The nonlinear Coulomb strength criterion is an extension of the classical Coulomb strength criterion，utilizing the uniaxial(compressive and tensile) strength of intact rocks as fundamental parameters. By employing the rock cohesion function and the internal friction angle function，a new rock strength criterion is established. The core value of this criterion lies in its capacity to calculate the triaxial strength of rocks under various stress states based on the uniaxial strength of intact rocks. This includes the peak strength，cohesive force，and internal friction angle of intact rocks，as well as the damage strength，cohesive force and internal friction angle of intact rocks，and the strength，cohesion，and internal friction angle of rock masses with different quality levels. Simultaneously，the corresponding rock strength curves were determined. A comparative analysis of 14 rock strength test results demonstrates that the theoretical calculations derived from the nonlinear Coulomb strength criterion align closely with the experimental results obtained from the Hoek-Brown empirical strength criterion. The key distinction between the nonlinear Coulomb strength criterion and the Hoek-Brown empirical strength criterion is that the former relies on the uniaxial strength of the rock，while the latter is contingent upon the experimental outcomes of the rock?s triaxial strength. Especially when the number of test samples is limited and only uniaxial strength testing is available，the nonlinear Coulomb strength criterion serves as the sole effective method for determining rock cohesion，internal friction angle，and strength curves.

Currently，wellbore instability in fractured formations significantly impacts the exploration and development of ultra-deep oil and gas resources. Traditional analytical models for wellbore stability struggle to account for the effects of drilling fluid plugging，and the computational efficiency of numerical simulation methods often fails to meet the demands of efficient drilling. To address these issues，a wellbore collapse mechanical model for fractured formations that incorporates the drilling fluid plugging effect has been established based on the mechanics of discontinuous media. Additionally，a method to verify the model?s computational results using the effective stress between fractured blocks has been proposed. The study analyzes the influences of drilling fluid plugging，block fragmentation degree，mechanical parameters，in-situ stress，hole size，and other factors on collapsing pressure. The results indicate that wellbore instability in fractured formations is primarily caused by shear slip of fractured blocks，which arises from unbalanced forces between these blocks. Unlike traditional theories，this study reveals that wellbore instability can also result from excessive drilling fluid density，and the direction of collapse may vary accordingly. Furthermore，the plugging effect of drilling fluid is positively correlated with wellbore stability，while the degree of block fragmentation is negatively correlated with it. Field verification demonstrates a significant improvement in wellbore stability following the application of drilling fluid density and plugging performance designed according to the new model. This research provides theoretical support for selecting appropriate anti-collapse drilling fluid densities and plugging performance in well drilling within fractured formations.

A rockfill dam with a crown height of 315 meters has been proposed as the water retention structure for a large hydropower station. The RS accumulation body is located approximately 5.4 km upstream of the dam，with an estimated volume exceeding 47 million cubic meters. The stability of this accumulation body is，therefore，a significant concern for the development of this project，prompting comprehensive studies. This paper presents a summary of the research conducted，focusing on the formulation and associated stability of the accumulation body. Site investigations have revealed that the accumulation body can be divided into two distinct parts based on the characteristics of the granular materials? compositions and structures. The lower part is formed from rock falls originating from the cliff behind，alluvial deposits along the river，and ice-water induced muddy flow from two nearby gullies. In contrast，the upper part exhibits an interbedded structure comprising granular materials，both glued and unglued，which is significantly different from that of the lower part and suggests a rare formation known as ice-snow sheet flow. These materials were transported by melting ice and snow and deposited interactively on the surface due to historical climate changes. Qualitative analyses and numerical modeling were subsequently conducted to assess the stability of the RS accumulation body. The findings indicate that the lower part is relatively stable，contributing positively to the overall stability of the entire structure. However，the potential for material sliding along the bedding in the upper part is expected to increase with the impoundment of the reservoir. The numerical model suggests a maximum failure volume of approximately 9 million cubic meters，which can be utilized as input for studies related to swell risk assessment and engineering design measures.

