Comparison of design wind speed using Extreme Value Distributions and IS 875 Procedures

Extreme wind speed is of significance in design process to study the structural safety particularly for tall structures like cooling towers, stacks, transmission line towers etc. Gumbel and Frechet distributions are used to estimate the design wind speed for different return periods. Method of least squares (MLS) and Lieblein technique based on order statistics approach are applied for estimation of parameters of the distributions. Anderson-Darling (A²) test is used for checking the adequacy of fitting of the method/distribution to the recorded wind speed data. The study shows that Gumbel distribution using MLS is found to be appropriate for estimation of design wind speed for Vadodara region. The paper presents that the values of estimated 3-sec average extreme wind speed adopting statistical distributions are compared with IS 875 procedures to arrive at a design parameter for the region under study.