Validated finite element techniques for quasi-static cyclic response analyses of braced frames at sub-member scales

Abstract

In this study, a numerically robust finite element procedure is described, which is based on explicit time-stepping, for high-fidelity simulations of inelastic and post-buckling cyclic responses of braced frame systems. The use of an explicit time-stepping method with properly chosen increments permits accurate results while avoiding (implicit) equilibrium iterations throughout the entire loading history, during which multiple yielding and buckling events occur. A number of essential techniques for properly calibrating the discrete models and to constrain their responses in order to obtain quasi-static outcomes are provided. The procedure is globally and locally validated (verified) using experimental data (implicit numerical simulations) from three types of specimens—namely, individual braces, and single and multi-story braced frame systems with diagonal and X-brace arrangements—under both monotonic and cyclic loading protocols. Results from these validation and verification studies indicate that the proposed simulation methodology can accurately capture sub-member (i.e., plastic hinges), member, and system behavior very accurately; and thus, it can be confidently used—e.g., as a virtual laboratory—to predict the responses of braced frames with configurations and dimensions other than those tested, and to seek optimum designs beyond those offered by basic guidelines.