A series of drained triaxial compression tests were performed on high-quality undisturbed samples of sandy soils for the determination of the angle of internal friction (φ_d). The high-quality undisturbed samples were recovered by the in-situ freezing sampling method (FS sample). The angle of internal friction for the high-quality undisturbed samples (φ_d(FS)) was compared with calculations from the empirical equations proposed by many investigators previously using the N-value of the standard penetration test (SPT). It was found that the angle of internal friction for the high quality undisturbed sand samples was much higher than that estimated from the proposed empirical equations. In order to take into account the effects of the confining stress on the SPT N-value, the measured N-value was normalized at an effective overburden pressure of 98 kPa (1 kgf/cm²) using the equation proposed by Liao and Whitman (1986). The normalized N-value, N₁, was induced to relate the angle of internal friction for the high-quality undisturbed sand samples. A fairly good correlation between the N₁ value and the φ_d(FS) was established. Finally, based on the test results, a simple equation (φ_d(FS)=(20N)^0.5+20) was proposed to relate the N₁ value and φ_d(FS) of sandy soils in the range of N₁ between 3.5 and 30, for this study. The relationship between the angle of internal friction for the high-quality undisturbed gravel samples (φ_d(GV)) and the penetration resistance was also discussed. A new parameter of N_L1 was introduced in order to relate to φ_d(GV), where N_L1 is the penetration resistance of the large scale penetration test (LPT) normalized at an effective overburden pressure of 98 kPa (1 kgf/cm²), using a similar equation to that proposed by Liao and Whitman (1986). Although data is limited, a correlation between N_L1 and φ_d(GV) similar to that proposed in this study for sandy soils was found.

View full abstract

A new analytical method is proposed for determining the inextensible grid reinforcement pullout resistance and pullout force/pullout displacement curve by using basic backfill soil and grid reinforcement properties. The pullout skin friction resistance/pullout displacement relationship is simulated by linear elastic-perfectly plastic model. A hyperbolic model has been proposed to represent the pullout bearing resistance/pullout displacement relationship in which the maximum bearing resistance of a single bearing member is determined using a new bearing capacity equation proposed in this paper. The influences of the grid bearing member spacing ratio, S/D, the bearing member deflection rigidity, and the pullout softening behavior on the mobilization of pullout bearing resistance are explicitly included in the proposed model. Good agreement has been obtained between calculated values and laboratory test results.

View full abstract

Repetitive liquefactions in angular and subangular gravels from an alluvial fan, in the case of two earthquakes with accelerations lower than that of a former which had caused liquefaction some months before, are described. The liquefaction took place from the surface, down to a depth of approximately 10 m. Water laden with sediments spouted out for some tens of minutes after the strong motion ceased. Knowledge of the grain size distribution of the sediments inside the body of the alluvial fan roughly defines the areas of the fan prone to future liquefaction. Non-repetitive liquefaction in overconsolidated fine fluvio-lacustrine sand of glacial origin at the unusual depth of 14 m to 16 m is also described. From the observation of sand lifted up for 10 m inside a large diameter well, and of that which spouted out from several microvents, it was estimated that extra-pore pressure was close to lithostatic conditions. Nevertheless, modeling of intrinsic pore pressure buildup on the DESRA2 programme gives lower values than the experimental one. Geotechnical characteristics of the 12 m-thick sandy stratum indicate that the upper part is unusually soft, either due to water circulation or past repetitive liquefactions which prevented reconsolidation. It is suggested that excess pore pressure generated by the earthquake was also due to redistribution of pressure from adjacent sediments.

View full abstract

In-situ geotechnical investigations were performed in normally consolidated gravels, which had repeatedly experienced liquefaction, and in overconsolidated sands (in part naturally remolded) which liquefied extensively at the unusual depth of approximately 14 m to 16 m from the ground surface. A rather good correlation exists between the measured SPT-N values and the effective overburden pressure for the gravelly site; on the contrary, a highly scattered correlation was found at the overconsolidated site. The stratum of overconsolidated sands present at this latter site between 13.40 m and 25.45 m depth has shear wave velocity, V_s, values measured in crosshole tests (CH) slightly increasing with depth, but completely uncorrelated with SPT-N values, whatever corrective factor is adopted. In addition, the current formulae used for predicting V_s from SPT-N overestimate V_s measured in overconsolidated sands in CH tests, but strongly underestimate it in overconsolidated gravels. This confirms that V_s is sensitive overall to the type of material and not to the stress-strain history. The current cyclic stress ratio approach, which uses the results of SPT tests, would have been able to predict liquefaction at both these unusual sites; attempts to adopt the V_s measurements to evaluate liquefaction potential were not successful. A third site was also investigated.

