Machine learning-enabled estimation of crosswind load effect on tall buildings

• DOI: 10.1016/j.jweia.2021.104860
• Publication Date: 2021-11-25T20:21:19Z
• Authors: Pengfei Lin; Fei Ding; Gang Hu; Chao Li; Yiqing Xiao; K.T. Tse; K.C.S. Kwok; Ahsan Kareem
• Keywords: Crosswind

This paper presents an approach to predict crosswind force spectra and associated response of tall buildings with rectangular cross-section based on machine learning (ML) technique random vibration-based analysis. An efficient ML algorithm, light gradient boosting (LGBM), was trained the by using database from Wind Engineering Research Center at Tamkang University embedded in aerodynamic NatHaz Modelling Laboratory. Furthermore, unsupervised K-means clustering, employed advance understanding spectrum characteristics buildings. The effects three factors, i.e., ground roughness, aspect ratio side ratio, were discussed clustering. To buildings, case studies carried out validate predictive accuracy LGBM model combined results demonstrate that proposed method multiple database-enabled design module for high-rise developed Laboratory Notre Dame is effective computationally provide fast accurate predictions responses