Experimental and Numerical Analysis on the Core Lateral Thrust in Bolted BRBs

The paper aims at presenting numerical and experimental results on the lateral thrust exerted by the inner core of a buckling restraining brace (BRB) when, after buckling in compression, it arrives into contact with the external restraining case. The BRB consisted of a plate core restrained by bolted U- shaped member, thus allowing an easy assembly of the BRB, as well as the inspection and the substitution of its core after a seismic event [1, 2]. The results of six tests on BRBs having a plate core with a cross section of 5x50 mm2 and a dissipative length of 560 mm are shown. The gap dimension was varied between 0.25 and 0.70 mm. Cyclic displacements of increasing amplitude were applied to the core up to a steel strain of 2%, adopting the loading history prescribed by AISC standards [3]. The instrumented bolts connecting the restraining elements and a thin tactile pressure sensor placed within the gap allowed to steadily monitor both the lateral thrust and the buckling shape of the plate core during the experiment. The results showed that the lateral thrust increased linearly with the gap dimension, thus confirming the trend provided by the analytical formulation proposed by Genna and Gelfi [4]. As reported in Table 1, for a maximum axial force of about 110 kN, the value of the transverse thrust ranged from 36.5 kN in the specimen 1 with a gap of 0.25 mm to 86.6 kN in the specimen 3 with a gap of 0.70 mm. Furthermore, the specimens 6, without stiffening plates in the web of the restraining U- profiles, showed the significant role of the local transverse deformation on the value of the lateral thrust, which was twice the thrust of the specimen 1, characterized by the same gap of 0.25 mm but with a stiffened case. The experimental results allowed to validate a non-linear 3D Finite Element model performed with the code ABAQUS [5]. The numerical analyses accurately predicted the cyclic behaviour of the tested BRBs in term of axial load, buckling shape of the core and lateral thrust action. The difference between the measured thrust and the calculated thrust at the end of the loading history (steel strain c=2%) ranges between -8% to +17%, while the 2D plane stress model proposed in [6] overestimated the lateral thrust of +85%, owing to the inability of the 2D model to describe the local transverse deformations of the case.