In- and Out-of-Plane Bending in Steel Through-Truss Bridges

Abstract

Steel through-truss is a very common configuration for railroad and highway bridges. There is a large number of such spans in surface transportation infrastructure worldwide, especially in railway systems. In design and evaluation, these spans’ in- and out-of-plane bending have been either omitted or approximated. A fuller understanding of this effect will be able to assist in reliably designing and assessing these structures for longevity and/or increasing allowable service load to facilitate economic developments. A full scale load test is performed in this paper on five such bridges of the Canadian National Railway (CN) using train load. The main truss elements prone to in- and out-of-plane bending are identified and strain-gauged including the hanger (L1U1 or L’1U’1). Out-of-plane bending is seen to produce more significant flexural stress than in-plane bending. Three-dimensional (3D) numerical simulation is also verified by physical tests and covers other uninstrumented members of the tested spans. These results are used to evaluate the accuracy of a new and simplifying two-dimensional (2D) analysis method for the most significantly bent vertical hanger out-of-plane. The 2D method is shown to capture a significant portion of the bending but still underestimate flexural stress. An empirical and hybrid approach is therefore developed and recommended to address the inadequately accounted out-of-plane bending for routine practice of design and evaluation. It is needed when resources for detailed 3D analysis are not readily available, and/or when a quick and reliable method is needed, e.g., for verification or calibration of another method. These results are also useful for stress range estimation for fatigue analysis, although fatigue is not a concern to these bridges and is therefore not specifically addressed in this paper. CN has adopted the recommended method and the other research findings in load-rating their existing through-truss bridges.