Civil Engineering Reference
In-Depth Information
induced damage assessment of geotechnical structures and structural vul-
nerability assessment due to ground shaking. Moreover, these advanced
analyses are an integral part of broader applications, such as decision analy-
sis under uncertainty and cost-benefi t analysis for seismic risk mitigation.
From the implementation viewpoint, an approach based on Monte Carlo
simulation is of advantage over conventional numerical integration, because
the former greatly facilitates fl exibility and transparency in the assessment.
It should be noted that the approach taken (numerical integration or Monte
Carlo simulation) is a matter of choice and convenience, and does not affect
the outcome of PSHA.
There are several PSHA issues that warrant improvements. Although the
concepts of PSHA are sound, its implementation requires considerable
experience and judgement to ensure reliable results. More research is
needed on how to properly characterise uncertainty and the relationship
between its epistemic and aleatory components. The division into these
components is somewhat artifi cial, and does not affect the mean hazard, but
needs to be carefully considered to avoid 'double-counting' of uncertainty,
which can artifi cially infl ate the calculated hazard. For example, a portion
of the scatter in GMPEs, as derived from regression analysis, is due to the
inclusion of multiple regions in the underlying database, and should not
really be treated as aleatory variability for a specifi c site (Atkinson, 2011).
Furthermore, the use of multiple GMPEs is not really an appropriate way
to characterise the uncertainty in what will be the mean ground motions at
a site for a given scenario. Alternative approaches to uncertainty in GMPEs
that examine the epistemic and aleatory components are needed (Atkinson,
2011).
The overall framework for PSHA can be modernised and streamlined by
integrating its various components and products more seamlessly. Ulti-
mately, the use of GMPEs to characterise ground motions will give way to
direct simulation of time histories for the events contributing to hazard (i.e.
simulating a time-history rather than invoking a GMPE). This will remove
the need to 'de-aggregate' a UHS to determine 'design events' for time-
history selection and scaling. Other methodological improvements will also
develop, facilitating further new applications.
This chapter has illustrated the step-by-step procedures of a simulation-
based PSHA approach by using a case study in western Canada. Two illus-
trative applications for incorporating PSHA into PLHA and PSRA were
used to demonstrate the potential of the method for more advanced appli-
cations in earthquake disaster management. Simply, any types of analyses
can be added to PSHA to conduct probabilistic assessment of buildings and
infrastructure subjected to extreme natural hazards (not necessarily limited
to ground shaking). For the specifi c examples discussed in this chapter,
these additional components are: liquefaction triggering analysis based on
Search WWH ::
Custom Search