| A theory for the buffeting of long-spanned suspension bridges has been proposed and a numerical computation based on this theory has been carried out. The computed results and the experimental measurements carried out at the National Research Council at Ottawa, Canada are in reasonable agreement in view of the required approximations implicit in two important parameters, namely the aerodynamic admittance function and the bridge vibration modes, which are taken here as uncoupled in bending, torsion, and sway.;It has been observed in wind-tunnel model studies that introduction of turbulence into a laminar stream delays the onset of flutter considerably, theoretical explanation being that for certain types of bridges, flutter is predominantly a torsional instability with the nearby modes "dragged" into the gross instability. The introduction of turbulence excites the structure into simultaneous responses of its major modes. Each of these modes has associated with it self-excited forces, including positive or negative damping, and aerodynamic coupling terms. Thus, buffeting brings in all these modes, and the actions of some, which are destabilizing, is countered by the actions of others, which are stabilizing. Overall, the result is the well-known delay in the onset of flutter with increasing wind velocity under conditions of turbulence.;An analysis of computed response spectra further clarify this point. As wind speed increases, at first there is a tendency for the structure to respond unstably in a predominant torsional mode close to the laminar flutter speed. This torsional mode decreases in importance as energy is spread out among other modes. Finally, above flutter speed, the system responds as a multi-mode excited structure under random loading. Thus, the effect of wind turbulence actually can destroy the onset of flutter by reversing the mechanics of flutter.;It is demonstrated that a section model of a full bridge can correctly predict the flutter speed based on previously established free-oscillation flutter theory emanating from Princeton. Moreover, when integration is carried out over the full span for mode shape effects, it substantiates the usefulness of the section model for full-bridge flutter prediction. The presently predicted flutter speed for a full bridge agrees very well with wind-tunnel results. |
