Based on the basic situation of rockburst prevention and control in China?s coal mines，the systematic research work and research results of rockburst for many years are combed and summarized. Then from the perspective of science，technology and management，the current situation，existing problems and development direction of rockburst prevention and control in China?s coal mines are analyzed and elaborated，hoping to play a role in promoting the improvement of rockburst prevention and control in coal mines in China. Among them，the analysis of the current situation of rockburst prevention and control includes rockburst occurrence mechanism and theory，rockburst risk evaluation and prediction，rockburst risk monitoring and early warning，rockburst prevention and control methods and technologies，rockburst laws and regulations and management system，and rockburst prevention and control effect. Besides，the existing problems including situation and standards of rockburst prevention and control，compound disasters，mine sesmic situation，mining roadway support，rockburst early warning，and rockburst monitoring technology are presented. Finally，the development directions are pointed out including strengthening basic research，reasonably determining the safety factor of rockburst design，developing the intelligence of rockburst prediction and control，implementing the integrated prevention and control of mine sesmic and rockburst disasters，perfecting the self-liberation mining technology of this coal seam，promoting the energy-absorbing hydraulic support technology of the whole roadway，and scientifically identifying rockburst mines.

Compared with the traditional fracturing technologies，the recently developed unlimited stage sliding-sleeve fracturing has been proved to have many advantages during unconventional development，including uniform stimulated reservoir volume and unlimited stages. However，the controlling factors of fracture propagation and induced stress distribution of the unlimited stage fracturing have not been fully understood. For this reason，taking a continental shale formation as an example，the multi-fracture propagation models are established based on the discrete lattice method. By comparing the fracture morphology of perforation fracturing and unlimited staged fracturing under different conditions(continental shale lithology，bedding dimensional parameters，and fracturing engineering parameters)，the qualitative and quantitative analyses can be performed concerning the fracture propagation law of unlimited stage fracturing. The research results show that the complex lithology，developed bedding interfaces，and the stress shadow effect are responsible for uneven fracture propagation regarding the multi-stage perforation fracturing. The bedding plane aperture，bedding density and vertical distance between bedding plane and wellbore would affect the propagation law of perforated fractures in continental shale formation，resulting in the difference in the stimulated volume of each cluster. In contrast，the unlimited stage fracturing can not only neglect the influence of the bedding interfaces，but also form more effective fractures and more uniform reservoir stimulated volume. The positive effect of unlimited stage fracturing can be strengthened depending on the lithology，cluster design，pumping rate and geometry of initial fractures. The stimulation effect of unlimited stage fracturing would get weakened under conditions of large cluster spacing，high pumping rate，and large radius of initial cluster. The magnitude and the distribution of the induced stress along the wellbore direction vary between the two technologies. For unlimited stage fracturing，the induced stress field of the former clusters will overlap and influence the following fracturing work. The research outcomes can provide theoretical support and technical guidance for the optimization of the fracturing technologies of continental shale formation.

In order to explore the law of deterioration and damage of roadway surrounding rocks and to clarify the water inrush mechanism of roadway spalling，laboratory tests on deterioration and seepage characteristics of sandstone under three stress paths of conventional triaxial(group C)，conventional discharge of confining pressure(group W) under different initial damage degrees，and cyclic loading and unloading confining pressure discharge(group X) were carried out based on Rock Top multi-field coupling tester. The results show that：(1) At the beginning of cyclic loading and unloading confining pressure of group X rocks，the axial deformation of rocks is slower than the lateral and volume deformation，but cyclic loading and unloading confining pressure can lead to “ strain accumulation ” of rock，which makes the stress drop point of rock in group X appears “hysteresis effect” compared with group W. (2) The initial damage significantly affects the rock strength and the sensitivity to the confining pressure. The initial damage degree of rocks in group X and group W is negatively correlated with the unified confining pressure drop parameter( )，and positively correlated with the incremental ratio of the strain and the confining pressure( ). The parameter   of group X is smaller than that of the group W，while   is greater than that of the group W. (3) Obvious brittle failure of rocks occurs in groups X，group W and group C，and the main failure mode is shear failure. When the initial damage degree is low，the rock failure degree of group X is large，accompanied by local macroscopic cracks. (4) The rock permeability is affected by the factors such as rock pores，framework structure and crack development. The rock seepage-strain relationship curves of groups X and group W can be divided into three stages. The first and third stages are consistent. In the second stage，the rock permeability of group X decreases first and then increases，while the rock permeability of group W increases as a whole. (5) The permeability fluctuation of group X tends to be obvious with the increase of the circulation grade. The higher the initial damage degree is，the more easily this phenomenon occurs. The number of cycles has little effect on this law，but the increase of the number of cycles has an inhibition effect on rock seepage，and the inhibition effect is more significant than that caused by the decrease of confining pressure.

