Wind effects on a cable-suspended roof: Full-scale measurements and wind tunnel based predictions: Fid-Bau Portal

Wind effects on a cable-suspended roof: Full-scale measurements and wind tunnel based predictions

Chen, Bo / Wu, Teng / Yang, Yilong / Yang, Qingshan / Li, Qingxiang / Kareem, Ahsan

Abstract Full-scale measurements of wind effects on a long-span cable-suspended roof were carried out. The wind-induced response data were analyzed in the time-frequency domain based on the wavelet transformation, and modal frequencies and amplitude-dependent damping ratios were identified with the random decrement technique. The damping ratios were noted larger with increasing response amplitude especially in the fundamental mode. The identified modal frequencies from the full-scale measurements were lower than those from the original finite element model (FE model). The identified modal parameters were then used to update the FE model with an artificial neural networks (ANNs) based scheme. The dynamic response from the roof was computed from the updated FE model of the structure where the wind loads input was obtained from simultaneous pressure measurement in wind tunnel tests. In view of the uncertainties surrounding both full scale measurements and wind tunnel based predictions, these comparisons show good agreements at least at two locations. Both the dynamic response from full-scale measurements and numerically predicted analyses demonstrated significant contributions from multi-mode effects. The computed response spectra at selected locations were similar to those from the full-scale measurements, suggesting a validation of the predictions.

Highlights • Wind effects from full scale measurement are compared to numerical simulation. • Identified modal frequency from full-scale measurement is lower than that from FEM. • The identified damping ratio increases with an increase in the response amplitude $ξ (%) = 4.7 x + 1.42$ . • Response spectra from FEM agree qualitatively with those from full-scale measurement. • Differences in RMS response are from inaccuracy in wind loads, structural model and damping.