Assessing the seismic vulnerability of masonry buildings - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 352

façades with porticoes at ground level, resulting in possible soft storeys. The
lower value of average limit displacement for near collapse in Nocera
Umbra is to be ascribed to a large proportion of buildings of rubble stone
masonry, failing condition (4) due to the small unit size. At the other end,
the lower lateral capacity of the Serravalle case for the combined mecha-
nism is associated with more slender façades.
The reliability of such results can be considered within the framework
set out in the Eurocode 8, whereby the level of knowledge associated with
the results of a seismic assessment of a structure is expressed as a function
of the confi dence factor. Eurocode 8 recognises three levels of knowledge
(limited, normal and full) and three fi elds of information (geometry, con-
struction details and materials). Data used in the FaMIVE approach are
collected by on site visual inspection with some measurement and in situ
detailed observation of construction details. However only limited in situ
non-destructive tests of materials are performed and material characteris-
tics are otherwise assigned based on past studies and surveyor experience.
The level of knowledge is thus regarded as superior to KL1, limited , but not
quite equal to KL2, normal . For this level of knowledge, a static nonlinear
analysis, such as the limit state mechanism approach, leading to a capacity
curve is deemed appropriate. Hence according to the recommended values
in the Eurocode 8, the confi dence factor CF should be in the range 1.2 to
1.35 depending on how close the actual knowledge is with respect to the
reference level KL2. The confi dence factor is then used to reduce the capac-
ity values as obtained from the assessment.
Although the EC8 approach recognises the importance of treating epis-
temic uncertainties, the level of knowledge is then translated to a safety
factor rather than a probability or possibility of a specifi c value occurring.
While this approach can be considered as acceptable for the assessment of
single buildings, it does not account for the natural, organic variation, given
the three remits of knowledge discussed above, when considering a large
number of buildings - even more so when the buildings are several decades
old and have undergone alterations, changes and natural decay.
In the FaMIVE procedure a different approach is proposed to defi ne the
reliability of the input data as well as the output. Depending on whether
data, in each section of the data collection form, have been: (i) collected
and measured directly on site, (ii) collected on site and confi rmed by exist-
ing drawings or photograph, or (iii) collected from photographic evidence
only, three levels of reliability are assigned: high, medium and low, respec-
tively. For the three reliability levels, the following confi dence ranges for a
given parameter are considered: 15%, respectively.

The confi dence levels associated with the output parameters, ultimate accel-
eration A u and limit state displacements

μ ± 5%, μ ± 10%, and μ ±

Δ I , are computed as weighted com-
bination of the confi dence range of the input parameters.