In order to investigate the physicochemical damage mechanism and dynamic mechanical properties of annular sandstone specimens under weak acid environment，the specimens of annular sandstone with outer diameter of 50 mm and inner diameters of 0，5，10，15，20 and 25 mm were corroded by weak acid solution(pH = 5) for 28 d，and the basic physical parameters，mineral compositions and microstructures were tested，and impact compression tests were carried out by using split Hopkinson pressure bar(SHPB) test device. The impact compression test was carried out using SHPB test device. The results show：Corrosion of the annular sandstone specimen by a weak acid solution produced Na+，Al3+，Fe3+，and precipitates such as silicic acid(H2SiO3)；Specimen mass loss rate，volume expansion rate and density reduction rate all increase with increasing pore size；with increasing pore size of annular sandstone specimens，the dynamic compressive strength and dynamic elastic modulus show a decreasing trend of exponential function，the dynamic peak strain and the average strain rate show a quadratic and linear trend of increasing，respectively，and the degree of specimen impact crushing is intensified；Compared with the specimens of annular sandstone specimens eroded by neutral solution(pH = 7)，the specimens showed a decrease in dynamic compressive strength and dynamic elastic modulus，and an increase in dynamic strain and average strain rate. The acid-rock reaction causes the sandstone to generate new substances，and the larger the pore size，the easier it is for acid ions to enter the interior of the specimen，and the more serious the deterioration of its structural damage is，resulting in a significant decrease in the kinetic properties.

At present，the design value of the grouting engineering quantity for large-scale water-sealed caverns in China is mainly determined by referring to the storage capacity(or investment scale) and in the light of the actual grouting amount of the already constructed projects. The lack of consideration of the differences in geological conditions among different projects resulted in a multiple-fold discrepancy between the actual and the designed grouting amount of most projects，and brought out negative impacts on the control of project investment，duration，and quality. The length of grouting tunnel，the amount of grouting cement taken，and the total length of grouting sections are the main indexes to characterize the amount of grouting works. To explore a method to improve the estimation accuracy of the amount of grouting works，hydrogeological test of 57 boreholes with a total of 852 sections during investigation phase，the grouting information of a construction tunnel with the length of 3 200 m in construction phase，and the grouting data of 384 grouting boreholes have been collected in this study. By comparing and analyzing the proportion of borehole sections with over-standard permeability and the proportion of tunnel sections those need grouting during the construction phase，a method for estimating grouting tunnel length is proposed. Drawing on the practical experience garnered from lots of such cavern projects，a method for estimating the total length of grouting sections corresponding to different excavation types of caverns is summarized. Through establishing the correlation between the cement consumption and the permeability before grouting，a method for estimating the cement consumption of the project is put forward. The established grouting engineering quantity estimation method system above，that takes the engineering scale，the geological characteristics and the excavation type into account，not only contributes to standardizing the grouting works in water-sealed cavern projects，but also possesses reference significance for grouting in other types of underground rock engineering.

In order to explore the mechanism of the influence of shield construction on the load deformation of the upper stratum，a model shield machine with a similarity ratio of 1∶10 was independently designed to simulate the shield construction. Only the influence of the shield excavation process on the load deformation of the upper soil was considered. At the same time，the settlement and soil pressure changes of the upper stratum were monitored and analyzed in real time. The results show that：(1) for the loose and weak stratum，the settlement of the upper stratum is not obvious in the initial stage of shield construction. With the increase of the excavation distance，the settlement of the soil gradually increases，and reaches the peak when the excavation exceeds about 0.7 times the excavation diameter of the corresponding section. (2) In the process of shield excavation，the soil pressure on the excavation face and on both sides shows a decreasing trend in a certain range. At the same depth，the soil pressure near the excavation face decreases more，and the soil pressure increases more in the later period. (3) There is an obvious soil arching phenomenon at the position of about 0.2 times the excavation diameter from the horizontal net distance of the excavation face，and the soil pressure increases obviously. (4) In the actual construction process，for the weak loose stratum，in order to reduce the formation loss，it is recommended to design the cutterhead diameter as the same as the shield shell diameter.

