It is a generally accepted fact that tensile failure takes precedence over shear failure when a material is transformed from a ductile state into a brittle one. In order to evaluate the growth of extensile cracks, two series of compression tests with samples from volcanic cohesive soil obtained from a low fill dam were performed. Results show that extensile cracks may generate prior to shear failure when the friction between the specimen and the pedestal was reduced. Under such circumstances, compressive strength does not decrease until the shearing cracks are generated. This is the same in the case when a confining pressure is acting. In a test using a composite specimen made of two samples of different water contents, extensile cracks occurred at the portion with a low water content. The compressive stress at the time when the extensile crack occurred was lower than that in the case of the low water content specimen. Moreover, finite element analysis was performed to calculate the stresses in the specimens at the generation of extensile crack. Results shows that the horizontal stresses that exerted on the crack surfaces were not tensile but compressive in the case when a confining pressure was acting.

View full abstract

A significant proportion of the surface deposits of the central and upper half of the North Island of New Zealand consists of volcanic ash soils. Given that New Zealand is situated within an active seismic region, the dynamic properties of these soils are an important geotechnical earthquake engineering consideration. A systematic investigation of three volcanic ash sites was performed to help characterise the dynamic properties of these soils. In situ down-hole and cross-hole seismic tests produced small strain shear modulus (G_max) values ranging between 17 and 300 MPa, with excellent agreement between the two test methods. Bender element and free vibration torsion testing of specimens trimmed from high quality samples produced G_max values which were 28% to 75% lower than those measured in the field. This disparity was considered to result from negative pore pressures potentially present in the field and irreversible changes in the structure of the soil that occurred during removal of the soil from its in situ stress environment. Normalised reduction curves (G/G_max versus shear strain) produced results which were comparable to those presented by other researchers for cohesive soils, though the influence of the plasticity index on the form of the reduction curves was observed to be negligible. Investigation of the G_max/S_u relationship from field investigations of volcanic soils produced average values that ranged between 560 and 890.

View full abstract

A series of tests was conducted by using small-scale and large-scale true triaxial apparatuses (where σ₁≠σ₂≠σ₃) in order to study small strain behaviours of dense granular soils including anisotropic deformation characteristics. These apparatuses are equipped with local deformation transducers to measure deformations of prismatic specimens in three orthogonal directions. Dense gravel specimens were prepared by manual compaction and tested in large-scale true triaxial apparatus, while dense sand specimens were prepared by air-pluviation and tested in both scales of true triaxial apparatuses. During isotropic and triaxial compression, at several stress levels, many very small vertical and horizontal cycles were applied to evaluate the quasi-elastic deformation property including Young's modulus and Poisson's ratio at small strain levels around 0.001%. Its inherent and stress-state induced anisotropy could be clearly observed for both cases of dense sand and dense gravel. The inherent anisotropy is affected by the bedding plane or heavy compaction during specimen preparation. The values of Young's moduli in the direction that was perpendicular to the bedding plane are smaller than or equal to those of Young's moduli in other orthogonal directions during isotropic compression for all the tested specimens. Discussions on anisotropic characteristics of dense granular soils including their hypo-elastic modeling are also made.

View full abstract

The pore pressure response of elastic stiff materials during undrained loading is a function of the volumetric stiffness of the skeleton and stiffness anisotropy. This paper presents experimental results examining the influence of these factors on the pore pressure response in conventional triaxial tests. In order to explore fully the undrained behaviour, the total stress path (TSP) was rotated to cover 180° in the triaxial plane. It was found that a unique linear relationship exists between the pore pressure coefficient (Δu/Δq) and the direction of the TSP (Δp/Δq) with a slope equivalent to the Skempton pore pressure parameter (B), which may be below unity for stiff materials. The intercept of this linear relationship reflects the effect of the material's anisotropy. It has been demonstrated that a pseudo anisotropy of repeatable and consistent nature was induced by membrane compliance in granular cemented soils. The results have shown that, unless eliminated or taken into account, the membrane compliance may lead to erroneous interpretation of pore pressure response. The paper discusses the implications of this finding.

