Earthquake risk management of civil infrastructure: integrating soft and hard risks - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 229

tainty and risk in engineering. It is used in practical performance-based
engineering, PBE (Ellingwood, 2008). Unfortunately, reliability theory is
partial, since it covers only the same uncertainties as simple safety factors
- albeit more coherently and rigorously - but omits many of the risks that
have been found to result in actual failures. It also does not directly address
any of the practical factors described in the previous section. The idea of
PBE is wider than reliability theory - it is 'to provide owners and designers
with the capability to select alternative performance goals for the design of
different buildings' (Hamburger, 1996). Ellingwood (2008) writes that PBE
is 'evolving to enable new building technologies . . . to better meet height-
ened public expectations'. The aim is to extend consideration and hence to
cover all risks. PBE as currently formulated in PEER (2010) does enable a
better consideration of the levels of expected performance under earth-
quake loading, ranging from 'minor damage' to 'incipient collapse', and the
idea that essential buildings such as hospitals remain functional even under
an extreme event. However, it still is inadequate for an integrated treatment
of risk. The problem is that PBE excludes many of the important processes
for managing engineering risks. As Hamburger (1996) states, 'perhaps the
most signifi cant barrier to the adoption of performance-based earthquake
resistant design is the lack of control exercised by structural engineers over
the building design process'. He pointed out that the role of the structural
engineer is typically limited to designing and detailing structural compo-
nents of a building. However, the problem is even deeper than that as we
shall see.
Within the limited paradigm of PBE, the role of reliability theory is even
more limited. We can illustrate this with an example. Seismic fragility is
defi ned as the probability of occurrence of a stated level of damage due to
a given intensity of ground motion. This is apparently a suffi cient statement
of a technical problem - that of capturing the chance of a level of damage
from a given seismic intensity. Unfortunately, there are two important quali-
fi cations that derive from restrictions that are axiomatic to the theory that
make it insuffi cient as a general treatment of seismic risk. The fi rst restric-
tion is that it is only possible to represent important uncertainty in the
quality of the theoretical model of the physical system by a series of random
parameters. This is simply too crude. Not only may the scientifi c model of
the behaviour of the structural components be lacking under the effects of
an earthquake but also the vulnerabilities and subtleties of structural form
are too important to be treated so simply. For example the use of non-
structural partial in-fi ll panels in a frame building can change the response
characteristics of a structure. Also as Agarwal and England (2008) and
Nafday (2010) pointed out lack of structural robustness is a property of the
form of a structure where the 'well-formedness' or conditioning of con-
nectedness is critical. Ghosn et al. (2010) have discussed the complexities