Comparative and sensitivity study of flutter derivatives of selected bridge deck sections, Part 2: Implications on the aerodynamic stability of long-span bridges

This study is the continuation of a comprehensive investigation on section-model aeroelastic coefficients for bridge decks (flutter derivatives) extracted from wind tunnel section-model tests. The original motivation emerged from the United States—Japan Benchmark Study on Bridge Flutter Derivatives, which promoted a series of systematic comparisons of experimental data extracted by two laboratories (Iowa State University, USA and Public Works Research Institute, Japan) as well as previous results available in the literature. Comparisons, which included both streamlined and bluff deck girder models, were summarized in a companion paper [Sarkar P, Caracoglia L, Haan FL, Sato H, Murakoshi J. Comparative and sensitivity study of flutter derivatives of selected bridge deck sections. Part 1: Analysis of inter-laboratory experimental data. Eng Struct 2009;31(1):159–69]. Differences in the flutter derivatives were mainly attributed to: distinct experimental methods in the wind tunnel (free or forced vibration methods), intrinsic variability between different laboratory environments and effects of amplitude dependency in the tests (for bluff sections).In this paper, a sensitivity study was performed to examine the implications of the perceived dissimilarities among flutter-derivative data sets discussed in [Sarkar P, Caracoglia L, Haan FL, Sato H, Murakoshi J. Comparative and sensitivity study of flutter derivatives of selected bridge deck sections. Part 1: Analysis of inter-laboratory experimental data. Eng Struct 2009;31(1):159–69], on the aeroelastic instability of long-span bridges.Numerical analyses were conducted to evaluate flutter instability boundaries of a set of long-span bridge configurations. Both single-mode and coupled-mode instability were considered, depending on the cross section type and characteristics. It is concluded that uncertainty in flutter derivatives occurring as a result of extraction method or intrinsic variability between different laboratories from negligibly small values to as much as fifty percent, as observed in [Sarkar P, Caracoglia L, Haan FL, Sato H, Murakoshi J. Comparative and sensitivity study of flutter derivatives of selected bridge deck sections. Part 1: Analysis of inter-laboratory experimental data. Eng Struct 2009;31(1):159–69], do not affect the variability in the predicted critical velocity in a proportional way. However, differences in the resulting critical velocities have been observed and estimated from as small as five percent to more than thirty percent, heavily depending on the type of bridge, the simulated conditions and type of instability, either dominated by a single mode or influenced by modal coupling.