hydropower station in Liangshan Prefecture，conventional bearing plate test was modified，which can be performed with multisize bearing plates flexibly and efficiently during construction of tunnel，to reasonably determine the deformation characteristics and parameters of lamellar and fractured slate. The results show that：(1) the dark black and extremely thin carbonaceous slate was poor integrity，the stability of surrounding rock was very poor in addition with groundwater softening；due to the compaction of stratification plane，weakly water softening and relatively strong block，the gray sandy slate was better stability. (2) the rock mass was significantly anisotropic on account of the extremely fractured macro-meso structural plane，which also resulted in the significantly size effective of tested deformation parameters by conventional bearing plate tests. The secant and envelope modulus， and ，decreased with the increasing of diameter of bearing plate D. (3) in order to diminish the size effect of bearing plate，it is suggested to perform tests with multi-sized bearing plates on the same position for the modified in-situ small tonnage and efficient bearing plate test. Then exponential function can be used to fit the measured deformation modulus and further determined the more reasonable deformation parameters of surrounding rock mass. Based on it，the proposed deformation modulus of carbonaceous slate and sandy slate of the diversion tunnel were = 42.9 MPa， = 60.8 MPa and  = 238.3 MPa， = 691.3 MPa，respectively. This study provides an efficient method for weak and fractured surrounding rock mass to rapidly determine the deformation parameters， and provides a reference for the similar excavating tunnel in the future.

In order to investigate the bonding performance between the distributed optical fiber sensor and rock mass of Grotto Temple，a series of surface bonding tests were carried out using two different types of Grotto Temple sandstone and three kinds of adhesives with different components. The interface bonding strength，deformation characteristics and failure mode were analyzed，and then the influence mechanism of adhesive and rock type on the bonding strength was discussed. The results indicate that：(1) for all different adhesives，the bonding strength between sandstone with compact structure and distributed optical fiber sensor was higher than loose sandstone?s. For the same sandstone samples，the bonding performance of cyanoacrylate adhesive is better than polyurethane and epoxy resin. (2) It was found that the expression of the load-displacement curves of distributed optical fiber sensor and sandstone bonding tests is very similar to the expression of the axial force of the bolt based on the nonlinear fitting. (3) The failure modes of the interfaces of distributed optical fiber sensor and sandstone are divided into fiber sensor sheath broken，rock surface rupture，fiber sensor and adhesive layer separation and fiber sensor and fixture separation. (4) The penetration range and bonding shape of adhesive are related to adhesive viscosity，curing time and porosity of rock sample. Bonding strength has a significant effect on the failure modes of the interfaces of distributed optical fiber sensor and sandstone. The research results can provide a technical support for the application of distributed optical fiber sensor in the field of rock mass stability monitoring of Grotto Temple.

It is difficult to characterize the degree of stress disturbance during underground tunnel excavation，especially to quantify the change of principal stress direction. Based on the distance between the stress tensors，the index of stress disturbed degree(SDI) is proposed，which can quantitatively describe the change of the magnitude and direction of the principal stress. Taking Mine-by Laboratory in Canada as an example，the variation characteristics of principal stress and stress disturbed degree during step-by-step tunnel excavation were studied by numerical simulation. The results show that during tunnel excavation，the degree of stress disturbance at different axial positions of the tunnel is related to the excavation face and has significant spatial distribution characteristics. The stress disturbance is more serious only in the area affected by the excavation face. The stress disturbed degree along the radial direction of the tunnel decreases gradually from the tunnel wall to deep into the surrounding rock. The depth of stress disturbance increases when the excavation surface is close to the monitoring section and decreases gradually after the excavation face leaves. SDI at the wall is the largest，but the disturbed degree of the principal stress direction is not necessarily the largest，and the peak value position varies with the position of the excavation face. The proposed index has a clear physical meaning，which can well describe the variation characteristics of the magnitude and direction of the principal stress. The disturbing degree of the magnitude and direction of the principal stress have the same dimension and are easy to compare. The position with a large degree of stress disturbance corresponds well to the fracture of surrounding rock，which can provide reference for the stability of surrounding rock in deep rock engineering and provide a scientific basis for support design.

Fracture behaviors of rock samples with discontinuities is difficult to quantify and predict. In this study，machine learning method is used to predict mode I fracture toughness and crack propagation mode of rock samples with an infilled fracture. Firstly，the feasibility of the method was verified by comparison of results of notched semi-circle bend test and numerical simulation based on discrete element method，and the dataset which filtered the outliers was generated by a lot of random numerical simulations. Multiple techniques were used in the preprocessing of dataset. Then four classical machine learning models were established. The optimal parameters of the models were obtained by grid search and five-fold cross-validation method. Multiple indicators and graphs were used to verify the prediction performances of the models. Comparison between the four models shows that Multilayer Perceptron(MLP) is the optimal machine learning model. The sensitivity of the MLP model was analyzed and the stability is good. The MLP model was applied to the test samples，which has high prediction accuracy. And a user interface program was developed. The data driven model of rock mechanics based on the machine learning method reduces time cost compared with the traditional test and numerical simulation，which provides a new method and thought for solving traditional rock engineering problems.

The coupling mechanism of rock mass fracture seepage and normal stress is an important basis for hydro-mechanical coupling analysis of fractured rock mass. The normal stress and permeability curves of rock mass fracture often show nonlinear characteristics，and there is a significant difference with the increase of normal stress：the fracture permeability decreases sharply under low stress conditions，while the fracture permeability decreases slowly under high stress conditions. At present，traditional mathematical models can not describe the difference of fracture permeability under low stress and high stress. Therefore，based on two-part Hooke?s Model(TPHM)，the fracture closure is divided into two parts：the hard closure and the soft closure. The influence of the soft closure and the hard closure on the deformation and seepage characteristics of rock mass fracture under low stress and high stress conditions is considered respectively，and the coupling mathematical equation of rock mass fracture seepage and normal stress is established. Three groups of hydro-mechanical coupling test data under loading condition of granite fractures were used to verify the theoretical model. The results show that the theoretical model is basically consistent with the experimental data，and better than the negative exponential model，power function model and logarithmic model，indicating the effectiveness of soft and hard aperture in quantitatively describing the nonlinear hydro-mechanical coupling characteristics of rock mass fractures. The main control effect of soft aperture on rock fracture permeability under low stress state is revealed.

