The accuracy of the aeroelastic model or truss girders has consistently been a decisive factor influencing the results of full-bridge aeroelastic wind tunnel tests. A novel design approach, termed the Multi-Spine Frame System, is introduced for the reduced-scale model of the stiffening girder. This system incorporates multiple spines, thinner beams and columns to minimize aerodynamic interference while accurately simulating overall stiffness, mass, and constraints. The design procedure is formulated as an optimization problem with the objective of optimizing the low-order natural frequencies (lateral bending, vertical bending, and torsion) of the aeroelastic model. Pattern search method and penalty functions are employed to identify a locally optimal solution that satisfies engineering applications for this optimization problem. This design method is applied to an actual project involving a four-cable suspension bridge with a truss girder, and the test results demonstrate the accuracy of this approach in simulating the dynamic characteristics of the aeroelastic model. Corresponding wind tunnel tests on flutter instability, as well as numerical multi-modal flutter analysis, are conducted to analyze the critical flutter speed, vibration shapes, and frequencies of this bridge. Furthermore, the impact of deviation in structural mode shapes and natural frequencies on flutter instability is investigated using ten design schemes achieved by altering boundary conditions, mass distributions, and stiffness distributions, were examined. Significant differences in critical flutter speed (up to 9%) are observed due to deviations in mode shapes, emphasizing the necessity of considering modal frequencies and other modal information, such as mode shapes, in aeroelastic model design. This approach contributes to the advancement of aeroelastic model design for bridges with truss girders by introducing an effective system and an optimization method, providing a basis for future studies on wind instability and structural dynamics and also underscores the importance of accurate mode shape representation in flutter analysis.
Read full abstract