The seismic behaviour of asphaltic concrete core dam under earthquake loads has been studied. The study consists of three parts; element tests on asphalt concrete, numerical analysis and some model tests. In the first part, a series of triaxial monotonic and cyclic tests were performed at the Norwegian Geotechnical Institute (NGI)'s laboratory to determine the geotechnical parameters of asphalt concrete. Then, numerical analyses were carried out for Garmrood dam in Iran with a height of 110 m. The different stages of construction and impounding were analyzed using a hyperbolic stress-strain model. The dynamic analysis was then performed using two different models; nonlinear elasto-plastic material model and an equivalent linear method followed by the Newmark approach to estimate the permanent displacements. Finally, some 1-g model shaking table tests were performed on a typical rockfill dam to have a qualitative picture of the modes of seismic behaviour of an asphaltic concrete core rockfill dam. Element tests result shows that the asphaltic concrete is resistant to earthquake excitations and the earthquake has to be very strong to cause any detrimental cracking or material degradation of the properties of a ductile asphaltic concrete. In addition, the obtained results from numerical analysis and model tests show the appropriate response of the dam during and after earthquake caused shaking.

View full abstract

According to the type of works and to the consolidation of the ground, soil-structure interfaces are orientated in various ways with respect to the stress tensor in the soil. For instance, before shearing, the normal stress applied to the interface could be the major stress in the soil or the minor one or some intermediate value. This paper shows how the anisotropy of confining stresses can influence the behaviour of sand/steel interface. To this end, a series of tests was conducted with the cylinder shear apparatus. The cylinder shear apparatus is dedicated to the study of the shear behaviour of ground/structure interfaces. This device controls the radial confining stress and an average axial confining stress. One is particularly interested in estimating the influence of anisotropy of confining stresses and to look for empirical laws clarifying this influence on the peak shear, the post-peak shear, the displacement needed for full mobilization. The influences of the roughness of the interface, of the density of the sample and of the level of confining stresses are also studied.

View full abstract

Although time domain reflectometry (TDR) has been widely accepted as a non-destructive in-situ method with high reliability, the effective use of TDR for water content (w) and electrical conductivity (EC) measurements in clay slurry has not been established. We developed a Teflon-coated (TC) probe with a 0.05-mm thick fluorocarbon resin coating, and conducted a calibration experiment for kaolinite slurry to indicate its applicability to the measurements. The TC probe could estimate w from the observed effective dielectric constant (ε_eff) in the range of w≥47% with a ±6% error range. The ε_eff includes dielectric constant of two components: material under test (ε_m) and coating material (ε_tef). The relationship between ε_eff and w for this particular probe design cannot be directly applied to other coated probes because each probe has its own specific dielectric characteristic. We defined a function expressing the characteristic, that is ε_eff vs. ε_m, according to the testing in various fluid media. By applying the function, the observed data sets of ε_eff were transformed for a soil specific value (ε_m). Owing to the sufficient agreement of the ε_m values with those for the uncoated probe, we obtained the data sets of ε_m vs. w and its empirical expression for kaolinite slurry.

View full abstract

Water unsaturated hydraulic conductivity (k_w) functions of two remolded fine-grained materials were measured over a wide range of degrees of saturation (S_r) with two methods. The instantaneous profile method (IPM) was used for S_r>50%. An original vapor equilibration method (also known as the vapor equilibrium technique, VET) was used for S_r<50%. Both materials compacted at the standard Proctor optimum water content and maximum density, have saturated hydraulic conductivities (k_sat)<10^-7 m/s. The VET couples the total soil suction (s) control from desiccators with saturated salt solutions with water mass measurements from a digital laboratory balance. The k_w measurements of the two techniques are consistent and complementary. The effect of hysteresis on the k_w functions at higher s values was also investigated. The experimental results suggest that the hysteretic effect on the k_w-S_r and k_w-s relationships cannot be neglected, and that the measured k_w are significantly dependent on the initial S_r. The VET tests on the specimens that were initially dried give the lowest values of k_w and the tests on the specimens that were initially saturated give the highest values of k_w. The relative hydraulic conductivity values are very small (k_rw<2×10^-5) in this saturation range (S_r<50%) for the tested materials.

