Behavior of stiffened compression flanges of trapezoidal box girder bridges

Steel trapezoidal box girder systems are seeing increased use based on their high torsional stiffness and aesthetic appearance. This dissertation research is part of a study sponsored by the Texas Department of Transportation investigating various aspects of design and behavior of trapezoidal box girder bridge systems. The design equations contained in the bridge specifications of the American Association of State Highway and Transportation Officials (AASHTO) for the steel compression flange of box girders are different in form than other bracing-type equations and design equations contained in other box girder specifications. This dissertation research investigated the basis and validity of the AASHTO design equations and the general behavior of stiffened compression flanges of steel box girder bridges.; The basis of the design equations in the AASHTO specifications was researched and documented. The validity of the equations was investigated using eigenvalue finite element analyses results. A new formulation of the plate buckling coefficient equation, which corrects an error in the current formulation, is proposed. During the finite element study, it was noted that alternate stiffener designs, utilizing torsionally stiff longitudinal stiffeners, could result in significant increases of the elastic buckling strength of stiffened plates. An experimental program was undertaken to test the performance of plates stiffened with torsionally weak stiffeners, typical of current stiffener designs, as well as torsionally stiff longitudinal stiffeners. Both steel and concrete longitudinal stiffeners were tested.; The results of the experimental program revealed that the basis of the AASHTO design equations did not predict the experimental behavior of the stiffened plates tested. The assumption that longitudinal stiffeners of sufficient size will form straight nodal lines, about which the plate will buckle, was not supported by the results of the experimental tests. Many factors, including initial imperfections, stiffener eccentricity, overall girder bending, and second order effects have been shown to contribute to producing stiffener displacement. The AASHTO design equations result in conservative estimates of the capacity of slender stiffened plates but may indicate unconservative capacities for some stockier plate geometries. A modified design approach, which includes the influence of factors such as initial imperfections and stiffener eccentricity, should be considered for incorporation into the AASHTO specifications.