Partitioning of wind pressure coefficients on a tall building under interference effect: a cluster analysis approach

Abstract

This study proposed a cluster analysis-based approach for partitioning wind pressure on tall buildings in densely-packed areas, considering interference effects. By utilizing a modified GK cluster algorithm and leveraging the coordinate positions of pressure taps along with their corresponding wind pressure coefficients (WPC), the study aimed to distinguish different parts of the building surface exhibiting varying wind load characteristics. The adoption of cluster validation indices (CVIs) allowed for the determination of the optimal number of clusters, enhancing the accuracy and reliability of the partitioning process. The resulting clusters effectively captured the distribution of wind pressure, providing valuable insights for wind-resistant design considerations. Overall, the proposed approach demonstrates promising potential in facilitating the design and analysis of tall buildings subjected to complex wind loads in urban environments.