PSI - Issue 78

Lorenzo Hofer et al. / Procedia Structural Integrity 78 (2026) 1927–1934 Author name / Structural Integrity Procedia 00 (2025) 000–000 7 branches weights. The effect of the two main branches (i.e. the GMPE and the non-linearity modelling strategy) is clearly identifiable, since dots tend to produce four graduals steps on the 3D representation of the reliability index matrix. Results shows that higher uncertainty levels are associated with higher damage states, with � �,��� varying between 5.7% and 6.7 % respectively for ds 1 and ds 4 . 1933

Fig. 8. Reliability index for each damage state, reliability index distribution and variation of � �,� and � �,� with one deterministic parameter (Hofer et al. 2023). Then with the aim of identify which parameters mostly contribute to the reliability index uncertainty, numerical simulations have been rerun by imposing each time a different parameter as deterministic (i.e. removing its uncertainty source). Fig 8 also show respectively the % variation of � �,��� and � �,��� with the assumption in turn of one deterministic parameter. The GMPE seems to be the most influencing parameter, since when it is assumed as deterministic, a significant reduction of � �,��� can be observed. Furthermore, the reduction is not constant for the four damage states, by it is higher for the ds 4 (about -33 %), and lower for ds 1 (-14 %). The second parameters mostly affecting � �,��� is represented by non-linearity modelling strategy: indeed, when the uncertainty arising from it is not accounted for, the � �,��� decreases of about 10 %. In this case, the higher reduction has been obtained for the � (- 15 %), while the lower reduction for � (-8.7 %). The third parameter that most influences the yearly reliability index dispersion � �,��� is the number of the accelerograms adopted for the execution of NLTHAs (reduction of -9 %, with little variation within the four damage states). The uncertainties related to the remaining investigated parameters (i.e. � , � , � , � and � ) seem to have a negligible impact on � �,��� . Finally, considering the GMPE and the non-linearity modelling strategy as deterministic, the reduction of � �,��� reduction is about -56 % for � and – 44 % for � . When the number of records ( A ) and the GMPE are assumed as deterministic, a reduction of – 47 % is obtained for � , while a reduction of – 31 % is obtained for � . Further considerations and a sensitivity analysis on the logic-tree branches weights are available in Hofer et al. 2023. 6. Result discussion This work aims to explore the influence of key uncertainty sources commonly encountered in the seismic reliability assessment of structures on the final reliability dispersion. To achieve this, a general framework is proposed to systematically account for all relevant uncertainties, and then is applied to a representative case study. In the case study, eight different uncertainty sources were considered and integrated using a logic tree approach, resulting in a total of 2916 branches. The highest dispersion of the reliability index (6.73%) was observed for the Complete Damage State ds 4 , while the lowest dispersion (5.74%) occurred in the Slight Damage State ds 1 . A detailed sensitivity analysis was then conducted to rank the influence of each uncertainty source on the overall dispersion of the reliability index. The results indicate that the uncertainty associated to the choice of Ground Motion Prediction Equation (GMPE) has the most significant impact, followed by the nonlinear modelling technique and the number of accelerometric records used in structural analysis. Collectively, these three factors account for nearly 80% of the total reliability index dispersion. Furthermore, this study showed that uncertainties associated with seismic hazard calculations are