Based on the field investigation and exploration, unmanned aerial survey and large-scale shaking table tests, the triggering types, characteristics and disaster-generating mechanism of seismic loess landslides are systematically studied. The results show that the earthquake-induced loess landslides have their distinctive characteristics in spatial distribution, single size, influencing area, plane modality, topographical and hydrological conditions, seismic intensity, deposit thickness and relations to seismic faults. They can be classified into three types from the perspective of the triggering mechanism: shear landslides, liquefaction landslides and seismic subsidence landslides. The shear landslides can be further classified according to the lithology of the sliding surface strata into three types: landslides within a loess layer, landslides on the contact surface between loess and mudstone, and landslides cutting into bedrock. The liquefaction landslides can be divided according to the location of the liquefaction layer into three types: deep liquefaction sliding type, surface liquefaction mudflow, and combined deep-surface liquefaction type. The seismic subsidence landslides can be divided into two types of landslides, subsidence slide and avalanche slide, according to the damage modes caused by seismic subsidence. This study may provide a scientific basis for the risk assessment, prevention and control of loess seismic landslides.

The dissolved rock mass is widely distributed in Southwest China. Under the action of karst, the continuous deterioration of structural plane strength is one of the important factors affecting the stability of rock mass. In order to explore the evolution characteristics of structural plane under dissolution and to reveal the influences of karstification on the shear mechanical properties of limestone structural plane, based on the example of the dissolution rock slope of Jiwei Mountain in Wulong, the apparent evolution patterns of limestone structural plane and the evolution laws of shear mechanical properties as well as the deterioration mechanism of structural plane are expounded by using the indoor seepage dissolution and direct shear tests on the structural plane and the three-dimensional morphology optical scanning technology. The results show that under the dual action of chemical corrosion and physical erosion, the limestone structural plane has experienced four stages: point selective dissolution, thin groove linear stable seepage dissolution, strong dissolution of dominant pipeline flow and wall slow dissolution. During the dissolution process, the surface roughness index and dissolution rate index of limestone structural plane increase with the increase of dissolution time, and exhibit a convergence trend. During the direct shear process, the corrosion structural plane shows two-stage characteristics of the initial locking and the later shear friction and sliding, and the longer the corrosion time and the higher the stress level, the more obvious the shear hardening characteristics. With the increase of the corrosion time, the main anti-sliding structure of the structural plane develops from a rigid stable microconvex to a fine solution groove and finally evolves into a deep karst pipeline, and its ultimate shear strength shows a trend of " first decreasing, then increasing". The prediction model for shear strength of limestone dissolution structural plane is established based on the Barton's formula.

Taking the composite ground improved by prefabricated vertical drains (PVDs) and cement mixing piles as the research object, an analytical model for the consolidation of multi-reinforcement composite foundation is established by considering the clogging effects of PVDs based on an assumption that their permeability coefficient decays simultaneously as an exponential function of time and a linear function of depth. Moreover, the following factors are also considered: the smearing effects of PVDs, the disturbance effects of cement mixing piles, and the radial and vertical seepages within the soil. The analytical solutions for the analytical model are subsequently deduced. After then, the solutions are degenerated to the cases with only the time- or depth-dependent well resistance. The correctness of the solutions is verified by comparing the predicted results with the measured data. Finally, the consolidation behaviors of the multi-reinforcement composite foundation are investigated through a series of the parametrical analysis. The results show the consolidation rate slows down when considering the time- and depth-dependent well resistance of PVDs by comparing that of the constant well resistance. Moreover, the reduction in the consolidation rate can be regarded as the sum of the influences of each single factor. The smaller the values of ${\theta _1}$ and the larger the values of ${\theta _2}$ and ${\theta _3}$, the slower the consolidation rate. When ${\theta _3}$ increases to a certain extent, the radial consolidation ceases, and the consolidation process will be completed only by the vertical flow at the later stage of consolidation.

