Collapse hybrid simulation for testing steel building columns subject to boundary condition changes

Abstract

AbstractThis paper develops a hybrid simulation to experimentally evaluate the ultimate behavior of steel columns subject to boundary condition changes and bending moment redistribution due to the progressive seismic collapse of building structures. The proposed experimental method combines sub‐structuring test techniques and refined finite element (FE) analysis methods in a hybrid scheme. The inelastic behavior of steel moment‐resisting‐frames (MRF) is computationally simulated while ground‐floor columns are physically tested online with the FE frame analysis. Boundary conditions at the interface node between the column test and the frame analysis are reproduced with the aid of three actuators that impose axial loads and bending deformations to the column test. Two square tubular steel columns are tested as part of a five‐storey five‐bay steel MRF designed to employ two different heights of the ground floor (i.e., 4.6 and 5.1 m, respectively). The hybrid simulation revealed that individual failures of the peripherical frame members significantly alter the distribution of the bending moment along the height of the column test, thus changing the boundary conditions at its top end. The imposed moment distribution was further influenced by local buckling initiated at the base of the column test and ground‐floor beams and could cause the top end of the upper‐storey columns of the frame to buckle. In the MRF with 4.6 m‐height ground‐floor, this resulted in a global collapse mechanism that involved the ground floor and the second storey. In the MRF with 5.1 m‐height ground‐floor, a weak‐storey collapse mechanism was formed within the ground floor. Compared to a corresponding isolated column tested as cantilever using conventional methods, the columns tested with the proposed hybrid simulation exhibited a more favorable inelastic behavior primarily due to the more realistic loads and changes of boundary conditions which are ignored in conventional test methods.