High performance steel (HPS) is rapidly gaining attention as a desirable material for highway bridge girders largely due to its superior toughness properties and high strength. However, the benefits of using steels with nominal yield strengths of 485 MPa (70 ksi) or greater is restricted by factors such as web stability, deflection, and fatigue design limits, which may govern the design and prevent the effective utilization of the material strength. Therefore, new and innovative bridge design concepts are needed to take better advantage of the enhanced properties of HPS. One design innovation that provides a means of optimizing bridge girders for high strength material utilizes I-girders with double web plates. The web is composed of two steel face plates connected internally by continuous longitudinal stiffening elements. The voids between the face plates may be grouted or ungrouted. The stiffeners permit thin webs to be used, while still allowing the material to reach stresses as high as the yield strength without buckling. In the case of grouted webs, composite behavior increases the out-of-plane stiffness of the web, although bond between the two materials may be unreliable. Nevertheless, it is shown that even in a debonded state, the presence of the grout enhances the buckling capacity of the face plates significantly. Using classical plate buckling theory, design criteria are proposed for bend buckling of the web face plates, considering both the g-routed and ungrouted cases. As a means of assessing the anticipated behavior of the plates, upper and lower bounds to the buckling strengths are established. In order to evaluate the ability of classical plate theory to predict the buckling of the face plates, tests were conducted on a series of web panels that simulate a portion of a girder web subjected to flexural compressive stresses. Two of the specimens were ungrouted, two were grouted with a cementitious material, and one was grouted with an epoxy grout. It was confirmed that the presence of grout increased the buckling capacity of the face plates and that the improved bond using epoxy grout served to delay buckling as well, although when the bond broke the failure was sudden. The experimentally determined buckling loads are used to validate the theory. (C) 2002 Published by Elsevier Science Ltd.