View full abstract

The purpose of the present study is to estimate horizontal subgrade reaction coefficient (k_h) of piles from SPT N-value that is required to design piles against horizontal loading. This information is much demanded recently due to the introduction of the limit state design of pile foundations where quantified uncertainty in the prediction of pile behavior is required. For this purpose 52 horizontal pile loading test results were collected from literatures and reports, which were compiled in a database. The actual number of test results put to the statistical analyses was 38 cases with pile top displacement and 21 cases with bending moment measurements along a pile. The statistical analyses were carried out in the two folds; First, the inverse analysis procedure was applied to each loading test result to obtain the coefficients in different soil layers. Then the obtained coefficient values were related to observed SPT N-values of the layers by the regression analysis. The pronounced feature of the regression analysis employed in this study is that it takes into account the magnitude of estimation uncertainty as well as the correlation structure of estimated k_h for every pile automatically. The mean value of k_h obtained from the recommended equation in this study gives very similar results to those of the major design equations used in Japan. The uncertainties associated with the recommended equation are presented, which are intended to be used in a reliability analysis of horizontally loaded piles.

View full abstract

A series of strain controlled undrained true triaxial tests on cubical specimens have been performed using a fully automated flexible boundary experimental setup, with real-time feedback control system to evaluate the three-dimensional mechanical behavior of kaolin clay. This paper concentrates on the comparative response of the soil elements under monotonic shear loading observed at various overconsolidation levels. A possible shape of the initial yield surface is explored based on the limit of purely elastic deformations. The discussion includes the influence of the relative magnitudes of principal stresses on the clay behavior observed at different overconsolidation ratio values. Failure conditions are shown to be governed by the onset of localizations developed within the specimens, and because of which the specimens show sudden failure response before reaching a perfectly plastic deformation conditions. A reference surface constraining the ultimate plastic yielding of clay is found to be different from the observed failure surface. Based on the experimental observations, a 3-D failure criterion is developed that grows in size as a function of pre-consolidation stress.

View full abstract

It has been reported in many research papers that considerable settlements are observed in the Osaka Pleistocene clay layers, even though their stress levels are less than their yield consolidation pressure (p_c). To study the consolidation behavior around the p_c value, Long Term Consolidation (LTC) and Inter-Connected Oedometer (ICO) tests were carried out. This paper presents the consolidation behavior measured by the LTC, while the companion paper describes the results from ICO test. The LTC test is conducted using the conventional oedometer, whose diameter and initial height are 60 mm and 20 mm, respectively. The maximum duration for the consolidation was half a year. The tested clays are mainly Osaka Pleistocene clays. Test results from these clays are compared with those from Holocene clays retrieved from various sites across the world. Relationship between the strain and elapsed time measured by this study has been analyzed in terms of the strain rate theory. Based on test results from the Osaka Pleistocene clays, the family of equi-strain rate lines is proposed, consisting of three regions. However, in case of other clays, whose OCR is probably created by stress changes, i.e., mechanically overconsolidated clays, the consolidation behavior is different from that of the Osaka Pleistocene clays : i.e., a certain pressure exists where no more strain is yielded even after the strain rate becomes small.

View full abstract

It has been known from the conventional oedometer test that considerable settlement is observed after the end of primary consolidation, i.e., under the constant effective stresses. One of the most controversial issues in the geotechnical engineering is whether such secondary compression is taking place during and/or after the dissipation of the excess pore water pressure. Using the Pleistocene clays recovered from the Osaka basin, Long Term Consolidation (LTC) and Inter-Connected Oedometer (ICO) tests were carried out. This paper presents the consolidation behavior measured by the ICO and the companion paper will deal with the findings obtained by the LTC. The main advantage of the ICO test is its ability to investigate the influences of drainage path length under conditions of relatively small friction between the specimen and the oedometer ring. It was found that dissipation of the excess pore water pressure follows the H² rule, regardless of the relative stress range, i.e., at over- or normally consolidated state, where the H is the thickness of the specimen. However, the generated strain at overconsolidated state does not follow the H² rule. These test results were tried to be explained by the strain rate theory.