In order to study the change of blasting damage range and fragmentation distribution of single free face rock，firstly，the reflection law of explosion stress wave on single free surface was studied by optical refraction Snell law，and the bottom radius of blasting crater was determined. Then，combined with the field blasting crater experiment of Baima iron mine，the geometric characteristics and rock fragment size of blasting crater under different burial depths and weights of explosive packet were analyzed. And the continuous-discontinuous element method(CDEM) was used to simulate the morphological characteristics of blasting crater and rock fragment size. The results show that the blasting crater volume under unit explosive weight increases in the power type of explosive packet ratio weight，and raises first and then reduces with the grow of the explosive ratio burial depth. There is a maximum point when the ratio buried depth is 1.01. The fragmentation of blasting crater narrows with the increment of the charge weight，and expends with the growth of the charge buried depth. The average particle size of broken rock decreases with −19/30 power of the charge weight，and increases with 1.42 power of the charge buried depth. Finally，using the verified numerical simulation method，the rock breaking effect of blasting design with the large hole distance and small resistance line at the Xia?an creek in the Three Gorges Project was reproduced. It provides reference for analysis and prediction of rock blasting fragmentation in practical engineering.

In view of the limitations and difficulties of the existing direct tensile method for testing rock tensile strength，such as high requirements for equipment，complex test process and low success rate of tests，the existing direct tensile method cannot be widely used in rock tensile strength testing. Therefore，based on the concept of “compression to tension”，an universal direct tensile method is proposed which can be performed by the ordinary testing equipment. In order to clarify the feasibility of the direct tensile method proposed in this paper，the Brazilian splitting method，which is recommended internationally，was used as the compared method，and three kinds of rocks such as marble，sandstone and granite were taken as the representatives of metamorphic rock，sedimentary rock and magmatic rock for testing and discussing. The results show that the method is simple and easy to use，and that the test results have good consistency. The direct tensile strength of the three types of rocks is lower than that of the indirect tensile strength，which is consistent with the existing research results. Furthermore，considering the factors potentially affecting the test results，the key factors that should be considered in the further optimization of the method in the future were proposed，such as seam width，seam bottom radian，the column height and bottom thickness，and the ratio of cylinder wall thickness to particle size.

To accurately assess matrix permeability and its role in the fluid flow through fractured rocks，flow-through experiments are conducted on both intact and split rock cores(including mudstone，sandstone and limestone) subjected to hydro-mechanically coupling conditions. By varying confining pressures(10–18 MPa) and fracture roughness(1.64 to 15.52)，the influence of the confining pressure and the fracture roughness on the fluid flow within fracture and matrix were explored. The experimental results indicate that the permeability of both matrix and fracture shows a decreasing trend with increasing the confining pressure，and the permeability of fracture(Kf) declines faster than that of matrix(Km)，resulting in a decrement in the ratio between them(Kf/Km). Specifically，when the confining pressure increases to 18 MPa，Kf/Km of mudstone reduces to 7–8(depend on fracture roughness)，reflecting fracture permeability is in the same order with matrix permeability，and then matrix permeability should not be neglected. Furthermore，the numerical simulations were performed on fractured rocks with fracture roughness of 2.58–17.4，and under confining pressures of 3–53 MPa. The fluid flow in matrix and fracture were assumed to obey Darcy law and Forchheimer law，respectively. The numerical results match experimental ones well，and it successfully reproduces the evolutions of fracture and matrix permeability with confining pressure and fracture roughness. If we take Kf/Km of 10 as a threshold，below which matrix permeability should be considered，then the critical confining pressure of 18，20，and 46 MPa can be determined for mudstone，sandstone and limestone，respectively. In addition，it is found that the contact area between rough-walled fracture surfaces increases with confining pressure and fracture roughness，leading to shrinkage in the flow channels，which consequently lowers the fracture permeability. Meanwhile，the velocity of flow within the narrowed channels will then accelerate，resulting in vortex inside the flow field，which further reduce the permeability of rough-walled fractures. This study systematically explores the fluid flow evolution within fracture and matrix，and provides a quantitative index Kf/Km to evaluate the role of matrix permeability in fluid flow through fractured rocks. 