The response characteristics of large-diameter monopiles under static and dynamic loads are obviously different from those of traditional small-diameter piles. To study their cyclic responses, the centrifuge tests on the large-diameter monopiles in soft clay under horizontal cyclic loading are carried out. Through a contrast study on the cyclic responses of monopiles under different working conditions, the laws of deformation characteristics, stiffness weakening and excess pore pressure accumulation of large-diameter monopiles are investigated. The test results show that the increase percentage of the bending moment caused by the number of cycles is less than 10% of the first maximum bending moment. With the increasing amplitude of one-way cyclic loads, the whole large-diameter pile-soil system can go through the elastic stage, elastoplastic shakedown stage and ratchet failure stage. The unloading stiffness, lateral cumulative displacement and excess pore pressure of soils around the piles are affected by the cyclic amplitude and number of cycles. Additionally, the unloading stiffness is also positively correlated with the pile diameter. The negative excess pore pressure can significantly accumulate at the pile toe of large-diameter rigid piles, which may offset the effects of soil weakening on the lateral behaviors of a monopile. When the amplitude ratio of the cyclic loads is below 68%, the whole pile-soil system is stable, and the lateral resistance of soils insignificantly weakens. It is recommended that the loading secant stiffness of p-y curve within the wedge soil flow zone should be reduced by 0.8 to consider the cyclic weakening effects.

In the previous seismic analyses of underground structures, only the peak structural response under the main earthquake is considered, but the damage mechanism of underground structures subjected to sequential ground motions has not been fully understood yet. Special attention is paid to the seismic damage analyses of subway station structures under sequential ground motions. The seismic damage evolution of underground structures under aftershocks and the feasible seismic performance indices are explored. Seven bedrock earthquake records, which are back-calculated from seven mainshock- aftershock sequences, are adopted. Considering the influences of the directivity and relative intensities of sequential ground motions, nonlinear dynamic time history analyses of soil-underground structure system are performed. The peak inter-story drift ratios, residual inter-story drift ratios and Park-Ang index are used to evaluate the earthquake damage of the subway station. The results show that the directivity of aftershocks has irregular influences on the structural damage. When the relative intensity between the mainshock and aftershock is large, the underground structures damaged during the mainshock are likely to transit to a severer damage state. In terms of the selection of seismic damage evaluation indexes for the underground structures, the residual inter-story drift ratios are significantly affected by the relative direction between the mianshock and the aftershock. While the peak inter-story drift ratios cannot reflect the damage to the structures caused by the aftershock intensity smaller than the main shock. Therefore, both the peak and the residual inter-story drift ratios are not suitable for representing the additional damage of the underground structures caused by the aftershock. The Park-Ang index can reflect the excessive deformation damage and the accumulated hysteretic energy dissipation, which can better illustrate the actual damage of the structures under the mainshock-aftershock sequences, and is more feasible as the damage evaluation index of the underground structures subjected to sequential earthquakes.

Based on the extensive field observations, the environmental deformation characteristics of a 31.3 m-deep excavation in Shanghai soft ground are investigated. The results show that compared with the general excavations with a depth ranging from 12 to 20 m, the ultra-deep excavation presents significant environmental effects and time-space distribution behaviors: (1) The influence zone of settlement near the long side of the excavation is related to the wall deflection distribution on the plane. The more gently the lateral wall displacement changes from the middle area to the corner, the more extensive the influence zone is. The faster the lateral wall displacement transits from the middle to the corner of the excavation, the more concentrated the influential zone is. (2) Due to the corner effects, the ground surface settlement decreases rapidly from the center to the corner, and exhibits a Gaussian distribution law, with the influence range extending to 1.5H_e (depth of excavation) behind the excavation corner. (3) The deformation of buildings exhibits distinct three-dimensional characteristics. The buildings located near the excavation corner have less settlement than those near the center of the excavation, accompanied by a certain torsional deformation. (4) When the buildings paralleling to the retaining wall cross the corner of the excavation, the most dangerous point is located within 0.5H_e near the corner of the excavation, and its damage degree depends on the relative location between the buildings and the excavation and their stiffness. (5) Compared with the conventional deep excavations, the ultra-deep excavation leads to a larger primary influence zone for ground surface settlement, reaching about 3H_e, but the location of the maximum surface settlement is closer to the wall, nearly 0.5H_e behind the retaining wall. (6) The maximum ground surface settlement δ_vm is about 0.03%~0.50%H_e, and the relationship between the maximum ground surface settlement δ_vm and the maximum lateral wall displacement δ_hm can be expressed by δ_vm=0.6δ_hm averagely.

