Hogging Moment Behavior of Steel–UHTCC Composite Bridge Decks: Flexural Test and Response Prediction Approach

Due to the superior tensile properties and excellent durability exhibited by ultrahigh toughness cementitious composites (UHTCC), utilizing UHTCC to the hogging moment regions of continuous composite bridge decks (CBDs) is a new solution to the cracking issue. Addressing the limitations of welded stud connectors, epoxy-aggregate (EA) bond connections were employed in this study as a viable alternative for hogging moment regions. First, five groups of the four-point flexural tests were conducted to investigate the quantitative effect of different interfacial shear connections and longitudinal reinforcement ratios on the flexural performance of steel–UHTCC CBDs under hogging moments. Then, test results including failure modes, deflection responses, and crack resistance performance were discussed in detail. The flexural capacity of the EA-connected CBD specimen, with an aggregate particle size ranging from 5 to 10 mm, was comparable to that of stud-connected specimen at half the critical stud number for achieving full shear connection within a shear span. The comparison results of the crack width developments between steel–UHTCC and steel–NSC (normal strength concrete) CBDs further demonstrated the advantages of utilizing UHTCC in hogging moment regions to enhance the crack resistance performance while further reducing the reinforcement ratio and slab thickness. Subsequently, to enhance comprehension of the stress state and deformation mechanisms exhibited by steel–UHTCC CBDs under hogging moments, a novel iterative-based analytical model was proposed for complete flexural response prediction. The model demonstrated practicality with reduced complexity, while incorporating analysis of a pure bending section, material nonlinearity (including UHTCC strain-hardening characteristics and yielding behavior of structural steel), as well as constitutive slip behavior of various interfacial connections. The accuracy of the response predictions was validated against tests and established finite element models. The validation results demonstrated satisfactory accuracy in predicting both vertical deflection responses and strain developments of the steel beam, UHTCC slab, and reinforcements.