The role of soil-structure interaction (SSI) in the seismic response of structures is reex-plored using recorded motions and theoretical considerations. Firstly, the way current seismic provisions treat SSI effects is briefly discussed. The idealised design spectra of the codes along with the increased fundamental period and effective damping due to SSI lead invariably to reduced forces in the structure. Reality, however, often differs from this view. It is shown that, in certain seismic and soil environments, an increase in the fundamental natural period of a moderately flexible structure due to SSI may have a detrimental effect on the imposed seismic demand. Secondly, a widely used structural model for assessing SSI effects on inelastic bridge piers is examined. Using theoretical arguments and rigorous numerical analyses it is shown that indiscriminate use of ductility concepts and geometric relations may lead to erroneous conclusions in the assessment of seismic performance. Numerical examples are presented which highlight critical issues of the problem.

This paper recognizes that soil layering may have a profound effect on the settlement of pile groups and that pile-to-pile interaction induces not only additional settlement at the head of each pile but also additional stresses along its shaft. A general analytical formulation is developed, based on the Winkler model of soil reaction, for determining the vertical interaction factors between two piles embedded in multi-layered soil. It is shown that such interaction factors depend not only on the displacement field arising from the settlement of a loaded ('source') pile but also on the interplay between the adjacent ('receiver') pile and the soil subjected to this displacement field. Such interplay, which has not been considered (at least explicitly) in currently available simple methods, is quantified here through an analytically determined factor ζ which lies between 0 and 1. The paper also develops closed-form expressions for pile stiffness and interaction factors in a two-layer stratum, and highlights crucial aspects of the problem. L'article reconnaît que la stratification du sol peut avoir de profonds effets sur le tassement des groupes de pieux et que l'interaction entre les pieux produit non seulement un tassement supplémentaire à la tête de chaque pieu, mais aussi des contraintes supplémentaires le long de son fût. Les auteurs ont mis au point une formule analytique générale basée sur le modèle de r´action du sol de Winkler pour déterminer les facteurs d'interaction verticale entre deux pieux enfoncés dans un sol à couches multiples. Ils montrent que ces facteurs dépendent non seulement du champ de déplacement engendré par le tassement d'un pieu chargé («source»), mais aussi de l'interaction entre le pieu adjacent («récepteur») et le sol soumis à ce champ de déplacement. Cette interaction, dont ne tiennent pas compte (du moins explicitement) les méthodes simples actuellement disponibles, est quantifiée ici &grave l'aide d'un coefficient ζ établi par analyse, qui se situe entre 0 et 1. L'article présente également des expressions de forme fermée pour la rigidité des pieux et les coefficients d'interacdon dans un sol à deux couches, et fait ressortir les principaux aspects du problème.

Simplified analytical models are developed for the lateral harmonic response of single piles and pile groups in layered soil. Pile-to-soil interplay is represented by a dynamic Winkler formulation based on frequency-dependent springs and dashpots. For pile-to-pile interaction, the wave field originating from each oscillating (“source”) pile and the diffraction of this field by the adjacent (“receiver”) piles are considered. The response of single piles and pile pairs is evaluated both numerically (through a transfer-matrix formulation) and analytically (introducing an efficient virtual-work approximation). Closed-form solutions are obtained: (1) for the impedance of single piles; (2) for the dynamic interaction factors between two piles; and (3) for the “additional” internal forces (“distress”) developing in grouped piles because of pile-to-pile interaction, a phenomenon frequently ignored in current methods of analysis. Both swaying and rocking vibrational modes are considered. The effect of pile length and soil layering on the impedances and interaction factors is studied. The predictions of the model are in agreement with earlier results, while its simplicity offers a versatile alternative to rigorous solutions.