View full abstract

Sea-side ground improvement is one of the techniques used for increasing seismic stability of sheet pile quay wall. A series of centrifuge shaking table test is conducted to investigate the seismic behavior of sheet pile quay wall stabilized with sea-side cement deep mixing (CDM). At 50 g centrifugal acceleration, the clay ground is consolidated and four to five stages of shaking are applied to the quay wall. The recorded input-output accelerations, deflection of the quay wall, earth pressures along depth and pore pressures were used to evaluate the behavior of the stabilized sheet pile quay wall. The results indicated that an improved area provides significant resistance against seismic loading. The ground improvement can reduce 30% to 60% bending moment during the seismic loading. The horizontal deflection of the quay wall decreases rapidly with increase in area of the CDM until it reaches a certain limit. Prediction by the numerical model agrees fairly well with the results of centrifuge model tests.

View full abstract

This article addresses the numerical analysis of the load-settlement relation of footing on a soil ground by the finite difference method incorporating the subloading surface model. The subloading surface model is the elastoplastic constitutive model describing most pertinently the plastic strain rate due to the rate of stress inside the yield surface so as to fulfill the continuity and the smoothness conditions. For the finite difference program FLAC3D is used, which adopts the explicit dynamic relaxation method without solving the total stiffness matrix equation and thus has the capability for the remarkable reduction of calculation time. The relevance of the present approach to the description of the load-settlement curve exhibiting the ultimate load is demonstrated by the numerical experiments. Further, the applicability to the prediction of real behavior of footing-settlement is verified by comparison with the test data for sand grounds with a high friction property.

View full abstract

It has been pointed out that there are two possible mechanisms that enhance liquefaction resistances of unsaturated sand. The first mechanism is where air in a partially saturated sand mass plays a role of absorbing generated excess pore pressures by reducing its volume. Okamura and Soga (2005) derived the influential factors of the liquefaction resistance for partially saturated sand from theoretical consideration and effects of the factors were examined through a series of triaxial tests on a clean sand. They found a unique relationship between liquefaction resistance ratios and the potential volumetric strain, which allows the estimation of the liquefaction resistance for partially saturated sand. The second is the matric suction of unsaturated sand which increases the effective stress and thus the strength of the soil mass. In this study two series of cyclic triaxial tests on non-plastic silt were carried out to observe the liquefaction resistance in both mechanisms. In the first series, a top cap with an accumulator tank was used to study the effect of compressibility of pore fluid on the liquefaction resistance. The empirical relationship derived by Okamura and Soga is found to be valid even for the silt provided that the matric suction is negligible. In the second test series an ordinary cap was used. The liquefaction resistance increased linearly with the matric suction, with the increasing ratio being higher than that for the net stress. A unique linear relationship is found between the normalized liquefaction resistance and the matric suction. Results are summarized in the form which can be easily applied to evaluate the liquefaction resistance of a partially saturated soil.

View full abstract

As part of a long-term research program to evaluate the rate effects on the stress-strain behaviour of geomaterials, the viscous properties of a poorly-graded relatively angular quartz-rich sand, Toyoura sand, under air-dried conditions, were investigated by performing a comprehensive series of direct shear (DS) tests at a fixed normal stress equal to 50 kPa. The tests were performed on loose and dense specimens subjected to the following different loading histories: a) monotonic loading (ML) at constant shear displacement rate (s) differing by a factor up to 100,000; b) ML at constant s including otherwise a number of step changes in s by a factor of 100; and c) a number of sustained loading (SL) stages during otherwise ML at constant values of s differing by a factor up to 1,000. Tests a) revealed that, with dense specimens, the peak shear strength is remarkably independent of s while the residual shear strength noticeably decreases with an increase in s. That is, the viscous property is the so-called TESRA type at the peak stress state, while it is the so-called Positive & Negative (P&N) type at the residual state. With loose specimens, both peak and residual shear strengths decrease with an increase in s, indicating that the viscous property is already the P&N type at the peak stress state and definitely so at the residual state. These results are qualitatively and quantitatively consistent with those from tests b), by which the viscosity properties were quantified in terms of the rate-sensitivity coefficient. The results from tests c) showed that creep shear displacement, Δs, increases with a decrease in the tangent stiffness at the immediately preceding ML phase or with an increase in the shear stress level during the SL stage. The value of Δs for a given period steadily increases with an increase in s during the immediately preceding ML phase. These trends of viscous behaviour are simulated all very well by a non-linear three-component model incorporating a general expression of viscous stress.