To study the dynamic response characteristics of the tunnel lining reinforced by multi-anchor piles in landslide area，the optimal seismic performance of multi-anchor piles and the dynamic damage evolution mechanism of tunnel lining are discussed. In this study，a series of shaking table tests were carried out by using energy-dissipation springs as the damping optimization device of anchor head. The acceleration and dynamic strain data of the multi-anchor piles and tunnel lining under different probability levels of earthquake were obtained. By analyzing the time domain characteristics of tunnel lining acceleration and dynamic strain，the dynamic response characteristics and damage evolution law of the tunnel lining are elucidated. Based on the response data of multi-anchor piles peak acceleration(PPA)，peak dynamic strain(PTS) and plastic effect coefficient(PEC) at typical measuring points of the tunnel lining，the dynamic damage evolution mechanism of the tunnel lining is revealed. The results show that：(1) Under the loading of seismic waves with different amplitudes，the tunnel lining exhibits a spatial continuous effect of damage evolution from local to global，and the tunnel lining acceleration amplification factor( ) exhibits the distribution characteristics of prominent reactions on the unoptimized side；(2) limited to whether the optimization of multi-anchor piles and seismic wave loading effect，there are regional differences in the damage site and damage pattern of the tunnel lining，and the invert and vault are prone to become the weak areas in the seismic design of the tunnel；(3) The evolution of earthquake cumulative damage catastrophe of tunnel lining is elucidated from the perspective of plastic deformation characteristics based on the PEC variation potential；(4) The multi-anchor piles with energy dissipating springs as damping devices weaken the seismic wave energy of the pile to a certain extent，but are prone to produce a greater dynamic response during the reciprocal action of the seismic inertia force compared to ordinary multi-anchor piles.

To effectively simulate the random propagation of hydraulic fractures and accurately describe the tensile，shear and mixed damage behaviour in naturally fractured shale reservoirs，a finite element method-discrete fracture network model coupled with seepage-stress-damage is established. The viscosity-dominated analytical model is utilized to verify the accuracy of the model. The effect of the approach angle，cluster spacing，horizontal stress difference and pumping rate on multi-cluster fractures propagation is investigated and the interaction mechanism between hydraulic and natural fractures is discussed. The results show that：(1) There are three main interaction mechanisms between hydraulic fracture and natural fracture: crossing through natural fracture，opening natural fracture and arrested by natural fracture；(2) With the increase of the approach angle，the natural fracture is easier to deflect the propagation direction of hydraulic fracture and even prevent the propagation of hydraulic fracture；(3) The increase of cluster spacing，horizontal stress difference and pumping rate can promote the uniform growth of multi-cluster fractures to a certain extent. The model can more accurately describe the propagation behaviour of multi-cluster fractures，and provide a new method for numerical simulation of hydraulic fracturing in naturally fractured shale reservoir.

When supercritical CO2 is used as drilling fluid to drill shale formation，wellbore stability is one key issue. The interactions between supercritical CO2 and shale are complex. When oil and gas resources are developed with supercritical CO2，the research is little for the influence of supercritical CO2 adsorption on wellbore stability of shale formation. In this study，for the multi-physics coupling process of supercritical CO2 seepage，the influence of supercritical CO2 adsorption on deformation and mechanical properties of shale are considered. Based on the transport and thermodynamic properties of supercritical CO2，the multi-physics coupling models of wellbore stability for supercritical CO2 drilling are established for considering adsorption effect. The finite element numerical method is used to analyze the distribution of formation temperature，pore pressure and stress with time and space in the process of supercritical CO2 drilling，meanwhile，the influence of supercritical CO2 adsorption on wellbore stability of shale are studied. The results indicate that the adsorption-induced solid deformation has little effect on formation temperature and pore pressure，whereas，the stress are influenced greatly by adsorption. The stress value will increase when adsorption-induced strain is considered，while the stress value will decrease when adsorption-enhanced elastic modulus is considered. If adsorption-induced strain is neglected，the risk of wellbore collapse will be slightly underestimated，whereas，if adsorption-enhanced elastic modulus is ignored，the risk of wellbore instability will be greatly underestimated. When the depth exceeds 1 200 m，the collapse pressure is greater than wellbore pressure，the wellbore collapse will occur. This paper can provide theoretical guidance for wellbore stability analysis for supercritical CO2 drilling.

Based on the unified phase field model, the mechanism of crack propagation and damage evolution of rock materials is analyzed. A numerical uniaxial compression test of high-strength gypsum is conducted，and combined with the DIC test results，the analysis presents the following facts. First，the crack propagation of a single crack rock sample is the result of the continuous accumulation of damage caused by the concentration of stress in rock. Second，a threshold value of crack driving force at crack initiation exists，which is related to the rock material，and can be used to evaluate the difficulty of rock crack initiation. Third，the curve of crack width and length has a high correlation with the rock failure process，so the rock failure can be divided into four stages：elasticity，crack initiation，macro crack occurrence，crack propagation and penetration. Fourth，the failure type of rock gradually changes from tensile failure，which is dominated by tensile crack，to shear failure，which is dominated by shear crack with the increase of crack dip angle. Fifth，the smaller the angle between the crack and the loading direction，the closer the crack initiation position is to the compressive stress concentration area at the end of the rock precast crack，resulting in the stronger restriction of the crack driving force，and the greater the load is required to reach the threshold of crack initiation，which explains the phenomenon that the rock microcrack with large crack inclination develops more fully and the crack length curve is smoother from the mechanism. Last，the crack propagation path of the double crack rock sample is restricted by the surrounding stress field，and whether the rock bridge area has through failure depends on the stress type of the rock bridge area. In conclusion，this study shows that the unified phase field model can be used to study the mechanism of crack propagation and the damage evolution of rock.

