The studies dealing with slope stability analyses within the framework of limit equilibrium can be classified into two major categories : the first is formulation of factors of safety i.e. the problem of how to calculate the factor of safety for a given slip surface, and the second refers to minimization of factors of safety i.e. the problem of how to determine the minimum factor of safety and the corresponding critical slip surface. From the latter aspect of the studies, an efficient and practical minimization approach for the 3-D slope stability analysis, based on Dynamic Programming (Baker, 1980) and Random Number Generation, is proposed in this paper. This approach can simultaneously provide the minimum factor of safety, and the corresponding critical slip surface for a general 3-D slope. The 3-D simplified Janbu method, which was developed by Ugai (1988) and Ugai and Hosobori (1988), is incorporated with the proposed approach. The slope stability analysis presented in this paper has two noticeable characteristics : 1) the potential slip surface is not restricted to any particular shape; 2) the approach is available for either inhomogeneous problems or cases in which no symmetrical plane appears. The developed analysis method was implemented into a computer program called 3D-DYRANUT. Using this program, several example problems, including a multilayered convex slope as well as an actual slope failure caused by an earthquake, were analyzed and the results are reported.

View full abstract

This paper identifies recursive symmetry-breaking bifurcation phenomena as major sources of complexity in soil shearing behavior. By means of the group-theoretic bifurcation theory, a complete rule for bifurcation is presented for a cylindrical domain made up of uniform geotechnical materials, such as soil, sand, and rock. The bifurcation behavior of soil has two major phases : i) the formation of diamond, oblique stripe and echelon modes with high spatial frequencies at an earlier stage, and ii) the deformation pattern change and shear-band formation at a later stage. This behavior is indeed a recursive loss of symmetry that enlarges and changes deformation patterns. The mathematical knowledge of recursive bifurcation provides an overall view of the soil behavior.

View full abstract

Three-dimensional (3D) and two-dimensional (2D) model tests of tunnel excavation problems with two different depths were carried out to confirm the analytical results obtained previously. 3D and 2D elastoplastic finite element analyses corresponding to the model tests were also performed. The model test results were compared with the results obtained from numerical analyses. Through comparison supported by the model tests and numerical analyses, it was determined that a clay foundation results in a large surface settlement and a much wider surface settlement curve than a sand foundation with the same displacement. It is also shown that the settlements are different whether or not the construction procedure is taken into account (3D or 2D) even if the displacements imposed are the same at the end of excavation. In addition, the earth pressures were predicted through numerical analyses. It was concluded that the distribution of earth pressures in the sandy subsoil near the bottom of lowering basement is influenced considerably by the smoothness of basement, dilatancy characteristic and the magnitude of the prescribed displacement.

View full abstract

A discrete element code has been developed to simulate assemblies of two-dimensional polygon-shaped particles to study the behavior of granular materials such as rockfill. Two different contact laws (overlap-area and linear) were introduced to model the real interaction between particles. The results of simulations show that both contact laws yield similar macroscopic behavior. The focus of this study is on effects of confining pressure on shear strength and dilatancy of simulated rockfill. It is demonstrated that shear strength and dilatancy decreases as confining pressure increase. The results are similar to data obtained from experiments on rockfills.

View full abstract

The mechanical behaviors of soils are remarkably influenced by their physical and chemical properties and by environmental factors. The physical properties are classified as primary and secondary categories; primary properties are permanent parameters of soil grains such as grain density, shape and size distribution, and secondary properties are non-permanent or changeable parameters of soil fabric. The long term objective of this study was to reveal the mechanical behavior of soils with different primary properties. The physical properties of sands with different primary parameters were investigated extensively. Some 200 granular materials including sands and a few artificial materials were prepared. The primary physical properties such as grain density, shape and size distribution of all samples were measured. Some index properties characterizing physical properties, such as maximum and minimum void ratios, crushability and angle of repose of sands, were also measured for all the samples. Based on the test results, the relationships between the primary properties and the index properties are discussed.

View full abstract

A numerical method, based on the smeared shear band technique, is presented for the localized deformation analysis. In contrast to the conventional shear band analyses, the present research considered two shear bands for a localized element. The constitutive relation was formulated by coupling the two bands, and was named Coupled Shear Band Method. This method was applied to simulate experimental research on the seismic active earth pressure developed against a retaining wall supporting dry sand as backfill. The method can capture the progressive failure characteristics of the soil mass efficiently. The numerical results show that the simultaneous mobilization of the maximum angle of internal friction along the failure surface does not take place. The nature of the failure zone formation was seen to govern the distribution of the active stress. Good agreement has been achieved with the experimental observations regarding various earth pressure parameters using the method of analysis presented as compared to the analysis that assumes continuity of stress in the backfill throughout the entire deformation field.

View full abstract

A new construction method is described which aims at making reinforced soil structures, such as a geosynthetic-reinforced soil bridge abutment, very stiff and elastic. To make the deformation of a reinforced soil mass nearly elastic, a large preload is first applied by introducing tension into metallic tie rods which penetrate the reinforced soil and are connected to top and bottom reaction blocks. The tensile force in the tie rods functions as a prestress increasing the confining pressure and the stiffness of the soil is kept high. Compressive or tensile surcharge loads applied to the top block are supported by both the tie rods in tension and the soil mass in compression. Prestress is introduced in the reinforcement by preloading the tie rods, which contributes to maintaining the integrity of the soil mass. A working example and the result of a full-scale test are presented.

