International Journal of Geomechanics

Recycling waste materials, such as steel furnace slag (SFS), coal wash (CW), and rubber crumbs (RC) for transport infrastructure is environmentally friendly and offers significant economic benefits. This paper presents a fundamental study of the geotechnical characteristics of this blended matrix (SFS + CW + RC). A semiempirical constitutive model for SFS + CW + RC mixtures is proposed within the framework of critical state (CS) soil mechanics and based on the bounding surface plasticity theory. A CS surface is formulated with the changing RC contents in the waste mixtures, and an experimental relationship between the total work input (W_total) and CS parameter (M_cs)$$is established to capture the energy-absorbing capacity of the matrix. The theoretical model is validated using two sets of data, i.e., very recent triaxial test data obtained by the authors and totally independent test results from a past study conducted on sand-RC mixtures.

Massive voids form in underground mines due to the continuous removal of valuable minerals. Backfill and pillars are commonly built into the voids to provide both local and regional support. The backfill material cannot rigidly support the upper load and acts as a secondary support system to the pillars. Confinement effect effectively enhances the strength of the pillar; however, the lateral stress mobilized within the fill is contingent upon deformation of the roof and pillars, which act as a passive support. Without large-scale volumetric compression, the backfill has no effect on the prepeak stage of the pillar. In this case, the friction resistance of the interface between the pillar and backfill plays a key role in increasing the load capacity of the support system. This paper presents an analytical solution based on the load transfer of friction force in the interface, which involves the shear behavior of the interface, the mechanical properties of the pillar, and the pillar dimensions. Experimental results showed that the bearing capacity of the support system was strongly dependent on the interface shear parameters, lateral stress, and fill ratio. Comparisons were made between the proposed solution and experimental results. The results may represent a workable understanding of the interaction mechanism between backfill and pillars and assist in the design and optimization of underground mine support systems.

Although accurate evaluation of pavement friction promises significant safety benefits to highway agencies, the development of such models has proven to be challenging due to the lack of complete and quality pavement surface data for friction studies. In this study, the state-of-the-art three-dimensional (3D) laser imaging technology and the Grip Tester, which is a type of continuous friction measurement equipment (CFME), are used to collect 1-mm 3D pavement surface data and friction data, respectively, at highway speed in the field; whereas the Aggregate Imaging System (AIMS) is used in the laboratory to analyze the surface characteristics of aggregates. Forty-five pavement sites, including six major preventive maintenance (PM) treatments and seven typical types of aggregates in Oklahoma, are identified as the testing beds. Multiple field data collection events have been performed from 2015 to 2017. Panel data analysis (PDA), which is able to investigate the differences of cross-sectional information (the time series changes over time), is conducted to identify the most significant influencing factors for pavement friction prediction model development. Statistical analyses indicate that the random-effects panel model outperforms the fixed-effects model and the traditional ordinary least-squares regression model. The panel model developed in this study could assist decision makers in the selection of PM treatments and aggregates for optimized skid resistance performance.

Vertical plate anchors are being widely used in the construction of earth retaining structures. In the present study, experimental and numerical investigations were carried out to develop an understanding of the behavior of geocell-reinforced vertical plate anchors in sand. The parameters studied include anchor embedment depth, strength of geocell reinforcement, and geocell–soil interface friction angle. It was observed that geocell reinforcement can significantly enhance the anchor capacity, both at shallow and deeper embedment. The unreinforced anchor, at an embedment depth of about seven times its height, tended to reach a critical stage beyond which the load-carrying capacity did not increase much. However, with geocell reinforcement, this limitation was overcome and the pullout capacity continued to increase further. Moreover, a shallow anchor with geocell reinforcement could perform better than the unreinforced anchor placed deeper. The central portion of the geocell mattress close to the anchor plate actively sustains the loading, and the end portions serve in a secondary manner to mobilize passive resistance from the surrounding soil. For better performance improvement, the geocell–soil stiffness ratio should be in the range of 5–30. With increases of the geocell–soil interface friction angle beyond 1.8 times the soil friction angle, further increases in performance improvement tend to be negligible.