Intermittent failure behavior often occurs in coal samples under uniaxial compression. The damage assessment of complex failure process is helpful to clarify the energy evolution accompanying the process，which is of great significance to the early warning and prevention of mining engineering disasters. To explore the damage evolution in the process of intermittent failure，the standard coal sample was subjected to conventional uniaxial loading，and its deformation characteristics were monitored in real time. It is found that when severe damage occurs to the local crack cluster of coal sample under external load，the axial load drops suddenly，and the radial deformation increases sharply. The closer the load is to the peak strength，the faster the growth rate is. The damage originates from the development and expansion of internal cracks，resulting in the rapid expansion of the sample along the radial direction. Thus a method of damage characterization based on radial deformation increment was proposed，which is more applicable to the nonlinear damage evaluation of heterogeneous media such as coal. Comparing the damage evaluation based on radial deformation increment with acoustic emission and the development and expansion evolution of fracture，the highly consistent evolution process of the above three proves the reliability of damage evaluation based on radial deformation. The results show that，the cracks are locally clustered or interconnected along the direction of the maximum principal stress to form large-scale fracture in the intermittent failure stage. The radial expansion of the sample causes local severe damage，which leads to a sharp increase in its damage variables. Intermittent failure behavior originates from repeated local severe damage in different areas of the sample. The zigzag stress-strain curve and the nonlinear damage evolution accompanying this process indicate that the damage mutation phenomenon exists in heterogeneous media such as coal under external load.

The coal pillar-backfilling composite structure jointly bears the stress in the pillar-side backfilling. The interface between the coal pillar and backfilling is the weakest area. The instability of the interface can induce the failure of the composite structure, and fiber-reinforced polymer(FRP) could be applied to enhance interface stability. Three sets of Brazilian splitting laboratory tests were carried out on coal-backfilling composite structures(without FRP wrapping for reinforcement，FRP wrapping for 1 circle and FRP wrapping for 2 circles) with different interface angles. The DIC and acoustic emission monitoring techniques were used to study the characteristics of the specimens. The results show that：(1) When the load on the interface exceeded the interface tensile or shear strength，interface cracks occurred；the load exceeded the tensile strength of the coal or backfilling element，resulting in tensile cracks；after FRP wrapping，the energy accumulated by the increased load on the specimen cannot be released through the interface separation，resulting in a concentrated damage zone on the coal element. (2) As the interface angle in-creased，the interface crack changed from tensile failure to shear failure. After FRP wrapping，the development of interfacial cracks was significantly reduced，and the brittle characteristics of the specimens were reduced. (3) The strain zone of the specimens expanded along the loading or interface direction. FRP wrapping increased the tensile and shear strength of the interface. The stability of the interface increased. (4) FRP wrapping increased the AE counts and reduced the damage degree of the specimens when the specimens were damaged. (5) The number of AF-RA data points for the specimens increased after FRP wrapping. It indicates that the specimens were able to experience more crack generation and expansion before instability. This study is expected to provide theoretical guidance for the prevention of splitting instability in pillar-side backfilling.

In the process of reservoir construction and management，the rock would be under an alternating state of drying-wetting owing to the seasonal rainfall and storage and drainage of the reservoir. In addition，the rock is commonly subjected to dynamic loadings such as earthquake，blasting，and tunneling，which have significant influences on the stability of the reservoir. To investigate the effect of drying-wetting cycles and dynamic compression on the mechanical behaviors of rocks，the dynamic test of sandstone undergoing different drying-wetting cycles was carried out using the split Hopkinson pressure bar system. The influences of wetting and drying cycle conditions on the rock dynamic characteristics，energy dissipation law，and fractal dimension of fragmentation were systematically analyzed. The microstructure of sandstone was observed using the scanning electron microscope to reveal the physical deterioration mechanism. The results show that drying-wetting cycles would weaken the ultrasonic characteristics，static and dynamic mechanical characteristics of sandstone. With the increase of drying-wetting cycles，the static uniaxial compressive strength，elastic modulus，P-wave velocity，dynamic compressive strength，dynamic elastic modulus，and dissipative energy of sandstone decrease，while the fractal dimension of rock fragmentation increases. In addition，with increasing strain rate，the dynamic compressive strength，elastic modulus，peak strain，dissipative energy，and fragmentation fractal dimension of sandstone show an increasing trend. The macroscopic analysis results illustrate that the deterioration of sandstone under drying-wetting cycles is caused by the significant change in microstructures，i.e.，pores increasing，particle breakage，cementation weakening，and crack spacing expanding. The research findings in this study could provide a theoretical basis for the design and construction of reservoir de-risking and reinforcement under drying-wetting cycles.

In the actual deep displacement monitoring，the“r”type inclinometer curve usually takes the maximum displacement mutation point as the position of the sliding surface. However，this method is easily affected by the sensor spacing and the observation scale of the coordinates，and there are problems such as unclear definition of the sliding surface position and numerical uncertainty. In order to overcome these shortcomings and improve the accuracy of slip surface identification，this paper divides the landslide into three parts：sliding body，sliding interval and inactive body，and proposes the theoretical generalization model of landslide with the characteristics of the curve based on the characteristics of the“r”type inclinometer curve. This paper analyzes the characteristics and laws of displacement rate of soil at different depths in the process of landslide movement. It is found that the sliding surface will cause the mutation of displacement rate at this position to form two groups of curve clusters which correspond to sliding body and inactive body respectively. The displacement rate has a step stage along the depth direction，which can accurately distinguish the range between the sliding body and the inactive body，and is conducive to identifying the position of the sliding surface. Further，this paper summarizes the distribution types of discrete points in the step stage，and constructs different mathematical calculation methods. This method can accurately determine the position of sliding surface and improve the reliability and accuracy of deep displacement monitoring，which has great practical value.

