Search

Following such landmark aerodynamic tests as Irminger''s in 1894, Flachsbart''s in 1932—the first to be conducted in boundary-layer flow—and University of Western Ontario (UWO) 1970s tests, considerable progress has been achieved in low-rise building design for wind. Nevertheless, the present state of the art remains inadequate. UWO tests were conducted at low angular and spatial resolutions. Their results were then used to create drastically simplifying standard aerodynamic tables and plots designed for slide-rule era calculations and entailing errors that can exceed 50%.We review material which shows that significant improvements in main wind-load resisting system and component design can be achieved by using database-assisted design (DAD) and associated structural reliability tools, thus accounting realistically for the complexity of the wind loading as well as for the stochasticity and knowledge uncertainties affecting wind effects calculations. We illustrate DAD''s capability to obtain, for the first time in a wind engineering context, realistic estimates of ultimate limit states due to local or global buckling failure. In the future other types of nonlinear behavior associated with ultimate limit states can be similarly dealt with. We note that DAD is ideally suited for use with data likely to be obtained in the future by Computational Fluid Dynamics methods. We discuss the need for assuring quality control procedures for wind tunnel testing so that inter-laboratory comparisons of test results and wind tunnel certifications can be conducted effectively. We also discuss the possibility of systematic database corrections based on full-scale test results, and the possibility of using tests in non-standard wind environments and/or on buildings with non-standard shapes. [Copyright &y& Elsevier]