View full abstract

An empirical equation in use for estimating the pseudo-elastic shear modulus, G_f, of subsoil, associated with shear strains less than 0.001% is proposed in this paper. In a series of in-situ seismic cone tests performed nationwide, the profiles of both G_f and the in-situ void ratio, e₀, with depth were successfully characterised at five sites, each comprising a soft clay layer deposited in the Holocene era. The database which comprised the original data from the field and laboratory tests, coupled with similar information on well-documented Holocene clay deposits in Europe, was statistically analyzed in attempts to determine a generalised relationship with which G_f of soft clay may be reasonably estimated only from routinely available borehole data; that is e₀ and the current geostatic effective overburden pressure, σ'_v. An empirical relationship, G_f=5, 000 e₀^-1.5√(σ'_v) (kPa), was derived from the statistical analysis applied to data from seven different clays worldwide, for which e₀ extended over a range between 1 and 5, and the overconsolidation ratio ranged roughly between 1 and 2. The applicability of the proposed relationship was evaluated for two case records, each in which the clay exhibited unusual behavior; i.e., the undrained shear strength remained more or less constant with depth due to the existence of artesian pressure at one site, and, at the other, G_f decreased, whereas e₀ increased, with depth. It was demonstrated that even in these clay deposits exhibiting exceptional profiles, the proposed relationship was capable of predicting G_f with a reasonable accuracy by determining the profiles of e₀ and σ'_v with depth. In addition, the prediction when compared to G_max from carefully performed laboratory cyclic tests, yielded a better estimate of G_f from the in-situ seismic survey. Despite the fact that the empirical relationship was initially designated to estimate G_f of soft clays, it may be equally applicable to sandy deposits. This was verified by comparing it to similar, and well-established, relationships developed for sands. A case record as such is also described for a loose sand deposit at Higashi-Ohgishima in Tokyo Bay which was placed in 1960's by land reclamation.

View full abstract

The response of anchor piles in clay subjected to cyclic loading was investigated in a series of tests on small size model piles, 40 cm long and 1.9 cm in diameter, jacked into a reconstituted and normally consolidated clay. Evaluation of the test results focusses upon the degradation of cyclic pile stiffness, the accumulation of permanent displacements, the degradation of cyclic pullout capacity, as well as the short time effects of cyclic loading on the post-cyclic static pile response. The cyclic response of the model piles was correlated to that of the surrounding soil elements, observed in a series of undrained cyclic triaxial tests. In addition, the influence of the small scale of the model piles on the experimental results was assessed through comparison to similar data from large scale model and field tests reported in the literature. Cyclic load and displacement magnitudes, number of cycles and cyclic preshearing were identified as the main factors which control the pile response. The small size of the pile influences the experimental results at low cyclic loading intensities when radial drainage leads to partially drained conditions during cyclic loading.

View full abstract

For a more accurate prediction of the failure of a granular material, it is essential to clarify the constitutive relations after its peak strength on which the initiation and development of shear bands play an essential role. Since the shear bands consist of granule column if viewed microscopically, micromechanical analysis is suitable to understand the nature of the shear bands. This paper proposes a new model of microstructure of the granular material; a rigid granular particle and surrounding empty void whose static and kinematics are rigorously formulated. A granule column is then regarded as an assembly of randomly piled grains and voids, and the average behavior is computed by conducting a Monte-Calro simulation of forming such assemblies. The strain localization due to the granule columns is well reproduced in this simulation. In particular, it is shown that while most assemblies grow as the load increases, fewer assemblies keep growing once the load attains a critical value. Simulation is done for various micromechanical parameters such as grain geometry and friction coefficients, and their effects on the assembly behavior are discussed.

View full abstract

To elucidate the engineering significance of natural water content w_n of mudstone disturbed into clay from the view point of Critical State Soil Mechanics, CD triaxial test (with top and bottom drainage only) and CD direct shear test were carried out and the unique relationship between log q_max (or log τ_max) and w and that between w and log p' (or log σ'_v) were obtained. Comparing the experimental results thus obtained, p' was found to be nearly equal to σ'_v in fully softened shear zone. Thus, if w_n at the depth of h is known, the effective confining stress p' at that depth can be known. Therefore the pore pressure acting at that depth can be calculated as u=γ_t·h-p'. Two example case histories, Mizunashi landslide and Nabetachiyama tunnel in Neogene mudstone of Shiiya formation in Niigata Prefecture are presented. In the former, by measuring w_n of boring core of 90 m, several number of fully softened shear zones were found to exist where artesian pressure was acting and, in the latter, in the section where tunneling was very difficult due to squeezing pressure, w_n of fault clay was found to correspond either to p' induced by the overburden weight or to p' decreased by the amount of methane gas pressure.

