Study on the influence of structural parameters and 3D effects on nonlinear bridge flutter using amplitude-dependent flutter derivatives

To estimate nonlinear flutter response of long-span bridges, this study established a method for identifying full set of amplitude-dependent flutter derivatives (FDs) from free vibration wind tunnel tests. Taking a typical double-deck truss bridge as a Case study, the amplitude-dependent FDs of the bridge deck at the whole wind speed regime are identified and cross-validated based on large-amplitude free vibration wind tunnel tests of its single degree of freedom (SDOF) torsional and 2DOF vertical-torsional section models. The influential mechanism of vertical DOF on nonlinear flutter was revealed by quantitatively comparing the nonlinear aerodynamic damping of the SDOF and 2DOF systems. The amplitude-dependent FDs are then used to calculate the nonlinear flutter responses of the 2D bridge section and a prototype long-span suspension bridge (1650m) with four main cables based on developed 2D and 3D nonlinear flutter analysis methods. Finally, the influence of structural parameters and 3D effects on nonlinear flutter are quantified and discussed. The results show that the 2DOF system has a lower critical wind speed and higher torsional stable amplitudes compared with the SDOF system since the participation of vertical DOF introduces the negative coupled aerodynamic damping to the system. The aerodynamic nonlinearity becomes stronger and stronger as the wind speed increases and it mainly leads to the significant amplitude dependence of the uncoupled aerodynamic damping, which is the key factor to cause the limit cycle oscillation (LCO)-type of flutter. While the coupled aerodynamic damping appears to be a relatively linear damping with weak amplitude-dependence within the studied wind speed and it mainly plays the role of reducing the stability of the system. The 3D effects of the vibrating bridge deck will reduce the system stability mainly by increasing the negative uncoupled aerodynamic damping. Therefore, the amplitudes of nonlinear flutter will be seriously underestimated if the 3D effects are ignored.