Site response analysis is a component of paramount importance in reliable seismic risk assessment in earthquake-prone zones. The theoretical approaches to derive the site free field response to earthquake excitation include linear elastic, equivalent linear elastic and nonlinear analyses. This work shows new evidence on advantages of using nonlinear site response in modelling shear wave propagation. The nonlinear approach is shown to be able to reproduce the presence of elastic waves released in soil of hysteretic-type nonlinearity when subjected to dynamic loading. Based on the recent findings a simple 2D numerical study of an ideal experimental setup is presented. Finally, potential consequences of the recent findings on wave propagation in soil experiencing nonlinear hysteresis-type behavior are briefly discussed regarding the seismic response in large earthquakes (e.g. Mexico City) and developments of novel conbined finite element and constitutive formulations.

It is well established that site effects and the interaction between the structure and the ground play a significant role in the structural damage observed after major earthquakes. The latter fact, along with advancements in computation geomechanics, has made more common the use of numerical dynamic soil-structure interaction (DSSI) analyses, for instance with the finite element method, to assess the seismic behaviour of structures in major engineering projects. However, the generation of the input motion at the base of the numerical model might not be a trivial task, particularly when the adopted motion is specified at the top of a nonlinear soil deposit. In this case, one-dimensional frequency domain analyses are usually employed for deconvolution, where nonlinear behaviour is accounted for through the equivalent linear approach. However, if a complex nonlinear elastoplastic constitutive description is adopted to characterise the behaviour of the soil, the original motion will not be recovered at the surface because of the very different approaches for representing the behaviour of the soil. In this context, the paper addresses the application of a time-domain deconvolution procedure which allows us to consider the same nonlinear behaviour of the soil that is intended to be used in the numerical DSSI analysis. The methodology was applied to the Treasure Island site, in San Francisco, during the 1989 Loma Prieta earthquake, where significant site effects were identified. Results show that the evaluated procedure can satisfactorily generate an input motion for the numerical model such that the adopted surface spectrum is recovered when propagated through the nonlinear soil deposit.

Sites with deep soil layers are well-known for their propensity to significantly amplify ground motion responses in the surface soil. Moreover, the prolonged fundamental period characteristic of these sites makes structures with extended natural periods susceptible to damage induced by resonance. The focus of this study is the Eureka vertical array site in California, located within the context of deep overburden. The study utilizes both equivalent linear (EL) and Pressure-Dependent Hyperbolic (GQH) nonlinear time-domain methodologies to simulate and analyze the dynamic response of the ground surface under varying seismic excitations, including both strong and weak conditions. Following these analyses, a comparative assessment is conducted using recorded ground motion data. The findings indicate that under weak seismic conditions, both analytical methods effectively capture and reproduce ground seismic responses. However, under intense seismic forces, the nonlinear time-domain analysis method outperforms the equivalent linear analysis method significantly in predicting the dynamic response of the ground surface. Furthermore, the results obtained from the equivalent linear approach show significant deviation from the data recorded by the array, indicating its limited suitability for seismic response analysis in deep soil strata sites exposed to intense seismic events.

Marine soils are primarily formed through the transportation of rock and soil particles from adjacent land areas to the sea/ocean by wind, ice, rivers, and rainwater runoff, which accumulate on the seafloor. Waterfront structures are continually being constructed globally, either directly on these soils or in conjunction with reclamation projects to create new commercial land. These soils vary from coarse-grained (gravelly sands, sands, and generally soils exhibiting sand-like behaviour) to fine-grained (clay-like behaviour), and their particle size distribution depends on the distance from the landmass, the mechanism of transportation, and the coastal processes that may affect them. A special and common category of these soils is the mixture of coarse-grained and fine-grained materials that exhibit, depending on the location investigated, either sand-like or clay-like behaviour, which can be challenging to differentiate. In this study, three different nonlinear dynamic analysis techniques are applied to assess the impact of such soils on reclamation and wharf waterfront structures. This paper compares these techniques and discusses the outcomes, while also proposing a method to reasonably simulate marine transitional soils for designing new waterfront and/or marine structures.

Several constitutive models have been developed over the years to capture the salient features of the undrained cyclic stress-strain response of sandy soils, that is of crucial interest for assessing the behaviour of geotechnical systems under seismic loading. Among them, the bounding surface plasticity theory turned out to be a valuable framework for reproducing the hardening soil response under both drained and undrained conditions. However, such models usually require a non-straightforward calibration of many parameters, employing also trial-and-error procedures that inhibit the physical meaning of the parameters and their eventual correlation. In this view, the present study proposes an optimised procedure for calibrating the trial-and-error parameters of bounding surface constitutive models for coarse-grained soils. The proposed approach is applied to a modified version of the model developed by Papadimitriou and Bouckovalas (2002), which was recently implemented in OpenSees by the Authors to investigate liquefaction-related phenomena. Based on the obtained results, a preliminary correlation between the critical parameters and a significant response quantity to identify liquefaction triggering is proposed with reference to a well-characterised soil. Considerations on the ongoing developments and of the extension of the method to other classes of models are finally provided.