The prefailure instability of the granular soil slope due to lateral stress relief under the drained condition has been investigated experimentally under the constant shear drained (CSD) path in the literature. However, consensus on the drained instability and associated instability stress ratio, ${\eta }_{\mathrm{I}\mathrm{S}}$, has not yet been reached due to the differences in experimental setup and interpretation techniques. Earlier studies applied inconsistent real-time specimen-area correction and used three different interpretation techniques for ${\eta }_{\mathrm{I}\mathrm{S}}$: (1) rapid generation of axial strain, (2) change of volumetric dilation to contraction, and (3) a modified second-order work criterion. In this study, an advanced triaxial testing system was used for the CSD path to apply real-time area correction, and it was found that the aforementioned three techniques estimated slightly different ${\eta }_{\mathrm{I}\mathrm{S}}$. The rate of reduction of confining stress ($\Delta {\sigma }_3/\mathrm{d}t$) and initial state were found to have a significant effect on ${\eta }_{\mathrm{I}\mathrm{S}}$. The intriguing observation is that unlike undrained instability, drained instability can be recovered to a stable state by changing $\Delta {\sigma }_3/\mathrm{d}t$. The drained stress path used in this study can be applied in analyzing and designing tailing dams (e.g., Fundão tailing dam). Further, to compare these findings, a critical state line (CSL) and instability curve were developed through a series of undrained and drained triaxial tests. The results show that ${\eta }_{\mathrm{I}\mathrm{S}}$ and the state parameter at the onset of drained instability exhibit a narrow trend, which also matches with the instability curve from undrained tests.

In this study, a new wetting deformation simulation method for core-wall rockfill dams is proposed based on a wetting model that considers the upstream damming rockfill as softening materials when they are immersed and saturated. In this simulation process, as the water level in the reservoir rises, the calculation of dam wetting deformation is realized by gradually reducing the deformation modulus of the rockfill materials. To verify the method’s reliability, the collapse settlement caused by the wetting deformation of the Guanyinyan rockfill dam during impoundment is numerically simulated, and the results show that the calculated wetting deformations agree well with the measured ones. The wetting of upstream rockfill materials would cause an adverse deformation trend, which may lead to crack occurrence at the crest and on the upstream slope.

The stochastic analysis of unsaturated rainfall-induced slope instability has been studied by many researchers. However, the applicability of random FEM (RFEM) has not been used enough to assess the effect of the uncertainty of soil parameters and suction on these types of slope. In this study, considering the uncertainties of soil parameters leads to an uncertain wetting branch of the soil–water retention curve (SWRC); therefore, a stochastic permeability coefficient curve with respect to suction is obtained. For this purpose, a computer program has been coded to determine the reliability indices of unsaturated slopes under different rainfall intensities and seepage flow using RFEM. In a subsequent part of the paper, the functions of slip surface are obtained for different simulations and rainfall intensities by fitting the points with maximum shear strain. Finally, the uncertainties of slip surfaces are analyzed.

In this paper, the seismic active resistance of a slanted cantilever retaining wall holding a cohesionless backfill is computed using the method of stress characteristics in association with the pseudodynamic approach. Dissimilar to the past investigations reported in literature adopting the limit analysis or the limit equilibrium method, a preordained failure mechanism is not assumed in this analysis. Various parameters such as inclination and roughness of the wall, angle of internal friction of the backfill soil, and phase difference of the seismic waves are considered in this study. The present results are compared with the results reported in literature. The present values of seismic active earth pressure coefficients are found to be higher than the values obtained from the pseudodynamic analysis, assuming the linear failure surface, but are found to be lower than the magnitudes derived from the pseudostatic analysis where constant ground acceleration is considered throughout the influence domain.

The spatial variability of mechanical properties of deep cement mixed (DCM) soil leads to uncertainty in performance of DCM column–supported embankments. While adequate safety levels can be achieved using a large factor of safety in a deterministic design, a reliability assessment helps to find an optimum safety level for the design by managing variability parameters. However, only few studies have been conducted on the performance of DCM column–supported embankments considering the variability of mechanical properties of columns. None of these studies have considered the strain-softening behavior of DCM columns along with the strength variability. This study investigates the reliability of a DCM column–supported embankment considering the spatial variability of mechanical properties and strain-softening behavior. Compressive strength, elastic modulus, and cohesion of DCM soil were considered as stochastic parameters in this study. The analysis was carried out using Monte Carlo method assuming three different cases of spatial variability. The reliability response of the embankment to changing coefficient of variation (COV), mean compressive strength, and the spatial correlation length are discussed. Results demonstrate that the reliability performance level of the embankment can be conservatively predicted by considering the variability of mean DCM column strength without considering the spatial variability. Based on the results of this study, a method is proposed to determine the overdesign factor required to achieve a target reliability for the embankment.

