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

127

5

Italy, different vulnerability models are proposed in the official NRA (Dolce et al., 2020). For this application, the models proposed by Lagomarsino et al. 2021 and the one proposed by Borzi et al. 2021 were adopted for masonry and RC buildings, respectively, while for Slovenia the model previously mentioned (Dol š ek et al., 2020; Babi č et al., 2021) was adopted as official national vulnerability model. It is worth noting that the Italian fragility curves were defined for 5 vulnerability classes (A, B, C1, C2 and D, ranked according to decreasing vulnerability level). Therefore, a procedure to obtain fragility curves for typological classes needed to be applied preliminarily. The latter is described in (BORIS, 2022) and basically consists of the linear combination of the fragility curves for vulnerability classes using the relative percentages of the exposure model as linear combination coefficients. Moreover, the rules defined in the previous paragraph are used to convert the 4 damage states of the HAZUS scale, adopted in the Slovenian vulnerability model, into the five grades of the EMS98 scale, adopted in this study. The weights used for combining Italian and Slovenian fragility curves are calibrated based on differences and similarities between building typologies widespread in the cross-border areas and in both countries nationwide. To do this, in each area (i.e., Italy at national level, Slovenia at national level, Italian cross-border area, Slovenian cross-border area) the incidence of the main vulnerability factors (e.g., diffusion of irregular or regular layout masonry buildings, RC frames or walls) for each given building class is investigated interviewing some experts on the area under study. The analysis of prevalent building typologies' diffusion shows that the vulnerability of oldest masonry buildings (i.e., built before 1965) is quite similar in all study areas, with a predominance of regular layout masonry structures with flexible slabs. This similarity leads to define almost equal weights for both vulnerability models. Considering RC buildings, differences are more significant due to the different types of structure mainly adopted in the two countries (i.e., frames/walls). In Fig. 1 the combined curves for a masonry and a RC building class for the Italian cross-border area are shown. Time-based risk assessment at the municipal level is performed. The unconditional risk in 50 years is obtained by the convolution of seismic hazard at the site with vulnerability and exposure of the assets at risk (e.g. Dolce et al., 2021). Fig. 2 shows the expected percentage of damaged buildings (in different damage states) referring to the oldest masonry building typology for the two cities of Gorizia and Nova Gorica; results are compared considering analyses performed with the Italian and Slovenian vulnerability models as well as the Heuristic model. It is worth noting that the percentage of buildings with damage level from D1 to D3 is greater adopting the Italian model. This is probably due to differences observed for this typology in the fragility curves of the two countries: the Italian curves show higher fragility for the lowest values of seismic intensity, values that mostly affect the unconditional risk analysis, while the Slovenian curves show higher fragility for higher PGA values (Fig. 1). Adopting the heuristic approach, the number of damaged buildings for this typology is intermediate with respect to the ones expected adopting the national models. The reason is that, as already mentioned, the similarities of masonry buildings built before 1965 in the two cross

border areas lead to assigning similar weights to the vulnerability models. (a)

(b)

Fig. 2. Percentage of masonry buildings in the damage states of the EMS98 scale for the city of Gorizia (a) and Nova Gorica (b) obtained by adopting the Italian and Slovenian vulnerability models and the Heuristic model.