Performance-based layout optimization framework of tall buildings subjected to dynamic wind load

In the pursuit of advancing structural design methodologies, this paper presents a comprehensive framework for the conceptual design stage to systematically identify the layout of the main wind force-resisting system (MWFRS) (i.e., shear walls) of tall buildings within a performance-based wind design (PBWD) perspective. The framework integrates computational fluid dynamics for wind load evaluation with a finite element method to perform a linear time history analysis. A deep neural network surrogate model is developed to reduce the computational cost required for objective function evaluation. The optimization problem is defined to minimize structural material usage while adhering to performance constraints based on the PBWD described in the ASCE prestandards through using a non-dominated sorting genetic algorithm-II to derive the multi-objective optimization process. Finally, a nonlinear time history analysis is employed to predict the inelastic behaviour of tall buildings based on the prestandards assumptions to quantify the differences that can be generated using the developed framework. To ensure the proposed framework's effectiveness and practicality, a case study of a standard tall building is presented. • Novel optimization framework for performance-based wind design of tall buildings. • Integration of CFD and FEM for wind loading evaluation. • Development of a deep neural network surrogate model to reduce computational costs. • Utilization of genetic algorithms for multi-objective optimization. • Nonlinear time history analysis for wind-induced tall buildings.