A Pre-Tensioned Bridge Bent System for Accelerated Bridge Construction

Abstract

Nearly all bridge bents (intermediate supports) are constructed of cast-in-place reinforced concrete. Such bridges have served the nation well in the past, but to meet current design expectations, they need to be improved in three areas: 1) speed of construction, 2) seismic resiliency, and 3) durability. Building on previous research at the University of Washington (Hieber et al. 2005, Wacker et al. 2005, Pang et al. 2010, and Haraldsson et al. 2013), a new pre-tensioned bent system has been developed to address these needs. The system consists of 1) precast technology that reduces construction time, 2) unbonded pre-tensioning that minimizes post-earthquake displacements, and 3) high-performance materials that extend the bridge’s life-span. Davis et al. (2012) tested a version of the system using conventional concrete in the plastic hinge regions. They found that pre-tensioning improved the system’s re-centering capabilities but led to earlier bar buckling and bar fracture than in previously tested RC columns. In order to delay bar buckling and bar fracture, the system was modified to include Hybrid Fiber Reinforced Concrete (HyFRC) in the plastic hinge regions. This composite concrete has been shown to exhibit superior durability and cracking resistance (Ostertag et al. 2007). The effect of the HyFRC on the pre-tensioned bent system was investigated both with quasistatic and dynamic tests. The quasi-static tests showed that using HyFRC in the plastic hinge region increased column ductility; in all cases the column maintained more than 80% of its strength up to a drift ratio of 10%. The HyFRC also delayed spalling of the concrete, but it did not significantly increase the drift ratios at the onset of bar buckling and bar fracture. The shaketable tests of a cantilever column, which was designed to re-center up to a drift ratio of 3.0%, showed that the new system had lower expected residual drifts than columns constructed with conventional cast-in place methods. The pre-tensioned column had a residual drift of 0.23% after experiencing a peak drift ratio of 5.5%. In contrast, the companion reference column, constructed using cast-in-place technology, had a residual drift ratio of 0.83% after experiencing a peak drift ratio of 5.7%. A numerical model in OpenSees was developed and calibrated with a set of 34 RC quasistatic, cyclic tests. This model was calibrated using a concrete constitutive model that takes into account concrete early reloading, developed as part of this research, and used commonly used steel constitutive models; Giuffre-Menegotto-Pinto’s (Steel02) model, and Moehle and Kunnath’s (ReinforcingSteel) model. The simulations showed improved accuracy in comparison to previous research (e.g., Berry and Eberhard 2007), and showed that the response of the system was affected more by the chosen steel model than by the concrete model. The results of these simulations were used to make predictions of the response of five columns tested on the UC Berkeley shake table. These simulations showed that models built using the proposed strategy predict peak displacements quite accurately, especially at the yield and design level, but do not accurately capture residual displacements.