Seismic performance and fragility functions of a 3D steel-concrete composite structure made of high-strength steel

Abstract

This paper provides insight into a probabilistic seismic demand analysis of a steel-concrete composite structure made of a novel type of high-strength steel moment resisting frame, to be used either in a seismic risk assessment or a fully probabilistic Performance-Based Earthquake Engineering (PBEE) framework. The application of the PBEE methodology with a full probabilistic character is able to rigorously evaluate the seismic risk to which a structure may be exposed, as well as to quantify economic losses, including both direct -repair, reconstruction costs, etc.- and indirect costs -downtimes, etc.-. In this respect, the knowledge of seismic fragility functions is paramount. Moreover, due to the dynamic complexity of the examined structure caused by irregularity in elevation and different lateral-force resisting systems in the two main directions -moment resisting frames (MRFs) and concrete shear walls- the seismic behaviour is not straightforward to foresee. Therefore, two separate 2D analyses along the building main directions may not suffice to identify the actual dynamic response and, consequently, a 3D comprehensive probabilistic seismic demand analysis was performed by taking into account the earthquake incident angle. In order to exploit the inherent overstrength of non-dissipative members, consistently with the capacity design philosophy, the structure, that is a representative example of a realistic office building, is characterised by a newly-conceived type of moment resisting frame made of high-strength steel circular columns filled of concrete and of mild steel beams. In this respect, a nonlinear 3D FE model was developed and calibrated on experimental tests performed on both beam-to-column and column-base joints that formed MRFs. A multiple incremental dynamic analysis (MIDA) was then performed with two groups of bespoke accelerograms characterised, on one hand, by large magnitude and large distance and, on the other hand, by near-source effects. The earthquake incidence angle was also considered and, to decrease the number of simulations, the accelerogram-incident angle pairs were selected by means of the Latin hypercube sampling (LHS) method. The relevant seismic analyses highlighted the need to include the incident angle to better characterise its dynamic behaviour. Hence, the seismic fragility functions were built both for damage and collapse limit states considering both the maximum interstorey drift ratio as engineering demand parameter and different intensity measures as well as the incident angle randomness. The results showed that peak ground displacement entails a more efficient probabilistic model because the dominant structural dynamic behaviour was governed by MRFs characterised by fairly long periods.