Studies of Davenport and others have provided the basis for a simple gust f
actor approach to evaluate the wind-induced structural responses. This appr
oach is based on the maximum responses that can be represented by their cor
responding mean maximum response. However, the uncertainty in the wind char
acteristics and in the material properties and geometry variables affects n
ot only the mean maximum responses but also their variability. By consideri
ng these uncertainties, estimates of the statistics of the maximum response
s can be obtained from the unconditional probability distributions of the m
aximum responses. Alternatively, they can be obtained directly from the unc
ertainty propagation analysis based on the conditional mean maximum respons
es. The fort-ner is accurate whereas the latter is an approximation. The ad
equacy of this approximation and its effect on the structural reliability a
re investigated.
Results reported in this study seem to support that the use of the conditio
nal maximum response equal to the mean maximum response is adequate for the
uncertainty propagation analysis. Results of the reliability analyses also
suggest that the consideration of the uncertainty in structural parameters
such as the fundamental period of vibration and the damping ratio is very
important for serviceability limit states, and that these uncertainties may
be ignored for the resistance based limit states as long as the structure
behaves linearly. (C) 2001 Elsevier Science Ltd. All rights reserved.