Parametric effects of turbulence on the flutter stability of suspension bridges

Despite extensive past research efforts, the influence of turbulence on suspension bridge flutter stability has not been fully understood yet. Moreover, the role of large-scale turbulence has been overlooked in predicting the critical wind velocity, despite experimental and numerical studies indicating that atmospheric turbulence can have either a stabilising or a destabilising effect. This study investigates the parametric effects of large-scale turbulence on flutter stability based on the 2D Rational Function Approximation model, considering the Hardanger Bridge, in Norway, as a case study. First, a Monte Carlo method is used to analyse the bridge stability under various turbulent wind conditions, considering turbulence intensity and integral length scale as key parameters. The results highlight the sensitivity of flutter stability to these turbulence parameters, as well as essentially its independence of the spanwise correlation of the parametric excitation. The role of the variation in the wind angle of attack is found to be largely dominant compared to that in the wind velocity magnitude. Subsequently, Floquet multipliers are employed to study the influence of periodic parametric excitations on bridge stability, assuming sinusoidal fluctuations of wind velocity magnitude and angle of attack. This simplified scenario helps us understand the various parametric excitation mechanisms and explain the results of the more realistic Monte Carlo approach. The study emphasises the key role played in flutter stability and buffeting response by the so-called "average parametric effect", associated with the mean of the aerodynamic derivatives as nonlinear functions of the slowly-varying angle of attack. The preponderance of this effect in most realistic turbulent wind scenarios also explains the negligible impact of correlation of the parametric variation in the angle of attack. Finally, an equivalent linear time-invariant model is proposed to account for the parametric effects of turbulence in a simple way, yielding good results and offering a new perspective on the use of classical self-excited force coefficients.