The flexural strength and ductility of steel girders is dependent upon many factors including cross sectional parameters, loading condition, and cross bracing. These effects are investigated for bridge girders made of high performance steel through the use of experimental testing and non-linear finite element analysis.
A parametric study was performed and the results were compared against the strength prediction provisions contained in Chapter 6, and associated Appendix A, of the AASHTO LRFD Bridge Design Specifications (2007). The main Chapter 6 provisions were found to be overly conservative for cases when use of Appendix A would otherwise be allowed. However, for cases when Appendix A is not allowed the main provisions perform very well. If Appendix A is used when allowed and the main provisions are used otherwise, the predictions of the AASHTO Specifications are within 5% above and below the observed results for 65% of the cases. The predicted strength was between 90% and 100% (representing a conservative prediction) of the observed values for 73% of the cases and exceeded the observed values by more than 5% in only 2% of the cases.
An explicit moment rotation curve for use in the refined method of the Appendix B provisions of the AASHTO Specifications was developed. The curve is based on a power function that can model both convex and concave responses to more accurately match the observed behavior compared to the fixed, second order curve contained in the specifications. A linear fit option is also presented that recognizes the large overall variance in the resulting moment rotation curves. Further, expressions for effective plastic moment to be used in the simplified method of the Appendix B provisions were developed. In addition to cross-sectional parameters similar to the existing provisions, the moment rotation curve and effective plastic moment relationships incorporate a dependence upon bracing length to account for observed behavior.
