The case of the seismic behaviour of Kalamata harbour quaywall during the large Kalamata's earthquake (Ms=6.2, 13.9.86) is used for a systematic critical review of the different design methods of gravity retaining walls. The overall wall behaviour is studied with a two dimensional finite element model. The classical procedures of a seismic design of gravity walls based on the "limited strength" criterion (Mononobe-Okabe method), and the more recent methods based upon the concept of "acceptable limited displacement" (Richards and Elms), have also been applied. All the results are compared with the existing measurements. The results of these comparisons and discussions are used for design considerations and for a general review of the accuracy and the limitations of each method.

View full abstract

This paper presents the results of direct shear tests on sand samples reinforced by inclusions. A large direct shear apparatus (1 m³) was used for the experiments in conjunction with Leighton Buzzard sand and different types of grid and sheet reinforcement. The results showed that the presence of the reinforcement layer inclined to the central plane of the box caused a significant reduction of shear strains developed along the central region of the sample and increased the overall strength of the sand. High normal stresses on the reinforcement plane prevented bond failure between soil and reinforcement, which limits the application of direct shear tests with inclined reinforcement as a way of measuring bond strength between soil and reinforcement. At large shear displacements the performance of polymer reinforcement was enhanced by the favourable deformed shape of the reinforcement. At peak shear load the deformed shape of the reinforcement had little effect on the test results. This suggests that the assumption of a deformed reinforcement layer at the interception with the failure surface may be unrealistic when analysing reinforced soil structures at peak conditions by limit equilibrium methods. As a consequence of the latter, it was also observed that, for the range of values tested, the influence of the reinforcement bending stiffness was negligible.

View full abstract

Lateral loading tests were carried out on model free-headed pile groups. The test piles were embedded in sand which was prepared by the boiling method. Data on group efficiency with pile spacing are presented for comparison with theoretical predictions based on the method developed by M.F.Randolph (1981). The predicted and measured values of group efficiency are generally in good agreement, with a maximum discrepancy of about 30%. Some parametric studies on the behaviour of laterally loaded prototype pile groups are also made using Randolph's equation.

View full abstract

Cyclic triaxial tests, cyclic torsional shear tests on isotropically consolidated specimens and cyclic torsional simple shear tests on one-dimensionally consolidated specimens, using saturated Toyoura sand, were performed under drained conditions in order to investigate the general tendencies of the stress-dilatancy characteristics under cyclic loadings. It was found that in each testing method a unique relationship between the stress ratio and the dilatancy factor (strain increment ratio) exists irrespective of void ratio and pressure level. In each testing method, the rate of dilatancy, as defined as the rate of volumetric strain increment to shear strain increment (positive for volume expansion), becomes negative by the reversal of loading direction. Then it increases continuously with shearing without showing any discontinuous behavior at the moment when the sign of shear stress changes (i.e., at the neutral stress point), and subsequently it becomes positive. It was also found that in the triaxial tests an over-consolidation stress history affects the stress-dilatancy relation only below a certain stress ratio in the first loading and has no effect in the subsequent shearing.

View full abstract

Based on the results of a series of cyclic triaxial tests and cyclic torsional simple shear tests on saturated Toyoura sand performed under drained conditions, stress-dilatancy equations under cyclic loadings were studied. It has been found from the experimental work that in each of the testing methods a unique relationship between the stress ratio and the rate of dilatancy (strain increment ratio) exists which is independent of void ratio and pressure level. It has also been found experimentally that the rate of dilatancy, as defined as the rate of volumetric strain increment to shear strain increment (positive for volume expansion), becomes negative by unloading after the reversing of loading direction, and it increases continuously with shearing without showing any discontinuous behavior at the moment when the sign of shear stress changes, and subsequently it becomes positive. Thus, at the same stress condition, two different rates of dilatancy are possible to exist depending on the direction of shear straining. Four different representative stress-dilatancy relations based on (i) the sliding block theory, (ii) the Rowe's theory, (iii) the Roscoe's energy dissipation theory, and (iv) the Taylor's energy dissipation theory, were modified to apply to cyclic loading conditions with reversed loading directions. It was found that after these modifications some of these theories can well simulate the stress-dilatancy relations under drained cyclic conditions obtained by the tests. Finally, curves of equal plastic potential in cyclic drained simple shear were portraied on a stress plane. It is a double plastic potential function in the sense that two different plastic potential curves are passing through any given stress point, which are for two different directions of shear-straining.

