PSI - Issue 44

Carlo Del Gaudio et al. / Procedia Structural Integrity 44 (2023) 259–266 Carlo Del Gaudio et al. / Structural Integrity Procedia 00 (2022) 000–000

261

3

inspected building based on the maximum observed seismic damage (e.g. Rota et al. 2008; Del Gaudio et al. 2017a; Rosti et al. 2018). Although alternative seismic intensity measures could be used (e.g. Rosti et al. 2020b), the peak ground acceleration (PGA) was employed for seismic input description at the sites of damage observations, making the proposed fragility model usable for territorial applications via the Italian national seismic risk platform (Borzi et al. 2021). PGA values were estimated at the different building locations by latest INGV ShakeMaps (Michelini et al. 2020). RC buildings were allocated to twenty-four building typologies (Table 1), identified based on the design level (i.e. buildings designed for gravity loads only and buildings designed for seismic loads), construction age (i.e. 1946-1970; 1971-1980; 1981-1990; >1990) and number of stories (i.e. 1, 2, 3 and ≥4 stories). Given that the majority of the Irpinia municipalities were not yet classified as seismic in 1980, RC buildings designed for gravity loads only refer to the Irpinia post-earthquake database. RC buildings accounting for some level of seismic design instead derive from the L’Aquila damage database, given that most of the L’Aquila municipalities were classified as seismic since 1915.

Table 1. Identification of RC building typologies Design Level Construction age

Number of stories

Reference damage database

Gravitational

1946-70 1971-80 1946-70 1971-80 1981-90 > 1990

1 1 1 1 1 1

2 2 2 2 2 2

3 3 3 3 3 3

≥4 ≥4 ≥4 ≥4 ≥4 ≥4

Irpinia (1980)

Seismic Pre-80

L’Aquila (2009)

Seismic Post-80

3. Seismic fragility assessment Sets of empirical fragility curves were derived for predefined RC building types (Table 1) by fitting an appropriate statistical model to observational damage data. In accordance with existing studies (e.g. Rota et al. 2008; Del Gaudio et al. 2017a; Rosti et al. 2021a, b) and consistently with the main features of the Italian national seismic risk platform (Borzi et al. 2021), the probability of reaching or exceeding predefined damage levels as a function of the selected seismic intensity measure, P ( ds ≥ DS i | PGA ), was approximated by the cumulative lognormal distribution: ( ≥ ! | ) = Φ / log( / "# ! ) 6 (1) where θ DSi is the median PGA value associated with damage level DS i and β is the logarithmic standard deviation. The multinomial model was used for approximating the subdivision of buildings in the different damage states, n ij , given the ground shaking (e.g. Rosti et al. 2021a, b): !$ ~ 9 $ ! !$ ! < = ! = $ > % !" %"# !&' (2) where N j indicates the total number of buildings at the j th PGA threshold and P( ds = DS i | PGA j ) is the conditional probability of occurrence of damage level DS i . Fragility curves were simultaneously fitted on empirical data points via maximum likelihood estimate approach (MLE), by enforcing a unique constant dispersion value ( β ) on damage state, number of stories, age of construction and design level.