Reservoir landslides are a serious threat to the safe operation of the reservoir area and the lives and properties of the residents on river banks. The inducing factors and evolution features are both vital for the safety monitoring and control. In this study，considering the characteristics of high water level impoundment and large water level variation in the southwest reservoir area，the Gapa landslide in Jinping reservoir area is taken as an example. The landslide deformation characteristics has been revealed based on multi-source data such as GNSS，inclinometer and InSAR，then the prediction and warning model has also been established. The field monitoring results reveal that reservoir water impoundment has a strong interference effect and cyclic fluctuation has a cyclic perturbation effect on the deformation of this landslide，respectively. In terms of prediction，if the landslide is under a special assumption of future failure，the deformation trend of the landslide in the next fifteen years is deduced by the Verhurst inverse function using the GNSS time series. A more practical three-stage deformation damage curves of the landslide and displacement prediction results can be obtained. In terms of early warning，due to the significant influence of reservoir water on the landslide，the thresholds for water level and rate of water level change are determined. Based on Artificial neural network，they are 1 820 m and 0.4 m/d，respectively. According to this dual-threshold can test the GNSS data，the threshold of landslide rate for daily frequency warning is determined to be 4 mm/d. Finally，the dual-threshold for the reservoir water and the threshold for the displacement velocity are formed. The prediction and early warning results can provide a reference for the safety monitoring of large landslides in the reservoir area with high impoundment and large variation in water level.

To investigate the thermal deformation and failure mechanisms of surrounding rock in medium-deep gasification cavities，the influence of high temperature and thermal stress on the thermal damage of the rock is identified on the basis of the testing of rock samples in the profile of the spontaneous combustion coal seam and the analysis of numerical simulation，the mechanical model of the gasification cavity roof was established based on the structural characteristics of medium-deep UCG cavities. Based on the results of multi-physical field coupling numerical simulation of the surrounding rock，the stress-deformation patterns of the surrounding rock is proposed. Finally，the thermal deformation damage process of surrounding rock is systematically described from three dimensions(rock，roof and surrounding rock). The research results show that：(1) temperature affects the microstructure and permeability of rock，when the temperature is 25 ℃–400 ℃，the percentage of micropores in muddy siltstone increases，when the temperature is 400 ℃–1 000 ℃，the percentage of micropores decreases，the percentage of mesopores increases and macropores begin to appear，when the temperature is more than       1 200 ℃，the percentage of macropores is the largest，and 600 ℃ is the threshold for the increase of permeability. The difference in thermal expansion coefficients of mineral components positively contributes to the tensile thermal stress. Under the joint action of rock thermal physicochemical reaction and thermal stress，the rock thermal damage shows the change rule of slow growth，rapid development and tends to be stable. In the stage of slow growth and stabilization phase，the thermophysical-chemical reaction is dominant，and in the stage of rapid development，the thermophysical-chemical reaction and thermal stress play a joint role. (2) Compared with the assumption of the beam model，the thin plate model of the gasification cavity roof is more consistent with the characteristics of the gasification cavity of medium-deep coal. After the roof meets the instability condition，the roof will have the destructive process of small-scale collapse of solid support model，large-scale collapse of simply support model, and the arch structure tends to be stabilized in sequence. (3) The surrounding rock in the gasification cavity shows a stress deformation pattern of“three deformation characteristic zones and five stress characteristic zones”，and the multi-field coupling effect leads to the increase of the scope of the“butterfly-shaped”shear plastic zone. The tensile damage mainly occurs around the gasification cavity，the top and bottom of the cavity are deformed downward and upward respectively，and the vertical stress of the surrounding rock on both sides of the gasification cavity increases by 30 %. The development height of the shear plastic zone in the overlying rock of the gasification cavity is generally 2–5 times of the coal thickness，and it should be ensured that the water barrier in the upper part of the coal seam meets the thickness requirement when selecting the site. The research results are of theoretical significance for the safe site selection of underground gasification of medium-deep coal.

To efficiently and accurately simulate fracture propagation in tight reservoirs with different types of natural fractures，a fully coupled hydro-geomechanical model based on embedded mesh is adopted. A time step adaptive adjustment strategy is proposed to improve the efficiency of main fracture propagation. The energy release rate method is employed to consider the effect of strongly cemented natural fractures. The damage model is applied to simulate the activation of weakly cemented natural fractures. After that，the influence of different types of natural fractures on the fracture propagation is investigated. The results show that：(1) the time step adjustment strategy can improve the computational efficiency of main fracture propagation by 50%. (2) If the orientation of strongly cemented natural fractures is relatively consistent，it is recommended to adjust the fracturing direction perpendicular or parallel to the natural fracture orientation appropriately. It can help reduce the likelihood of inter-well interference，making it suitable for closely spaced multi-cluster fracturing. (3) On a small scale，the hydraulic fracture geometry and fluid pressure field are greatly affected by the distribution of weakly cemented natural fractures. In addition，these natural fractures affect the tensile strength of the matrix rock. The lower the tensile strength，the larger the stimulated reservoir volume，the stronger the fluid loss effect，and the shorter the main fracture length. (4) On a large scale，when natural fractures of different cementation strengths exist，hydraulic fractures tend to form curved main fractures with damage zones. The method provides an efficient tool for fracturing simulation and optimization in fractured tight reservoirs.