A technology was proposed to construct Grid Ground Model, which is a general-purpose mediation model using ground information such as borehole data and soil test results. By building a common data structure once, flexible data manipulation can be performed. In existing study, the authors developed a model building function based on this technique that can be used for more advanced applications, including finite element code, by providing a high level of generality in material parameters setting and geometry processing. In this study, two functions were developed to enable the use of this technology for seismic response analysis. The first is a function to automatically construct a model of ground around the target area for use in lateral boundary conditions. The second is a function to input seismic wave data to the bottom of the model. With these functions, input data for seismic response analysis with boundary conditions can be obtained immediately based on ground and seismic data. When a simple model for seismic response analysis of the FEA software DACSAR-I is constructed using the developed technology, it was confirmed that the man-hours required could be reduced to about 1/50 of those required in constructing the model by Excel. The developed technology is useful for constructing FEA models for earthquake damage estimation and seismic design of construction works.

In order to improve the stiffness and stability of the debris flow pile forest structure, this paper applies the CFRP-Concrete-Steel tube combination pile, which is used in explosion and blast resistant engineering, to the debris flow barrier project; the dynamic responses of concrete pile (CP), concrete-steel tube pile (CTP) and CFRP-Concrete-Steel tube combination pile (CCP) under the impact of large boulders are simulated by finite element software ABAQUS respectively. The Von Mises stress cloud diagram, equivalent plastic strain cloud, pile acceleration and pile bottom support reaction force time course, block velocity time course and impact force time course of the pile were compared and analyzed under different conditions. The simulation results show that the CCP is not only stiff, but also has more distributed stresses and is less prone to damage due to stress concentration when subjected to impact loading, and has the best overall structural stability after being impacted by the boulder. CP and CCP have plastic strain only at the impact site, and the area damaged by the impact of CCP is the smallest; the pile bodies of the three structures are subject to different degrees of oscillation along the impact direction of the debris flow after the impact of the boulder, and the contact with the boulder occurs once or twice during the whole process; the energy transferred from the boulder to the pile body is mainly consumed by the strain and oscillation of the pile body.

Ground reconsolidation following strong earthquakes can severely damage buildings and infrastructure. However, the nonlinear void ratio–effective stress relationship during reconsolidation remains poorly characterized owing to limitations of laboratory testing. The discrete element method is utilized to examine reconsolidation characteristics in K₀-consolidated granular materials. The results demonstrate that the residual effective stress is the core indicator of the volumetric strain during reconsolidation. Particular focus is given to the post-liquefaction reconsolidation process, which can be divided into two stages: liquefied portion and subsequent solidified portion. The former stage accounts for a substantial proportion of total volumetric strain. For specimens with identical average size, particle size distribution has a major influence on volumetric strain in the liquefied portion while a negligible impact in the solidified portion. This study also finds a higher proportion of floaters in the polydisperse (PD) versus monodisperse (MD) specimen. Furthermore, pore uniformities in both specimens increase during undrained cyclic shear. During the reconsolidation process, pore uniformity rises in the MD specimen while the PD specimen exhibits variable changes.

This paper discusses the results of a preliminary study of numerical simulations investigating wave propagation during the installation of in-situ driven concrete piles (type Franki pile). Franki piles are cast-in-situ piles that are installed by driving a steel casing into the ground using heavy ramming of a cylindrical hammer. It is a dynamic pile installation process where the hammer directly transfers the dynamic forces to the soil within the installation tube causing high wave propagation through the soil. When installed in groups, the vibrations caused by the driving process of one pile may result in damages to the early-age concrete of adjacent piles. Such complex soil-structure interaction problems can be represented numerically by modelling the pile installation process using realistic parameters (ground conditions, ramming energy etc.). As a preliminary study, a single pile installation is simulated using the FEM software Abaqus. The simulation is based on a Coupled Eulerian-Lagrangian (CEL) approach where the soil is modelled using the hypoplastic constitutive model. The aim is to investigate parts of the installation process regarding the effect of discrete hammer drops. Within the CEL method, the pile hammer is modelled as a Lagrangian part, while the soil is treated as a Eulerian part. The Lagrangian part can move freely through the Eulerian mesh until it encounters Eulerian material. As a result of the simulations, a realistic amplitude pattern can be observed. This can be used as basis for the next phase of the study where the numerical analysis will be validated with the data recorded from field tests.