Earthquake risk management of civil infrastructure: integrating soft and hard risks - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 238

specifi c context but sub-optimal in others. It is possible that a structure may
become vulnerable (as in the case of an I-beam) and further a highly inter-
connected system or structure may become prone to unzipping or cascading
failure. Of course one can argue that we simply require an optimum over
all possible contexts - but because we now know that complex systems
contain unknown unknowns and hence future unknown contexts in which
an optimum may fail dramatically. We need to aim for robustness rather
than optimisation.
Integrating disparate sources of information to improve the management
of risk can also bring direct positive benefi ts for opportunities to gain
synergy, robustness and resilience. Grundy (2011) has described how a
headmaster in Japan successfully took special measures to retrofi t his school
building and to prepare, educate and rehearse his children and staff. Robust-
ness is strength in all possible modes of failure of a system including the
people. Robust systems are sturdy and hardy; they endure and are tolerant
of damage. A system is robust if it resists damage and, if damaged, resists
the propagation of that damage. Progressive collapse or unzipping occurs
when, after some initial damage, further damage progresses, without extra
inputs, through many limiting conditions and damage states of increasing
scope, breadth and consequences. Vulnerability is susceptibility to initial
damage. It is the ease with which damage propagates. A vulnerable system
may suffer progressive collapse where small damage leads to disproportion-
ate levels of damage. Vulnerability is suffi cient but not necessary for lack
of robustness. For example, lack of robustness is also caused by undue sen-
sitivity to details such as design assumptions, the accuracy of manufacturing
or construction methods, the quality of details such as connections and pos-
sible perturbations such as ground movements. Resilience is the ability to
recover from damage. It is being capable of returning to an original state
after change. For example, linear elastic strain energy is resilient because it
is recoverable - there is no permanent change. Hollnagel et al. (2006) have
proposed the idea of 'resilience engineering' to provide a new paradigm for
considering the totality of requirements for a system and to capture a way
of thinking about safety which enhances the ability at all levels of organisa-
tions to be robust yet fl exible and to use resources proactively to manage
processes to success. They distinguished between resilience and adaptability,
which is a capability of being able to adjust to different conditions and still
be fi t for purpose. Systems that are highly interconnected may not be robust,
resilient, or adaptable. Notice that these defi nitions do not involve risk. The
2011 earthquake in Japan illustrates vividly the interconnections between
seismic risk, risks to individual structures and parts of the infrastructure,
tsunamis and consequent damage to whole regional areas, damage to
nuclear power plants and radiation leaks, social consequences to the spirit
of a whole nation, and impact on the nation's economy and possible impacts