Random flutter stability of bridges in turbulent flow via Lyapunov analyses

Assessing the flutter stability of suspension bridges in turbulent wind is a challenging task, since turbulence-driven fluctuations in the angle of attack modify self-excited forces, introducing parametric effects and making the system time-varying. Here, Lyapunov exponents (LE) and moment Lyapunov exponents (MLE) are used to quantify the stochastic stability of the Hardanger Bridge, the largest suspension bridge in Norway. Self-excited forces modulated by angle-of-attack fluctuations are modelled through a two-dimensional Rational Function Approximation (2D RFA), which reproduces the associated changes in aerodynamic damping and stiffness and is validated against wind-tunnel measurements. Monte Carlo simulations are then performed to estimate LE and MLE and to map stability across different turbulence intensities and integral length scales, representing distinct noise levels and wind-energy distributions. The results for a 2D model of the Hardanger Bridge show that turbulence reduces flutter stability, especially for high turbulence intensity and large integral length scales. In most conditions, the maximum LE is sufficient to characterize the bridge stability. However, in some cases (for instance, for high turbulence intensity and small integral length scales), the flutter stability limit is not univocal and MLE are necessary for a complete assessment.