Effect of soil–structure interaction and spatial variability of ground motion on seismic risk assessment of bridges - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 590

on these issues still remain approximate and require improvements, as the
phenomenon is often treated as either benefi cial or is completely neglected
(Mylonakis and Gazetas, 2000). One reason for this oversimplifi cation is
based on the anticipated period elongation of the structure with respect to
the fundamental period of vibration of a fully fi xed system and the corre-
sponding monotonic decrease of spectral accelerations foreseen, at least on
the basis of the design spectrum. Another reason is the energy dissipation
that is known to take place at the foundation level due to the radiation of
seismic waves and hysteretic damping, both contributing to the reduction
of seismic loading when the presence of soil is considered (Gazetas, 2006;
Makris et al. , 1994). The above arguments support the common perception
that essentially any structure (bridges included) can be conservatively con-
sidered to be fi xed at its base during analysis, design and assessment.
Contrary to the above state-of-practice, however, current research has
shown other evidence (Finn, 2005; Gazetas, 2006; Pecker, 2007; Mylonakis
and Gazetas, 2000; Wolf and Preisig, 2003). The foundation is indeed fl exible
and dissipates energy but interacts with the surrounding soil and the super-
structure in such a way that it fi lters seismic motion ( kinematic interaction ),
while it is affected by inertial forces generated by the vibration of the
superstructure in a complex manner ( inertial interaction ). This depends on
various factors (Mylonakis and Gazetas, 2000; Wolf and Preisig, 2003):
intensity of ground motion, dominant wavelengths, angle of incidence of
the seismic waves, stromatography, stiffness and damping of soil, and size,
geometry, stiffness, slenderness and dynamic characteristics of the structure.
This complex and coupled interaction may lead to a signifi cant increase of
displacement and ductility demand under certain circumstances (Finn, 2005;
Gazetas, 2006; Pecker, 2007).
Bridge structures are also sensitive to the assumptions made regarding
the selection of an earthquake ground motion scenario in terms of fre-
quency content, duration, amplitude and directivity. Moreover, spatial vari-
ability of earthquake ground motion (SVGM) may be signifi cantly different
at support points of long bridges, due to (a) the traveling of the waves at a
fi nite velocity (Der Kiureghian et al. , 1997), (b) the loss of their coherency
in terms of statistical dependence due to multiple refl ections, refractions
and superposition of the incident seismic waves (Zerva, 2009), (c) attenua-
tion of motion due to geometrical spreading of the wave front and the loss
of kinematic energy during propagation, (d) the effect of local site condi-
tions, (e) pier-dependent kinematic interaction at the soil-foundation inter-
face (Sextos et al. , 2003a) and (e) non-uniform liquefaction along the bridge
length (Elnashai et al. , 2009).
The SVGM phenomenon can lead to signifi cant spatial and temporal
variations of seismic motion in terms of amplitude, frequency content and
arrival time, thus inducing signifi cant forces and deformations under certain