Field measurements and corresponding FEA of cross-frame forces in skewed steel I-girder bridges
University of Delaware
While it is known that bridges have the capacity to easily sustain loads greater than their design loads, a codified method for quantifying this reserve capacity that accounts for the three-dimensional structural behavior does not exist. This state of practice is the overall motivation for this research. Cross-frames have been shown to significantly influence the load distribution behavior that leads to significant reserve capacity. Furthermore, it is also known that the role of cross-frames becomes more significant in skewed and curved bridges and also that the skew angle influences the reserve capacity. Thus, this research aims to quantify the forces in cross-frames of two in-service skewed, steel I-girder bridges and calibrate corresponding finite element models that accurately capture these forces. Two bridges of varying skews, SR 1 over US 13 and SR 299 over SR 1, both located in New Castle County, Delaware, were selected for field testing. Cross-frames and girder locations were instrumented and the bridges were load tested with a weighed truck. Overall between the two bridges, the field tests captured data for 11 cross-frames and 6 girder locations. For the bridge with less skew, SR 299 over SR 1, the maximum bottom flange stress was 1.7 ksi while the maximum cross-frame stress is of similar magnitude, 1.5 ksi. For the more-heavily skewed bridge, SR 1 over US 13, the maximum bottom flange stress was 1.5 ksi while the maximum cross frame stress is more than double this value, 3.6 ksi. This suggests that the potential for cross-frame yielding is an important consideration in determining the reserve capacity of steel bridges. Finite element models of each bridge were created and calibrated based on results from the field tests in order to accurately capture the forces in the structure. The finite element model for SR 1 over US 13 predicted stresses at bottom flange girder locations within 20% of the field test results. Cross-frames with a pinned connection to the stiffener were shown to result in the best representation of the cross-frame stresses, but future work is needed to further explore this connection. Hand calculations of the expected stress in the bottom flange according to American Association of State Highway and Transportation Officials specifications matched the finite element model for SR 299 over SR 1 well, but the bridge behavior captured during the field testing differs from conventional expectations. Further work is needed to identify the source of this unexplained behavior and more accurately calibrate this model. The knowledge gained from these efforts can be used in future work to more broadly study the three-dimensional behavior of steel I-girder bridges. Specifically, through calibrating a FEA technique in this work, additional finite element analysis exploring additional variables can be carried out in future work.