The stability analysis problem of the composite earth dam with multiple weak joints composed of soils in different consolidation states will be encountered when adopting the method of filling earth from upstream deposited silt which be common in northwest loess hill region for the reinforcement of the reservoir. Based on the upper bound theorem of plastic mechanics limit analysis，two destruction mechanisms have been constructed with the slope of filling earth from deposited silt as the research object that according to the structural type of composite earth dam and the operating characteristics of the reservoir. Then，considering the influence of the static water pressure of the dam slope and the pore water pressure in the earth caused by the dynamic change of the reservoir water level，the safety factors have been analyzed and optimized by using strength reduction technology and SQP method. On this basis，taking the reservoir reinforcement project of filling earth from upstream deposited silt from Ningxia southern mountainous area as a case，the upper limit solution of this paper and the numerical simulation results of Optum-G2 and Geo-studio have been compared and analyzed. The research results show the upper limit solutions of the two destruction mechanisms are very close to the numerical simulation values and the relative error of the safety factor is less than or equal to 5.7% when the reservoir water level drops sharply from the design water level to the dead water level. The failure mode change with the width of the weak interface of silt and new fill. The slope safety factor shows a nearly linear relation with the weak interface width variation that the safety factor will be increased about 2.7% as add every 1 m of interface width when the dam slope fails in translation. And the slope safety factor also has a nearly linear relation with the slope ratio variation that the slope safety factor will be reduced about 8.6% as add every 0.1 of slope ratio under rotation-translation combined failure conditions. The upper bound solution that presents in this paper is simple，convenient and suitable for the type of problem.

The jacking force changes dynamically due to the continuous decreasing height of overburden soil in vertical tunneling method. To study the variation law of jacking force in vertical jacking construction，combined with the spherical cavity expansion theory and shear failure principle，two calculation models of jacking force，spherical cavity expansion-sliding friction model and shear failure-sliding friction model，are proposed，and a new and complete system of theory of jacking force is established. Based on the actual monitoring data of drainage engineering in Beihai and the test data，the correctness and suitability of theoretical calculation formula is verified. The results show that the theoretical calculation method in this paper is in good agreement with the measured results and test data.

At present，the academic circles have different understanding of the gas migration path after gas injection operation and reinforcement mechanism of air-booster vacuum preloading. In view of this problem，several groups of air-booster vacuum preloading model experiments were designed to explore the gas migration path after gas injection operation and reinforcement mechanism of air-booster vacuum preloading. By comparing and analyzing the test results of drainage，pore water pressure，settlement，vane shear strength and water content，the following conclusions were obtained：in the process of treating ultra-soft soil by air-booster vacuum preloading，the gas moves to the soil surface first after gas injection. At the same time，by considering the real-time pore pressure change during gas injection operation，it is speculated that the main mechanism of gas injection operation to improve the drainage and consolidation effect of soil is as follows：during gas injection operation，the rise of pressurized gas will drive the deep pore water to migrate to the shallow layer，so that the deep pore water can be quickly absorbed and discharged by the drainage plate in the shallow layer，so as to improve the reinforcement effect of deep soil and the overall drainage efficiency.

The results of classical ultimate bearing capacity calculations developed by Terzaghi et al. and the results of in-situ plate load tests show that the ultimate bearing capacity of coralline sand foundations can exceed up to 4 000 kPa，which differs significantly from its general understanding of low bearing capacity. Since the coralline sand has high internal friction angle，based on the classical limit equilibrium calculation model，the influence of internal friction angle on the size of general shear failure range is analyzed to clarify the mechanism and conditions of high ultimate bearing capacity of coralline sand. Through the in-situ plate load test of the coralline sand sites at the Port of Sudan before and after treatment，the phenomenon of high ultimate bearing capacity of coralline sand foundation is confirmed. The analysis and results indicate that under the high internal friction angle and general shear condition，the coralline sand foundation could have a larger shear failure depth and wider shear failure range，which result in higher ultimate bearing capacity according to the ultimate equilibrium equations. General shear failure is a prerequisite for high ultimate bearing capacity of coralline sand foundation，and the foundation treatment of layered dynamic compaction and layered dredging can meet this condition. The actual shear failure range of coralline sand foundation may differ greatly from that of classical ultimate bearing capacity calculation method，and the ultimate bearing capacity calculation formula applicable for coralline sand is necessary for further study.

In order to investigate the effect of root content on the shear strength of the root-soil-rock composite  and analyze the variation rules of the cohesion and internal friction angle of the root-soil-rock composite，the root distribution rule of the Malus halliana Koehne with the increase of buried depth was obtained according to the field profile，and the rock content along the shear surface was quantified based on the digital image analysis technique，and the modified direct shear tests were conducted under four different normal stresses(50，100，150，200 kPa) combined with the distribution characteristics of the root area ratio and rock content in the field. The results showed that：(1) The shear strength of the soil-rock mixture increases with the increase of rock content，the increment shows a change rule of“slow-steap-slow”for increasing rock content under the same normal stress. (2) The shear strength of the root-soil composite and the root-soil-rock composite can be improved with the root of the Malus halliana Koehne，and both of them increase first and decrease later with the increase of root area ratio and the maximum value is reached when the root area ratio is 0.2%. (3) When the rock content is 30%，the root content has little effect on the shear strength of root-soil-rock composite，and the roots mainly play a role in crack resistance and reinforcement. When the rock content is 50%，the shear strength increases significantly when the root area ratio is 0.2%，and the rocks around the roots produce arc slip and some rocks subject to extrusion damage during the shear process. (4) The shear strength of the root-soil-rock composite conforms to Coulomb strength criterion，and the internal friction angle is mainly controlled by the rock content，i.e.，it increases with the increase of the rock content. The cohesion is mainly controlled by the root content，i.e.，it increases firstly and then decreases with the increase of the root area ratio，and the maximum value is reached when the root area ratio is 0.2%. The results are helpful to further explore the interaction between the root system and the soil-rock mixture，and provide some theoretical basis for the plant reinforcement of the soil-rock mixture shallow slopes.

The Guangdong-Hong Kong-Macao Greater Bay Area is the most economically developed regions in China and has an important strategic position in national development. However，due to the extensive distribution of marine soft clay，it has brought many challenges to engineering construction. Based on the theoretical and numerical analysis and monitoring data in Nansha，Guangzhou. This paper discusses the cause of failure and deformation characteristics of a utility tunnel excavation. The research results show that the failure mode of this excavation belongs to the combined damage of kicking and dumping，mainly kicking damage. The horizontal displacement of the pile top points to right. The reasons for this phenomenon are that the distance between the vehicle and the excavation is too small and trucks are overload. Due to the support by the cross brace，the horizontal displacement and vertical displacement of steel sheet pile are still relatively small and stable before the excavation is destroyed，and cannot effectively reflect the safety of the excavation. Therefore，attention should be paid to the monitoring of the surrounding ground settlement and deformation of the deep horizontal displacement. The maximum deep horizontal displacement is at a depth of 6–11 m，which is below the bottom of the excavation. This deformation is different from that in the non-soft soil area. The deformation of the deep horizontal displacement reaches a depth of about 18 m，indicating that the deformation of the supporting pile caused by the excavation reaches 12 m below the bottom，and the ratio of deformation depth to the excavation depth is 3. The damaged location of the excavation should be backfilled in time.