The landslide dams formed by rock-ice avalanches are a common geological hazard in the alpine and cold mountainous areas in the southwest of China. The effects of ice content and spatial location of ice-rock materials on the formation morphology，dam failure process，and stability of debris dams were researched by large-scale flume experiments. The results indicate that ice debris enhance the mobility of debris material，leading to increased blockage in the riverbed. Ice debris are influenced by particle segregation，primarily accumulating on the upper part of the landslide dams and on the opposite bank of landslide. Under the impact of water flow，ice debris within the landslide dams melt rapidly，thereby increasing the porosity of the dam body and weakening the resistance to erosion，resulting in the formation of deeper and wider breach outlets and accelerating the dam failure process. Additionally，a dimensionless stability evaluation method( ) for landslide dams containing ice is proposed. Evaluation results demonstrate a linear decrease in the stability of the landslide dam with increasing ice content.

Taking granite as the research subject，a uniaxial cyclic loading and unloading test scheme based on axial strain control is proposed to investigate the strength and deformation evolution of rock samples under cyclic loading. The uniaxial compression test reveals significant variability in peak strength and Poisson?s ratio，while peak strain and elastic modulus remain relatively stable，at 0.43% and 0.37%，respectively. The stress-strain curve was analyzed and fitted in segments，with the axial strain control values determined through slope analysis，yielding 0.37% and 0.32%. During cyclic loading and unloading，as the number of cycles increases，the stress-strain curve forms a hysteresis loop that evolves in the direction of increased strain. The difference in elastic modulus between the first loading and unloading cycle is the largest；thereafter，this difference gradually decreases，both trends demonstrating a nonlinear downward trajectory，while Poisson?s ratio exhibits a nonlinear upward trend. The cyclic loading-unloading response ratio serves as an indicator of the degree of damage to the rock sample，which continues to increase. In the case of uniaxial compression failure of damaged rocks，there are more primary fracture surfaces compared to conventional rocks，and with the increase in the number of cycles，peak strength displays a nonlinear accelerating downward trend. The elastic modulus shows a linear downward trend，while Poisson?s ratio exhibits a nonlinear decelerating upward trend.

In contrast to conventional hydraulic fracturing，the recently proposed methane deflagration fracturing exhibits characteristics of transient low overpressure，ease of repeatability，and low cost. This study investigates the enhancement of fracture permeability and its influence based on pressure-timing data. Physical simulation specimens were prepared using rock-like materials，and synchronized experiments were conducted on methane deflagration，the phase transition of supercritical carbon dioxide，and deflagration fracturing involving an activating agent. Permeability measurements were taken from different regions of the rock samples before and after the fracturing experiments. The results indicate that all fracturing methods can enhance permeability，with methane combustion showing the most significant effect，followed by the activating agent. Notably，there is no apparent correlation between peak pressure and permeability；however，the rate of pressure increase significantly affects permeability variations. Higher pressure ramp rates result in more substantial increases in the permeability of the model samples. The experimental findings reveal that fracturing methods characterized by significant pressure oscillations after reaching peak pressure exhibit similar changes in permeability for both proximal and distal samples. This suggests that the oscillation performance of the pressure-time curve effectively broadens the range of the permeability enhancement effect. The research outcomes hold positive significance for the efficient development of shale gas，providing a foundation for the design and engineering application of combustion fracturing in shale gas reservoirs.

