Ground motion selection for seismic risk analysis of civil infrastructure - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 77

4
Ground motion selection for seismic risk
analysis of civil infrastructure
B. A. BRADLEY, University of Canterbury, New Zealand
DOI : 10.1533/9780857098986.1.79
Abstract : This chapter presents a ground motion selection methodology,
based on the generalized conditional intensity measure (GCIM)
approach, which enables ground motion selection as part of a
probabilistic seismic performance assessment. The basis of the approach,
which involves the construction of, and selection of ground motions
from, the conditional multivariate distribution of ground motion
intensity measures (so-called GCIM distributions) is discussed. A case
study is used to illustrate the methodology input requirements and
salient features. Continuing developments facilitating the routine
practical implementation of the methodology are discussed.
Key words : ground motion selection, generalized conditional intensity
measure (GCIM) approach, seismic response bias, seismic demand curve.
4.1 Introduction
The continuing evolution toward the seismic design of engineered facilities
based on their expected performance places an increased emphasis on the
use of computational models to predict the seismic response of such facili-
ties. A key consideration in time-domain computational models is selection
of the input ground motions for seismic performance assessment. Many
ground motion selection and/or modifi cation methods have been proposed
in literature over the past decade, and yet a lack of consensus remains (e.g.
Katsanos et al. , 2010, and references therein).
This chapter presents a ground motion selection methodology, which
unlike the majority of ground motion selection methods, has a theoretically
sound basis. The methodology enables ground motion selection as part of
a probabilistic seismic performance assessment, and can be considered a
generalization of the conditional mean spectra approach of Baker and
Cornell (2006). The basis of the approach is the construction of a condi-
tional multivariate distribution of a general set of ground motion intensity
measures. Ground motions are then selected from random realizations of
this multivariate distribution, with a problem-specifi c 'weight-vector' used
to emphasize particular ground motion features of importance in the selec-
tion process. Subsequent to the use of the selected ground motions in
79