In this paper, based on a elastoplastic model that can properly take into consideration the influences of the density, the structure formed in depositary process and the stress-induced anisotropy of soils, 2D and 3D finite difference-finite element analyses considering the soil-water coupling problems are conducted to simulate element test of sand specimen subjected to cyclic load under different loading conditions in triaxial test using the code named as DBLEAVES. In the element test, through stress-strain relation is regarded as uniform within the specimen of soil, it is usually non-uniform in reality. Due to limitation of element test devices for soils in laboratory test, a perfect element test is impossible in reality due to some factors such as an initial imperfection of test specimen, friction between loading plates and specimen, gravitational force of specimen and etc. It is, therefore, necessary to answer the question whether the results of an element test commonly used in laboratory test are convincing in determining the mechanical behaviours of the soils. If the answer is yes, then how much is the influence of these factors on the tests which is usually regarded as an element behaviour. In the simulation, the element is considered as a boundary value problem. The influence of those factors such as amplitude and frequency of cyclic loading, confining stress is also considered in detail. In the simulation, all values of material parameters are kept the same, which makes the numerical simulation being meaningful.

View full abstract

This paper presents a study on time effects on total suction of bentonite-based materials under constant water content conditions. Three different types of bentonite (i.e., MX80, Calcigel I, and Calcigel II) and a natural expansive clay (i.e., London Clay) were used. Total suction was measured using chilled-mirror hygrometer technique. The measurements were performed on specimens aged 1 hour, 6 hours, 1 day, 2 days, 1 week, 2 weeks, 1 month, and 2 months. The results show that the true equilibrium state for two types of bentonite (i.e., MX80 and Calcigel I) used is time dependent. Total suction increases with time and this behaviour is controlled by the characteristics of the bentonites. Total suction of the other specimens (i.e., Calcigel II and London Clay) is not affected by curing time. The changes in total suction with time as the specimen's age are attributed to non homogeneity of the water content distribution in the micro- and macropores of the bentonites and inaccuracy of the device used.

View full abstract

In this paper, we study the influence of the scaling functions in Perzyna's type elasto-viscoplastic models for predicting time-dependent behaviour of natural clays. The constitutive models were developed based on the overstress theory with different scaling functions and on the elastoplastic model S-CLAY1S which accounts for induced anisotropy and gradual degradation of apparent soil bonds. Laboratory tests were simulated on natural clays under one-dimensional and triaxial conditions to evaluate the scaling function influence in the modelling of the strain-rate effect on soil strength and creep effect. The influence of scaling functions on modelling the time-dependent behaviour under pressuremeter condition was also numerically examined. This test is treated as an example of a boundary value problem, which also allows us to see if this in situ testing is capable of deriving the parameters controlling the time-dependent behaviour.

View full abstract

Mechanically stabilized earth (MSE) structure has been widely accepted as a retaining structure. Its construction cost is mainly controlled by backfill materials, which are generally coarse-grained soils, and reinforcement type (steel volume). The present paper introduces a new cost-effective reinforcement, designated as “Bearing Reinforcement”. It is composed of a longitudinal member and transverse (bearing) members. The longitudinal member is made of a deformed bar, which exhibits a high pullout friction resistance. The transverse members are a set of equal angles, which provide high pullout bearing resistance. The maximum pullout bearing resistance of a single isolated transverse member, σ_bmax, can be determined by using the plasticity solution based on the modified punching shear failure mechanism. Influential factors governing the mobilization of pullout bearing resistance are spacing, S, leg length, B, and numbers, n of transverse members. The larger the S/B, the lower the transverse member interference. The S/B ratios of <3.75 and >25 are referred to as full and free interference, respectively. The relationship between normalized average pullout bearing stress, σ_bn/nσ_n and pullout displacement, d, where σ_bn/n is average pullout bearing stress of the bearing reinforcement with n transverse members and σ_n is applied normal stress, is practically identical for the same level of transverse member interference. This relationship can be modelled by hyperbolic function. From this finding, a suggested procedure for estimating pullout characteristics (maximum pullout resistance and pullout force versus displacement relationship) of the bearing reinforcement for any level of transverse member interference (any S, B, and n) based on a one point test on the bearing reinforcement with a single isolated transverse member is proposed. Good agreement has been obtained between the predicted and the measured pullout characteristics. This suggested method is useful for the internal stability analysis of MSE wall in terms of engineering and economic viewpoints.