当地下结构穿越液化土层，场地液化可能会对结构的动力响应产生影响。基于已开展的局部液化夹层场地地下结构离心机振动台模型试验，建立二维局部液化场地土–结构动力相互作用的有限元模型。计算模型考虑砂土液化后易产生剪切大变形的特点以及饱和两相介质与结构接触的非线性特性，通过与试验结果对比，验证数值模型的正确性，进一步分析土–结构接触非线性特性对结构地震响应的影响。研究结果表明：孔压的增长是由于地震作用的累积效应；近场砂土由于土结相互作用的存在超孔压比上升的更快，且由于土与结构变形不协调而更易消散；对于局部液化夹层场地，土结捆绑接触会放大结构的地震响应，随着地震动强度的增大，砂土层液化程度提高，场地非线性特性增强，土结刚度差异增大，土结界面采用捆绑接触所带来的误差越大，故应合理考虑土结界面接触非线性问题。

Complex hydro-mechanical coupling properties of unsaturated cohesive soils appear under external loading. In this paper，an elastoplastic two-surface model which can describe the hydro-mechanical coupling character has been formulated based on elastoplastic theory，adapting the suggested plastic hardening criterion extended to unsaturated soils，considering the arbitrary stress path variation in the(p?-q-s) space. The model reflects the coupling effect of stress and suction through mean skeleton stress and modified suction. Two groups of loading surfaces and bounding surfaces were used to describe the plastic yield induced by coupled hydro-mechanical during loading and/or dry-wetting processes respectively. The fully coupled description and solution of the hydro-mechanical characteristics of unsaturated clay are realized. Compared with the conventional triaxial shear test results of unsaturated clay，the model can better simulate the stress-strain characteristics of unsaturated clay under coupled hydro-mechanical.

To solve the problem that there is no calculation method for the design of artificial freezing wall in seepage stratum in the relevant design codes，the coupled temperature field-seepage field calculation of seepage strata was realized by using COMSOL Multiphysics finite element software. The calculation results were verified by indoor tests. The development law of the temperature field with the change of seepage velocity，freezing pipe spacing and refrigerant temperature is studied by numerical calculation. The results show that when the groundwater seepage velocity is greater than the critical velocity，the freezing wall does not cross-loop and there is a good linear relationship between the critical velocity and the freezing temperature gradient. The calculation method of the critical velocity is proposed. Referring to consolidation theory，a dimensionless closure factor is proposed. The closure factor - seepage velocity curve under different freezing pipe spacing and refrigerant temperature conditions is well fitted by using the growth function curve. On this basis，the calculation method of artificial freezing closure time under seepage conditions is established. The research results can provide reference for the design and calculation of artificial freezing construction in seepage stratum.

The soil thermal resistivity is an important parameter to determine the thermal properties of geotechnical materials. In order to accurately predict the soil thermal resistivity，a prediction model of soil thermal resistivity is established based on the multivariate distribution model. The prediction model is verified，analyzed，and compared with the traditional empirical model to clarify the performance of the proposed model. The results show that the established multivariate distribution model can accurately predict the soil thermal resistivity. With the increase of input parameters，the prediction accuracy of multivariate distribution model is significantly improved，the is increased from 0.719 5 to 0.899 5，the value is reduced from 1.077 8 to 1.037，and the value is reduced from 0.367 5 to 0.212 1. The model has the best prediction accuracy. The prediction accuracy of multivariate distribution model is obviously better than that of traditional empirical model. For the prediction of thermal resistivity of different soil types with significant differences in engineering properties and complex sedimentary environment，it is suggested to select different types of multivariate distribution models according to the data to evaluate the thermal resistivity.

In order to address the frost heave problem of railway embankment in cold regions，an artificial heating system and a distributed heating scheme based on ground source heat pump technology was proposed. A full scale half-width embankment test platform of class I single-track railway was built，and the temperature field of embankment was monitored. The effective heating influence radius of the heating system was defined as the longitudinal range within which the embankment freezing depth is controlled below the harmful threshold within a certain time，and a prediction method of the effective heating influence radius was put forward. The test results show that the heating temperature of the heat pump system can reach above 30 ℃，the coefficient of performance(COP) can be as high as 6.9，and both the effective heating capacity and COP gradually decrease with time. The temperature rise amplitude of soil decreases with the distance from heat source，and the temperature increment rate in the vertical direction around the heating pipe is greater than that in the longitudinal direction along embankment. Artificial heating has a significant inhibitory effect on the impact of cold wave weather. At the end of the test，the freezing depth within 2.0 and 3.0 m on both sides of the heating system is controlled within 30 and 35 cm respectively. The heating influence radius is proportional to both heating time and the harmful freezing depth threshold. When the harmful freezing depth threshold is 30，35，40 cm，the effective heating influence radius is 0.44 m，0.64 m，0.83 m respectively when heating for 5 days；and if the heating influence radius needs to reach 1.5 m，the heating time should be 53，41，35 d respectively. In application，the longitudinal layout spacing of heating systems along the embankment should be reasonably designed according to the harmful freezing depth threshold and the required treatment time after the appearance of frost heave，so as to ensure the merge of thawing range of adjacent heating systems within the specified time.

