PSI - Issue 44

Sabatino Di Benedetto et al. / Procedia Structural Integrity 44 (2023) 1901–1908 Di Benedetto et al. / Structural Integrity Procedia 00 (2022) 000 – 000

1902

2

1. Introduction The traditional approaches to design steel structures are based on adopting full-strength joints to dissipate the seismic input energy through the plastic engagement of beam ends. Nevertheless, the brittle fractures experienced by many welded beam-to-column connections during Northridge and Kobe's seismic events pointed out the welds' low ductility (FEMA 351, 2000). In order to overcome such trouble, the Reduced Beam Section (RBS) or dog-bone solution was proposed (Moore and Malley, 1999; ANSI/AISC 358-16, 2016). It weakens the beam ends into areas located sufficiently far from the column, ensuring lower stress concentrations in welds and higher ductility. However, an alternative design approach based on partial-strength beam-to-column joints has been explored in the last decades. This strategy relies on the first identification and design of the weakest joint component, representing the element required to provide the joint's ductility and energy dissipation capacity. Instead, all the remaining joint components have an elastic behaviour. An example of such a strategy is the possibility of including friction dampers in steel connections, as proposed by Yang and Popov (1995) and Piluso (2018), achieving two targets: easy replaceability and high dissipation capacity. In particular, the second aim is attained by developing wide and stable hysteretic cycles that confine wearing only to the friction pads. In the 90s, the Sliding Hinge Joint (SHJ) was the first device conceived with this philosophy by Butterworth and Clifton (2000); however, in the last years, another solution, known as FREEDAM (FREE from DAMage) connection, has been widely studied at the University of Salerno, as discussed in the works proposed by Latour et al. (2015, 2018). In both cases, the beam bottom flanges are endowed with friction dampers which have three purposes: controlling the joint resistance by adequately regulating the tightening torque of pre-loadable high-strength bolts; settling the joint ductility by appropriately designing the length of the slotted holes; uncoupling the stiffness and the resistance of the connection thanks to the high initial stiffness provided by the friction shims, and the constant yielding resistance at the attainment of the slippage force. Another example of a partial-strength strategy is represented by the double split T-stub joint, which has been widely studied, as well demonstrated by the works proposed by Leon and Swanson (2000) and Piluso et al. (2001). A couple of T-stubs characterizes this connection typology to connect the beam flanges to the column. The T-stubs, if properly designed, can act as seismic dampers with levels of ductility and energy dissipation capacity that can be easily calibrated in the design phase. Additionally, the geometry of the T-stubs can be easily modified to provide additional energy dissipation and ductility by extending the basic concepts of ADAS devices to the T-stubs (Whittaker et al., 1989). For example, at the University of Salerno, many monotonic and cyclic tests have been performed by Latour and Rizzano (2009, 2011) on beam-to-column sub-assemblies endowed with X-shaped T-stubs which have been weakened to induce a uniform yielding of the flange plate and large energy dissipation capacity. Many research works have proposed traditional and innovative joint typologies' behaviour (Iannone et al., 2011). Nevertheless, although several tests have been carried out on sub-assemblies of beam-to-column joints, very few efforts have been devoted to experimentally defining how these connections can affect the structural response of seismically loaded MRFs. This paper compares the results of three campaigns embedded in an ongoing experimental program at the STRENGTH Laboratory of the University of Salerno. The planned testing activity deals with the seismic simulation of a large-scale two-storey steel building by considering five accelerograms applied through the pseudo-dynamic method. The program is performed by equipping the same MRFs, designed according to capacity design principles, with traditional (RBS) and innovative (FREEDAM and dissipative double split tee) beam-to-column connections to assess how they affect the local and global structural behaviour. Since the main results of the individual campaigns have already been discussed in detail by Di Benedetto et al. (2020, 2022a, 2022b), the purpose of this paper is to compare the global and local behaviour of the structure equipped with the three connection typologies mentioned above. 2. Design of the connections and the building mock-up The tested structure is shown in Fig. 1a; it is characterized by a one-bay two-storey steel structure composed of two MRFs that withstand the seismic input and two transversal bracings, which prevent accidental torsional effects. The longitudinal and transversal span lengths are 4.00 m and 2.00 m, respectively, while the interstorey height is