Earthquake risk management of civil infrastructure: integrating soft and hard risks - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 232

buildings in narrow streets mixed with modern developments. Bernadini
and Lagomarsino (2008) classifi ed such buildings as: palaces, churches, mon-
asteries/convents, mosques, towers, obelisks, theatres, castles, triumphal
arches, and arch bridges. Clearly such classifi cations are imprecise so that
the task of dealing with these structures does not (as they pointed out)
readily fi t within the precise scope of reliability theory. The structure of a
palace, church or monastery may indeed be quite similar, because each does
not 'know' it is what it is supposed to be. As far as 'mother nature' is con-
cerned, they are buildings that may be quite similar except in the way they
are adapted for a particular human purpose. Another obvious reason for the
lack of fi t into a precise theory is the sheer paucity of data and dependable
models of any single structure. Statistical data from damage assessments
after earthquakes are available and contribute to planning at regional and
national levels. However, any predictions from such data concerning indi-
vidual buildings will be at best approximate. Bernardini and Lagomarsino
(2008) therefore have suggested a methodology based on the European
macroseismic scale (EMS) in which the classes of vulnerability summarise
the observed damage to types of building subjected to past earthquakes.
However estimations of the local intensity of an earthquake from
recorded damage are clearly affected by the building type. The EMS 98
(1998) scale used by Bernadini and Lagomarsino (2008) defi nes the overall
damage to buildings, both masonry and reinforced concrete (RC) buildings,
using a scale of six qualitative judgements taking into account the severity
of damage to structural and non-structural elements. Grade 1 damage is
defi ned as negligible to slight, grade 2 is moderate, grade 3 is substantial to
heavy, grade 4 is very heavy and grade 5 is destruction. They then used
qualitative rules to summarise the vulnerability of a building not by the
type but by its membership in a defi ned class of vulnerability ranging from
A (the most vulnerable) to F (the least vulnerable) to cover the full range
of the expected behaviours from traditional non-engineered structures to
structures designed for seismic protection. These behaviours are expressed
as probabilities of different degrees of damage (six levels) for different
values of macroseismic intensity (four values) in damage probability matri-
ces (DPM).
Since these defi nitions are not precise, Bernardini and Lagomarsino
(2008) derived imprecise interval probabilities to estimate bounded inter-
vals of particular functions of the damage. They recognised that the assump-
tion in classical probability theory - that every member of the total sample
space is known and can be allocated a probability measure - needs to be
relaxed. However, they also note that the widespread distribution of instru-
ments providing direct records of the strong motions due to earthquakes
may seem to have reduced the signifi cance of this kind of macroseismic
scale and observational approach. They recognised that the HAZUS pro-