| description abstract | The Y-plan architectural configuration, characterized by three limbs converging at a central core, is widely adopted in commercial construction due to its ability to enhance privacy, ventilation, and external visibility. This design is prevalent in structures of varying scales, from iconic skyscrapers like the Burj Khalifa and Jeddah Tower to smaller buildings such as the Queensbridge Public Houses in New York. This study provides a detailed analysis of aerodynamic coefficients for Y-shaped buildings, focusing on aspect ratios defined by height-to-length (H/L) and length-to-depth (L/D) ratios, varied within the ranges of 0.5 to 7 and 0.5 to 4, respectively. Computational fluid dynamics (CFD) simulations, based on Reynolds-averaged Navier-Stokes (RANS) equations, are conducted for selected aspect ratios determined through nested Audze–Eglais Latin hypercube (NAELH) sampling. The resulting pressure, force, and moment coefficients from the simulations are used to train artificial neural network (ANN) models. These ANN models are validated against existing literature, after which they are employed for predictive analysis. The ANN-generated data identifies maximum and minimum coefficients for different angles of attack (AOA), leading to the creation of contour plots for these coefficients with H/L and L/D ratios as variables. The maximum force coefficients range from 0.84 to 1.3, with lower values corresponding to reduced L/D ratios, suggesting potential benefits of lower L/D ratios. This study provides structural engineers with critical data for determining extreme aerodynamic coefficients across various aspect ratios under worst-case AOA conditions. | |
