Behavior of steel tub girders with modified cross-sectional geometry
Access full-text files
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Steel trapezoidal box girders, also referred to as steel tub girders, have been an attractive design option for long-span horizontally curved highway bridges. The completed composite bridge system presents an aesthetic appeal profile as well as large torsional stiffness. However, during construction, the open U-shaped steel girder section is relatively flexible in torsion and requires extensive bracing. A recent application on straight bridge system in Waco, Texas showed potential wider utilization of tub girders for span length normally reserved for precast concrete beams. Current design and fabrication practices have several relatively inefficient aspects regarding the section geometry and bracing details. These details include the restrictions on the slope ratio of the webs and the top flange-web attachment. Due to the use of empirical equation for live load distribution factor, the slopes of tub girder webs are restricted to be no greater than 1 Horizontal: 4 Vertical in current AASHTO specification. Significant economy can be achieved by utilizing flatter webs. While keeping the width of bottom flange constant, the use of flatter webs increases the tributary width of individual girder. This leads to potential reduction of the required number of girder lines to support traffic live loads and considerable savings on fabrication time and cost. Additionally, the webs of the tub girder are usually attached at the mid-width of the top flanges. This leaves limited flange width to connect the top lateral braces directly with the flange. Therefore, large gusset plates are frequently used to provide sufficient space for the connection. However, the use of gusset plates leads to poor load transfer and unnecessary fabrication cost. If the top flanges are allowed to offset toward the inside of the box, more flange width would be available for simple bolted bracing connection without using gusset plates. Since these proposed details are not currently permitted by AASHTO Specification, a TxDOT-sponsored research project has been conducted at UT Austin to demonstrate the impact of these improved tub girder details using large-scale experimental study and finite element analyses. This dissertation presents part of the results of this research project to provide better understanding of tub girder behavior as well as design recommendations to improve the efficiency and economy of the steel tub girder system