View full abstract

Related to the construction of a new type bridge abutment having the backfill of cement-mixed gravel, effects of moulding water content, w_i, compaction energy level, cement-to-gravel ratio by weight, c/g, and curing period on the stress-strain behaviour of compacted cement-mixed gravel and their correlations were investigated. A series of drained triaxial compression tests were performed on a well-graded gravel of crushed sandstone from a quarry mixed with ordinary Portland cement. Possible effects of inherent anisotropy produced by compaction in different directions were evaluated to a limited extent. The following results were obtained. With a fixed c/g, the maximum compressive strength is attained when compacted at w_i=the optimum water content, w_opt, for a given compaction energy level, while the compressive strength is not an unique function of compacted dry density of solid, ρ_d (or compacted void ratio, e). These trends of behaviour are analysed in terms of positive effects of higher ρ_d (or lower e) of cement-mixed gravel and an associated increase in the cement amount per total volume for a fixed c/g as well as effects of the amount and strength of cement paste, both being a function of moulding water content. The benefits of compaction at w_i=w_opt for a fixed compaction energy level in achieving higher compressive strengths increase with an increase in c/g and the curing period. The pre-peak stress-strain behaviour tends to become more linear with curing period. The effects of in-herent anisotropy are not as significant as those with air-pluviated unbound sand.

View full abstract

There are some statistical models on the strength of brittle materials such as rock and concrete. Weakest link model or bundle model is a representative one. Scale effect on the strength of cement-treated soils is studied theoretically using a new statistical model in which the weakest link model and the bundle model are combined. The proposed combined model is represented by a function with variables of diameter, height and coefficient of variance of the specimens. This model expresses not only the difference in volume of the specimens but also the differences in diameter and height, individually. In order to confirm the validity of this model, the unconfined compression and undrained triaxial tests are performed on the cement-treated soils in different sizes. The combined model is compared with the laboratory test results of the treated soils and the previous results of model mixing test with different mixing level. The validity of this model is also confirmed by comparing with test results on large size specimens sampled from the improved soil columns by Deep Mixing Method. The average strength of cement-treated soils in different sizes is predicted from the average strength and coefficient of variance of specimen in standard size.

View full abstract

Soil suction is an essential parameter for describing the mechanical behavior of unsaturated soils. For analytical purposes, its value is usually inferred from the soil-water characteristic curve. However, the expressions developed up till now to describe this curve, do not properly account for the physical facts involved in the hysteresis phenomenon. It is generally accepted that hysteresis arises because pores of different sizes are interconnected generating the “inkbottle” effect. In this paper, a twofold model that simulates the porous structure of soils is used and the mathematical expressions describing the hysteresis of the soil-water characteristic curves are obtained. This model considers that the voids in the soil mass are composed of two kinds of elements : the sites (cavities) and the bonds (throats), each kind showing its proper size distribution. Based on these premises and making use of the probability theory, the equations corresponding to the main hysteresis loop during wetting or drying are established and it is shown that the extent of the hysteretic behavior depends on the parameters describing the size distributions. In addition, the equations describing the specific internal (scanning) paths for cyclic wetting and drying are obtained. Finally, these results are compared with experimental results that include the main hysteresis loop, some scanning curves and the pore size distribution.

View full abstract

A phenomenological constitutive model, 'BWGG', is developed for the non-linear one-dimensional ground response analysis of layered sites. The model reproduces the nonlinear hysteretic behaviour of a variety of soils, and possesses considerable flexibility to represent complex patterns of cyclic behaviour such as stiffness decay and loss of strength due to build-up of pore-water pressure, cyclic mobility, and load induced anisotropy. It also has the ability of simultaneously generating realistic modulus and damping versus strain curves, by a simple calibration of only three of its parameters. The model is implemented through an explicit finite-difference algorithm into a computer code which perform integration of the wave equations to obtain the nonlinear response of the soil. The code, 'NL-DYAS', is then applied to study the seismic response of a soft marine normally-consolidated clay. The results are compared with those of widely used codes. Finally, the records of the Port Island array during the Kobe 1995 earthquake, are utilized, and the model is shown to “predict” the observed response with sufficient accuracy.

View full abstract

The superposition errors in bearing capacity factors, which are inherently involved in Terzaghi's bearing capacity formula, are examined based on the numerical analyses assuming the method of characteristics developed by Sokolovski (1960). The numerically calculated values of the bearing capacity factors obtained from the present study are compared with the classical formula by Terzaghi (1943) and the revised formula by Hansen (1970), and it is found that the superposition errors lie within tolerable limits for most practical applications.

View full abstract