BEHAVIOR OF A 2-SPAN CONTINUOUS PLATE GIRDER BRIDGE DESIGNED BY THE ALTERNATE LOAD-FACTOR METHOD

A large experimental test program to evaluate the behavior of a continuous plate-girder bridge with precast prestressed deck panels, designed according to Alternate Load Factor (Autostress) procedures, has been conducted at the FHWA Turner-Fairbank Highway Research Center in McLean, Virginia. The project was sponsored jointly by the American Iron and Steel Institute (AISI) and the Federal Highway Administration (FHWA). A 0.4 scale model of a two-span continuous plate-girder bridge was subjected to a series of tests at each of the three AASHTO load levels--Service Load, Overload, and Maximum Load. At the Service Load level, elastic lateral live-load distribution was studied. At the Overload and Maximum Load levels, the adequacy of the Alternate Load Factor Design limit-state criteria to satisfy related structural performance requirements was analyzed. Deck panel behavior was studied at all three load levels. The prototype bridge had two equal spans of 140 ft, an overall width of 48 ft with three girders spaced at 17 ft, and 10-in. thick composite modular precast deck panels pre-stressed in the transverse and longitudinal directions. The girders were designed using Alternate Load Factor Auto-stress) procedures. Alternate Load Factor Design (ALFD) is a limit-states design approach that more realistically approximates the actual behavior of continuous steel members at higher loads than present design procedures. ALFD recognizes and takes advantage of the ability of continuous steel members to adjust automatically for effects of controlled local yielding. An AASHTO guide specification presently permits the use of ALFD for braced compact sections. This model bridge study is part of a comprehensive research program to extend ALFD procedures to non-compact girders with slender webs. The model bridge consisted of three plate girders, with two 56-ft spans, transversely spaced at approximately 6 ft 9 in. and supporting 4-in. prestressed modular deck panels with approximately 2 ft 10 in. overhangs. Each plate girder was approximately 28-in. deep and was made composite with the deck panels using stud shear connectors. Preliminary test results from the model bridge study are presented. Experimentally determined elastic girder wheel-load distribution factors at the Service Load level are compared to factors computed from finite-element results, recently developed empirical formulas, and present AASHTO procedures. Limit-state criteria introduced to continuous-bridge design in ALFD, such as the formation of automoments and the shakedown phenomenon at Overload, are illustrated experimentally, along with the ability of these criteria to satisfy the structural performance requirements at Overload. The available reserve strength at Maximum Load is analyzed, as well as the adequacy of the mechanism analysis allowed in the ALFD procedures for strength prediction.