Pseudo‐dynamic testing of a full‐scale two‐storey steel building with RBS connections

Abstract

AbstractAfter the earthquakes of Kobe (1995) and Northridge (1994) many research efforts have focused the attention on the seismic behaviour of partial strength beam‐to‐column joints because, in these seismic events, brittle fractures of welds of full‐strength joints belonging to MRFs have been observed. One of the most popular solutions developed in the aftermath to overcome the drawbacks of the typical welding joint details consisted in the adoption of the so‐called Reduced Beam Section (RBS) strategy. With this strategy beam ends are weakened into areas located sufficiently far from the column in order to reduce the stress concentration in the welds. Nevertheless, while many experimental efforts have been devoted to this connection typology, the available experimental results on full‐scale structures are still quite limited.Within this framework, aiming to provide a contribution to fill this knowledge gap, an experimental program concerning the pseudo‐dynamic testing of a full‐scale two‐storey steel building equipped with five different connection typologies is currently undergoing at the University of Salerno. In this paper the results of the first experimental campaign, dealing with MRFs with RBS connections, are presented. The mock‐up has been subjected to a sequence of seven accelerograms and the results have confirmed that, if properly designed, RBS connections provide large energy dissipation capacity and plastic rotation supply. Nevertheless, the failure of welds due to damage accumulation can be still possible, even adopting full penetration butt joint welding details conforming to Eurocode 3 and Eurocode 8 provisions.The experimental results have been complemented with numerical analyses developed with the software SeismoStruct. The comparisons between the numerical simulations and the experimental results show that the applied mixed distributed and concentrated plasticity approach, based on the adoption of phenomenological spring models, can capture accurately the global response of the structure, but it predicts with lower accuracy the local response parameters.