View full abstract

A series of centrifuge model tests for a typical rockfill dam (central core zone type) was conducted in order to investigate behaviour of the dam due to cycles of reservoir water levels. The dam models consisted of comparatively well-compacted core zone and loosely compacted rock zone. A typical result among the model tests was also simulated by using a consolidation analysis method coupled with an elastoplastic model for unsaturated geo-materials. The main behaviour seen from the centrifuge tests was that large amounts of settlements due to wetting within the upstream rock zone were monitored; the crest of the dam at first moved toward upstream and then returned toward downstream in the first reservoir filling; and cracks along the dam axis were observed on the upper parts of the upstream and downstream slopes. A typical result of the dam model tests during reservoir filling was simulated. The simulation results showed that the settlements calculated agreed mostly with those measured; the settlements due to wetting could also be estimated within the upstream rock zone; the horizontal displacements toward upstream increased with the cycles of up-and-down water levels and they concentrated to a shallow region near the upstream surface; tension stresses were calculated at the places where the cracks were monitored in the model tests; and just after construction, arching actions in both the vertical and horizontal effective stress distributions could be seen on both the upstream and downstream boundaries between the core and rock zones but the arching action on the upstream boundary disappeared with water levels going up. The simulation method presented here provided an effective approach to analyze the behaviour of the rockfill dam during reservoir filling periods.

View full abstract

The continuous behaviour of a ground-embankment system from the stage of deformation up to failure was predicted with respect to an actual test embankment that had been constructed in stages on a soft ground made up of peat and calcareous soil. The behaviour prediction was carried out by employing soil-water coupled finite deformation analysis, which also included simulation of the embankment construction process. The information used in the analysis was limited to such things as the results of soil tests on soil materials (peat and calcareous soil) sampled from the ground, including their sensitivity ratios, and the embankment's construction history. The SYS Cam-clay model was used in the constitutive equations of the soils to determine the material constants of the soils and the initial conditions of the ground, and the computations were performed under plane strain conditions. As a result, the computed profiles of W-type ground settlement and of slip surfaces running through the embankment were found to be in good overall agreement with the actual profiles measured at the site. Furthermore, we found that this slippage is attributable to the undrained shear response of the soil elements in the calcareous soil layer, where slippage begins to occur during embankment loading. In other words, the slippage is caused by the rapid softening behaviour caused by the degradation of structure after the effective stress ratio reaches the vicinity of the critical state line.

View full abstract

A series of 1-g shaking table model tests were carried out to study the behavior of pile groups embedded in sloping ground subjected to lateral flow of liquefied soil. Two different configurations of pile groups: large (6×6 and 11×11) and small (3×3), were considered. The models were subjected to the liquefaction-induced large ground deformation to investigate the effect of several parameters on the response of pile groups and mechanism of lateral flow. These parameters comprise amplitude, frequency, and direction of input motion; density and slope of ground; and the thickness of non-liquefiable layer at the surface. The outcome of this parametric study reveals the importance of above mentioned factors which should be taken into account for analysis and design purposes. In addition, the results from the experiments clearly illustrate that in sloping ground configuration, both front (in upstream) and rear (in downstream) row piles receive greater lateral forces than middle row piles. This finding is attributed to the distribution of soil motion (displacement and velocity) of the liquefied soil in the model. As a result, installation of additional pile rows in front and behind an existing pile foundation can be considered as an effective retrofitting technique. Finally, soil-pile interaction was evaluated by running experiments with different pile spacings, and reliability of the JRA 2002 design manual in estimation of liquefaction-induced lateral force on piles is evaluated.

