Seismic risk analysis using Bayesian belief networks - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 191

7.5 Vulnerability and damage assessment of
individual buildings
7.5.1 Background
The quantifi cation of earthquake-induced building damage requires inputs
related to seismic hazard (e.g., spectra acceleration, liquefaction susceptibil-
ity) and parameters that contribute to building vulnerability (e.g., building
type, vertical and plan irregularity, construction quality). Various techniques
are used to assess building vulnerability and resultant seismic loss, however,
albeit similar techniques, different taxonomies are reported. The reported
techniques can be grouped into (FEMA249 1994; Boissonnade and Shah
1985): empirical/statistical method; heuristic method; and analytical/ mecha-
nistic/ theoretical method. At times, where a lack of information exists, a
combination of these methods can be used (Kappos et al. 1998).
The mechanistic methods involve building models that consider the
dynamic characteristics of a structure, and are commonly used to predict
damage for a single structure. The advantage is that they correlate the
seismic demand with typical physical parameters used in seismic design.
However, the main drawback is the challenge involved in model develop-
ment and computational efforts (Villaverde 2007). Lack of available data
and incomplete knowledge base, for example, necessitate the use of heuris-
tic-based models based on the intuitive knowledge of the engineers. The
empirical methods involve gathering and correlating ground motion and
damage information. They are based on fi eld observations of damaged
buildings during earthquakes. This method gives an average value of the
damage ratio for a group of buildings having similar structural systems and
materials of construction (Boissonnade and Shah 1985). Consequently, this
method is prone to high uncertainty. The empirical method is suitable for
regional damage estimation; however, it cannot be used for an individual
building.
The complex problem of building damage assessment can be handled
through a simple and manageable hierarchical structure (Tesfamariam and
Saatcioglu 2008, 2010). The hierarchical structure follows a logical order
where the causal relationship for each supporting argument is further sub-
divided into specifi c contributors. Figure 7.12 shows a four-level hierarchical
structure. Level 1 of the hierarchy is the overall goal of the analysis, i.e.
building damage . The building damage is computed by integrating the
parameters at level 2 that are site seismic hazard and building vulnerability .
The site seismic hazard (level 2) is quantifi ed through the fundamental
period ( T 1 ) of the structure and response spectra. The response spectra
are obtained either through a site-specifi c design response spectrum or
existing representative earthquake record. Soil type is used to modify the