In order to locate the hidden leakage of underground diaphragm walls, the numerical simulation studies on the detection of leakage defects of underground diaphragm walls are carried out by using the cross-hole quadrupole method and the tripole method. The results show that the curve of change rate of the apparent resistivity is more effective than that of the apparent resistivity to identify the location of the hidden leakage. A quadrupole method is proposed to accurately detect the limit power supply pole distance and measuring pole distance of hidden leakage. When the power supply pole distance exceeds a certain distance, the quadrupole method can no longer detect hidden leakage. The tripole method can accurately detect the limit power supply distance and measuring distance of the hidden leakage. When the power supply distance increases, the tripole method can perceive the location of the hidden leakage, but the measuring accuracy of the tripole method decreases gradually. Finally, the feasibility of the cross-hole electrical method to accurately locate the leakage hidden danger of underground diaphragm walls is verified through the laboratory tests, which provides the effective reference for the subsequent similar projects.

The underground utility tunnel is inevitably affected by ground fissures when crossing active ground fissures. The distribution of soil stress around the utility tunnel is studied through a large physical model test by monitoring the contact pressure between the soil and the structure around the utility tunnel. The monitoring positions are located at the bottom axis of the test box, bottom axis of the utility tunnel, bottom side line of the utility tunnel, side wall of the test box and top axis of the utility tunnel respectively. The results show that the cracks first appear near the preset ground fissure with a shear state, leading to tension state at the top face of soil. The effects of the deformation of the utility tunnel on soil stress are mainly concentrated at the end of the hanging wall and the position near the ground fissure. The stress state of the soil in hanging wall is relatively stable. However, the stress redistribution is mainly at the bottom of the utility tunnel, and the effects of the deformation of the utility tunnel on the surrounding soil decrease from the bottom axis to the sides gradually. The research results can be used for disaster prevention and foundation treatment of prefabricated utility tunnels in the areas with frequent ground fissure disasters.

The accumulation of the excess pore water pressure of saturated sand under various cyclic loadings is the cause of soil liquefaction, and it is a relatively new research idea to treat the liquefiable sand as a fluid. A comprehensive experimental investigation of the liquefaction flow characteristics is performed for the saturated coral sand subjected to the jump of 90° and the continuous rotation of the principal stresses with cyclic loading frequency (f) at isotropic consolidations. The test results show that the relationship between the normalized apparent viscosity (η/η₀) and the excess pore water pressure ratio (r_u) is significantly affected by the cyclic stress paths, the degradation of η/η₀ with r_u is a progressive process, and a positive exponential correlation exists between the average flow coefficient ($\bar \kappa$) and r_u. The correlations between η/η₀ and r_u and between $\bar \kappa$and r_u seem to be independent on the cyclic stress ratio (CSR = 0.25~0.40) and f (= 0.1~1 Hz). Another significant finding is that the apparent viscosity gradient and the average flow coefficient gradient both increase first and then decrease with the increase of r_u regardless of the jump of 90° and the continuous rotation of the principal stresses, and r_u approximately equal to 0.9 at the reversal point can be regarded as the threshold value of the excess pore water pressure ratio (r_uth) at the phase transformation state from the solid state to the liquid one, by denoting the corresponding $\bar \kappa$ as ${\bar \kappa _{{\text{th}}}}$. The data points of $\bar \kappa /{\bar \kappa _{{\text{th}}}}$-r_u for all testing conditions are distributed in a narrow band, and a virtually positive exponential relationship exists between $\bar \kappa /{\bar \kappa _{{\text{th}}}}$ and r_u.

Based on the response deformation method, a longitudinal seismic design method for underground cross utility tunnels is proposed. By changing the phase angle of the displacement function, which deforms the site for one period, the periodic ground deformation input is realized. The deformation history of T-type precast utility tunnels is analyzed, as well as the most unfavorable modes for structural deformation and internal forces around the cross node. Through the orthogonal test analysis, the seismic responses of the T-type underground utility tunnels under different site parameters and seismic input parameters are studied, as well as the sensitivity analysis for the most unfavorable modes. The results show that the periodic ground deformation input can catch the most unfavorable modes of structural deformation and internal forces, and the incident angle of seismic wave controls the most unfavorable modes of structural deformation and internal forces. The proposed method can be directly used for the longitudinal seismic design of underground cross structures.

