Journal of Geotechnical and Geoenvironmental Engineering

An extensive suite of three-dimensional finite-element simulations was carried out to examine the response of pile-raft-soil systems under seismic excitation representative of far-field events. The numerical procedure was first validated by a series of seismic centrifuge tests performed mainly on $4\times 3$ pile-raft systems embedded in soft clays and was subsequently extended to study the influence of various factors on the computed raft acceleration and maximum bending moment near the pile head. The factors considered were pile length, pile flexural stiffness, peak base acceleration, superstructural mass, soil stiffness, soil frictional angle, pile density, and soil-layer thickness. With the help of dimensional analysis, the results of the parametric studies were processed and interpreted using regression analysis to derive correlations for predicting the maximum pile-group bending moment and peak raft acceleration. Such correlations are highly useful in the practical seismic design of pile-raft systems embedded in soft soils, because they provide quick first-order predictions for making initial engineering judgments prior to performing more-detailed and rigorous numerical simulations.

In centrifuge modeling, it is important to evaluate the in-flight strength and stiffness parameters of the soil model to understand soil behavior. Both the shear-wave velocity ($V_s$) and cone tip resistance ($q_c$) are commonly used to determine the stiffness and strength characteristics of soils. If direct $V_s$ measurement is not available, the $q_c$ profile measured in the centrifuge can be used as an alternative method of estimation, although direct $V_s$ measurement is preferred. In this paper, a series of carefully controlled cone penetration tests and $V_s$ measurement tests are simultaneously conducted in uniform silica sand during centrifuge operation. The tests are undertaken in loose to dense sandy samples under both dry and saturated conditions. Laboratory resonant column tests are performed to validate the small-strain shear modulus ($G_0$) obtained from centrifuge tests. Measured $q_c$ profiles at shallow penetration depth are corrected to consider the difference of cone diameters between a field standard cone and centrifuge miniature cone. Finally, new empirical correlations between the $V_s$ and $q_c$ for normally consolidated, uncemented silica sand are proposed and compared with previously reported in situ relationships.

Steel-reinforced high-density polyethylene (SRHDPE) pipe uses steel ribs to carry loads and plastic cover to prevent the steel ribs from corrosion so that it overcomes the disadvantages of steel and high-density polyethylene (HDPE) pipes. The SRHDPE pipe can be classified as a flexible pipe. Compaction is essential for flexible pipes to ensure their load carrying capacity. Numerical modeling is a useful tool to investigate the performance of the SRHDPE pipe during installation. However, two problems must be solved to properly simulate the installation of the SRHDPE pipe using the numerical method: (1) compaction pressure cannot be removed after the completion of the compaction; and (2) the SRHDPE pipe is a composite material (i.e., steel and plastic). In this study, the hardening soil model included in the finite-element software PLAXIS 2D was used to model the backfill material, which allowed the removal of the compaction pressure. A pressure of 80 kPa was applied on the surface of a 1.5 m wide and 1.4 m deep trench backfilled with aggregate to simulate compaction. The aggregate was modeled using the hardening soil model to estimate the residual pressure after the removal of the compaction pressure. An equivalent modulus method (EMM) was proposed to simulate the SRHDPE pipe in the numerical model. Parallel plate test results of the SRHDPE pipe were used to compare with the numerical and theoretical results to demonstrate the validity of the EMM. Finally, the test results of the SRHDPE pipe installation in the field were used to verify the effectiveness of the numerical method proposed in this study.

Geotechnical engineers face serious problems when building on collapsible soils. These types of soil show considerable strength when dry and experience excessive and sudden settlement when inundated. Due to the expansion of modern cities in arid areas, highways and roads may pass through pockets of collapsible soil. If unrecognized, it may cause pumps or deep sinkholes, which leads to traffic accidents and loss of lives. Furthermore, due to the increasing use of water for irrigation and domestic purposes, roads built on collapsible soils may experience uncalculated settlement. This paper presents the results of an experimental investigation of a rigid strip footing on collapsible soils subjected to inundation due to raising the groundwater table. A prototype experimental setup was designed to test these footings on homogeneous collapsible soils, partially replaced collapsible soils with cohesionless material with and without geotextile reinforcement layer at the interface. It is of interest to report that the case of footing on partially replaced collapsible soil with cohesionless material showed slight improvement, while a combination of partially replaced collapsible soil with geotextile layer at the interface showed a significant improvement in reducing the collapse settlement. Design procedure is presented to predict the collapse settlement for a given collapsible potential of the soil, loading conditions, replacement thickness, and strength of the geotextile material used.

