Structural vulnerability analysis of civil infrastructure facing seismic hazards - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 205

8
Structural vulnerability analysis of civil
infrastructure facing seismic hazards

J. A G A RWA L, University of Bristol, UK

DOI : 10.1533/9780857098986.2.209

Abstract : Managing risks to structures or infrastructure systems requires
a thorough analysis of uncertainty. Structural reliability methods are
suitable for dealing with parametric uncertainty, but many other aspects
of system uncertainty remain unaddressed. These require an analysis of
the form of the structural system as well as the structural response to
hazards, such as material defects, loading conditions, and accidental
damage. An approach to identify structural risks due to failure scenarios,
which may remain hidden during a normal response analysis, has been
developed. Such an analysis leads to identifi cation of failure scenarios
with high vulnerability. Different actions on a structure might contribute
differently to a failure scenario; an interval probability theory can be
used to combine the evidence associated with different actions.
Structural risk is obtained as a function of probability of the identifi ed
failure scenarios and their consequences. The approach has been
generalized to other infrastructure systems.

Key words : safety, vulnerability, robustness, risk analysis, uncertainty,
disproportionate failure.

8.1 Introduction

One of the main objectives of a design process is to ensure the safety and
serviceability of the product. Traditionally, the safety of components or
systems has been expressed by safety factors as the ratio between capacity
and demand (see e.g. Pugsley 1966). Although easy to understand, it does
not account for uncertainties in material and loading. Reliability analysis
tools (see e.g. Thoft-Christensen and Baker 1982; Melchers 1999) have
proved useful to explicitly account for statistical variations in material and
demand parameters. However, they do not address all aspects of the system
uncertainty. For example, it is common to start with a predefi ned model of
load and to focus on failure scenarios that are most likely to occur. In some
cases, this may lead to a false sense of safety if there exist failure scenarios
with relatively low likelihood but with large consequences. The realization
of these low-probability and high-consequence failure scenarios cannot
be ignored for important structures (Taleb 2007; Government Offi ce for
Science 2012). In particular, human error and sabotage may contribute

209