High-locality rockfalls occur frequently combined with strong dynamic fragmentation phenomenon.  However，the insufficient understanding of rockfall fragmentation often causes a large error between the hazard assessment and the actual situation. Therefore，based on the movement characteristics analysis of typical high-locality rockfalls in Guizhou Province，the dominant factors affecting fragmentation are determined. The low-strength and high-brittleness model materials are designed for the physical model test in rockfall dynamic fragmentation. The fragments? movement and deposit features under different conditions are observed and the relationship between fragmentation and its maximum movement distance is evaluated quantitatively. The results show that：(1) The falling height，volume，impact angle and joint setting are the main factors influencing rockfall dynamic fragmentation. (2) The fragmentation phenomenon is concentrated in the first contact with the slope surface. The movement of generated fragments is characterized by stratification. There are small particle-size blocks with “ejection” movement in the upper layer. The large block is rolling in the middle layer，and the fine particle is sliding at the bottom layer. (3) The block distribution characteristics have a concentrated deposit and scattered blocks. But，there is no significant block size distribution along the movement direction. (4) The fragmentation is most sensitive to the impact angle，and the fragmentation degree is positively correlated with the maximum movement distance. The empirical formula is carried out as well. (5) The problem of energy distribution in rockfall dynamic fragmentation needs to be further studied from in-situ experiments and real-time observation.

In order to reveal the influence of microstructural features on the fracture characteristics of rock beams with a U-shaped notch，the theoretical relationships of the rock apparent fracture toughness KIN with the inherent fracture toughness KIC，the critical distance L，as well as the notch radius ? were established based on the theory of critical distance. A new approach to calculate KIC and L was proposed. Then，the discrete element method was used to built two groups of rock grain-based models，one of which is different in grain size，and another group is different in heterogeneity. Both of them were employed to conduct three-points bending simulation test to obtain the fracture mechanical parameters of KIN，KIC and L. Furthermore，the influences of the grain size and heterogeneity of rock on KIC and L were studied. The results show that，the obtained KIC and L from numerical tests have a good agreement with the values calculated by the theoretical equation，which indicates that the numerical experiment is highly reliable. The microstructural features have a significant influence on the rock fracture characteristics. With the increase of grain size，the rock fracture toughness KIC gradually decrease，while the critical distance L gradually increase. Additionally，KIC and L are both decreased with the increasing of rock heterogeneity.

The phenomenon of supporting structure cracking due to the expansive soft rock is common in tunnel engineering. Taking the non-uniform stress of tunnel surrounding rock structure caused by expansive soft rock as the starting point，the stress characteristics of the supporting structure of expansive soft rock tunnels are studied，and a swelling force release method of expansive soft rock tunnel supporting structure is established. The granular bentonite is implanted between the primary support and the secondary lining as the buffer layer，and the model test is carried out. The three-dimensional numerical model is established by MIDAS GTS NX software，and the mechanical characteristics of the expansive soft rock tunnel support structure before and after the buffer layer are comparatively analyzed. According to the tunnel mechanics theory，the swelling force slow-release model of the expansive soft rock tunnel supporting structure is established，and the expression of release coefficient is obtained. The practicability of the slow-release method is verified. The results show that the deformation of the supporting structure and the swelling force of the surrounding rock after adding the buffer layer are significantly smaller than those before adding the buffer layer. The deformation characteristics are characterized by the sinking of the vault，the characteristics of arch bottom upward uplift and the arch waist inward convergence. The buffer layer reduces the swelling force of surrounding rock acting on the supporting structure by resisting，weakening and transferring. Consequently the single concentrated force system is transformed into the uniform load system. The slow-release coefficient calculated by the slow-release model is 30%–50%，indicating that the implantation of granular bentonite as a buffer layer between the initial support and the secondary lining can slow down the swelling force. This research can provide a reference for solving the swelling problem of soft rock tunnels.