The field test of bridge pile foundation passing through the cave 25.5 m in height were carried out to explore the load transfer mechanism of the pile foundation of highway bridge passing through the super-large cave. The vertical bearing characteristics and load transfer mechanism of the pile foundations passing through the caves with different height were studied when the backfilling method was used to treat the karst cave. The maximum value of negative friction caused by consolidation settlement of the backfill materials and the proportion of its distribution range were proposed. The results show that：when the load is small，the axial force of the pile foundation passing through the super-large cave increases in a certain range，and the negative friction resistance of pile foundation is generated after the cave is backfilled. The vertical ultimate bearing capacity of the pile foundation decreases with the increasing height of karst cave. When the height of the cave increases from 5 m to 30 m，the maximum negative friction resistance under ultimate load increases by 102.7%，and the proportion of its range increases from 0% to 24%. When the height of the cave exceeds 25 m，the negative friction range accounts for more than 20%. If the backfill method is used to treat the cave，it was recommended that when the height of the cave was 5 m to 10 m，the influence of negative friction resistance could be ignored. When the cavern is very large，the adverse effect of negative friction resistance on vertical bearing characteristics of pile foundation should be considered. When the cavern height is 15，20，25，30 m，it is recommended to calculate the bearing capacity of the piles according to negative friction resistance within 0.092 H，0.145 H，0.213 H，0.240 H(cavern height) from the cavern top.

The dredged soil was solidified with soda residue，ground granulated blast furnace slag and carbide slag. Dynamic triaxial tests were conducted under different solidifier contents，water contents，confining pressures and consolidation ratios. The dynamic stress-strain relationship，dynamic elastic modulus，damping ratio and empirical formula were studied. The results show that the increase in solidifier content and the decrease in water content contribute to increase the dynamic elastic modulus and decrease the damping ratio，but when soda residue content is higher than 40%，the dynamic elastic modulus decreases and the damping ratio increases. The increase in confining pressure and consolidation ratio help to increase the dynamic elastic modulus and decrease the damping ratio. From the viewpoint of dynamic properties and economy，the appropriate contents of soda residue，ground granulated blast furnace slag and carbide slag are 35%，10% and 4%，respectively. The dynamic stress-strain relationship of solidified soil is in accordance with the Hardin-Drnevich model. The Darendeli model can be used to describe the dynamic elastic modulus and stain relationship for solidified soil. On the basis of Hardin?s maximum dynamic shear modulus formula，the modified empirical formula for solidified soil is proposed. The maximum dynamic elastic modulus of the solidified dredged soil is 86.96–157.23 MPa，which is close to or exceeds the lower limit of general solidified soil. The solidified soil can be used as general engineering filler and exhibits good dynamic deformation properties.

Under the condition of measuring suction，a series of three-dimensional swelling pressure tests，three-dimensional swelling pressure tests for controlling vertical deformation and three-dimensional swelling pressure tests of expansive soils after drying-wetting cycle were carried out on intact expansive soils by using an improved unsaturated soil true triaxial apparatus to investigate the variation characteristics of three-dimensional swelling pressure. The effects of matric suction，vertical deformation and drying-wetting cycle on the three-dimensional swelling pressure were studied. The results show that three-dimensional swelling pressure increases with a decrease in suction and gradually stabilizes. In this process，there is an obvious critical suction value ，and which increases with an increase in initial suction. When ，the growth rate of three-dimensional swelling pressures and the decrease speed of suction are faster，when ，the growth rate of three-dimensional swelling pressures and the decrease speed of suction are slower. Three-dimensional limit swelling pressures increase with an increase in initial suction，and the vertical limit swelling pressure increases faster with the initial suction. The small change of vertical strain will lead to the rapid decrease of three-dimensional swelling pressures. With the continuous development of vertical strain，the three-dimensional swelling pressures gradually decrease and tend to be stable. Under different vertical strain conditions，there is no significant difference in the attenuation of the two horizontal swelling pressures，but the attenuation of the vertical swelling pressure is always larger than that of the horizontal swelling pressures. When the vertical strain is constant，as the initial suction increases，the attenuation of the three-directional swelling pressures increase. After drying-wetting cycles，the moisture content and suction of expansive soils decrease with an increase in the number of cycles，and the three-dimensional limit swelling pressures of intact expansive soils increase with an increase in the number of drying-wetting cycles.

In order to discuss the construction effect and bearing capacity abnormality of dynamic driven pile in mudstone foundation，the stress and end resistance of pile were monitored based on low temperature sensitive fiber optic strain sensor and full section spoke pressure sensor of pile end. The variation of internal force and end resistance of test pile with penetration depth was clarified. The bearing characteristics of each test pile were analyzed by vertical compressive static load and high strain variation measurement. The influence of hammer energy on the driving effect of prefabricated piles was studied by using hammer number and hammer weight. The influence mechanism of dynamic pile driving on the strength of mudstone around pile was revealed by means of the distribution of hammer number of standard penetration test of pile edge before and after pile driving. The test results show that the time-history curve of the end resistance measured by different sensors is consistent during the pile driving process. The ultimate bearing capacity of test pile obtained from static load test is higher than that obtained from variable test，and all test piles show the characteristics of end bearing pile. The strain variation range of heavy hammer driving pile is obviously larger，and the pile driving energy is related to the vertical bearing capacity of pile. High energy driving pile will lead to aggravated damage of mudstone structure and reduce the bearing capacity of driven pile. Compared with before pile driving，the hammer number of mudstone around pile after pile driving decreases obviously，which indicates that dynamic pile driving causes damage to mudstone structure around pile to different degrees in macro view. It can be considered that mudstone structure damage is one of the main reasons for the reduction of dynamic drive pile bearing capacity.

The soft soil with high compressibility has typical geometric and material nonlinearities. However，the solutions for the large strain consolidation of soils with vertical drain，in which the time-and depth-dependent well resistance are considered，are rarely reported in the literature. Considering the nonlinear compressibility and permeability，and the time-and depth-dependent well resistance，a large-strain nonlinear consolidation model of soils with vertical drain is established in the Lagrangian coordinate system，in which the excess pore water pressure serves as a variable. The numerical solution for this consolidation model is derived by the finite difference method. The numerical results are compared with the existing small-strain analytical solution and large-strain numerical solution for the vertical drain ground to verify the correctness of the presented method. Then，the consolidation behavior analysis of the large-strain consolidation of soils with vertical drains under different conditions is performed，and the results show that the space-time variation of well resistance significantly reduces the consolidation rate. The time-dependent well resistance has a greatly impact on the consolidation rate. However，the final settlement of vertical drain ground can not be affected by the time-and depth-dependent well resistance. The consolidation rate under the uniform distribution of initial effective stress is less than that under the linear distribution. The final settlement values of the vertical drain ground in the two cases are not equal，and the differences between the two settlement curves are greatly large. The assumption that the discharge capacity flowing into the vertical drain at any depth are constant results in the faster consolidation rate，and overestimates the actual consolidation rate of the vertical drain. The promotion effect of vertical flow on the consolidation of vertical drain decreases as the ratio of the length to the influential radius of drains increasing.

