Page 242 - Forensic Structural Engineering Handbook

7.26 CAUSES OF FAILURES

structures that span a fault will experience severe differential foundation displacements dur-
ing an earthquake.

Determination of Earthquake Loads
The assumption of inelastic behavior complicates the seismic analysis of structures. First,
it is not necessarily valid to assume that loads can be superimposed linearly. For instance,
generally it is not adequate to treat gravity loads separately from seismic loads in accurate
analyses. It is important to know which structural elements are taken beyond elastic limits
by the combination of applied forces, because effective stiffnesses change when this hap-
pens. Analyses that do not consider the impact of non-seismic loads may overlook build-
ing response characteristics that are important to a structure’s survival.
The assumption of inelastic behavior also complicates analyses because accurate analy-
ses for such conditions are nonlinear. Furthermore, our understanding of post-elastic, cyclic
response of building elements generally is not as developed as that for elastic response.
Most conventional seismic design approaches dictate construction detailing that is intended
to ensure a level of ductility in the completed structure that allows engineers to apply sim-
plifying assumptions in analytical procedures. For these reasons, investigators face a
dilemma when they must analyze structures that are not detailed in conformance with pre-
scribed standards, and when analyses must be performed with high accuracy.

Earthquake Data. The earthquake data needed for an investigation of a seismic event
normally come from monitoring stations in the vicinity of a site of interest. Seismographs
at various sites in and around earthquake zones are poised to collect data on ground motions
during all serious events. Records acquired at these sites normally will contain data neces-
sary to perform any of the common calculations for estimating earthquake response: time-
history analyses, empirical calculations, and modal analyses.
It is common to base seismic analyses on a response spectrum that is calculated from
earthquake records using 5 percent damping to be representative of typical building
16
response. To the extent that the characteristics of a structure under investigation are not
represented by these parameters, adjustment is needed.
Ground motions are site-specific. Not only do seismic motions attenuate and change
character with distance from the epicenter, but also there may be soil conditions at any par-
ticular site that will amplify motions. Therefore, to use data collected off-site, conditions
should be evaluated at both sites to judge whether data must be adjusted.
If the structure of interest is much closer to the ground motion recording site than to the
fault that generated the earthquake, it sometimes is possible to use motion records without
adjustment for attenuation. However, consideration should always be given to seismic
ground motion attenuation models that can be used to improve the applicability of seismic
ground motion data to the site of interest. In general, published attenuation models 17–22 are
based on regional geology and seismicity and therefore are of limited applicability outside
the study region.
Subsurface materials normally are assigned to one of six classifications for seismic
response evaluations, based on the conditions in the top 100 ft (30 m) of the layer:
A. Hard rock, with an average shear wave velocity, v, greater than 5000 ft/s (1500 m/s)
B. Rock, with v greater than 2500 ft/s (760 m/s) and less than or equal to 5000 ft/s (1500 m/s)
C. Very dense soil and soft rock with v greater than 1200 ft/s (360 m/s) and less than or
equal to 2500 ft/s (760 m/s)
D. Stiff soil with v greater than 600 ft/s (180 m/s) and less than or equal to 1200 ft/s (360 m/s)