Wakota Bridge Thermal Monitoring Program Part I: Analysis and Monitoring Plan

Abstract

In this work, a common refined design method is evaluated with respect to a recently constructed bridge. Two finite element models of the Wakota Bridge in South St. Paul, Minnesota, were produced, one using a design level program (SAP2000) and the other using a research level program (ABAQUS). These models were verified with respect to each other using linearly elastic materials and were found to behave similarly. After this verification, an arbitrary temperature load was applied to each model and the refined design method was evaluated for accuracy of reduced section properties with respect to the more descriptive progressive cracking solution simulated by ABAQUS. The refined design method was employed using two, four, and six stiffness segments at which stiffness is evaluated along the height of the pier walls. It was seen that accuracy increased as the number of stiffness segments increased and that four segments seemed to balance accuracy and time-commitment by the engineer adequately. A staged construction model of the Wakota Bridge was also built, using the design level program, which incorporates all time-dependent effects of the construction sequence as well as locked-in forces. A pile analysis was performed and appropriate rotational springs were found for Foundations 2 and 3. A simplified method for the determination of the rotational springs is discussed, and a range of effective lengths was found for use with this procedure. The staged construction model is used for field data correlation in Part two of this report. The staged construction model was also used to evaluate the different design options as described in the AASHTO LRFD. The two options given for accounting for reduced section properties were evaluated and compared. The refined analysis option and gross section option were compared for the Wakota Bridge and are shown to correlate to within about 10%. The two temperature application methods (Procedure A and B in the AASHTO LRFD) were also compared. As expected, Procedure B produced much larger design moments than that of Procedure A.