As a kind of green and renewable resources，the scale and efficient development of hot dry rock(HDR) geothermal energy has always been a worldwide research hotspot. In order to solve the problems of small heat exchange volume and poor heat exchange effect in the development of intact HDR geothermal by Enhanced Geothermal System(EGS)，the fracture distribution characteristics of natural deep HDR body were studied. Through field observation，it was found that structural fractures in deep granite bodies were universally filled by hydrothermal fluid or magma，forming fractured-subsequently-filled granite. Through micro-observation test，it is concluded that under the influence of high temperature action of backfill，changes in the number of thermal-cracking fractures of fractured-subsequently-filled granite versus the distance from the cementation interface can be divided into a zone in which the number of fractures fluctuates gently，a zone showing a sharp increase in the number of fractures，and a zone in which the fluctuation in the number of fractures decreases. The fractured-subsequently-filled granite mainly produces micro-fractures with length＞50 μm and length＞100 μm，and the position of the maximum number of thermal-cracking fractures with different length scales is located in the parent rock at a certain distance away from the cementation interface. When the fracture backfill thickness is 3.4cm，the unilateral influence range of the high temperature action of the fracture backfill is about 1 m. The widespread fracture backfill result in the formation of a huge weak-plane structure in the granite parent rock. Through hydraulic fracturing test，it is concluded that the weak-plane structure can be used as fracture channels to construct artificial reservoir through hydraulic fracturing，and can greatly reduce the breakdown pressure，so as to achieve large-scale and efficient construction of HDR artificial reservoirs. On this basis，the research on new technology of artificial reservoir construction in HDR geothermal exploitation constitutes a new research direction for HDR geothermal development.

In deep engineering，the development and application of novel metallic materials has attracted many attentions. However，the influence of prestressing process and rock loading process on high-strength and high-toughness steel for rock bolts remains still unclear，and relevant research is urgently required. Muzhailing highway tunnel with extremely high in-situ stress area was taken as the engineering background，combined the physically based crystal plasticity model，and NPR steel was taken as an example to study the effects of high prestress and rock loading rate on the plastic deformation behaviors of high-strength and high-toughness steel. The results show that the rock loading rate plays a major role on affecting the strain capacity of bolt steel，and the ultimate strain and deformation relaxation of bolt steel will increase with the decrease of the rock loading rate. It is found that the large deformation capacity of bolt steel can be significantly enhanced when the rock loading rate is controlled lower than 10－2 MPa/s. In addition，there will be obvious strain relaxation phenomenon for bolt steel as the prestress is higher than 600 MPa. The bolt steel will exhibit stress relaxation as the prestress increases higher than 800 MPa，which may result in the loss of prestress for rock bolt and should be paid attention. The research results provide theoretical basis and technical guidance for practical application of high-strength and high-toughness steel for rock support engineering in high in-situ stress field.

