LAUSR

Abstract Reliably predicting brace fracture is crucial in seismic evaluation of post-fracture performance of concentrically braced frames. This paper evaluates existing fracture models, proposes a new fracture initiation model based on a large amount of experimental data, and compares all models with the test results. The existing ductility-related empirical models and strain-related fiber models are evaluated by experimental tests. A comprehensive database of previously tested square hollow structural sections from different research programs is established to conduct a regression analysis leading to development of a fracture initiation model. The study concludes that the existing fracture models are unable to predict fracture events observed during physical tests. Furthermore, the existing strain-related fracture models, accessible to all users in OpenSees, fail to simulate the cyclic response of braces beyond initial inelastic deformation. The fracture initiation model proposed by this study depends on the reliable measure of ductility, and it has consistently predicted fracture initiations observed during the tests. The fracturing process concept, along with the proposed fracture initiation model, is able to simulate the cyclic response of the braces, whether they are tested as single brace members or in actual braced frames, throughout inelastic deformation, fracture initiation, fracture propagation, and terminal fracture stages, providing a much-improved tool for numerical simulation of the post-fracture response of concentrically braced frames.