View full abstract

A boundary element analysis based on an elastic theory is performed to analyse the behaviour of piled raft foundations subjected to vertical load. Characteristics of settlement and load transfer for piled raft foundations whose raft rests on a homogeneous isotropic elastic half-space soil are contrasted with free-standing pile groups and single piles. The critical length of pile groups for a vertical loading (the length beyond which settlement shows little further decrease) is longer than that of single piles whose dimensions are the same as the piles in the groups. The reduction of the settlement caused by the presence of the raft is very small, although the raft transmits 20-40% of the applied load direct to the soil. A significant difference of pile load is recognized at the upper part of the shafts between two kinds of pile group with and without raft, and only small difference is found at the remaining part of the shafts. The ratio of long-term settlement to total final settlement of pile groups is greater than that of single piles. The contact earth pressure on the raft is relatively uniform in the inside area surrounded by piles.

View full abstract

High-quality undisturbed samples of clean sands were obtained by in situ freezing methods from a man-made fill, and from natural deposits at three sites in Niigata. Conventional tube samples were also obtained from the sites in Niigata. The SPT N-values corrected for a vertical effective stress of 1 kgf/cm² (98 kPa), N₁, of the in situ frozen samples ranged between 2 and 30. Undrained triaxial tests were conducted on cylindrical specimens prepared from the samples after they had been thawed, saturated and consolidated. The results of the field and laboratory tests indicate that : (1) the liquefaction resistance of the high-quality undisturbed samples shows a good correlation with N₁ in such a way that the liquefaction resistance increases abruptly as N₁ exceeds a certain critical value, which increases with an increase in the number of load cycles ; (2) the liquefaction resistance vs. N₁ relationship of the high-quality undisturbed samples is consistent with the previously published relationship based on field records of liquefaction during earthquakes ; and (3) the liquefaction resistance of the tube samples shows a good correlation with N₁ over a range between 18 and 30, although the liquefaction resistance is considerably lower than that of the in situ frozen samples.

View full abstract

A similitude is derived for the shaking table tests on saturated soil-structure-fluid model in 1g gravitational field. The main tool used for deriving the similitude is the basic equations which govern the equilibrium and the mass balance of soil skeleton, pore water, pile and sheet pile structures, and external waters such as sea. In addition to the basic equations, an assumption is made upon the constitutive law of soil ; i.e. the stress-strain relation is determined irrespective of the confining pressures if appropriate scaling factors are introduced for the stress and the strain for taking the effect of the confining pressures into account. Applicability of this assumption is examined by using the presently available data under the confining pressures ranging from 0.05 to 1kgf/cm² (from 5 to 98kN/m²) and 0.05 to 4kgf/cm² (from 5 to 392kN/m²). The results indicate that the assumption is applicable within the intermediate strain levels ; i.e. the strain levels which are lower than the strains at failure. In consequence, the similitude derived in the present study is applicable to the model tests in which the major concern is directed toward the deformation, rather than the ultimate state of stability, of the soil-structure-fluid system.

View full abstract

An isotropic hardening elastoplastic constitutive model for sand is developed by extending the model for clay proposed before. This model for sand (named tij-sand model) can precisely take into consideration the influence of intermediate principal stress on the deformation and strength characteristics by using the concept of the mechanical quantity tij, and the influence of stress path on the direction of plastic flow by dividing the plastic strain increment into two components, in the same way as the previous model for clay (tij-clay model). In the present paper, a new "plastic work based on tij" W^*p is defined, and it is experimentally shown that W^*p is proper as the quantity of state for sand. Then, the model for sand is formulated by employing W^*P as the strain hardening parameter, in place of the plastic volumetric strain ε_ν^p in the model for clay. The validities of the model are confirmed by various shear and consolidation tests under triaxial compression, triaxial extension and three different principal stresse conditions and stress probe tests under triaxial compression and extension conditions. All of the soil parameters of the model can be determined by a conventional triaxial compression test after loading-unloading-reloading isotropic consolidation.