To study the mechanical mechanism and damage characteristics of the new fracturing device gas producing device weakening hard roof，the dynamic model of multistage weakening hard roof with gas producing tool is established，the effective stress and damage degree of the roof during the weakening process are calculated，and the morphological characteristic change and damage characteristic evolution of the roof are obtained. The results show that the high-pressure gas pre-expands the primary cracks in the rock mass around the borehole during the combustion stage of gas producing tool；the detonation stage occurs when the pressure reaches the strength of the pressure-sensitive material，due to the combined action of shock wave，stress wave and explosive gas，the near field crushing，mid field fracturing and far field vibration of the roof are caused. The damage characteristic area of roof weakened by gas producing tool fracturing technique is compared with blasting fracturing technique，the crushing zone radius is decreased by about 22%，the fracture zone radius is increased by about 23%，and damage zone radius is increased by about 5%，compared with hydraulic fracturing technique，the crushing zone radius is increased by 133%，the fracture zone radius and damage zone radius are reduced，but the damage uniformity is significantly improved. The gas production tool is suitable for stable roof with less joint cracks，which can give full play to the effect of high-pressure gas，which is positively correlated with the strength of pressure-sensitive materials and the length of primary cracks，and negatively correlated with the initial confining pressure of rock mass. Combined with the field test results，the effectiveness and superiority of the gas producing tool weakening hard roof are verified.

Relying on tunnel engineering crossing active fault fracture zone in high intensity regions in Western China，shaking table test of tunnel crossing multi-rupture surfaces was carried out. Based on acceleration，dynamic strain，displacement responses at measurement points and the state of cracks，the energy and damage characteristics of tunnel segmental lining structure and surrounding rock under incremental seismic wave excitation were studied. On basis of HHT transform method，the Hilbert marginal spectrum and transient energy spectrum demonstrated the damage trend of the surrounding rock and tunnel structure. The results are as follows：(1) The coseismic dislocation of the model soil surface is significant under ground motion，the major and minor rupture surfaces in fault fracture zone can be defined according to the dislocation displacement on both sides of rupture surfaces. (2) The damage evolution of tunnel structure near major rupture surface in footwall lagged behind that in hanging wall，the decrease in peak values of primary frequency and marginal spectrum of tunnel structure near the major rupture surface reaches the maximum of 29.1% and 87.1% under 0.4 g seismic excitation，and the extent of damage is much more serious than that in other parts. (3) Tunnel structure is mainly subjected to tensile cracking. The tensile cracking occurs in invert near major rupture surface in hanging wall under 0.2 g ground motion，while damage in arch waist appears under 0.4 g ground motion. (4) Surrounding rock in hanging wall decreases 34.7% in primary frequency under 0.5 g ground motion. The damaged degree significantly exceeds that of footwall and middle part of fault fracture zone. (5) According to the damage pattern of tunnel structure，invert is the weakest part which is easily cracked to varying degrees. The seismic design coupling damping design of tunnel invert are needed to focus on strengthening. Tunnel structure near interface between hanging wall and fault suffers most under ground motion，while tunnel in the middle of fault fracture zone suffers less. The seismic partition and fortification of tunnel engineering through fault fracture zone is required. The conclusions can provide certain reference basis for the seismic design of tunnels in high intensity seismic zones.

In the process of utilizing water injection heat exchange for the geothermal development of high-temperature rock mass，the surrounding high-temperature rock mass of the convective channel in the heat exchange will experience repeated cycles of cooling and heating. Under the cyclic temperature effect，the deformation of the fracture wall surface，fracture flow，and heat transfer characteristics of the high-temperature rock mass will be affected to a certain extent. To reveal the granite fracture seepage characteristics and forced convection heat transfer law in this complex process，the water seepage and heat transfer experiments were carried out in the laboratory under the cyclic cooling-heating process(300 ℃→250 ℃→200 ℃→150 ℃→300 ℃). The results showed that：(1) the injection of low-temperature water could induce damage and fracture of the high-temperature granite fracture surface. The roughness coefficient JRC of the fracture surface changed from the initial 14.51 to 21.03 after the cyclic process，and the maximum height difference ξ of the fracture surface profile increased from 2.2 mm to 3.21 mm. The flow-conducting fracture of the high-temperature rock mass became more tortuous and the roughness increased with the temperature change. (2) When the rock temperature decreased from 300 ℃ to 150 ℃，the fracture permeability decreased exponentially from the initial 1.63 Darcy to 0.53 Darcy. However，when the rock temperature increased from 150 ℃ to 300 ℃，the fracture closed due to the thermal expansion of the rock matrix，and the permeability further decreased. With the increase of the cooling-heating cycles，the overall fracture permeability fluctuated to a certain extent，eventually leading to a significant decrease in permeability. (3) A higher initial rock temperature and an increase in the injection water pressure for heat exchange were beneficial to the forced convection heat exchange effect of the high-temperature granite fracture. However，with the increase of the cooling-heating cycles，the effect of convection heat exchange weakened. This study has certain guiding significance and value for the efficient development of geothermal resources in the high-temperature rock mass and the control technology of convection heat exchange.