To investigate the correlation between fragment size distribution and the internal pore structure of muck-piles，Nuclear Magnetic Resonance(NMR) testing was employed to measure and image the pore structure of muck-pile analogues. The complexity of the internal pore structure within these models was quantified and analyzed using the Box-counting method，revealing the underlying patterns of muck-pile fragment size distribution. Through simulations and indoor analogue experiments，the distribution of internal pore structures within muck-piles was elucidated. The results indicate that：(1) compared to intact rock，the porosity of granular materials(muck-pile constituents) is significantly higher，reaching a maximum of 81.55%. The connectivity of pores within larger gravel particles is superior，with the ratio of small and medium pore T2 spectrum peak areas to the total peak area exceeding 80%. The size of gravel particles significantly influences the number of pores and the pore size distribution in granular materials. (2) The internal pores of granular materials adhere to fractal principles. The standard deviation for extremely small-sized gravel specimens exceeds 0.01，and the fitting values of the fractal dimension R2 for these specimens consistently fall below 0.9，suggesting a close relationship between the uniformity of internal pores and the content of extremely small-sized gravel. (3) Granular materials exhibit a sedimentation effect，with extremely small-sized particles accounting for over 90% at the ends of the specimens，resulting in a distinct stratification phenomenon. By integrating on-site mine observations with simulated muck-pile fragment size distributions，the vertical distribution pattern of“from low to high and from small to large”within muck-piles was verified. The sedimentation effect of extremely small-sized gravel significantly impacts the fragment size distribution of muck-piles.

In order to solve the problem of fracture intersection interference between deep shale gas wells in southern Sichuan，a numerical model of shale gas zipline fracturing is constructed based on finite-discrete element method，which considers the influence of natural fracture-fault structure. The model uses joint units with different mechanical properties to characterize natural fractures and faults，which can realize continuous simulation of multiple operation processes of“response well fracturing，response well shut-in，and excited well fracturing”，so as to accurately characterize physical behaviors such as fracture intersection，pressure and stress interference. The KGD analytical solution model of single fracture propagation and the experimental data of intersection between hydraulic fracture and natural fracture are used to verify the proposed model. Furthermore，a comparative case study has been conducted for typical well sections. The research results show that compared to natural fractures，the pressure drop response dynamics are more significant when intersecting with faults and hydraulic fractures. Under high stress difference，natural fractures are mainly subject to shear failure，while the matrix is mainly subject to tensile failure. At the initial stage of shut-in，natural fractures or faults continue to extend due to the net pressure of the remaining fluid in the fractures. After the fracture intersects between wells，the high pressure fluid of the exciting well preferentially flows into the response well through the high diversion channel，resulting in the increase of the response well pressure，the increase of filtration loss and the repeated reconstruction of some fractures，which will lead to the slow increase of the fracture length and the decrease of the overall fracturing efficiency.

Accurately understanding the loading characteristics on blasthole wall is important for rock blasting engineering. Many complex factors，such as the charge structure，initiation location，and rock wave impedance，have significant influences on the explosive loading characteristics in rock blasting. Therefore，the distribution of explosive loading on blasthole wall，as well as the transmission characteristics of blasting pressure between the explosive and the rock mass，are still unclear. In this study，based on the one-dimensional flow model of detonation products，the detonation loading characteristics of cylinder charge were investigated under different initiation locations. Subsequently，the transmission characteristics of blasting pressure between the explosive and the rock mass were studied，considering the propagation of shock waves and the expansion of detonation products. Furthermore，onsite blasting experiments at the Baima Iron were conducted to verify the influence of the charge structure and rock property on the impact loading curves. Finally，using numerical model with single blasthole，the influence of the initiation location，decoupling coefficient and rock mass on the loading characteristics was investigated. The results indicated that due to the directional effects along the charge，the blast loading on blasthole wall is non-uniform. In cases of bottom initiation and middle initiation，the peak explosive loading on blasthole wall is lowest at the initiation locations. However，as the detonation distance increases，the peak explosive loading gradually reaches a steady state. When explosives are initiated simultaneously from the top and bottom，the collision of shock waves occurs in the middle section of the charge，resulting in a sudden increase in blasting pressure at middle of the charge. Furthermore，the explosive type，charge structure，and wave impedance have influences on the transmission characteristics between the explosive and the rock medium. As the decoupling ratio increases，the pressure increase ratio gradually increase，while the pressure after the incident wave decreases. This results in a reduction of the explosive loading on blasthole wall. Under different rock masses，the pressures after the incident wave are at the same level. The pressure increase ratio gradually increases with the wave impedance，resulting in an increase in the explosive loading on blasthole wall. In practice，it is advisable to select the explosive type，charge structure，and initiation location according to the characteristics of the rock mass，so as to adjust the loading characteristics on blasthole wall，further improving the blasting excavation effects.

