Comparison of Experimental and Analytical Load-Rating Methodologies for a Pony-Truss Bridge

Vehicle and environmental loading, combined with the effects of deicing salts and inadequate maintenance, have resulted in a large number of structurally deficient bridges in the United States. Because of the limited resources of county and district bridge owners, aging structures must be evaluated and rated to facilitate suitable bridge repair, replacement, and posting strategies. A 1937 riveted pony-truss bridge in Pennsylvania was studied to evaluate the effect of increasingly refining rating parameter inputs to the analysis of the bridge. The study considered results from experimental field testing of the bridge, a detailed sectionloss inspection, and a refined analytical modeling using STAAD III, a commercially available software package, to increase the accuracy of the rating analysis. The research team evaluated the relative effect of using (a) allowable stress versus load-factor design methodologies; (b) experimentally derived versus code-specified impact and distribution factors; (c) experimental stress measurements versus manual- and computer-stress calculations; and (d) member fixity and composite action for both inventory and operating rating with an AASHTO HS20-44 and ML-80 rating vehicle. Results for selected truss members, floor beams, and stringers are presented. The investigators found that the rating increased with increasingly refined parameter input, with the rating factor increasing as much as 60 percent over a standard manual rating analysis when an experimentally based rating analysis was used. The best agreement with experimental results was obtained by using reduced deck stiffness to model deck deterioration, fully composite action between stringers and the deck, pinned stringer-floor beam connections, and fixed floor beam-to-truss connections. The study procedures can be applied to other similar bridge structures.