Under the state of chamber pressure, the pore pressure of conditioned soil will be generated, and it can balance the water pressure in front of the excavation to reduce the level of groundwater infiltration. Therefore, one-dimensional compression calculation of conditioned soil is of great significance for guiding the safe excavation of shield. Therefore, based on the Boyer' s law and the discontinuous particle accumulation theory, the model for calculating the undrained pore pressure considering gradation parameters and soil-conditioning parameters is established. According to the deformation characteristics of foam-conditioned soil, the method for the initial compression modulus E_s is proposed, and the compression model for the foam-conditioned soil is established. Furthermore, the compression model is introduced into the pore pressure model, and a simple model for the pore pressure is proposed. In order to verify the reliability of the proposed model for the pore pressure and compression, the undrained one-dimensional compression tests on the foam-conditioned coarse-grained soil with different gradations are carried out by using the self-designed large-scale compression devices, and the measured values of pore pressure and compression are obtained. The comparison between the measured and calculated values shows that the pore pressure model and compression model can describe the variation of the pore pressure and compression of foam-conditioned soil under different gradations. The foam injection ratio and gradation have a significant impact on the pore pressure of the foam-conditioned soil.

The temperature field is the basis for assessing the mechanical state and water-sealing performance of the frozen wall, which is an important research direction of the artificial freezing theory. For the freezing pipes in the form of a closed circumferential arrangement, there are only analytical solutions under regular annular conditions, including single-circle and double-circle models. However, the rectangular arrangement of freezing pipes is also very common in practical projects, especially for the subway station projects that use frozen concealed excavation, and the temperature field has not yet been answered. According to the geometric consistency of rectangular and annular layouts, based on the four-pipe model, a method of "replacing squares with circles" is firstly proposed for the rectangular problem. Furthermore, considering the boundary separable properties of the steady-state heat conduction control equation and the superposition principle of potential functions, the analytical solutions of the temperature field for rectangular arrangement with eight pipes and the generalized rectangular arrangement with multiple pipes are solved. By comparing with the transient numerical results the model test ones, the correctness and the applicability of the analytical solutions are verified. The results show that the temperature field exhibits a highly rectangular distribution characteristic near the pipe layout line, and the isotherm gradually transforms to a circular shape as it moves away from the freezing pipes. The inner side of the rectangular freezing wall develops faster than the outer side, and the temperature field inside and outside the 0℃ line is significantly affected. The influences of the freezing pipe arrangement on the geometric characteristics of the freezing wall should be reasonably considered in the design of freezing scheme.

In order to establish an optimal model for estimating the uniaxial compressive strength (UCS) of rocks as well as its reasonable estimation, a fully Bayesian Gaussian process regression method (fB-GPR) is proposed by combining the Gaussian process regression (GPR), Bayesian framework and Markov Chain Monte Carlo (MCMC) simulation. The proposed fB-GPR approach is compared with different model selection methods, such as the Akaike information criterion (AIC), Bayesian information criterion (BIC), deviation information criterion (DIC), Kullback information criterion (KIC), etc. The results show that the proposed fB-GPR method performs better than other methods. In 100 random trials, the probability of M-7 being selected as the optimal model by fB-GPR method reaches 100%, and the accuracy of selecting the optimal model is far higher than other model selection methods. When the measurement noise reaches 50% of UCS standard deviation, the proposed fB-GPR can still achieve model selection accurately, which shows that the fB-GPR approach is robust and accurate, and is less affected by the measurement noise associated with UCS, comparing with other model selection methods. The proposed fB-GPR therefore provides a new way for establishing the optimal estimation model as well as reasonable estimation for the key geotechnical parameters in practice.

The integrity of rock mass is an important parameter in evaluating the quality grade of hydraulic rock mass. The traditional methods often adopt a single index that cannot fully reflect the comprehensive influences of geological conditions such as structural plane, groundwater, and unloading on the evaluation results. An intelligent evaluation method for the integrity of hydraulic rock mass is proposed by coupling with multi-source survey information. Firstly, the synthetic minority oversampling (SMOTE) algorithm is used to balance the survey information data to improve the data set structure. Then the random forest algorithm is used to predict the original rock mass integrity data and the pre-processed data, respectively. Based on the data of actual projects, the validity and applicability are verified, and the predicted results are discussed and analyzed according to different factors affecting the integrity of rock mass. The results show that the proposed method can effectively improve the evaluation accuracy of the integrity of a few rock samples by balancing the data sets. By coupling and mining the deep information of different integrity indexes, the intelligent evaluation of rock mass integrity can be realized, which provides a new method for further assisting the evaluation of rock mass quality.