One of the most basic topics in soil mechanics is the geometrical features of the particles and the soil arrangement. Although these features are of high relevance, many definitions and processes used to evaluate them are not as precise as they could be. The conventional laboratory tests can require a certain amount of effort and lack objectivity. Also, the geometrical properties are important to understand the physical phenomena that occur in the medium and characterize the soil. Thus, this study presents several image processing algorithms and mathematical formulations to analyze sandy material through microtomography. These algorithms are capable of generating supervised information about the characterization process of soil morphology by using mathematical definitions. The algorithms are capable of generating the grain-size distribution, the soil tortuosity, the void ratio, and the sphericity of grains. Moreover, this work considers the fractal nature of the soil using non-Euclidean approaches for a broader comprehension of the soil geometric properties.

The anchor trapdoor provides a simple and economical system widely used to study the stability of similar geotechnical engineering structures, such as foundations and tunnels. In this study, the passive and active failure mechanisms were proposed to estimate the uplift capacity for anchor-trapdoor design. The nonhomogeneous soil properties were investigated with the help of the slice method, which was inspired by interpretations of integration. Within the framework of the upper-bound theorem, a nonlinear slip-line layout optimization procedure with classical tools of variational calculus was presented for a single layer to provide the basis for multilayer analysis. This approach was extended to put forward an innovative failure mechanism describing the curvilinear function for circular anchor plates in a three-dimensional space. To investigate the effects of nonhomogeneity on the stability of the anchor trapdoor, case studies were conducted for soils with linear changing cohesion and layered soils. To make safe guidance for engineers facing geotechnical limit-state problems, considering both linear and nonlinear failure modes, the stability factor calculated by a few examples was validated by lower-bound solutions and results in commercial packages.

The small strain shear modulus (G_max) is an important parameter in geodynamic problems. To predict the G_max of unsaturated soils that are normally subjected to complex drying and wetting processes, the effect of hydraulic hysteresis needs to be evaluated. Although several equations have been proposed in recent years, limitations still exist, requiring more research studies in this field. In this study, G_max was investigated in a multistage test during several drying-wetting cycles and a loading-unloading cycle of net stress. The results revealed four key factors that directly influence the magnitude of G_max: the void ratio, net stress, matric suction, and degree of saturation. Although variations of the void ratio, net stress, and matric suction cause persistent responses of G_max (i.e., if all other factors remain unchanged, G_max would then be reversely proportional to the void ratio and directly proportional to the net stress and matric suction)_, variations in the degree of saturation result in different responses. A decrease in the degree of saturation may induce a reduction or growth of G_max because, on the one hand, it reduces the effect of matric suction, whereas on the other hand, it increases the total effect of van der Waals attractions and electric double-layer repulsions. At the same stress state, a reverse trend, induced by an increase in the degree of saturation, will occur with a growth in the effect of matric suction and a reduction in the combined effect of van der Waals attractions and electric double-layer repulsions. An analysis of the results showed that hydraulic hysteresis occurred in all the stress loops, and it directly influenced the response of G_max. The effect of hydraulic hysteresis can only be captured if the van der Waals attractions and electric double-layer repulsions are considered. A model to estimate G_max while incorporating the van der Waals attractions and electric double-layer repulsions was developed, and it was found to provide a good agreement with the experimental measurements.

To investigate soil-arching behavior under localized loading, trapdoor tests were conducted in this study with a model ground subjected to repetitive static and cyclic loads considering the influence of loading magnitude, loading frequency, loading width, and fill height. An analogical soil of aluminum rods was used as the fill material in a test box to create a two-dimensional condition. During the trapdoor tests, the particle image velocimetry technique was employed to analyze the deformation patterns of the fill. Four trapdoor tests with four different fill heights under fill self-weight were conducted as baseline cases for comparisons with the previous studies in the literature and the trapdoor tests under surface loading in this study. The theoretical solutions in the literature were used to calculate the soil-arching ratios for comparison with those from the trapdoor tests. The test results indicate that the application of surface loading caused degradation of soil arching, and cyclic loading had more effect on this degradation. A thicker fill section had more resistance to surface loading in the degradation of soil arching. The degradation of soil arching caused by surface loading increased with an increase in the loading magnitude and loading frequency and a decrease in the loading width.

