Shape-dependent characteristics of full-scale wind profiles

Numerical investigations on vortex excitation of tall and slender structures, like guyed masts or towers oscillating with higher eigenmodes, indicate that it can be beneficial to account for realistic shapes of the mean wind speed profile. In order to use shape-dependent profile characteristics for a prospective stochastic vibration analysis of structures subjected to vortex excitation, the occurrence of different profile shapes is derived from full-scale wind data, which has been continuously recorded at the 344 m measurement mast Gartow since 1989. After the shape-dependent classification of the profile data with an artificial neural network, the classified wind speed profiles and the corresponding turbulence characteristics are described statistically. The classification shows that the frequency of occurrence of constant profile shapes is only 18.4%. Since these are usually used to analyze the vortex-induced fatigue, accounting for realistic wind speed profiles in a vortex excitation analysis can enhance the predicted fatigue safety for structures that oscillate with multiple modal nodes. For such an analysis, the shape-dependent profile characteristics can be well represented independently of the wind direction with Weibull and Gaussian statistics, where the latter is a reasonable approximation for the conditions at the measuring site. For the turbulence intensity, common models apply within each profile class.

► With a neural network, full-scale wind speed data is assigned to six profile shape classes. ► Constant profile shapes occur only at 18.4%, but are used in building codes for vortex excitation. ► For the measuring site, Weibull and Gaussian statistics well represent the profile classes. ► Direction-independent statistics provide a reasonable approximation. ► A deterministic description of the turbulence simplifies the statistical model.