In order to investigate the influences of the attenuating characteristics of reaction parameters including adsorption and degradation in the clay liners under the geomembrane on the antifouling performances of composite liners, the attenuation of reaction parameters is expressed as a specific function. The one-dimensional semi-analytical solution for transport of organic contaminants in composite liners is obtained by the Laplace transformation. The semi-analytical model is validated through the field test data. The results of dimensionless analysis show that when the reaction parameters in CCL decrease rapidly with the increase of depth (e.g, β=0.1), the breakthrough time can be reduced by 68%. The bottom concentrations and fluxes can be reduced by 40% when the diffusion and degradation dominate transport of contaminants (e.g., Pe₂≤1 and Q≥10). The attenuation effects of the reaction parameters can be ignored when the advection is the dominant process (e.g., Pe₂≥10 and Q≤1). Without considering the attenuation of reaction parameters, the breakthrough concentration of hydrophilic organic contaminants and hydrophobic organic contaminants can be underestimated by 61% and 37% respectively. The field monitoring data can be better fitted by the proposed model, and it can be used to evaluate the effectiveness of landfill liners. It can also be used for the design of composite liners and the verification of complex numerical models.

The horizontal permeable reaction barrier (HPRB) is a passive remediation technology for the volatile organic compound (VOC) vapor in the vadose zone. It has the advantage of consuming less energy during the operation. And it can be used as a long-term risk control measure for the large-scale VOC contaminated sites. A transient analytical model is proposed in this study to simulate the VOC vapor migration in the layered soil containing a layer of the HPRB. The Laplace transformation is adopted to derive the general solution in the Laplace domain, and then the Laplace inversion of the numerical Talbot method is adopted to derive the semi-analytical solution of the VOC vapor migration. It is found that the HPRB is more suitable for the contaminated sites with low effective diffusivity soil. The large depth of the HPRB is not conducive to the removal of the VOC vapor. The increase of thickness of the HPRB enhances the removal of the VOC vapor. The increase of the source concentration decay rate can reduce the peak value of the VOC concentration in the contaminated sites. The neglect of the source decay can lead to the excessively conservative design of the HPRB. Finally, the design procedure of the depth and thickness of the HPRB is proposed.

Design of stability of soil slopes inevitably involves many uncertainties, but the deterministic design method is difficult to properly consider various uncertainties. In contrast, the reliability-based design (RBD) can quantitatively consider the uncertainties in geotechnical design. Different safety criteria are adopted in the deterministic design and RBD, and possible designs with the same safety factor may have different levels of reliability and resulting in the inconsistency of feasible design domains of the two design methods (i.e., the safety criterion is not equivalent), hampering the applications of RBD in practice. The ratio of safety margin and the generalized reliability ratio of safety margin provide a useful tool to bridge the design criterion of the deterministic design method and the RBD. The generalized reliability ratio of safety margin is applied to RBD of slope stability, and the sufficient conditions for the equivalence between the deterministic design and RBD of slope stability are proposed. Based on different random field models, the equivalence between the deterministic design and RBD for two soil slope examples with one layer and two layers is studied, respectively. The results indicate that the one-layer slope satisfies the sufficient conditions when considering spatial variability, and the same feasible design domain can be obtained by the deterministic design and RBD. On the contrary, the sufficient conditions are not satisfied for the two-layer slope. The equivalence between the safety criterion of the two design methods for the two-layer slope example considering the spatial variability cannot be held.

The particle breakage of sand under high stress is obviously different from that under the normal stress. The existing studies on the evolution rules and models of particle breakage for quartz sand under high stress are relatively limited. A series of consolidation drained (CD) and consolidation undrained (CU) triaxial shear tests are conducted under the confining pressures of 2~8 MPa to investigate the evolution rules of particle breakage for quartz sand and its effects on sand strength. The relationship curves of deviated stress-axial strain under various stress levels are obtained as well as the relative breakage during shear process. Then the evolution rules of particle breakage are analyzed, the Hardin's, Lade's and Wang's particle breakage models are adopted to describe the rules, and the applicability of each model is discussed. Finally, the critical relative breakage of quartz sand that affects sand strength under high pressures is given based on the relationship between the relative breakage and the effective failure internal friction angle.

The dynamic load test (DLT) and the rapid load test (RLT) on piles both pursue the full activation of pile capacity during the impact event and evaluate the pile capacity by measuring and analyzing the impact excitation and response signals of pile top. However, the both tests require 3~5 times different weights of drop mass or inertial reaction mass as well as distinct mechanical principles. Based on the hammer-pile-soil interaction model established under the combination of the traditional pile-soil interaction mode and the drop mass system, the analysis of wave mechanics, the proof of momentum theorem and the comparison of test cases show that the momentum and energy transfer of the hammer are related to the compressive resistance of the pile and influenced by the impedance ratio or mass ratio of the hammer and the pile and the side/tip resistance distribution of the pile. The influences of different side/tip resistances and impedance ratios on the energy transfer and impulse change and the dynamic phenomena relating to the level of the pile capacity, "pile-to-hammer bump", "pile run-away"and "higher efficiency of tip resistance mobilization for piles resting on hard bedrock than that of the static load tests under the equivalent energy consumption circumstances" are analyzed, which enriches the dynamic load tests on the piles. The principle defects of RLT are analyzed, and the integration of DLT and RLT is suggested.