Dilatancy (volumetric expansion) during yielding is an important aspect of geomaterial behavior that can be mathematically defined using a dilation model. A common way of modeling dilation has been to use a dilation angle, ψ, which is derived from the Mohr-Coulomb potential function assuming ψ is a constant. Theoretically, the constant ψ value is directly related to a ratio of plastic strain increments using an equation. Several researchers have found, however, that the ratio of plastic strain increments as measured in laboratory tests both changes during the deformation process and depends on the confining stress within a material, meaning that in general, ψ cannot be constant. In this study, we demonstrate the origin of this inconsistency and derive the relation between the dilation angle and the ratio of plastic strain increments acknowledging confinement dependency of ψ. Then, we show how an iterative scheme can be used to calculate the confinement-dependent dilation angle values from a series of triaxial tests. The dilation angle values calculated using the novel approach are found to be consistently higher than those obtained using the conventional approach that assumes a constant ψ as part of the calculation.

A new method is proposed for water inflow prediction during heavy rain while tunneling through karst fissured zones. The proposed prediction method is a semianalytical and seminumerical method. First, an analytical determination method is put forth for a dynamic water pressure head. The main factors in the determination method include the surface catchment area, rainfall, porosity, surface infiltration coefficient, and initial seepage field (as initial pressure head and initial water inflow). Then a numerical prediction model is built based on the determination method of the dynamic water pressure head analysis. Stationary and time-dependent studies are used to solve the prediction model. Scenarios with both stable and variable rainfall are simulated. Results with stable rainfall indicate that water inflow always increases with rainfall, and water inflow under the smallest rainfall will first reach its limit. Variable rainfall is applied to simulate actual tunnel engineering. Variable rainfall is set through imitating an actual rainfall depth-duration curve. It is found that the trends of numeric and actual results are basically consistent. The water inflow reaches the peak point after a rainfall. The delay time of the numerical simulation results agrees with this under actual conditions.

The creep characteristics of the key unit rock in a slope potential slip surface play an important role in the evolution of slope slip. In this research, a systematic restrictive shear creep experimental study was conducted to investigate the restrictive shear creep characteristics of the key unit rock in a slope potential slip surface. A new plastic nonlinear model (PFY model) was developed to characterize the progressive-failure creep characteristics reflected in the process of the restrictive shear creep of rock. A variable-parameter restrictive shear creep model was established to reflect the progressive-failure characteristics and describe the whole process of restrictive shear creep of the rock and the two failure mechanisms of the rock, which can provide the reference base for the prediction of the time of shear creep failure of the rock. The constitutive equations characterizing the shear creep characteristics of the rock were established, and the methods to solve model parameters were determined. In addition, according to the obtained experimental data, the model parameters were solved, and the theoretical curves of the model were fitted with the experimental curves. The new model was validated, and the sensitivity of the model parameters to creep-failure time were analyzed. The results show that the overburden pressure and the maximum creep strain of the key unit rock in a slope play a leading role in the progressive failure and instability of the rock, and the shear creep rate with time fits the power function well. The theoretical curves of the new model match well with the experimental curves, and the differences in the creep-failure time between the model calculation values and the experimental values are small, which indicates the new model is appropriate in describing the creep characteristics of the rock. For the shear creep-failure time, the viscous coefficient ${\eta }_2$ is the most sensitive, the accelerating point strain ${\gamma }_a$ comes next, and the progressive-failure coefficient $\beta$ is the least sensitive.