The existing landslide susceptibility analysis methods have the defects of ensemble modelling strategy and false negative errors phenomenon. Firstly the stacking algorithm integrating random forest(RF) and extreme gradient boosting(XGBoost) to predict the spatial probability of landslide occurrence was innovatively proposed in this paper. Then，the small baseline subset interferometry technology was used to measure 104 Sentinel-1A data from January 2018 to September 2021，and the deformation velocity along the line-of-sight(Vlos) was re-projected to a new velocity along the steepest slope direction(Vslope). Finally，the empirical matrix was considered to combine the susceptibility and deformation rate classification to achieve the landslide dynamic susceptibility map. The results indicate that the stacking based RF-XGBoost model has better prediction and generalization ability than the decision tree(DT)，RF，XGBoost，Bayesian Network(BN) model，and multilayer perceptron neural network(MLPNN). The landslide dynamic susceptibility map has a better identification ability on areas with strong deformation，reducing the proportion of low-susceptibility by 3%–8% and increasing the proportion of high- and very-high-susceptibility by about 2%. The field investigation verified that this method could improve landslide susceptibility and reduce false negative errors in areas with strong engineering activities. Real-time monitoring should be strengthened in the very-high dynamic susceptibility area of Dazhou Town (along the Yangtze River and the northern Fenghuang village). It is concluded that the ensemble framework and landslide dynamic susceptibility mapping strategy proposed in this paper has high spatial identification and early warning accuracy and can be used as a new method for regional planning of landslide disasters.

The liquefaction characteristics and anti-liquefaction treatment measures of coral sand sites are one of the key issues in the construction of tropical ports. Based on the dynamic centrifuge model parallel test，the seismic liquefaction characteristics of coral sand and standard sand in free fields are analyzed from the aspects of excess pore water pressure ratio，site amplification effect，surface subsidence，shear modulus and damping ratio. The results show that the load intensity triggering liquefaction of coral sand site is significantly higher than that of standard sand site，and the liquefaction depth is significantly lower than that of standard sand site under the same conditions. The acceleration amplification coefficient of coral sand site is obviously smaller than that of standard sand under the loads of 0.2 g and 0.3 g sine wave. The cumulative surface subsidence of coral sand site is about 2/3 of that of standard sand site under 0.05 g–0.3 g sequence loads. In liquefied soil layer，the shear strain of coral sand site is higher than that of standard sand site and the shear stress is lower than that of standard sand site. In the non-liquefied soil layer，under the strong loads of 0.2 g and 0.3 g sine wave，the shear stress and shear strain of coral sand site are smaller than that of standard sand. The attenuation rate of shear modulus of coral sand site with shear strain is faster than that of standard sand site under the same conditions. The damping ratio of coral sand site is slightly smaller than that of standard sand site under 0.05 g–0.3 g sequence loads. The research results provide important guidance basis and rich data for understanding the liquefaction characteristics of coral sites and designing the anti-liquefaction behavior of port engineering.