The existing researches on response and dynamic tensile failure of asphalt concrete core wall under spatial oblique incidence of seismic waves have great shortcomings. By considering the arbitrariness of SV-wave incident azimuth and oblique incident angles and constructing the non-uniform free field on foundation boundary based on the wave field superposition principle, an input method for spatial oblique incidence of SV waves is established. Then, an empirical formula for the change in instantaneous tensile strength of asphalt concrete with strain rate is established based on the test results. A new method for the safety evaluation of core wall based on instantaneous tensile stress and strength is proposed. Finally, the influences of incident azimuth and oblique incident angles on the acceleration and stress distributions of core wall are analyzed. The damage mechanism of core wall caused by tensile stress surge caused by spatial oblique incidence is revealed. Using the proposed method, the error of the traditional static strength judgment method for core wall damage is demonstrated. The distribution characteristics of tensile failure zone of core wall under different incident modes are clarified. The results show that compared with those under vertical incidence, the acceleration of core wall in water flow, dam axis and vertical directions can be increased by 54%, 9.2 times and 5.2 times at most under spatial oblique incidence. The tensile stress of core wall can be increased by a maximum of 14.2 times at most. Neglecting the spatial oblique incidence severely underestimates the accelerations and stresses of core wall. The more the incident direction deviates to dam axis direction and the larger the oblique incident angle, the more easily the tensile failure at the wave-facing side of core wall occurs. The traditional static strength judgment method leads to a large error of tensile failure of core wall.

Under the action of mining stresses and impact dynamic loads, the development, expansion and instability of the primary fissures in the surrounding rock of deep roadways are an important cause for the dynamic disasters of the deep surrounding rock, and increasing the integrity of broken surrounding rock and inhibiting the re-expansion of the primary fissures by using the anchor supports are an important method to prevent the dynamic disasters in the surrounding rock. To this end, the impact dynamic load tests are performed on CCNBD specimens of prestressing anchor under different preloading moments of end anchor, and three anchoring methods of end anchor, full anchor and yield pressure + end anchor are adopted using the split Hopkinson pressure bar test system. The effects of various factors on the dynamic fracture toughness, crack initiation time and crack expansion rate of the anchored rock mass are analyzed to reveal the crack-resistance effects of the prestressed anchors under impact loads. The results show that: (1) The crack-resistance effects of the prestressed anchors improve the macro-dynamic fracture toughness of the anchored rock mass and delay the micro-crack initiation time and reduce the expansion rate. (2) Both the high preloading end anchorage and the full-length anchorage help anchors develop the crack-resistance effects. Their strengthening order is full-length anchorage > end anchorage > yield pressure + end anchorage. The third method weakens the crack-resistance effects because the yield pressure structures compensates for the crack expansion space. (3) The deformation and damage process of anchored rock mass under impact loads has three stages: crack breeding stage, crack expansion stage and anchor bearing stage. The axial stress of the anchor rod with high preloading moment and full-length anchoring has the most apparent growth rate in the crack expansion stage, and the crack-resistance effects are the most obvious. The research results have some theoretical guidance and reference significance for the anchorage support projects of deep dynamic load roadways.

The loose earth pressure of shallow shield tunnel is closely related to the soil arch effect and the progressive failure of the loosen zone. Based on the ellipsoidal theory, the elliptic loosen zone model is established, and the relationship between the ground loss and the loosen zone height is proposed. The process of progressive failure and the limit state are defined. Considering the soil cohesion and the ellipse shape of the loosen zone, the lateral earth pressure coefficient under arbitrary dip angle of slip surface is obtained by means of the large principal stress trace method. The formula for calculating the loose earth pressure at tunnel top is derived and verified. The parameter analysis is carried out for the limit state and non-limit state, and the research results show that: (1) The loose soil pressure at tunnel top decreases with the increase of the internal friction angle and cohesion. (2) With the increase of the loosen zone height, the loose earth pressure at tunnel top decreases sharply first, then increases gradually, and finally tends to be stable.