Civil Engineering Reference
In-Depth Information
given separately for the three cities, and presented in terms of fatalities as
a fraction of total population. City B, as expected, has the highest rate. All
curves present a smooth behaviour. This is not the case for the MAFs of
displaced population, shown in plot (b). Similarly to the previous plot, rates
are given by city and in terms of the ratio of displaced to total population.
All cities, but in particular cities A and C, present a sharp transition close
to the 100% value. This is due to the fact that, as already seen in Fig.
18.13(d), displaced population is infl uenced by utility loss and for this simple
example, the EPN is either fully functional or completely non-functional.
In the latter case, independently of direct damage, buildings, even usable
ones, become non habitable and occupants are forced to evacuate. This is
more apparent in cities A and C where direct physical damage is less likely
due to the larger distance from the seismogenic sources.
To gain further insight into this effect of interdependencies captured by
the model, plot (d) shows the MAF of normalised displaced population for
the most affected city B under different modelling assumptions. The solid
curve is the same shown in plot (b), i.e. the case where utility loss is con-
sidered, and the lifelines are dependent (WSS on EPN). The dashed line
shows what would happen if utility loss was still considered in the evalua-
tion of building habitability, but the interaction between the two systems
was disregarded. The MAF would consistently decrease, and, most impor-
tantly, the sharp kink at around 100% would be attenuated. This attenuation
would be due to the reduced vulnerability of the WSS in case the possibility
of EPN-induced failure of the pumps was removed. This is clearly shown
in plot (c) where the MAF of the water supply system serviceability index
is shown in the two cases. The kink in the displaced people MAF would
completely disappear and the curve would more closely follow the trend of
the fatalities MAF in plot (a) if displaced people were computed based only
on direct physical damage (the 'no lifelines' case).
18.9 Futuretrends
The development of the presented model is still ongoing. Considerable
research efforts are aimed at completing the implementation of the tax-
onomy of systems in Fig. 18.2, with particular attention to critical facilities.
Future research will focus on methods to effi ciently compute sensitivity
measures within the simulation framework, on one hand, and on introduc-
ing Bayesian inference in the
Analysis
class on the other.
18.10 Acknowledgements
The model described in this chapter is the result of work carried out by the
authors mostly within the collaborative research programme SYNER-G
Search WWH ::
Custom Search