View full abstract

Construction of major civil engineering works in the weatherable Piraeus marl, Greece, raised several questions focusing on its "rock-like" or its "soil-like" behaviour. The research in this paper presents a method of verifying the degradation mechanism of the marl due to assumed weathering processes. For this purpose, a large number of laboratory tests have been reevaluated using correlations between selected index properties of the material. It is shown that the Piraeus marl is controlled by the degree of solution of diagenetic calcite bonds. The remaining amount of calcite, probably infiltrating downwards, is the key factor over large variations in strength, deformability and dry density, and over minor variations in mineralogy and plasticity. On weathering, the marl has increased in voids and in water content, with a general reduction in the intensity of particle bonding, mainly in the upper 15m. In that zone, strength can be attributed mainly to the effects of overconsolidation. On a rock engineering basis, the Piraeus marl can be classified at best as a very low strength-low modulus ratio material.

View full abstract

A triaxial testing system developed for testing soils under automatically controlled stress paths and strain paths is described. This system consists of a micro-computer and a pneumatic pressurizing system, together with a mechanically simple system consisting of an ordinary axial displacement-controlled loading frame and an ordinary triaxial cell having a small-diameter piston. Further, since only the cell pressure is adjusted so as to follow the prescribed stress path or strain path at a controlled constant rate of axial deformation of specimen, an accurate peak value together with post-peak behavior can be obtained. Some typical test results of sand obtained by means of this system are presented, which are (1) monotonic drained traixial compression tests of sand at a constant mean principal stress or at a constant axial stress, (2) a cyclic drained triaxial test of sand at a constant mean principal stress involving many intermediate small hysteresis loops of unloading and reloading, and (3) anisotropic compression tests of sand, and (4) cyclic constant-volume tests of sand with taking into account the membrane penetration effects. These results show that this system is a useful and versatile tool for accurately measuring strength and deformation properties of soils loaded under a wide variety of stress paths and strain paths.

View full abstract

The usual method of measuring the permeability of a homogeneous soil in the laboratory is a constant head or a falling head test. Of the two methods, the falling head test is more suitable for measuring the permeability of homogeneous clay as very small flow quantities are involved. The proposed method is to keep the conventional falling head test unchanged except to measure the falling head automatically through a very accurate differential pressure transducer. The low pressure end of the transducer is opened to the atmosphere to correct for changes in the atmospheric pressure over the 24 hour duration of a typical permeability test. By means of several repeated falling head permeability tests at different initial heads for the same effective stresses, it is demonstrated that reasonably reliable and repeatable measurements of permeabilities of clay can be made, provided that the applied head difference as a proprtion of the current vertical stress is small, so that no significant consolidation or swelling is induced by the seepage head.

View full abstract

Coefficient of dilatancy of clays is shown to be linearly related to their plasticity index. Undrained test results on twelve soils comprising natural reconstituted clays and clay-sand mixtures form the basis of the proposed correlation. Different correlation is obtained with variation in either consolidation history and stress path during undrained shear.

View full abstract

Laboratory tests were carried out to simulate pressuremeter testing so that the generation and dissipation of pore water pressure during pressuremeter tests could be studied. A modified triaxial cell was used to accomodate 150mm hollow cylinder specimens prepared with known stress history. Pore pressures on the cavity wall as well as in the soil mass were monitored during the expansion. The results show that a large magnitude of excess pore pressure gradient is set up around the 'model' pressuremeter where the maximum value is at the cavity wall. This is then followed by radial dissipation of pore pressure, the rate of which depends on the pore pressure gradient.

View full abstract