GNGTS 2024 - Atti del 42° Convegno Nazionale

Session 2.2 GNGTS 2024 Average spectral acceleraCon is defined as the geometric mean of spectral acceleraCons over a range of periods (Eads et al., 2015): The period range is selected to including the effects of higher modes and period elongaCons due to damage accumulaCon. Five periods are selected, including modes with a mass parCcipaCon greater than 10%: Where is the mean of the second mode periods in the two orthogonal direcCons. Results The findings encompass accelerograms, intensity parameter distribuCons, regressions, and fragility curves derived from four disCnct configuraCons. Various crustal and local models, yielding four configuraCons: CR1-L1, CR2-L1, CR1-L2, and CR2-L2. While the seismic source remains constant, 100 variaCons in rupture processes are considered to capture variability. Accelerograms resulCng from the combined models exhibit significant variaCons in shape and amplitude. DistribuCons of ground moCon parameters, including Arias Intensity, Significant DuraCon, PGA, PGV, and Response Spectra in AcceleraCon, underscore the impact of the chosen configuraCons. The Arias Intensity distribuCons highlight the influence of crustal model CR2 on signal energy, while the coupling of local model L2 with CR1 increases Arias. Signal duraCon, crucial for structures with cyclic strength degradaCon, shows greater dispersion in configuraCons coupling CR2 and L2. Notably, PGA, a historically pivotal intensity measure, displays no clear increase when transiConing from fast to slow B soil, emphasizing the intricate interplay between crustal and local models. The fragility curves, represenCng the structure's vulnerability, reveal intriguing insights. For the sake of brevity only results for PGA are reported, hence for regressions of parameter MIDR (Maximum Interstorey Drip RaCo) as a funcCon of PGA GMRotD50 (median value of the geometric mean of the two horizontal components rotated through all nonredundant period- dependent angles (Boore et al., 2006)) in the bi-logarithmic plane. Regressions associated with crustal model CR1 appear less steep due to higher standard deviaCon, indicaCng greater uncertainty. Conversely, the reliability of esCmates is higher for crustal model CR2, influenced by a more consistent distribuCon of intensity measures (Fig. 2). S a , avg ( T i ) = [ n ∏ i =1 S a ( T i ) ] 1/ n T i = [ T 2 m ž , min [ ( T 2 m ž + T 1 m ž )/2,ž1.5 T 2 m ž ] , T 1 m ž , 1.5 T 1 m ž , 2 T 1 m ž ž ]