View full abstract

The paper describes observed behaviour of a model tunnel embedded in dry sand subjected to cyclic ground shear deformation in a centrifuge, as well as the behaviour of the model ground during shear deformation. Detailed data on earth pressures acting on the tunnel lining and the sectional forces of the lining are presented during ground shear deformation. The data suggest that the earth pressure at tunnel crown before ground shear deformation is smaller than the full overburden pressure probably due to the formation of arch action and the arch action may deteriorate with the cyclic ground shear deformation, resulting in an increase in the earth pressure at crown and changing the distribution of the sectional forces, which are largely influenced by conditions between tunnel lining and invert.

View full abstract

The paper describes and interprets watertightness problems that occurred in two Italian earth dams: the Polverina Dam, a zoned earth dam with vertical clay core, and the Zoccolo Dam, a homogeneous embankment with upstream facing of bituminous conglomerate. Measurements of seepage flows, pore water pressures and displacements in different stages of the dams' lifetime were collected and suitably interpreted to detect the effectiveness of measured physical quantities in revealing watertightness problems. The paper focuses on the role of pore water pressure measurements to such a scope and enhances how the effectiveness of such measurements may be strongly affected by the particular stage reached by the consolidation process within the dam. The analysis of the two case-studies proved that pore water pressure measurements may result effective in revealing hydraulic problems only when significant re-equilibrium processes have taken place in the dam zones where the measurement points are located. Measurements result, instead, ineffective during a nearly undrained stage. The paper also discusses the effectiveness of seepage flow measurements to detect possible hydraulic problems.

View full abstract

Catastrophic landslides are not uncommon in the Three Gorges area in China during rainy seasons. It is anticipated that the frequency of landslides will increase as a result of the impoundment of the reservoir of the Three Gorges Hydropower Station, and this has prompted geotechnical researchers to pay special attention to the problem. This paper introduces a landslide model test system, which allows studying of landslides induced by the combined effect of reservoir impoundment and rainfall. The system consists of a large flume with hydraulic lifting facilities, a set of computer-controlled surface sprinklers and pipes to simulate rainfall and reservoir impoundment, as well as a comprehensive instrumentation and data acquisition system. A model test carried out to study the Shiliushubao landslide, which occurred in the Three Georges Reservoir area, is presented to demonstrate the potential application of the test system. It is shown that the horizontal and the vertical displacements are abruptly increasing, when the reservoir impoundment and rainfall are combined, which cause the failure of the model slope.

View full abstract

When the water table rises in a granular soil mass, a large additional settlement of footing resting on such a soil mass is commonly expected. Laboratory model tests show that when the water table rises to the footing level, there can be an additional 400 to 500% of the settlements compared to when the soil is dry. To understand the mechanics of the additional settlements of footings resting on the granular soil masses caused by water table rise, an investigation into the change in the Young's modulus of soil was therefore made using oedometer tests. A relation between the saturated and the dry Young's moduli of the granular soils is presented for use in the elastic analysis of footing settlements. The findings of the investigation were then used to predict the additional settlements of a model footing when subjected to a water table rise, using elastic analysis. The theoretical predictions of the additional settlements are lower than the actual additional settlements observed in the settlement tests conducted in the laboratory using model tests. The model tests show that the additional settlements are larger in loose sands than in dense sands.

View full abstract

Shaking table tests on a saturated sand specimen in a laminar shear box were performed to measure the coupled shear strain-pore pressure response of sands subjected to horizontal shaking. Three methods of shear strain evaluation techniques, which use soil motions recorded by embedded sensors within the soil specimen and external sensors on frame, were utilized to evaluate the shear strain-time histories in one-dimensional shaking cases. Coupled shear strain-pore pressure responses throughout the liquefaction process are presented based on evaluated shear strains and measured pore pressures at corresponding depths. More insights of the interactions between the induced shear strain and generated excess pore pressure are also given, including the temporal variations of induced shear strain prior and after the initial liquefaction, development of liquefied zone, and duration of liquefaction state. These spatial and temporal variations of coupled responses reveal that the induced shear strain amplitude is not only an important property in triggering soil liquefaction but also a better index in defining the time of initial liquefaction. Identifications of these coupled features will be beneficial for verifying and interpreting testing data in physical modelling for liquefaction studies.

View full abstract