The traditional stability calculation method for bedding rock slopes mainly simplifies the slope into a plane sliding-tensile fracture failure model，conducting overall stress analysis. However，it neglects the progressive failure process along the polygonal sliding surface during the failure of bedding rock slopes. Study the failure mode，scope，and stability state of bedding rock slopes. Based on the limit equilibrium theory，the nonlinear Barton-Bandis failure criterion for rock joint surfaces was introduced. Considering the hydraulic action of bedrock fractures at the rear edge of the slope，the instability range of polygonal sliding surface bedding rock slope under the instability mode is calculated. Based on the redistribution model of shear stress in the excavation loose area，and according to the determination of shear stress on the sliding surface and the peak strength and residual strength of rock mass，identify different failure modes of excavated rock slopes. It includes progressive failure mode and multi-stage multi-block failure mode with stepped sliding surfaces. Provide corresponding failure ranges and calculation methods for stability coefficient Fs to determine the stability state of the slope. The results demonstrate that the theoretical calculation method proposed in this study aligns well with the failure extents and stability coefficient Fs obtained from engineering examples and three-dimensional discrete element numerical simulations. The theoretical calculation method is capable of reasonably analyzing the progressive failure mode of bedding rock slopes along folded sliding surfaces. Therefore，it provides valuable reference for related research.

According to the construction technology process of pre-stressed pipe piles in geotechnical engineering，the phenomenon of static compression of pre-stressed pipe piles with axisymmetric constraints is analyzed，and the stress state of the surrounding soil is proposed. (1) First，the RMNLD strength criterion is introduced，which effectively describes the failure and yield characteristics of clay，sand，rock，and other geotechnical materials. Based on this criterion，a transformation stress method for a two-dimensional stress model with parameters p and q is proposed. (2) The soil around the pile is regarded as elastic-plastic material，and the hardening and softening processes of the soil are effectively described using the UH model. To enhance the failure description capability for various rock and soil types，the stress generalization of the UH model is achieved through the transformed stress method based on the RMNLD criterion，leading to the derivation of the corresponding three-dimensional soil constitutive equation. (3) A simplified analysis method is proposed to describe the static compression process of pipe piles，utilizing an exponential function of the distance-to-diameter ratio to represent radial displacement at a given radial distance. Leveraging the self-similar properties of the forces and deformations of the soil surrounding the pile，the relationship between the distance-to-diameter ratio and radial strain，volume strain，etc.，is established using the state parameter . The corresponding soil stress can then be determined using the aforementioned strain values as boundary conditions. The validity and applicability of the proposed method are confirmed through comparisons with predictions，test results，and numerical simulations.

The excavation of subway foundation pits inevitably leads to the settlement of the surrounding ground surface and buildings. Therefore，real-time and accurate monitoring of ground surface and building settlement is crucial. To thoroughly investigate the spatiotemporal characteristics of building settlement resulting from subway foundation pit excavations and enhance the precision of settlement estimations，a convolutional gated recurrent unit neural network(ConvGRU) model for the building settlement estimation was proposed based on deep learning. The spatiotemporal matrix of the building settlement was built using the building settlement data obtained from the subject and adjacent monitoring points. The convolutional neural network(CNN) was employed to capture the spatial patterns of the building settlement data，while the gated recurrent unit(GRU) neural network was utilized to extract the temporal patterns. This dual approach allows for a comprehensive analysis of the spatiotemporal characteristics of the building settlement data. The estimation performance of the ConvGRU model was compared with that of the history average model and those of four existing deep learning models using the data of the building settlement around a subway foundation pit in Shenzhen city of Guangdong Province of China. The results indicate that the estimation error of the ConvGRU model is reduced by 12.86% to 39.00% when compared to the five existing models. This research demonstrates that the ConvGRU model achieves higher accuracy and better generalization，providing high-quality building settlement estimation data for subway construction and serving as a warning mechanism for ground surface and building settlement associated with subway foundation pits.