Considering the well resistance varies with time and depth during consolidation of soils with vertical drain，a consolidation model of soils with vertical drain under vacuum combined surcharge preloading was developed based on the equal strain assumption，aiming at the situation of multi-stage surcharge preloading and single stage application of vacuum preloading and incorporating the different decay modes of horizontal coefficient in smear zone. The analytical solution of the consolidation model is obtained by using the homogeneous boundary conditions and solving ordinary differential equations. Analytical solutions are verified for validity by comparing them to the existing analytical solution and the field data，and the consolidation behavior is investigated by the parametric analysis. The results indicate that the variable well resistance theory can theoretically support the assumption that the final negative pressure reduces linearly with the increase of depth. The degree of vacuum attenuating with depth at a specific time will grow as the attenuation coefficient of drainage capacity increases and the final drainage capacity of the vertical drain decreases. The larger the attenuation coefficient of drainage capacity is，the slower the consolidation rate of the vertical drain is. The larger the final drainage capacity is，the faster the consolidation rate of the vertical drain is. Based on the consolidation model in this paper，an appropriate vacuum loading value is selected before construction to carry out vacuum combined surcharge preloading，which can simultaneously take soil consolidation rate and economic benefit into account.

Based on the dynamic governing equations of unsaturated soils，and considered the effects of matrix suction，three-phase(solid phase，liquid phase and gas phase) compression characteristics，as well as the effect of the degree of saturation on dynamic shear modulus，the vertical vibration response of end-bearing pile in layered unsaturated soils was investigated. Combined with Euler-Bernoulli bar theory，the vertical vibration response model of an end-bearing pile in layered unsaturated soils under three-dimensional axisymmetric harmonic load was proposed. The governing equation of unsaturated soils was derived by utilizing the Laplace transform，potential function and operator decomposition methods. In the frequency domain，the analytical solution of the dynamic impedance of the pile top was solved through the pile-soil coupling condition and impedance function transfer method. The semi-analytical solution of the pile top was obtained by using the inverse Fourier transform and convolution theorem in the time domain. After the correctness of the solution was verified，the effect of modulus ratio and thickness ratio of layered soil on the dynamic response of pile was investigated. The results shown that when pile in upper-hard-and-lower-soft two-layered soil，the fluctuation of dynamic stiffness and damping of the pile decreased，and the admittance had the phenomenon of“small peak sandwiching large peak”，the peak became larger. Conversely，the admittance presented“large peak sandwiching small peak”. Compared with the upper and lower homogeneous soil，the reflected signal at the pile bottom was obviously weakened. The thicker the soft soil layer was，the greater the oscillation amplitude of the pile. When the thickness of the soil layer was different，there was a reflection signal at the interface of the soil layer in the velocity time history curve，and the reflection position was closely related to the boundary position of the soil layer.

Uplift pile foundation is widely used in underground space in coastal areas. The uplift characteristics of structures are the engineering indices that must be evaluated. Many scholars have studied the prediction methods of the uplift bearing characteristics of pile foundation based on CPTU test，and proposed the empirical formula for specific site. These empirical formulas have great limitations when applied to local sites. Therefore，a method based on CPTU is proposed to predict the uplift bearing characteristics of pile foundation in soft soil layer in coastal areas. Model tests were carried out to study the influence of the length-to-diameter ratio of pile foundation and the roughness of pile surface on the bearing characteristics of the uplift pile. The mechanical mechanism of the uplift pile in the working process and the interaction mechanism between pile and soil were studied. The results of the model tests were further verified by numerical simulation analysis. The CPTU test was carried out in the model bucket. The empirical relationship between pile side friction resistance and CPTU test parameters was established. The empirical relationship proposed in this paper is applied to a soft soil engineering site in Wenzhou. The error between the uplift pile bearing capacity calculated by the proposed method and the measured results is only 7.7%，lower than other methods of calculation results. The calculation accuracy is higher and it has a high applicability in predicting the bearing characteristics of uplift pile in soft soil.

In order to investigate the effects of freeze-thaw cycles on the shear performance at silty clay-concrete interface，direct shear tests, electron microscopy tests and scanning electron microscope(SEM) tests were conducted at the silty clay-concrete interface with different numbers of freeze-thaw cylces. The meso-damage of soil in the interface area and the macroscopic shear performance of the silty clay-concrete interface were studied by combining macroscopic and microscopic views. Furthermore，the effects of normal stress and the initial moisture content of soil were also considered. At the microscopic level，the pores of silty clay in the interface area gradually develop and the cohesiveness gradually decreases under the action of freeze-thaw cycles. The porosity of the silty clay in the interface area is positively correlated with the number of freeze-thaw cycles and the contact mode of the soil changes from surface-surface to point-surface and point-point. While at the macroscopic level，the shear properties of the silty clay-concrete interface gradually deteriorate under the action of freeze-thaw cycles. The shear strength，cohesion and internal friction angle of the silty clay-concrete interface decrease with the increasing freeze-thaw cycles but the shear stress-displacement curves are still in strain-hardening form. The deterioration mechanism of the shear performance under the action of freeze-thaw cycles was investigated in combined macroscopic and microscopic views on the basis of laboratory testing results. It provides a reference for the study of the shear performance at the soil-concrete interface in seasonal frozen soil regions.

