Date of Graduation

5-2011

Document Type

Thesis

Degree Name

Bachelor of Science in Civil Engineering

Degree Level

Undergraduate

Department

Civil Engineering

Advisor

Selvam, R. Panneer

Reader

Grimmelsman, Kirk A.

Second Reader

Coffman, Richard A.

Third Reader

Hale, W. Micah (William Micah), 1973-

Abstract

Wind effects on long span bridges are a major concern for bridge design. Wind acts as a dynamic load on the bridge deck causing the bridge to react in ways not accounted for in static load design. Attention to these effects increased tremendously after famous bridge collapses, such as the Tacoma Narrows Bridge in 1940. Due to this collapse and others, engineers were made painfully aware that wind phenomena, such as flutter and vortex shedding, can cause major structural damage if not considered in the design process. The solution for many years since has been to conduct wind tunnel tests on small scale models of bridges, which are both time consuming and expensive. In the last few decades, the capabilities of computers have increased exponentially to a point where it is feasible for engineers to develop a computer process that will produce the same values that could be found in a wind tunnel test. This thesis begins to explore the time efficiency and viability of using a computer model based on the finite difference model (FDM) for solving the governing equations. The most typical computing method is the finite element method which has been thoroughly tested and produces reasonably comparable data to the wind tunnel tests that have also been performed. The Great Belt East Bridge (GBEB) is the subject of all the modeling done in both the wind tunnel and the computer simulations. The time required to run a finite difference method program is less than that required for the finite element method. The Strouhal number and coefficient of drag are observed to ascertain whether the FDM analysis results are sufficiently close to wind tunnel numbers for the GBEB. The results of this study indicate that FDM does not appear to generate reasonable values for Cd and St at the grid refinements simulated in this work.

Share

COinS