PSI - Issue 44

Greta Agata Venneri et al. / Procedia Structural Integrity 44 (2023) 291–298 Greta Agata Venneri et al. / Structural Integrity Procedia 00 (2022) 000–000

297

7

of the Case study 1 has lower probabilities of exceeding the limit states, except for the Immediate Occupancy (IO) performance level, where the curve of the semi-rigid model is higher for PGA values higher than 0.45 g. In addition, the Collapse Prevention (CP) performance level is not attained by any model, at least for the considered ground motions. On the contrary, the fragility curves in Fig. 4(b), Figs. 5(a) and 5(b) show that, for both buildings, the probability of exceeding the limit states is higher in the model with semi-rigid joints. In particular, the two case studies exhibit a similar behavior in the H2 direction, where the rigid model does not even reach the LS and CP performance levels (Fig. 4(b) and Fig. 5(b)). Moreover, for the Case study 2, the probability of occurrence of these limit states increases in the H1 direction, for both the rigid frame and the semi-rigid frame, even if this probability is lower in the rigid model (Fig. 5(a)). 5. Conclusions In this paper, the seismic response of two real steel multi-storey MRFs with semi-rigid joints was investigated. To this purpose two benchmark buildings, named “Case study 1” and “Case study 2”, were considered. Finite element models of the as-built frames, with semi-rigid joints, and of the theoretical model, with rigid joints, were set up, and their seismic performances were evaluated through nonlinear time history analyses and compared. The fragility curves constructed for three performance levels (IO, LS, CP) were then plotted. The obtained results showed that the presence of semi-rigid joints influences the global seismic performances of the analyzed multi-storey steel buildings. In particular, by comparing the fragility curves of the rigid models with those of the semi-rigid models, the following points emerge: • With regard to Case study 1, in the H1 direction the fully rigid frame results more vulnerable than the semi rigid frame, except for PGA values higher than 0.45 g at the IO limit state, for which an opposite trend is observed. While in the H2 direction, the semi-rigid frame results more vulnerable for all the considered limit states. • With regard to Case study 2, in both seismic directions the probability of exceeding the limit states is higher in the model with semi-rigid joints. • Both buildings exhibit good behavior at the LS and CP performance levels in the H2 direction: the semi rigid models reach this limit states with moderate probabilities, while the rigid models do not even reach them. In the H1 direction, for the Case study 2 the probability of LS and CP occurrence increases with the increasing of the PGA values, while for Case study 1 the CP performance level is not even reached by any models. • Therefore, considering the actual stiffness and strength of joints, which are found to be semi-rigid and partial-strength through the application of the component method, the buildings could be more vulnerable and, thus, less safe than expected in the design process. • However, only two case studies do not allow general conclusions to be drawn. Therefore, the work would need further investigation. Acknowledgements This research was developed with the support of a RELUIS-DPC 2019–2021 project, funded by the Italian Department for Civil Protection. The design data of the benchmark structures were provided by Eng. Gennaro Di Lauro from AIRES INGEGNERIA S.r.l. in Caserta (Italy). References Adam, C., Tsantaki, S., Ibarra, L. F., and Kampenhuber, D., 2014. Record-To-Record Variability of the Collapse Capacity of Multi-Story Frame Structures Vulnerable To P-Delta. In Second European Conference on Earthquake Engineering and Seismology . Instabul. Awkar, J. C., and Lui, E. M., 1999. Seismic Analysis and Response of Multistory Semirigid Frames. Engineering Structures 21 (5): 425–41. https://doi.org/10.1016/S0141-0296(97)00210-1. Baker, J.W., 2015. Efficient Analytical Fragility Function Fitting Using Dynamic Structural Analysis. Earthquake Spectra 31 (1): 579–99.