Development of three-dimensional finite element software for curved plate girder and tub girder bridges during construction
Because of its ability to be easily shaped, steel is an attractive material for curved girders. Plate girder and tub girder bridges, for example, are often the preferred solution for direct connectors in highway networks. This flexibility in fabrication, however, presents challenges for structural engineers because of the difficulties associated with accounting for combined bending and torsion with curved geometry. The potential presence of skewed supports is a further source of complexity. In fact, no commercial structural engineering program currently addresses the evaluation of plate girder and tub girder bridges while modeling them to the full extent of their three-dimensional configuration. Most engineers, for example, use a two-dimensional bridge representation, which is often accurate for typical design of a complete bridge but may also be unconservative in many cases. The few programs that allow a full three-dimensional representation require extensive knowledge of finite element theory as well as significant time to model any complex structure. This dissertation presents the assumptions, methodology and calculations involved in the programming of a new structural engineering program designed to assess the behavior and stability or curved plate girder and tub girder bridges during erection or deck placement. It then illustrates the capabilities of the program for various structural systems subjected to a variety of loads, from self-weight to wind and temperature loads. In addition to a linear elastic analysis, multiple types of analysis are offered to the engineer: a geometrically nonlinear analysis provides a more accurate behavior for flexible systems, a linearized buckling analysis yields an upper bound evaluation of the stability of the structure, while a modal dynamic analysis estimates the free vibration modes of that structure.