UC San Diego

In the past decade highly ductile buckling-restrained braces (BRBs) have become popular in building seismic design. Design provisions and testing protocols considering far-field earthquakes have been developed by AISC. To extend the application to bridges, especially long-span bridges located near major seismic faults, research is lacking to support the development of bridge design provisions. Viscous dampers used to retrofit the Vincent Thomas Bridge (VTB) in Long Beach, California have been leaking due to traffic and ambient vibrations. With the potential of using BRBs to replace the dampers in a case study, VTB finite element simulations lead to the development of Near Fault loading protocols for prequalification test of BRBs. A testing program with six full-scale specimens is carried out to verify the ability of currently available BRBs to sustain a deformation demand about twice that commonly tested for building applications. This research identifies inconsistencies in current BRB testing and design conventions, so a rational methodology is proposed for both far- and near-fault ground motions. In addition to the commonly used mild steel BRBs, four stainless steel (SS) BRBs are also tested in this research; the latter reveals significant cyclic strain hardening. Going beyond the current practice of testing BRBs pseudo-statically, dynamic testing was also conducted. A 20% increase in force response due to the high strain rated effect should not be ignored for near- fault applications. Design and testing recommendations for bridge applications are proposed. For numerical simulation of BRB response, the commonly used bilinear model is shown to be insufficient. A modified Menegotto-Pinto (MP) material model, intended for OpenSees, that incorporates the following features is shown to provide excellent correlation to test results : (1) a larger strain hardening in compression than in tension, (2) appropriate isotropic hardening relation for SS BRB that includes the effect of cumulative ductility, (3) the instantaneous strain rate effect. Buckling in the gusset connection of one BRB specimen, and the incipient buckling of another, observed in this research confirms the potential shortcoming of current practice in designing gusset connections when stiffeners are not used. An alternate model that considers the rotational restraint of the gusset and initial out-of-straightness is proposed