Saline soil is widely distributed in Xinjiang area，where using saline soil as foamed lightweight soil in subgrade backfill can effectively utilize excavated waste soil，and improve the durability of subgrade in Xinjiang cold and dry area due to the good thermal insulation performance. From this perspective，the influence of density，soil content and water to dry-material ratio on engineering characteristics of saline soil-based foam lightweight soil was studied in this paper. Firstly，a series of tests were conducted to compare the thermal conductivity，fluidity，water absorption and stability of different ratios. Secondly，the unconfined compressive strength of samples under various conditions(e.g.，freeze-thaw cycle and dry-wet cycle) was obtained，in which the mechanical properties index and energy absorption rate were also studied. Finally，the pore microstructure of foamed lightweight soil was deeply analyzed by SEM and image processing software. The experimental results show that moisture and foam content greatly affect the fluidity of foamed lightweight soil. The water absorption decreases gradually with the increase in density. The foamed lightweight soil presents good thermal insulation performance. In detail，the thermal conductivity is about 1/6‐1/10 of the ordinary compacted soil. Also，there is a well-linear relationship between the density and the thermal conductivity. Under the freeze-thaw cycles，dry-wet cycles and immersed water conditions，the unconfined compressive strength of the samples decreases by 40%‐50%. Regarding microstructure characteristics，the number and roundness of pores increased with the increase in density. The structure of pores can serve as a“bridge”to establish a connection between its microstructure and macroscopic mechanical properties. In the study，it is recommended that the density of 800‐1 200 kg/m3，the salt soil content is not more than 40%，and the water-material ratio is 0.45‐0.55 can be regarded as the optimum ratio. The unconfined compressive strength of the described above ratio is about 0.6‐1.0 MPa，meeting the requirement of subgrade strength，thermal insulation，and durability.

In order to clarify the impact of basement-addition on existing pile foundations，a geotechnical centrifugal was carried out to simulate the basement-addition excavations in sandy strata. Considering existing working load at pile head in the geotechnical centrifugal，the effect of bearing characteristics of existing pile foundation during basement-addition excavations was studied. A reference for the assessment and redesign of the pile foundation for basement-addition construction of the existing building is provided. The results show that：(1) Considering the influence of the existing load，the load test is carried out after the load at pile head is held for about 9 days with 0.94 times the bearing capacity characteristic value at the prototype state. The bearing capacity of the test pile is about 6% higher than that of the test pile without considering the existing working load. In the redesign of existing pile foundations，duo to soil compaction at pile side and pile end caused by existing load at pile head，the improvement of bearing capacity can be properly considered. (2) For a friction pile with length of 20 m under the existing load，the bearing capacity loss rate is 11.5% and 37.9% when the excavation depth is 20% and 40% of the pile length，respectively. When the excavation depth is 40% of the pile length，the pile changes from friction end-bearing pile to friction end-bearing pile，and the axial force and the settlement of the pile increase greatly during the process. In practical engineering，the compressive strength of pile should be checked and the risk caused by the increase of settlement should be avoided. (3) When the excavation depth is 40% of the pile length，the side friction resistance decreases. This may be related to the softening of the side resistance，or the reduction of vertical soil stress around the pile caused by excavation. The specific reason needs further verification. In the estimation of bearing capacity of existing pile foundation，this effect should be considered.

The electroosmotic grouting method has become an ideal reinforcement method for the scaling off earthen sites due to its advantages in uniform penetration radius and extended penetration depth. However，due to the fact that this method is still in the experimental stage，there are important problems such as low electro-osmosis efficiency and unclear effective permeability range，which restrict its high-quality promotion and application. The best electroosmotic grouting method was selected through the electroosmotic grouting reinforcement experiments of five kinds of power supply-liquid supply methods，and then the field reinforcement experiments with different on-off liquid time ratios were carried out by the preferred power supply-liquid supply method. Based on the experimental results，a two-dimensional model of effective permeability range was established by multiple linear regression. On the basis of establishing the control equation of scaling off of electroosmotic grouting reinforcement，the numerical simulation results，field experimental values and model calculation values were compared. The results show that the ideal electroosmotic grouting method is continuous power supply-intermittent liquid supply，and the best on-off ratio is 6∶1，the effective permeability range is affected by the total permeability amount of the solution，the porosity of the soil，the on-off ratio and the electrode spacing. The effective permeability range is equivalent to the electrode spacing within 0–1 cm，while the effective permeability range within 1–10 cm obeys the linear law of different slopes with 4 cm as the segmentation point.

