FLUTTER IN MULTI-BOX SUSPENSION BRIDGES: THE CASE-STUDY OF THE XIHOUMEN BRIDGE

Multi-box suspension bridges represent an innovative structural typology, characterised by multiple decks that offer greater stability, loading capacity and reduced torsional stresses. Until the construction of the 1915 Çanakkale Bridge, the Xihoumen Bridge (completed in 2009) was the longest multi-box suspension bridge in the world. Its deck’s configuration was studied to provide much better wind resistance performance, considering the stricter stability requirement of 78.4 m/s wind speed. With this structural innovation, the critical flutter speed reaches approximately 90 m/s. The aim of this Thesis is to investigate the flutter instability of the Xihoumen Bridge adopting the finite element software ANSYS APDL and an analytical model implemented in a MATLAB code. An extensive aeroelastic analysis is carried out studying the combination of different structural models and different aeroelastic forces defined by various sets of flutter derivatives. The first part of the Thesis presents the theoretical and bibliographical background of flutter analysis, necessary to understand the objectives of the following sections. This is followed by a description of the case study, analysing its structural characteristics, design challenges, and context. Next, a chapter is dedicated to the different methodologies used for calculating flutter derivatives: Theodorsen’s simplified method and Andersen’s method based on the superposition of flat plates are tested by the comparison of the derivatives provided with those obtained experimentally through wind tunnel tests. The concept of gap width sensitivity and its impact on the calculation of flutter derivatives is explored, underlying its fundamental role in understanding the differences in the results obtained. In the final part of this Thesis, Finite Element Models of the case study are presented. Preliminary zero-wind modal analyses are performed and the results provided by the various models are discussed and compared with literature data. Subsequently, using the previously discussed flutter derivatives, the flutter analyses are carried out to determine the critical flutter frequency and the critical wind speed at which instability occurs. In conclusion, all results are presented and discussed. The comparison of the results obtained with literature values shows that the 1-axis central span model provides the best estimate of the bridge’s dynamics which consequently translates in the best agreement of predicted flutter wind speed (difference less than 4%). Another relevant result for this model is that similar conclusions are reached using both experimental flutter derivatives and those obtained through Andersen’s simplified approach, when applying the gap-width scaling factor. This is an interesting outcome, highlighting both the importance of gap-width sensitivity and the potential to use simpler and faster approaches to predict flutter derivatives.