Evaluation of brace fracture models in seismic analysis of concentrically braced frames

Brace fracture is included in seismic analysis and developing seismic provisions of concentrically braced frames. Reliability in predicting brace fracture is crucial in seismic performance evaluation of these structures. This paper focuses on evaluating current brace fracture models by comparing them with a comprehensive database of laboratory-tested steel braces, and inspecting their reliability by post-fracture seismic analyses. The comprehensive database includes tested square hollow structural sections that suffered low cycle fatigue-induced fractured under cyclic loading, a typical failure pattern expected under earthquake ground motions. The brace fracture models were first used to predict low cycle fatigue – induced brace fractures observed in laboratory tests and then in post-fracture seismic analysis of a two-story braced frame under a strong earthquake ground motion. The study finds that these fracture models are unable to predict fractures observed during laboratory tests and that the existing strain-related fracture models implemented in OpenSees fail to simulate cyclic response and brace fracture. When these fracture models were used for seismic analysis of a two-story frame, the seismic responses vary dramatically from 5% to 10% story drift ratio, depending on the employed fracture models. •Existing fracture models perform poorly in predicting tested seismic braces.•Seismic evaluation of braced frames based on existing fracture models are not reliable.•Seismic provisions based on these seismic evaluations need to be reexamined.