View full abstract

Based on the kinematic approach of yield design theory, a method for the evaluation of seismic effects on the ultimate bearing capacity of shallow foundations on Mohr-Coulomb soils has been developed and successfully tested for both theoretical and experimental results. A new kinematic mechanism has been introduced, allowing for the uplift of the foundation under strong load eccentricities. An empirical formula which approximates the results of the kinematic approach has been derived for a simple investigation of the effects of inclination and eccentricity of seismic actions, as well as of soil inertia. The importance of these effects on the reduction of bearing capacity is discussed thoroughly, leading to the conclusion that soil inertia plays a negligible role compared to the seismic actions transmitted by the structure. Furtherly, taking load eccentricity into account in a pseudo-static approach may lead to very high safety factors in the design of shallow foundations.

View full abstract

During the Hyogo-ken Nanbu Earthquake which occurred on January 17, 1995, liquefaction occurred extensively throughout the Hanshin area (around Kobe and Osaka in Japan). On Port Island, about 20 m in thickness of a manmade fill consisting of gravelly sand, which is a type of weathered granite widely distributed in the Rokko mountain range (locally called Masado), was liquefied. In order to investigate the cause of the liquefaction of the Masado layer, a series of liquefaction tests were performed on high-quality undisturbed gravelly samples recovered from this hydraulic fill on Port Island using the in-situ freezing sampling method, about five months after the earthquake. The reliquefaction potential and the settlement of the hydraulic fill soil following the liquefaction were also investigated. Based on the laboratory test results obtained in this study and field observations made after the earthquake, the following can be concluded. 1) The mean diameter, D₅₀, of Masado fill is in the range of 1.7 to 3.7 mm, the maximum diameter is about 37.5 to 101.6 mm, and it has a large amount of gravel (almost more than 50%). Based on these physical properties, the Masado fill can quite reasonably be classified as gravel. The dry density ranges between 1.7 and 2.0 g/cm³. This value is nearly equal to that of natural gravel deposits, and is much larger than the maximum dry density of dense Niigata sand and Toyoura sand. 2) In spite of its large dry density and gravel content, the liquefaction strength of Masado fill was found to be 0.15 to 0.23, which is very low, and nearly equal to that of Toyoura sand with a relative density of about 70%. The liquefaction strength is almost consistent with that of clean sands obtained from undrained cyclic triaxial test indicated by Yoshimi et al. (1989) for the same N₁-value. This result suggests that the liquefaction strength of Masado fill obtained in cyclic triaxial test can be roughly determined by the simplified procedure using the N₁-value. 3) Considering the test results, which indicated that the artificial Masado fill has a large particle size but low liquefaction strength, the range of the particle size of the soil whose liquefaction potential should be investigated as described in some of the design standards, is hoped to be reviewed. 4) A correlation between the post liquefaction volumetric strain (ε_v) and the maximum shear strain (γ_max) of the undisturbed Masado samples was found in laboratory tests. Based on the correlation between γ_max and ε_v, the total settlement of the Masado fill layer at the sampling site was estimated to be about 32 cm, this value is almost consistent with the observation results (30 to 40 cm), as reported by Tokimatsu et al. (1996). 5) The effects of the strain history (liquefaction) on the liquefaction strength of Masado fill was known to be negligibly small.

View full abstract

This work focuses on the mechanical properties of cohesive soils at ultra-low stresses. It consists of an experimental program to establish the water content-effective stress-undrained shear strength parameters of freshly-deposited cohesive soil. The effective stress levels in this program are lower by three orders of magnitude or more than in traditional geotechnical practice. The experimental program was conducted on fully consolidated Boston Blue Clay beds of various ages, 1.7 to 11 cm in thickness deposited from concentrated slurries. It consisted of water content profile measurements at 0.5 cm intervals using a vacuuming technique, and undrained shear strength profile measurements at 1 cm intervals using a specially designed Automated Fall Cone Device. For effective stresses below 400 Pa and void ratios between 2.5 and 8, a non-unique void ratio-effective stress curve dependent on the bed height was found. The water content increased at the same depth for increasing bed thickness. The undrained shear strength, which varied between 10 and 120 Pa, increased with depth for all beds. The strength was mostly water content dependent at early age, then became mostly effective stress dependent as the bed aged. At the same effective stress and water content, the strength at 6 days was higher than at 3 days. This observed behavior is attributed to thixotropy.

View full abstract

A series of permeability tests on high-quality undisturbed sand samples was performed using a triaxial cell. The high-quality undisturbed sand samples were recovered by the in-situ freezing sampling method from four sites. The effects of anisotropy, confining stress, void ratio and relative density on permeability were investigated. Next, the effects of the physical properties of the undisturbed samples on the permeability coefficient were evaluated. The coefficient of permeability of the high-quality undisturbed samples was also compared with those estimated from Creager's proposal and Hazen's equations, using the soil gradation characteristics of the undisturbed samples. Based on the test results, the following conclusions were noted. (1) For the sands tested, the coefficient of permeability in the horizontal direction (H-specimen) was found nearly equal or only slightly larger than that in the vertical direction (V-specimen). The maximum difference of the coefficient of permeability between the H and the V-specimens was only about 70%. (2) The permeability of the sands was slightly influenced by the confining stress, such that the permeability decreased with increasing confining stress. A small decrease in the permeability due to the increase in confining pressure, can be interpreted to be caused by the minute decrease in the void ratio caused by an increase in the confining stress. (3) No relationship between the void ratio/relative density and the permeability coefficient, which is observed for all types of sand, was found. (4) A good correlation between the coefficient of permeability and the percent fines content of in-situ sandy soils was noted. Although the data was limited, a simple tentative equation was proposed to relate the coefficient of permeability and the fines content. (5) It is possible to estimate the coefficient of permeability of sandy soils roughly for design purposes, both from Hazen's equations and Creager's proposal, in the range of D₁₀ larger than 0.5 mm or D₂₀ larger than 0.1 mm, respectively.

View full abstract