Revisiting moment-based characterization for wind pressures

The estimation of peak wind pressures is important in the reliability-and performance-based design for low-rise buildings. Typically, Davenport's formula is widely used to determine the peak factor if the pressure approximately follows Gaussian distribution. Recently, the moment-based Hermite polynoMial model (HPM) is becoming popular to estimate the peak factor when the non-Gaussianity of wind pressure exists. However, their performances deserve further study based on the appropriate wind tunnel data. In this study, Davenport's formula and moment-based HPM are reviewed. The peak value of wind pressure is determined using very long time histories of wind pressure data to evaluate the performance of moment-based HPM and Davenport's formula. Results suggest that moment-based HPM should be adopted in the peak value estimation for wind pressures when the skewness and kurtosis of a process are sufficient to capture its non-Gaussian properties. Results also show that Davenport's formula may cause noticeable errors in the peak factor estimation for the wind pressure data close to Gaussian process while HPM provides a robust estimation for these data. (C) 2016 Elsevier Ltd. All rights reserved.