Tropical-cyclone-wind-induced flutter failure analysis of long-span bridges

Abstract

Continuously growing number and span length of long-span bridges in coastal region of China require a robust aeroelastic design to against the tropical-cyclone (TC) -induced flutter instability. Instead of using the deterministic approach, this study presents a Monte-Carlo-technique-based framework to analyze the flutter reliability of a long-span suspension bridge subjected to TC winds. A limit state function formulated as the difference between the flutter capacity and the extreme gust wind, termed as the product of a random gust factor and the TC mean wind hazard is employed. The flutter fragility curve in terms of the critical wind speed is derived using 2D and 3D flutter analysis models accounting for the structural modal and damping randomness as well as experiment-induced errors of aeroelastic flutter derivatives. The TC wind hazard curves at the height of the bridge deck described as the probability of exceedance at any given years of interest versus the extreme 10-min mean wind speed are estimated through generating a large quantity of synthetic tracks around the bridge site. The probabilistic solutions of the gust factor associated with any gust duration of interest are determined utilizing the statistics of observed TCs coupled with a theoretical model. The TC-induced flutter failure probabilities of present bridge are then predicted in different combinations of gust duration, surface roughness length and flutter analysis models, which enables the assessment of flutter risk subjected to TC winds as compared to code-suggested target reliability indices. The present reliability analysis provides some basic information to assist and enhance the flutter-resistant design of long-span bridges due to TC winds during the preliminary stage.