Abstract
Recent earthquakes have induced significant damage on reinforced concrete (RC) school buildings due to seismically deficient details, which can cause bonding failure between steel reinforcements and surrounding concrete due to low confinements. To mitigate the failure mechanism, a proper modeling method to capture the bond-slip effects is needed. This study aims to investigate hysteresis behavior of finite element (FE) models with three bonding modeling approaches (perfect, linear-elastic and nonlinear-inelastic bonding models) for a seismically vulnerable RC school building frame. The models were developed through three-step processes: (1) column model; (2) beam-column joint model; and (3) frame model. To quantify the simulation variation of bond-slip effects for key parameters of hysteresis curves (effective stiffness, maximum strength, and energy dissipation), the simulated responses were compared to the experimental results measured from quasi-static cyclic loading tests in each modeling process. While the perfect bonding model used by structural engineers overestimated all key parameters (e.g., 53.7% in energy dissipation) of the hysteresis curves, the nonlinear-inelastic bonding models reproduced the most accurate hysteresis behavior (less than 5.0%). Based on this investigation, uses of improper bonding models can exaggerate the seismic performance of the seismically vulnerable RC frames.