2839
Analysis and Design of Piles for Dynamic Loading
Analyse et conception de fondations par pieux en chargement dynamique
Ray R.P., Wolf Á.
Széchenyi István University, Győr, Hungary
ABSTRACT: With the acceptance of Eurocode 8 in Hungary a new level of seismic design is now necessary. This paper outlines
some of the past history and present implementation of foundation design for seismic loading as practiced in Hungary. It shortly
describes the possibilities of modeling foundations (fix support, linear elastic support, non-linear elastic support) during the design of
the superstructure, and introduces the Hungarian practice. The influence of the different support’s methods for the bearing
forces/stresses of a superstructure is analyzed on a typical reinforced concrete office building using SAP2000 finite element software.
The results of the calculation are compared through the moment of a column, the deflection and the support reaction.
RÉSUMÉ : Il était nécessaire de modifier les principes de dimensionnement des fondations par pieux, vis à vis des charges sismiques,
depuis l’entrée en vigueur en Hongrie de l’Eurocode 8. L’étude donne un apercue des méthodes précédentes et actuelles du
dimensionnement dynamique des fondations. Elle montre aussi les possibilités de modelisations des fondations profondes pratiqués en
Hongrie (rigide, élastique linéaire, élastique non-linéaire). La comparaison des differentes techniques de fondation est réalisé par
elements finis sur une structure en béton armé et à l’aide du logiciel SAP2000. La comparaison des différents résultats est montré en
terme de moment, de déplacement et de force de réaction d’appuis.
KEYWORDS: pile foundation, seismic design, Eurocode-8
1 INTRODUCTION
Acceptance of Eurocode 8 in Hungary has lead to a great many
changes in how engineers think about designing for seismic
loading. Since Hungary experiences only moderate seismic
events over several hundred years, it is difficult to implement a
practical yet thorough design procedure for seismic (and other
dynamic) loadings. Much of the research and development in
seismic design is driven by post-earthquake evaluation and
large-scale testing. For regions with lower seismicity and
modest research budgets, this approach becomes problematic. A
more urgent problem is that of practical implementation:
designers must do something. In order to develop a reasonable
and rigorous approach to the problem, the Structures and
Geotechnics Department at Széchenyi István University has
pursued methods to link the latest approaches in earthquake
engineering to standard structural and geotechnical design
practice. To that end, part of the difficulty is pile design for
seismic loading. This paper describes some of the ongoing
efforts to use sophisticated analysis and testing in “code-based”
design practice.
2 BRIEF REVIEW OF RESEARCH
Seismic behavior of piles and their contribution to structural
response has been recognized for over 50 years. However, due
to the difficulty in modeling this behaviour, only very simple
models could be implemented. One early approach by Penzien
(1970) illustrates that the complexity of the problem was
acknowledged by even the best analysts. The early methods
used a subgrade reaction approach where the soil is replaced by
(usually linear) springs. Dashpots may be added as well to
simulate both material damping of the soil as well as radiation
of energy from the foundation. Additional refinements in the
approach to modelling piles included visco-elastic media and
analytical methods with relaxed boundary conditions (Novak
and ElSharnouby 1984, Dai and Roesset 2004). Concurrent with
the development of the relaxed continuum methods, finite
element approaches became more practical and required less
expensive computing resources. This rather complex evolution
of hardware, programming and theory resulted in what are now
the de-facto analysis methods for assessing seismic behavior.
In addition to these classes of analysis, four levels of
progressively “complete” SSPSI analyses can be described
(Wolf, 1985). The basic level consists of a single pile kinematic
seismic response analysis, normally incorporating nonlinear
response and performed as a pile integrity evaluation. A pseudo-
static method for pile integrity evaluation consists of
transforming the horizontal profile of soil displacement (derived
from a free-field site response analysis) to a curvature profile,
and comparing peak values to allowable pile curvatures. This
method assumes piles follow the soil perfectly, and that no
inertial interaction takes place. Alternatively, a displacement
time history may be applied to nodal points along the pile in a
dynamic pile integrity analysis. The kinematic approaches
consider the difference in stiffness between piles and
surrounding soil as seismic waves travel through both. The
mismatch leads to stresses generated in the pile system that
become more pronounced as the soil becomes softer (the
mismatch becomes greater). This behavior is indeed difficult to
measure, or even qualitatively assess in the field.
In a second level of analysis, pile head stiffness or
impedance functions may be condensed from linear or nonlinear
soil-pile analyses and assembled into a pile group stiffness
matrix for use in a global response analysis. Secant stiffness
values at design-level deformations are normally ascertained
from nonlinear soil-pile response analyses. Third, both inertial
and kinematic interaction may be evaluated from a sub-
