Abstract
The design of Long Span Bridges involves complex analysis of the interaction between fluid and bluff body (Fluid Structure Interaction). In the past, the aerodynamic characteristics needed for the design of long span bridge deck sections have been obtained via wind tunnel tests. Recent advances in turbulence modeling, computational fluid dynamics and the increasing affordability of computers have made numerical modeling of these complex studies possible. Much research has been carried out on the applicability of CFD in the study of bluff body aerodynamics, and less relating to long span bridges. Unfortunately, due to computational costs and sometimes lack of complete details from the wind tunnel test results, these studies have been limited in scope; usually the work is 2-dimensional and often limited to the basic section without the parapets and equipment that are part of the super structure. Also, some experimental work has been done on shaped bluff body sections, such as rectangular cylinders, which has provided useful but limited application to a bridge deck section. The work described in this theses consist of modeling and simulation of the sectional wind tunnel test of the Carquinez Strait bridge in California, a real long span bridge deck section. The modeling incorporates the often ignored but important details such as parapets, barriers and most importantly, the effects of the shape of different edge details on aerodynamic characteristics such as lift, drag and moment coefficients, as well as the flow pattern created by the different edge details in the shedding of vortices in their wakes. The simulations were carried out using the kappa-o based Shear Stress Transport RANS (Reynolds Averaged Navier Stokes) turbulence model at an average wind velocity of 3.2 m/s with angles of attack of +/-10°. The basic deck section of the Carquinez Strait Bridge is of trapezoidal box girder with sharp edge detail type, this cross section was modified by modifying the edge detail and replacing it with three different types of details; a round edge detail, an oval and a triangular shaped edge type. Additional studies include the removal of the barrier, parapet and equipment to see their effect and the roles played by them in the aerodynamic static force response and the flow physics. Two grid types were explored to determine the most accurate; tetrahedral and hexahedral dominated meshes. Next, determining the appropriate RANS turbulence model, from the matrix of grid and turbulence model emerges the numerical simulation. Once the wind tunnel test results were corrected for errors, the results from numerical modeling compares very well with the static wind tunnel test, thereby validating the choice of turbulence model and grid type, and demonstrating the viability of CFD in long span bridge design. The results of the fluid flow around the differently modified edge details shows how the mechanics of vortex induced vibration develops off of the recirculating air underside the exterior web at the trailing edge, because of the variation in the velocity of air in this region due to the different edge details, it is reasonable to make deductions on stability. In the simulations where the parapets, barriers and equipments are removed off of the deck sections, the response are markedly different, revealing that they are critical and as important as the edge detail chosen during the preliminary design.