One of the major concerns of designing offshore wind turbine foundations is to address the effects of cyclic loading coming from, especially, wave loading. This paper describes the results from a test series on a small-scale mono bucket foundation installed in water-saturated dense sand exposed to cyclic loading with 1.0 Hz. A macro model has been calibrated to fit the test results, and it has been shown that the load frequency has a major influence of the measured rotation of the foundation. Compared to previous tests, the presented test program has been conducted under partly drained conditions due to a higher load frequency, and in addition the foundations have been preloaded by cyclic loads simulating small wave impacts prior to major storm events. The calibrated model shows, in contradiction to previous findings for drained conditions, that two-way loading leads to the largest rotations of the foundation. The reason for this is the partly drained or undrained soil response during testing. The paper also presents a cyclic load response diagram, which is a tool to be used in a predesign of a mono bucket foundation. The diagram gives the designer a quick overview of the cyclic load response and the effect of increasing or decreasing the foundation size.

This paper proposes approaches to estimate effective friction angle from cone tip resistance with incorporation of matric suction. Cone penetration tests (CPTs) are performed in unsaturated residual soil layers, where matric suction is measured with depth. Unsaturated triaxial testing is performed on retrieved samples adjacent to the CPT soundings. In addition, the proposed model incorporates a more robust stress state through the use of coefficient of earth pressure at rest ($K_0$) and overconsolidation ratio (OCR). Results indicate that reducing the CPT tip resistance data without incorporating the matric suction effect leads to 6–9° overestimations of the friction angle (the matric suction was approximately 80–90 kPa in type MH soil). The incorporation of a more representative stress state using the mean effective stress in the CPT data reduction leads to a 2–3° reduction of the estimated friction angle compared with traditional approaches for OCR values equal to or greater than 2.

This paper presents the study of possible failure mechanisms and stability of large-diameter tunnels in cement-treated soil surrounds. The failure modes and critical tunnel support pressure of tunnels with varying strength and thickness were observed from centrifuge model tests. The results of the centrifuge model tests were used to develop a numerical model in geotechnical finite-element software. Both centrifuge and numerical studies were conducted for the plane-strain condition. The results show that the improved soil surround may fail before a generalized plastic collapse occurs in the soft soil. Three failure modes were identified using centrifuge and numerical modeling: shear failure in weak and thin tunnels, multiple cracks or rupture in medium-strength tunnels, and tension failure in higher strength and thicker tunnels. Taken together, these failure modes define a transition from shear failure in thin, weak improved soil surrounds to tension failure in thick, strong improved soil surrounds. A failure map and stability chart have been proposed to enable some form of preliminary design for such cement-treated soil surrounds.

Reliability analysis and reliability-based design (RBD) via the first-order reliability method (FORM) can overcome some limitations and ambiguities in the load and resistance factor design (LRFD) approach, for example in situations where loads contribute not only destabilizing but also stabilizing effects, situations where load and resistance sensitivities vary due to different geometric and other details, situations with spatial variability and parametric correlations, and situations with different targets of probability of failure. Examples of RBD are presented to show that under these circumstances RBD can offer interesting insights apart from providing valuable guidance on what load and resistance factor values to use in LRFD. Specifically, this study investigates RBD of a shallow footing subjected to both vertical and horizontal loads in different proportions, laterally loaded piles with depth-dependent nonlinear p-y curves and different cantilever lengths, and an anchored sheet pile wall in which the load effects and resistance are nonlinear functions of soil shear strength, unit weight, geometry, and parametric correlations. This study demonstrates that RBD can automatically reveal critical design scenarios and obtain design values of loads and resistance which reflect different parametric sensitivities across different scenarios.