View full abstract

Infiltration and soak tests were conducted at different pH levels of artificial acid rain to experimentally simulate the erosion process on stabilized soils, which were used for subgrade and base course materials. The authors used three types of lime and cement stabilized soils, and focused on the physico-chemical and engineering behaviors of the eroded soils. Results obtained show that the pH values of the lime and cement stabilized soils generally decreased with continuous infiltration or soak and that these decreases were greater under a higher acid condition. A large reduction of the unconfined compressive strength obtained from the specimens eroded by the higher acid rain was observed in comparison to erosion from the weaker acid rain. In particular, the unconfined compressive strength of the lime stabilized decomposed granite soil (LDGS) eroded by the acid rain with pH 2.0 decreased to 1/6 of its initial value after a small flux of infiltration. The X-ray diffraction patterns show that the straetlingite (2CaO·Al₂O₃·SiO₂·8H₂O) was decomposed, but the gypsum (CaSO₄·2H₂O), calcite (CaCO₃) and unknown chemical crystals were formed in the lime stabilized soils eroded by acid rains with pH 2.0 and 3.0.

View full abstract

Swelling pressure often experienced in deep tunnels driven through soft mudstone or its fault clay has long been explained to be caused by swelling of clay minerals (such as Montmorillonite) due to their sucking up of water. This is the reason why the swelling test of clay is specified by "Standard design specification for tunnels" published by JSCE (Japan Society of Civil Engineers). Both Case study and experimental and analytical research based on Critical State Soil Mechanics' concept have been undertaken to investigate if this explanation is really correct. Incorporating shear strength parameters in terms of total and effective stress, for short-term and long-term problems respectively, into modified Kastner's classical approach, clear evidence has been given that the predominant cause of large earth pressure and convergence encountered in the so-called swelling tunnels is the plastic flow of the broken material and not the swelling of it, indicating that the competence factor and strength parameters are the decisive factors, the result obtained by swelling test of clay being of no use for designing purpose. Based on the result thus obtained, a smiple design procedure of a circular tunnel driven through mudstone or its fault clay is tentatively proposed.

View full abstract

The limit equilibrium analysis applied to a plane failure surface is utilized for the evaluation of the active earth pressure on reinforced-earth walls under different loading conditions, such as seismic loading, pore water pressures into the fill, vertical and horizontal loads acting on the top at some distance. Due to the complexity of boundary and loading conditions considered, a limit equilibrium method with a plane failure surface has been used because this allows to deduce closed form solutions that could be quite helpful in design reinforced-earth vertical walls. The analysis presents two different loading conditions on a reinforced earthfill and two solutions in terms of earth pressure coefficient are given. Earth pressure coefficients are a function of non-dimensional parameters such the ratio λ between the distance d of the applied surcharge and the height of the wall H, or the ratio m between the horizontal and the vertical surcharge q_h and q_v respectively. Pore water pressure effects on earth pressure have been taken into account by means of the pore pressure ratio r_u=u/γH. Seismic loading have been taken into account in a pseudo-static way, in terms of horizontal and vertical seismic coefficients. In a conventional design procedure, the analysis allows to define spacing or number of reinforcement as well as their length according to the failure wedge predicted.

View full abstract

A simple and useful technique whereby the membrane penetration which causes errors in the effective confining stress during an undrained triaxial test on a coarse-grained soil can be substantially reduced has been developed. In this method, a thin membrane penetration reducing layer (MP-reducing layer) is made on the surface area of a triaxial specimen with Toyoura standard sand, whose accurate membrane penetration characteristics have been determined by many researchers and have been established in laboratory undrained tests. The results of a series of cyclic undrained tests indicated that the proposed method can provide test data only with a known membrane penetration amount of Toyoura standard sand, when forming an MP-reducing layer with a thickness approximately the mean grain size of test material. This method has advantages over existing methods because it only requires placement of Toyoura standard sand around the test specimen and does not require special devices nor is it based on questionable assumptions.

View full abstract

The effects of membrane penetration on the cyclic deformation properties for undrained cyclic triaxial tests of granular soils are discussed in this paper. The soils used are six types of sands with different grain size distributions. Test results indicated that the relationship between the shear modulus at small strain obtained from a cyclic torsional test and an undrained cyclic triaxial test, and between the Young's modulus obtained from the drained and undrained cyclic triaxial tests are dependent on the grain size distribution of the sands. It is assumed that the main reason for this is caused by a decrease of the Poisson's ratio under undrained cyclic triaxial condition due to the membrane penetration effect. The Poisson's ratio in undrained cyclic triaxial tests has been estimated from the membrane penetration value and the pore pressure fluctuation under undrained cyclic loading. Consequently, it was determined that the Poisson's ratio of the soil skeleton in undrained cyclic triaxial tests decreases from 0.5 at small strain, as the grain size increases. In addition, the shear modulus calculated from the estimated Poisson's ratio and the measured Young's modulus in undrained cyclic triaxial tests is almost equal to those in cyclic torsional tests irrespective of grain size.

View full abstract