Title:
Collapse Assessment of Reinforced Concrete Building Columns through Multi-Axis Hybrid Simulation
Author(s):
M. J. Hashemi, H.-H. Tsang, Y. Al-Ogaidi, J. L. Wilson, and R. Al-Mahaidi
Publication:
Structural Journal
Volume:
114
Issue:
2
Appears on pages(s):
437-449
Keywords:
axial load variation; collapse assessment; hybrid simulation; limited-ductility reinforced concrete buildings; ratcheting behavior
DOI:
10.14359/51689438
Date:
3/1/2017
Abstract:
One of the major challenges in collapse assessment of reinforced concrete (RC) structures has been the lack of realistic data obtained from reliable experimental loading protocols that are capable of accurately quantifying the reserve capacity of RC structures beyond the design level to the state of complete collapse. Until now, quasi-static (QS) symmetrically cyclic or monotonic tests with constant axial load have been commonly used, which are not adequate to accurately capture the actual response of a collapsing RC structure in real earthquake events. Hybrid simulation (HS) can be considered an attractive alternative to realistically simulate more complex boundary conditions and improve response prediction of a structure from elastic range to collapse. This paper presents a comparative experimental study on two identical, large-scale, limited-ductility RC columns that are tested to collapse through QS and HS, respectively. The RC columns serve as the first-story corner-column of a half-scale symmetrical five-byfive-bay five-story RC ordinary moment frame building structure. A state-of-the-art facility, referred to as a multi-axis substructure testing (MAST) system, is used that is capable of controlling all six-degrees-of-freedom (6-DOF) boundary conditions in mixed load and deformation modes. The load protocol in the QS test includes constant axial load combined with bidirectional lateral deformation reversals, while in the HS, more realistic boundary effects including fluctuation in axial load and the ratcheting behavior (that is, asymmetrical lateral deformation prior to collapse) are simulated. The hysteretic response behaviors obtained from the QS and HS tests are then used for calibrating the analytical models employed in a comparative collapse risk assessment. The results show that the improved interface boundary effects lead to significant changes in hysteretic response and the calibration parameters and, as a result, estimating the probability of collapse. This highlights that the credibility of collapse assessment results relies to a great extent on the application of correct boundary interface on RC columns.