To investigate the effect of pile-side soil softening on the thermal response of energy piles，based on the load transfer method and the energy balance principle，a nonlinear analysis method for the thermal response of energy piles considering the softening of pile-side soil is proposed. In the method，the relationship between pile-side soil shear stress and displacement is described by a skin friction softening model，the relationship between pile-tip reaction and displacement is captured by a hyperbolic model，and the Masing criterion is introduced to construct the loading and unloading function of soil，which is compared with the existing experimental and numerical methods to verify the correctness of the method in this paper. Finally，the influence of relevant softening parameters on the thermal response of the energy piles is analyzed，and the influence of the pile head stiffness coefficient on the null point is explored. The results show that：(1) both the failure ratio of skin friction   and the ultimate shear displacement ratio  have an impact on the thermal stress，where the effect of both on the thermal stress presents different change rules under the same thermal load，and the thermal stress caused by both increases with the increase of the thermal load. (2) When the superstructure load is small(≤10% Pu)，the value of the thermal stress is larger，and the influence of parameter   on the thermal stress is more significant compared with parameter  ，but the difference decreases with the increase of the superstructure load. (3) With the increase of the pile head stiffness coefficient  ，the null point gradually approaches the pile head，and its position corresponding to the heating condition is always slightly lower than that of the cooling condition.

Based on the piecewise-linear finite difference approach，an one-dimensional large strain nonlinear thermal consolidation model for saturated soil，called TCS2，is presented. The model couples soil heat conduction with consolidation deformation，and accounts for the soil self-weight，large strain，variable hydraulic conductivity and compressibility，and nonlinear variation of soil parameters during the thermal consolidation process. The Fortran was used to compile the calculation program，and then the model was validated. The numerical solutions of TCS2 involving small strain are in good agreement with the analytical solution of thermal consolidation based on the assumption of small deformation. Under the large strain nonlinear condition，the calculation results of TCS2 are basically consistent with the measured values of laboratory tests. Then，the effects of soil self-weight，large strain and stress path on thermal consolidation of saturated soils were further investigated using example problems. Neglecting the soil self-weight，consolidation deformation and the simplified compressibility and hydraulic permeability relationships will lead to significant errors in a thermal consolidation analysis.

In recent years，water-induced failures of check dams frequently occurred due to short-time heavy rainfalls in Loess Plateau in China. The research foundation on check dam breach simulation is weak，while the erosion characteristics of compacted loess and the head-cut erosion mechanism under the influence of different factors are still unclear. Qualitative and quantitative tests on microscopic pore structures of loess material，and 14 groups of head-cut erosion flume tests of compacted loess have been conducted. In this study，the influences of initial water content and compactness on head-cut erosion process are the main concern，and the influence mechanisms of compacted loess microscopic pore structure on the head-cut erosion mode and rate have been revealed. The model test results show that the water-holding state of compacted loess has an essential impact on head-cut erosion mode and rate. Water molecules change the macroscopic erosion law by affecting the microscopic pore structure of compacted loess. When the soil is in a low water content state，the head-cut is dominated by layered erosion，and the “false cohesion” provided by matrix suction has little impact on the erosion rate. With the increase of water content，the head-cut erosion mode has changed into retrogressive erosion towards upstream accompanied by intermittent instability of blocks. However，the change of compactness does not affect the head-cut erosion mode，but the head-cut erosion rate by adjusting particle skeleton and contact area. The increase of compactness can significantly enhance the erosion resistance of loess. The research results can provide theoretical support for the study of dam breach mechanisms of loess check dams.

With the increase of high-speed railway construction in cold and arid areas and the increase in engineering service time，the problem of subgrade frost heave caused by vapor migration has been widely concerned. To deeply explore the hydrothermal distribution and frost heave characteristics of coarse-grained graded packing，a soil column test device considering hydro-thermo-mechanical coupling was developed，and unidirectional freezing tests with different water recharge types，fine particle contents，and fine particle types were carried out. The results show that the position of the freezing front is significantly affected by the water recharge method and the content of fine-grained soil，while the influence of the type of fine-grained soil is relatively small. With the increase in the freezing time，both liquid-vapor recharge and vapor recharge lead to the increasing water content of the specimen，while the former changes faster. In addition，the fine-grained soil content and type can significantly affect vapor migration，resulting in significant differences in water distribution. It is difficult for liquid water in coarse-grained soil to directly rise to the freezing front through suction，and vapor migration is an important factor in the development of frost heave. In the long-term operation of high-speed railways，the problems of subgrade water accumulation and frost heave caused by vapor migration cannot be ignored.