Extreme value distribution and peak factor of crosswind response of flexible structures with nonlinear aeroelastic effect

The crosswind response at the vicinity of vortex lock-in wind speed becomes an increasing concern in structural design of super tall buildings and other dynamically sensitive structures. Because of nonlinear aeroelastic effect of vortex shedding, the extreme value distribution and peak factor of crosswind response are distinctly different from traditional wind-induced stochastic buffeting response. This study establishes an effective approach for predicting extreme value distribution and peak factor of crosswind response of flexible structures on the basis of an improved understanding of the underlying mechanism. It is pointed out that the improved extreme theory of Gaussian processes with consideration of narrow-band feature cannot interpret the extreme crosswind response. This study, at the first time, reveals that the unique characteristics of extreme value distribution of crosswind response is primarily attributed to its hardening non-Gaussian distribution. An extensive analysis of simulated crosswind response illustrates that the extreme value distribution and peak factor can be estimated from those of an underlying Gaussian process based on the translation process theory. The translation function between non-Gaussian crosswind response and Gaussian process can be expressed in terms of Hermite polynomial model based on the first four statistical moments or quantified by mapping of cumulative distribution functions and curve fitting. Closed-form formulations are developed for calculating Hermite translation model coefficients, which permits the estimation of extreme value distribution and peak factor directly using the response kurtosis and bandwidth parameter. The characteristics of crosswind response and the proposed approach are also validated using full-scale vibration measurements of a wind-excited traffic-signal-support structure. This study not only clarified the mechanism responsible for the unique extreme value distribution of crosswind response with nonlinear aeroelastic effect but also established an effective simple approach for its estimation using the characteristics of response process.