View full abstract

The correction factor (η_ie) for the ultimate bearing capacity of a footing placed adjacent to a slope, taking into account the combined effect of horizontal and vertical seismic loads, represented by horizontal and vertical seismic coefficients, k_h and k_v, respectively, was derived using a modified Janbu's slice method. The influence of slope angles ‘α’ on the values of ‘η_ie’ are studied here. It was found that the values of η_ie can be expressed as an exponential function of ‘k_h/(1-k_v)’ and ‘α’, with a measurable interdependency between ‘α’ and ‘η_ie’. The influence of ‘α’ on the value of ‘η_ie’ increases as the input value of ‘k_h/(1-k_v)’ increases. Equations derived based on the analytical results are proposed to account for this effect. Based on the analyses of 11 near-fault seismographers obtained in the 1999 Chi-Chi earthquake in Taiwan, a ratio between the vertical and the horizontal seismic coefficients, λ, of between ±0.25 is suggested for including the combined effect of vertical and horizontal seismic forces in evaluating the seismic bearing capacity of footings located in near-fault areas.

View full abstract

In order to investigate the effects of different geological ages on liquefaction properties of sandy deposits, a series of undrained cyclic triaxial tests was performed on three kinds of in-situ frozen and their reconstituted samples which were retrieved from Holocene (Tone-river sand) and Pleistocene (Edo-river B and C sands) deposits. The specimens were subjected to isotropic consolidation at a specified confining stress which is equivalent to the in-situ overburden stress at the depth of sampling, and small strain shear moduli were measured before and during the undrained cyclic loading tests. The liquefaction properties and the small strain shear moduli were affected by not only the natural aging effect of the specimen but also the inter-locking effect that was enhanced by applying drained cyclic loading before the undrained cyclic loading tests. During liquefaction, different tendencies of degradation in the small strain shear moduli which would reflect the aging effects of the specimen were observed between Tone-river Holocene sand and Edo-river B and C Pleistocene sands. The applicability of reconstituted samples as substitutes for in-situ frozen samples was confirmed with Tone-river Holocene sand that has no cementation effect between soil particles, whereas it seems difficult to simulate fully the liquefaction behaviour of Edo-river B and C Pleistocene sands which have higher cementation effect.

View full abstract

An empirical approach is developed to analyze the nonlinear torsional behavior of free-standing pile groups with rigid pile caps. In this approach, the lateral and torsional responses of individual piles in a pile group are modeled by p-y and τ-θ curves; the interaction among lateral resistances of the individual piles is predicted through Mindlin's elastic solutions; the interactions between the torsional and lateral resistances of the individual piles are described through Randolph's solution; and the coupling effect of lateral resistance on torsional resistance of the individual piles is quantified using an empirical factor “β”. The proposed approach is capable of capturing the most significant aspects of pile-soil-pile interactions and coupling effect in pile groups subjected to torsion. The proposed approach is verified using results of centrifuge model tests. In general, the applied torque-twist angle response and the transfer of applied torque in pile groups can be reasonably well predicted and are sensitive to the pile group configuration.

View full abstract

This study attempts to develop a simplified procedure to evaluate earthquake-induced residual displacements of conventional retaining walls (e.g., gravity, leaning, and cantilever retaining walls). Based on the results from a series of shaking table model tests, the study found that subsoil and backfill deformations could be significant even in cases of dense model ground that cannot be evaluated by using Newmark's rigid sliding block analogy. In the proposed method, therefore, the effects of subsoil and backfill deformation were taken into account as well as the movement of the wall as a rigid mass, which was evaluated using Newmark's method. The deformation properties of the subsoil and backfill were modeled based on results from a series of laboratory soil tests and moment load tests. Simulations of shaking table model tests using the proposed method revealed that computed residual displacements agreed well with the measured ones even though the model tests were conducted under a wide variety of conditions.

View full abstract

Performance-based design has slowly been adopted for the use in geotechnical design. In the future, various types of foundations suitable for a variety of soils and applications should be accepted as an alternative to conventional flat foundations for the increase of bearing capacity and the savings of materials. The concept of shells is not new in foundation design, considering the construction with inverted brick arch foundation. Shell foundations are economic alternatives to conventional flat foundations where heavy superstructural or lateral loads are to be transmitted to weaker soils. Various types of foundations with different geometrical shapes have been extensively investigated in the structural design. However, the corresponding studies on the geotechnical design in terms of bearing capacity and deformation are scare. As such the advantages of various types of foundations have not yet been clarified in terms of the geotechnical design. The objective of this paper is to examine the overall geotechnical performance of various types of foundations on sand using model loading tests and the numerical limit analysis. The general superiority of various types of foundations has been revealed by comparing the loading tests with the analytical results.

View full abstract