Abstract
In precast bridge columns for accelerated bridge construction (ABC), grouted duct connections must provide reliable anchorage while accommodating reinforcement deformation near the base. Previously reported full-scale pullout and CIP/PGD-S tests, together with newly reported PGD-D/PGS tests, were examined using a local bond–slip-based finite-anchorage model for ultra-high performance concrete (UHPC)-grouted ducts. Bonded and unbonded lengths were varied at a fixed total bar connection length of 14db. Increasing recess depth from 10 to 40 mm raised initial stiffness by 12.6% but peak resistance by 4.35%. Among B8-U6 (B8), B10-U4 (B10), and B12-U2 (B12), peak resistance and cumulative hysteretic energy differed by at most 1.0% and 2.8%, respectively. At 6.38% drift, unbonded elongation contributed 53.2% and 32.9% of connected-bar cyclic deformation in B10 and B12, respectively. Relative to B10, the zero-unbonded-length B14-U0 (B14) model developed 0.6% lower peak resistance and dissipated 7.1% less cumulative energy. At 6.38% drift, B10 exhibited 5.2% lower spatially averaged and 14.0% lower peak reinforcement strains above the joint than B14; the coupled length changes preclude isolating the unbonded-segment effect. Reconstructed maximum local bond stresses remained below 38% of the adopted peak bond stress. Thus, length allocation affects local force transfer and reinforcement strain demand more strongly than global lateral resistance.