Limitations of the flutter derivative model

Abstract

Flutter derivatives identified from transient (free-decay) vibrations might not be suitable for the analysis of bridge flutter. The application of transient flutter derivatives in the flutter analysis relies on two assumptions: (1) transient flutter derivatives and steady-state flutter derivatives are equivalent and (2) aeroelastic effects are superposable. Both assumptions are challenged in this article. It is shown through transient vibration tests that (1) the aeroelastic-coupling between heaving and rotational motions may switch from one pattern to another as the wind speed varies and (2) some of the transient flutter derivatives may be time-varying. The former implies that the predicted flutter type based on transient flutter derivatives may not be unconditionally consistent with the experimentally observed flutter type; the latter implies the transient flutter derivatives may be physically different from the steady-state flutter derivatives. These two issues undermine the basic assumptions of the flutter analysis of bridges. A possible corollary to this study is that if free vibration is used to predict bridge flutter, we should resort to the steady-state (flutter state) vibration instead of the transient vibration of the sectional model. A revision to the aeroelastic force model is proposed to facilitate the discussion.