Through a self-improved temperature controlled triaxial apparatus，the consolidated undrained shear test after freeze-thaw(FC mode) and unconsolidated undrained shear test after freeze-thaw(CF mode) were carried out on the remolded muddy soil，considering the combined effects of freeze-thaw confining pressure，times and freezing temperature. The effects of the above factors on the mechanical properties of the remolded soil were analyzed. The influence mechanism of freeze-thaw confining pressure on the mechanical properties was revealed by SEM. The results show that the confining pressure during freeze-thaw cycles reduces the number of pores in the soil after freezing and thawing，thus reducing the weakening effect of freezing and thawing on the strength and modulus in FC mode. While in CF mode，the freezing and thawing confining pressure increases the pores in the soil after freezing and thawing，and the excess pore pressure in the soil increases without dissipation during the soil thawing，which further aggravates the weakening effect of freezing and thawing on the strength and modulus of the soil. In two modes，the lower the freezing temperature and the greater the number of freeze-thaw cycles，the more obvious the strength weakening effect of the muddy soil, and the higher the excess pore water pressure when the soil is yield. The variation rules of mechanical index deterioration coefficient of the muddy soil with freeze-thaw confining pressure，times and freezing temperature can be described by a growth function.

Open caisson foundations are widely used in the construction of various large structures，and the inclination of an open caisson is one of the most important indexes of its sinking attitude. Accurate prediction of the inclination is conducive to ensuring the sinking safety and steady of the open caisson and preventing potential construction risks. Based on two ensemble learning techniques，bagging and boosting，the random forest algorithm and XGBoost framework are applied for the inclination prediction modeling. The monitoring data of the structural stress at the bottom of the open caisson are used to predict the longitudinal height difference and transverse height difference. The reliability of the prediction model was verified by applying it to the super-sized open caisson foundation of the main bridge pylon in the Changtai Yangtze River Bridge Project，and the proposed model was compared with the prediction models applying other single machine learning algorithms. Then，the important parameters of the ensemble learning model were analyzed to study their influence on prediction accuracy. The results show that the prediction model in this paper can accurately predict the longitudinal height difference and transverse height difference and reasonably determine the inclination of the open caisson foundation. With fast operating speed and strong practicability，the proposed model has higher prediction accuracy than other single machine learning models. In addition，the prediction accuracy increases with the number of base learners and the maximum tree depth. The research results achieve the real-time prediction of the inclination of the open caisson foundation during the sinking process，which can provide an important reference for the monitoring of similar foundations.

Pressuremeter test and standard penetration test are two common in-situ test methods. Pressuremeter modulus，limit pressure，and standard penetration blows are the parameters to describe the mechanical properties of soil. The standard penetration test is a simpler and more cost-effective method. Pressuremeter test can obtain multiple mechanical parameters of soil，but the operation is complicated and expensive. This paper aims to reveal the characteristics of the loess silty clay pressuremeter modulus，limit pressure，and standard penetration blows with the change of natural moisture content of soil，and then to explore the experience expressions of pressuremeter modulus and limit pressure regarding to standard penetration blows. For this purpose，a railroad loess subgrade in Zhengzhou，Henan region was selected. 47 sets of pressuremeter tests and standard penetration tests were carried out. The empirical relationship among pressuremeter modulus，limit pressure，standard penetration blows，and natural moisture content of the soil was obtained by regression analysis. Hypothesis testing was carried out based on statistical analysis methods to verify the validity of the empirical relationship. The research results show that：with the increase of natural moisture content，the standard penetration blows，pressuremeter modulus，and limit pressure of loess silty clay in Zhengzhou area decrease linearly，respectively. When the natural moisture content increases to the plastic limit，the linear reduction rate of the standard penetration blows，pressuremeter modulus，and limit pressure all increase，respectively. A linear relationship between the pressuremeter modulus and the standard penetration blows is existed，and a linear relationship between the limit pressure and the standard penetration blows is existed. The hypothesis testing method F-test verifies the validity of empirical expressions for pressuremeter modulus and limit pressure. Compared with previous scholars' research，the empirical expressions obtained in this experiment has a higher coefficient of determination. The empirical expression fitting analysis method of the correlation between PMT and SPT used in this experiment is more applicable in loess silty clay. The research results provide a reference for the accurate and efficient acquisition of loess mechanical parameters in the survey and design of Zhengzhou loess area.

Large-scale direct shear tests were conducted to investigate the shear behavior of geogrid reinforced coarse grained saline soils under normal cyclic loading. The effects of static loading，the initial stress, cyclic amplitude，and cyclic frequency of normal cyclic loading on the shear behavior of two geogrid reinforced coarse-grained saline soils were studied. The results showed that the shear strength of the coarse grained saline soils could be improved by geogrid. The shear strength of geogrid reinforced coarse-grained sulfate saline soils was generally higher than that of geogrid-reinforced coarse-grained chloride saline soils. The shear strength of the geogrid reinforced coarse-grained chloride saline soils decreased with increase of frequency. However，there is no obvious correlation between the shear behavior and amplitude. The lower initial normal stress reduced the shear strength of geogrid reinforced coarse-grained sulfate saline soil. The smaller the frequency, the lower the shear strength of the geogrid reinforced coarse-grained sulfate saline soils，and the smaller the shear displacement of the peak strength.

The water content of solidified soil often changes periodically when it is used as artificial fill or subgrade material for resource utilization due to the rainfall and evaporation，ground water level fluctuation and so on. Study on the water retention characteristics of solidified soil has theoretical and engineering application significance. There are few researches on soil-water characteristic curve(SWCC) of solidified soil，the unified and clear SWCC of solidified soil within the full suction range has not been obtained until now. The law of water retention of solidified soil is not clear yet. The SWCC of solidified soil in full suction range under drying path and wetting were measured by pressure plate method，filter paper method and vapor equilibrium technique. In combination with scanning electron microscope scanning test and mercury pressure test，the SWCC and the water retention characteristics of solidified soil with different cement contents were analyzed. The results show that the pore distribution of the solidified soil becomes more uniform with the increase of cement content. With the increase of cement content，the large pores between soil particles and the total pore volume gradually decrease，resulting in a higher water retention capacity of the solidified soil. The SWCCs of solidified soil under drying and wetting paths show significant hysteretic behavior. With the increase of cement content, the surface roughness of soil aggregates increases due to the increase of hydration products. The hysteresis of SWCC becomes more obvious with the increase of cement content.