Horizontally Curved Steel I-Girder Bridge Response under Uniform Temperature Variation and Thermal Gradient in Deck and Exterior Girders

Horizontally curved steel I-girder bridges are common options for transportation efficiency in congested areas despite their complex construction and in-service responses, including under thermal loads. Prior thermal evaluation mostly focused on uniform temperature variations ( Δ T ), while the effect of thermal gradient was sparsely analyzed; uneven heating of exterior girders has not generally been considered for bridge design and analysis. 3D numerical simulations of a single-span bridge with a radius of curvature of 70 m (230 ft) were conducted in ABAQUS (using modeling approaches that were validated with existing field measurements) considering Δ T of up to + 50 ° C ( + 90 ° F ) or − 67 ° C ( − 120 ° F ), which represent the possible difference between environment and construction (nominal unstressed) temperatures. Three thermal gradient cases were analyzed: (1) a standard case along deck-girder depth for concrete heating specified by American Association of State Highway and Transportation Officials LRFD Bridge Design Specifications ( AASHTO 2024 ), and (2) two cases that considered direct sunlight heating bottom portion of either (outer or inner) exterior girders in addition to the standard thermal gradient profile. The bridges were analyzed with various support conditions, including guided bearings, elastomeric bearings, and integral abutments; idealized fixed bearings were modeled for comparison. Under thermal gradient, the peak bottom flange lateral bending stress was about 69 MPa (10 ksi) in the heated exterior girder when its exposure to direct sunlight was considered, while the peak stress was only 28 MPa (4 ksi) when only the standard thermal gradient for deck heating (consistent for all girders) was applied. Responses induced by exterior girder heating should be considered in bridge analysis for applicable site conditions. Under Δ T , the responses for bridges with idealized fixed bearings were considerably larger than the other bridges, which indicated that providing support flexibility was important for releasing temperature-induced stresses. The integral abutment bridge that was simulated with appropriate substructure behavior (including soil and pile stiffnesses) was observed with only slightly larger responses than the bridges with guided or elastomeric bearings.