Wind pressure field reconstruction based on unbiased conditional kernel density estimation

Abstract

Detailed surface wind pressure information plays an important role in the wind-resistant design of structural cladding and roof covers. Wind tunnel tests are an important way of obtaining this information. However, due to the limitations of measuring equipment capacity and the internal space of rigid experimental models, it is difficult to obtain sufficient information at all locations of interest. To this end, this study develops a nonparametric wind pressure interpolation method based on unbiased conditional kernel density estimation. Based on the sample data, this method is able to directly predict the Probabilistic Density Function (PDF) of the wind pressure coefficient features (e.g., time histories, mean, standard deviation, skewness and kurtosis) in the target location without the requirement of any parameters. These features can be effectively captured by the predicted PDF and then exhibited via the deterministic single value and probabilistic prediction interval. A case study based on data measured in wind tunnel tests is used to systematically illustrate the capability of the proposed method. The results demonstrate that the method is reliable and may be an effective tool for the data recovery of wind pressure fields in application.