Incremental dynamic analysis (IDA) applied to seismic risk assessment of bridges - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 570

collapse capacity. This is because including three different earthquake
types increases the record-to-record variability of the results. It is interest-
ing that the median collapse capacity predictions considering three differ-
ent earthquake types were similar to those predicted using crustal records
only, as shown in Fig. 21.5a and 21.5c and presented in Table 21.2a and
21.2c. In a study by Tehrani and Mitchell (2012b) the median collapse
capacities of different bridge confi gurations with different period ranges
were evaluated using different earthquake types in IDA. The results indi-
cated that the median collapse capacities obtained using crustal events
only were typically close to those obtained considering three different
earthquake types in IDA, whereas the record-to-record variability was
much larger when three different earthquake types were considered. This
may suggest that seismic evaluations using IDA for Vancouver could be
carried out using crustal events only and to include the effects of different
earthquake types the record-to-record variability could be increased. The
record to record variability in such cases could be approximately predicted
using the elastic response spectra of all records from different event types.
More research is needed on this subject before a conclusion can be possi-
bly made. The use of the CMS-based record selection resulted in somewhat
lower predictions of the collapse capacity as shown in Fig. 21.5b compared
to the predictions using the epsilon-based method (Fig. 21.5c). However,
the CMS-based method was less sensitive to the number of records consid-
ered in the IDA (Tehrani et al. , 2012). The IDA results obtained using the
UHS-based record selection are demonstrated in Fig. 21.5d and Table
21.2d. A comparison of the IDA results obtained using the UHS-based
method with those obtained using the epsilon-based method resulted in an
80% difference in the median collapse capacity. This demonstrates the
importance of the record selection on the IDA results. Therefore, neglect-
ing the spectral shapes and epsilon values in the record selection can result
in highly conservative predictions of the median collapse capacity of the
structure.
The direct use of the epsilon-based method to account for the spectral
shape of the ground motion records is often time-consuming and may be
complicated in practice. A simplifi ed method has been developed by Hasel-
ton et al. (2011) to modify the collapse capacity predictions for the effects
of epsilon and spectral shapes in IDA predictions. For this purpose, a fi xed
set of records can be used to perform IDA. The predicted capacities obtained
using IDA are then modifi ed for the effects of spectral shapes using a spec-
tral shape factor (SSF). To compute the SSF, a regression analysis needs to
be carried out to derive a relationship between the natural logarithm of
the collapse capacities versus ( T 1 ) values for each record. The results from

the simplifi ed method are in good agreement with those obtained using the
epsilon-based record selection method.

ε