Seismic risk analysis using Bayesian belief networks - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 171

7
Seismic risk analysis using Bayesian
belief networks
S. TESFAMARIAM, The University of British Columbia,
Canada and Z. L I U, Toyota Technological Institute, Japan
DOI : 10.1533/9780857098986.2.175
Abstract : Seismic risk assessment involves different levels of complex
models and requires consideration of uncertainties. The confi dence
level on seismic risk assessment can be enhanced by considering soft
computing techniques that account for different sources of uncertainty.
Soft computing is a conglomerate of computing techniques that include
fuzzy-based methods, neuro-computing, probabilistic reasoning, genetic
algorithms, chaotic systems, belief networks, and learning theory. In this
chapter, application and utility of Bayesian belief network (BBN) in
liquefaction assessment and seismic risk assessments at regional as well
as individual building levels are provided. For the liquefaction, the BBN
structure is generated from historical data through different machine
learning algorithms. For the risk assessment, a heuristic-based BBN is
proposed and an example is provided.
Key words : risk assessment, Bayesian belief network (BBN), liquefaction,
decision making.
7.1 Introduction
7.1.1 Problem statement
Reported damage from recent global earthquakes, 1994 Northridge earth-
quake in the USA, 2004 Sumatra earthquake in Indonesia, and 2008 Sichuan
earthquake in China, for example, highlighted the vulnerability of existing
buildings and the importance of seismic retrofi tting. The building vulnera-
bility is mainly due to older building design codes, poor design/construction
practice and poor code enforcement. Most of these buildings are currently
operational and need to be further assessed and upgraded to minimize
seismic damage and improve life safety. From a city manager's decision
making perspective, comprehensive retrofi tting/upgrading of all buildings
is not economically feasible, and it is desirable to perform risk-based pri-
oritization of defi cient buildings (e.g. Tesfamariam and Saatcioglu 2008,
2010; Ellingwood 2001).
A risk-based seismic assessment can be undertaken by integrating site-
specifi c seismic hazard, building vulnerability (likelihood of failure), and
175