Steel plate girder bridges make use of traditional cross-frame diaphragms to stabilize the compression flange of girders. These braces are required during construction, especially during deck placement, to prevent lateral torsional buckling of bridge girders. Girder buckling capacity is a function of cross-frame diaphragm spacing as well as strength and stiffness. Recent developments in bridge design may cause the governing girder limit state to shift from one of strength to one of stability. These developments include the elimination of in-plan bracing, composite girders, High Performance Steels, and phased deck replacements. In addition, the American Association of State Highway and Transportation Officials (AASHTO) has changed its code requirement for cross-frame diaphragm spacing in the 1998 AASHTO LRFD Bridge Design Specifications. The requirement for 25-foot maximum brace spacing has been removed. The current requirement is for a "rational analysis" to determine cross-frame diaphragm spacing. Explanations of the problems these changes cause in design are discussed. A case study is presented of a bridge that suffered construction difficulties during deck placement. This investigation found that the cross-frame diaphragms were not stiff enough to brace the plate girders during the deck placement. Suggestions are given as to an efficient, economical design and spacing for cross-frame diaphragms on plate girder bridges.

Cross-frame and diaphragm systems are important structural elements in steel I-girder bridges. These members enhance the lateral-torsional buckling resistance in straight girder systems by reducing the unbraced length. The critical stage for the braces from a stability perspective often occurs during construction of the concrete bridge deck when the non-composite steel girders must resist the entire construction load. The braces not only enhance the lateral-torsional buckling resistance of the girders, but are also necessary to resist the torsion applied to the girders due to the deck overhang construction, and distribute lateral loads across the structure from sources such as wind. In horizontally curved bridges, the braces are primary structural elements in the superstructure and engage the girders across the width of the bridge to behave as a structural system to resist the torsion that develops as a result of the curved geometry. For straight girder systems, the specifications of the American Association of State Highway and Transportation Officials (AASHTO) have generally provided little guidance in the sizing of the braces other than recommended connection plate (web stiffener) details or slenderness limits. While there have been many advances in recent years towards improving the understanding of the behavior of cross-frame systems, there has not been sufficient research carried out on the proper loading conditions for assessing the fatigue performance of cross-frames. The work outlined in this thesis is part of a larger study on the behavior of cross-frames in steel bridge system. The study is funded by the National Cooperative Highway Research Program (NCHRP 12-113). The fundamental goals of the research investigation are to produce methodologies and design guidelines for the following: evaluation of fatigue design stresses in cross-frames in straight and horizontally curved steel I-girder bridges; calculation of minimum cross-frame strength and stiffness requirements for stability bracing of I-girders during construction and in-service; development of improved methods to account for the influence of end connection details on cross-frame stiffness that extend beyond and improve upon the suggested guidance currently provided in Article C4.6.3.3.4 of the AASHTO LRFD Bridge Design Specifications. This work includes field monitoring and parametric FEA studies. The field studies are focused on three bridges: 1) a straight bridge with normal supports, 2) a straight bridge with skewed supports, and 3) a horizontally curved bridge. The field studies include rainflow monitoring of fatigue induced stresses in select cross-frames and the girders for a period of approximately 1 month as well as live load tests using trucks of known weights. This thesis focuses on some of the background studies, a survey of bridge owners around the United States, as well as the instrumentation and live load tests on the straight bridge with normal supports

Skewed bridges in Kansas are often designed such that the cross-frames are carried parallel to the skew angle up to 40°, while many other states place cross-frames perpendicular to the girder for skew angles greater than 20°. Skewed-parallel cross-frames are longer and may require different connections than cross-frames oriented perpendicular to the girder. Therefore, it is important to determine whether the cross-frames and their corresponding connecting elements placed in a parallel-to-skew configuration are sufficiently designed to resist lateral torsional buckling demands using current KDOT practices. The authors have performed a study to investigate the effect of cross-frame orientation, skew angle, and cross-frame connection upon bridge system behavior and cross-frame stresses. In a suite of detailed 3D, solid finite element analyses models of skewed bridge systems, cross-frame layout, connection thickness and type, and skew angle were varied. Skewed bridge systems with cross-frames placed parallel to the skew angle as well as systems with cross-frames arranged in a staggered configuration were considered. Varying bent plate connection thicknesses and a half-pipe connection were also analyzed. Cross-frame spacing of 4.6 m [15 ft] and 9.14 m [30 ft] were examined; severe cross-frame spacing of 13.7 m [45 ft] was also considered to examine behavior at very long unbraced lengths. The findings of this study showed that skew angle, skew configuration, and connection type all influenced the strength and stiffness of system. The data showed that cross-frame placed parallel to skew up to an angle of 40° performed similar or better than cross-frames oriented perpendicular to skew for every given skew angle and connection type.