PSI - Issue 78

Stefano Bracchi et al. / Procedia Structural Integrity 78 (2026) 745–752

751

that, since LPM parameters have been evaluated only for IML 2-4-6, such limited calibration is used also for different IML. In particular, LPMs calibrated for IML 2 are adopted for levels 1 and 2, the ones for IML 4 for levels 3, 4, 5 and, finally, the ones calibrated for IML6 are extended to higher IMLs. Hence, the seismic demand for each stripe is then compared to the capacity of the corresponding model. Fig. 5 shows the exceedance frequencies of the two considered performance conditions, for the different considered return periods of the seismic action. It should be noticed that, for the three stripes for which LPM were calibrated (IML 2, 4, 6, corresponding to a return period T r equal to 50, 250, 1000 years), the same frequencies of exceedance are obtained considering or neglecting SFSI at both UPD and GC performance states; the only difference is in UPD at IML 4 ( T r of 250 years), for which a larger frequency of exceedance is evident when considering SFSI. On the contrary, some differences are present in other stripes. Generally, a trend of decrease of frequency of exceedance when adding SFSI is evident at both UPD and GC performance states. However, these differences appear to be slight in all the cases.

(a)

(b)

(c)

Fig. 5. Exceedance frequencies of the thresholds corresponding to the considered performance conditions, for the different return periods of the seismic action, without (a) and with (b) SFSI; fragility curves with and without SFSI (c).

Fragility curves are finally derived using the software R2R-EU (Baraschino et al. 2020). Fig. 5 shows the fragility curves, obtained using the maximum likelihood approach, together with the comparison with the fixed-base configuration. The consideration of SFSI through LPM allows a slight reduction of vulnerability at larger values of the intensity measure; this reduction is more evident at the GC performance state. The beneficial effect would have been possibly larger if the calibration of LPM parameters was performed also for the highest IML. On the contrary, at lower values of the seismic action, a slightly larger vulnerability of the configuration with SFSI is evident. 5. Conclusions This paper investigates the effect of SFSI on the seismic response of URM buildings. First, a strategy to model SFSI based on LPM is developed. A case study building, representative of an existing structure in the city of Naples, is selected; for the considered location, a detailed soil characterization is available. LPMs for the foundation of each masonry panel are then calibrated and applied at the base of the numerical model of the building, derived adopting the equivalent-frame strategy implemented in the software TREMURI. The lateral capacity of the entire building is assessed by pushover analyses, allowing the estimation of relevant thresholds of a properly selected EDP, referring to two different performance states, related to global collapse and usability of the building. Seismic demand is then obtained by means of multi-stripe nonlinear dynamic analyses. Finally, the comparison between demand and capacity is performed, allowing to understand the effect of considering SFSI on the seismic performance of the case study. Results show that SFSI is effective in reducing the vulnerability of the considered building, although the effect is limited, at least for the structure and site considered. A larger effect is expected when SFSI is modelled through more refined macroelement models, able to better capture the complex phenomenon of SFSI.