The deformation monitoring data including the horizontal and vertical displacements from Chengchao Iron-ore Mine were gathered over 10 years. Combined with the field macrofailure characteristics at the ground surface, the time-dependent ground movement behaviors induced by the underground orebody excavation were investigated. Results show that the time-dependent deformation behaviors within the surrounding rock masses are very different with the distance to the mined-out areas increasing. Close to the mined-out areas, the time-dependent deformation behavior can be described by two distinctive stages termed as the “primary steady-state creep” and “progressive” stages. Slightly away from the mined-out areas, the time-dependent deformation behavior can be divided into three distinctive stages termed as the primary steady-state creep, “regressive,” and progressive stages. Far away from the mined-out areas, the surrounding rock mass creeps with a low deformation rate, and the time-dependent deformation behavior is still at the primary steady-state creep stage. The analysis shows that the groundwater can have a remarkable effect on the deformation of the rock masses by causing the movement rate to become faster. The main reason why the regressive stage occurs in the southeastern area of the main transportation tunnel can be attributed to the rainfall, and the period for the maximum deformation rate within the surrounding rock masses corresponds to the rainfall season of the year. The onset-of-failure (OOF) point within the displacement–time curves can be defined as when the failure surface begins to form at the deep part of the rock masses. The time-dependent ground movement behavior of the mine can be divided into four stages: the initial caving of the roof strata, the large-scale caving of the roof strata, when the failure surface begins to form at the deep part of the rock masses around the stack-site and the mine road, and when the failure surface begins to form at the deep part of the rock masses around the southeastern area of the main transportation tunnel.

The response of structure–foundation systems subjected to static, cyclic, or dynamic load depends on the interaction between them. The behavior of such systems is strongly nonlinear and inelastic and is affected by factors such as the interface behavior, e.g., relative motions, pore water pressure, and surface roughness. This two-part paper presents new and unique constitutive modeling, testing, and validations for interface behavior affected by factors such as surface roughness, confining pressure, amplitude, and frequency of cyclic loading. The aim of this work is to gain an improved understanding of the mechanical behavior of saturated interfaces between structural and geologic materials subjected to cyclic loading with emphasis on the effect of surface roughness of structural material; this is important for safety and to improve analysis and design of geotechnical structures. Monotonic and cyclic simple shear experiments are conducted on Ottawa sand–steel interfaces under undrained conditions using the cyclic multidegree-of-freedom shear device (CYMDOF-P) with pore water pressure measurements. The effect of various parameters such as normal stress, surface roughness of steel, type of loading, and the amplitude and frequency of the applied displacement in cyclic loading are investigated in Part I. The constitutive modeling for the interfaces are presented in Part II.

The data from comprehensive interface simple shear tests on saturated interfaces, presented in Part I of this paper, were used to calculate the parameters in the disturbed state concept (DSC) model. The resulting parameters were then used to verify the DSC model by back-predicting tests from which parameters were determined and by independent tests that were not used in parameter determination. The model predictions, in general, were found to provide highly satisfactory correlation with the observations. In the context of DSC, the concept of critical disturbance was developed to identify initiation of liquefaction in saturated Ottawa sand–steel interfaces; this method is based on microstructural changes in the deforming material leading to instabilities like liquefaction. A new and unique constitutive modeling approach, testing, and validations for interface behavior affected by factors such as surface roughness, confining pressure, amplitude and frequency of cyclic loading are presented in this two-part paper.

A series of uniaxial compression experiments was conducted on rocklike specimens containing a large-opening crack to investigate the effect of inclination angle and length of crack on the uniaxial compressive strength (UCS) and failure mode. A particle flow code, based on the discrete element method, was used to simulate the propagation process and reveal the fracture mechanism of a preexisting crack with a large opening. The failure of these large-opening crack specimens was mainly caused by the development of shear secondary cracks and not by tensile wing cracks, although the shear cracks initiated from the crack tip after the tensile wing cracks. The preexisting large crack remained open after compression failure. The mean rupture angle was approximately 63.6°, and it was independent of the crack inclination angle and length. An increase in inclination angle from 0° to 90° or a decrease in crack length from 30 to 0 mm led to a nonlinear increase in UCS owing to the increase in effective shearing resistance area. The experimental results are in accordance with the numerical results, and they provide a significant understanding of the UCS and failure characteristics of rocks with large-opening cracks.

This study used adaptive finite-element limit analysis (AFELA) to calculate the undrained bearing capacity of strip footings placed adjacent to two-layered slopes. Both the upper bound (UB) and the lower bound (LB) of the dimensionless undrained bearing-capacity factor N_s were obtained, with the relative errors being within 6%. The results are presented in the form of design tables and charts for a wide range of parameters. Meanwhile, the effects of parameters on N_s were also investigated, such as the undrained shear stress ratio c_u1/c_u2, the dimensionless strength ratio c_u1/(γB), the normalized thickness of the top clay layer D/B, the slope height H/B, and the slope angle β. In addition, comparisons with previous studies were also conducted, and the failure mechanisms were determined.