There is limited research on the constitutive model of highland lake-accumulated gravel soils. To more accurately reflect the stress-strain relationship of gravel soils，a large-scale triaxial consolidated-drained test was conducted to analyze the shear strength of the highland lake-accumulated gravel soils. The applicability of the Duncan-Chang constitutive model was also analyzed based on the triaxial test results，and improvements to the model were proposed to address its shortcomings. The results indicate that under different confining pressures，the highland lake-accumulated gravel soils exhibit significant strain hardening and contraction. The cohesion and internal friction angle are 9.4 kPa and 41°，respectively. The analysis of the Duncan-Chang model's applicability shows that for highland lake-accumulated gravel soils，neither the E-B model(where E is the modulus of deformation and B is the bulk modulus) nor the E-? model(where ? is the Poisson's ratio) can accurately reflect their stress-strain relationship. By establishing a functional expression based on the parabolic relationship between the square of the axial strain and the lateral strain，an improved Duncan-Chang model was developed. The fitting of the experimental results is good，accurately reflecting the stress-strain relationship of highland lake-accumulated gravel soils. The research findings can provide valuable references for practical engineering analysis.

In order to study the distribution of coupled physical fields in the subgrade of saline soil，relying on the subgrade of Qinghai—Xizang Railway around Qinghai Lake，based on the modified traditional salt field equation，a water-salt-heat-force coupled mathematical model was established using the COMSOL custom partial differential equation and solid mechanics field module. By analyzing the physical field characteristics of groundwater buried under the condition of water and salt recharge at the boundary of open system，the evolution laws of moisture，temperature，salt，and displacement field within the soil mass were obtained. The results show that the soil moisture content shows a four-stage development trend of decreasing firstly，increasing rapidly，then decreasing slightly，and stabilizing finally in the region from the subgrade surface to the depth of －0.4 m，and the peak moisture content increases by 2.65 % compared with the valley value. The salt content of subgrade increases initially and then stabilizes in the region from －0.4 m to －2 m of subgrade，the maximum salt content increases by 2% per meter on average，and the growing rate is proportional to the depth. The response of soil temperature to external temperature change has obvious lag phenomenon under －2 m depth of subgrade，and the lag time is about 30 d for each 1m depth increase. The maximum expansion displacement and salt frost heave height difference are 1.42 cm and 2.64 cm respectively，and the winter expansion displacement accounts for more than 32.2% of the salt frost heave displacement.

Mechanical properties of contact interfaces have significant dynamic fluctuations due to freeze-thaw cycles and rheological effect，which directly affects the long-term service performance of piles in permafrost environments，and remains to be further studied. To employing numerical simulation，firstly，a visco-elastic-plastic constitutive model for frozen soil considering the rheological effect was constructed. Secondly，the conventional Kelvin model was improved and a visco-elastic constitutive model for the contact interface between frozen soil and pile was constructed. Furthermore，a thermal-mechanical fully coupling model was established considering the rheological effect of both frozen soil and the contact interface. Verified by laboratory model test，this model was employed to reveal the evolution law of long-term service performance of pile. Results show that：The proposed contact interface model can reflect the influence of stress level on rheological effect. During the freeze-thaw cycle，the shaft resistance gradually decreased(39%) on the upper pile as the ground temperature rose，while accompanied by a corresponding increase(20%) on the lower pile. The reverse is true when the ground temperature drops. In addition，under the long-term rheological effect，the bearing capacity gradually exerts in the upper pile and the shaft resistance increases(50%)，while a 14% decrease in the deep pile correspondingly. The neutral point is at 2/5 pile length. In the end，there is a strict linkage mechanism between tip resistance and shaft resistance. In the whole pile length，the distribution of shaft resistance also has an up-down linkage mechanism，and the change of shaft resistance at a certain depth is governed by the stress state and its development trend of the whole pile. The coupling effect of rheology and freeze-thaw cycles affects the depth distribution of shaft resistance，which results in a dynamical changes in bearing mode of the pile，and ultimately has a significant influence on the service performance of the pile. Research results reveal the significant influence of rheological effect on the long-term service performance of piles in permafrost ground，which would offer a reference and basis for further simulation research，and provide theoretical support for the design，construction，operation，and maintenance of piles in permafrost regions.