GRAVITY DESIGN OF VERTICAL ELEMENTS USING K-MEANS CLUSTERING

Abstract

For building structures to be safe, effective, and economical, the gravity design of vertical structural elements is essential. Manual grouping and conservative assumptions are common in traditional gravity design methods, which can result in inefficient material utilization, especially in big and complex buildings. A hypothetical, data-driven framework for gravity design of vertical elements using the K-means clustering method is presented in this paper. A simulated multi-story building model is used to extract important gravity load metrics and geometric features of columns and shear walls, which are then processed using unsupervised machine learning algorithms. In order to provide standardized yet optimal gravity design solutions for each cluster, vertical pieces with similar load behavior are clustered using the K-means algorithm. The findings show that clustering enables rationalized section sizing, lowers design variability, and successfully distinguishes parts according to gravity need. A balanced distribution of load categories and design sections throughout the structure is further demonstrated by percentage frequency analysis. According to the results, incorporating K-means clustering into gravity design can preserve engineering judgment and code compliance while increasing design efficiency, improving material optimization, and offering insightful structural information.