Comparison of probabilistic methods for the effects of wind direction on structural response

Abstract

This paper compares three methods for predicting extreme wind-load effects on buildings, allowing for the varying direction of the wind, as well as the windspeed itself. These are the ‘out-crossing’, ‘multi-sector’ and ‘direct calculation’ methods. Wind-tunnel data from a generic tall building, assumed to be located in Melbourne, Australia, at eight different orientations to the climate, have been used to assess the methods, and to compare the results with direct calculation of base moments. The out-crossing method gives reasonable results for about half the building orientations, but underpredicts the responses when the dominant northerly wind direction is not the major contributor to the response calculations. The multi-sector approach although slightly conservative for predictions at low values of average recurrence interval (ARI), is accurate for all responses examined at high values of ARI.