PSI - Issue 44

Maria Polese et al. / Procedia Structural Integrity 44 (2023) 123–130 Maria Polese et al. / Structural Integrity Procedia 00 (2022) 000–000

125

3

2.2. Exposure model

The exposure model was developed based on the available exposure data. Concerning buildings in Slovenia, building-by-building data are available, which include information on the predominant material of the load-bearing structure, the number of storeys, the year of construction and the net usable surface area. This building-specific information was provided by Real Estate Register (GRS, 2008) for the entire country. Concerning population, the average number of people per housing unit in each municipality was established baased on the detialed Central Population Register, which it is not publicly accessible. In Italy, exposure data are produced by ISTAT (National Institute of Statistics) which provides publicly accessible information on buildings, dwellings and population. The last available census database (ISTAT, 2011) includes buildings' information on construction material, number of storeys and construction period at the municipality level. Although in literature, several methodologies to improve exposure models integrating different sources of information are available (Gamba et al., 2014; Pagani et al., 2014; Polese et al., 2020; Tocchi et al., 2022), the typological-based building classification is adopted for assembling building inventory. Thus, twelve building typological classes were defined considering different materials of the load-bearing structures (masonry, reinforced concrete), construction periods (before 1965, 1965–1982, after 1981) and numbers of storeys (1–3, 4 or more). Buildings with the predominant structural material other than masonry and reinforced concrete were not included due to their low number. The construction periods were determined according to the evolution of the seismic codes in Slovenia. Moreover, for harmonization purposes, the exposure model was defined at the municipality level. Thus the Slovenian building-by-building data were grouped consistently with the data classification in Italy wherever possible. 2.3. Vulnerability model The harmonized vulnerability model for cross-border risk assessment was developed based on existing national vulnerability models in Italy (Zuccaro et al., 2021; Lagomarsino et al., 2021; Donà et al., 2021; Rosti et al., 2021a; Rosti et al., 2021b; Borzi et al., 2021) and Slovenia (Dol š ek et al., 2020; Babi č et al., 2021). The model was defined by a set of PGA-based fragility functions defined at the level of building classes. Typology-based building classification was selected, as it is the lowest common denominator of the classifications used in both countries (vulnerability- and typology-based classifications were primarily used in Italy and Slovenia, respectively). The EMS-98 damage scale (Grünthal, 1998) was selected, comprising six damage states, including the no-damage state (from D0 to D5). Therefore, the HAZUS damage scale (FEMA, 2015), used in Slovenia, needed to be converted to the EMS-98 damage scale. This conversion was performed based on the recommendations from Lagomarsino and Giovinazzi (2006). In addition, it was assumed that only a percentage of buildings reaching the complete damage state from the HAZUS scale reaches the destruction damage state from the EMS-98 scale, with the percentage determined based on the type of structural material and number of stories, consistently with FEMA (2015) and as described in (BORIS, 2022). Another important difference between the two national vulnerability models was found in the parameters of fragility curves. This discrepancy can be attributed to the differences in the methodologies for vulnerability assessment in Italy and Slovenia, and to the typological differences between buildings from the two countries allocated to the same building class. However, it is worth noting that the typological differences mainly depend on the typological characterization of the built environment at the national level. In cross-border areas, however, buildings located on different sides of the border may not be so different, due to a similar historical and cultural backgrounds. This issue was tackled by using a heuristic approach in which a weighted average of both national vulnerability models was defined for each side of the border and each building class. The weights were determined by (1) assuming that the methodologies for vulnerability assessment in both countries are equally credible and (2) considering that buildings on a given side of the cross-border area have similarities with buildings typical of both countries nationwide. These similarities were quantified separately for each building class based on the comparison of vulnerability factors characteristic of (1) buildings in Italy at the national level, (2) buildings in Slovenia at the national level, (3) buildings on the Italian side of the analyzed cross-border area, and (4) buildings on the Slovenian side of the analyzed cross border area.