Application of Neural Dynamics Model for Optimization of Diagrid Structural System for High-Rise Steel Buildings

In high-rise steel building design and construction, diagrid structural systems are valued for their material efficiency, spatial flexibility, and lateral stability. However, due to the lack of specific guidelines for diagrid configurations such as diagrid angle and density in existing design building codes, this presents challenges and limitations in design optimization. To address these gaps, this study applies a Neural Dynamics model focusing on diagrid angle configurations and cross-sectional variations to enhance structural efficiency under lateral and seismic loads. A Multi-Layer Perceptron (MLP) neural network was used to simulate diagrid structures of varying heights and loading scenarios, predicting critical parameters that balance material use with structural performance. The results reveal that increasing diagrid angles in taller buildings significantly improves lateral stiffness and reduces deformation, particularly at angles between 65° and 74°. Additionally, by incorporating a selection of steel sections, varied cross-sectional areas along the building height enhance vertical stability without substantially increasing overall weight. The findings of this study highlight the critical role of diagrid angles and material distribution in optimizing structural efficiency. The neural network model demonstrated high accuracy in predicting key structural parameters, confirming the potential of computational intelligence in structural design. This research provides actionable insights into optimizing diagrid systems, offering practical guidelines for integrating neural dynamics into design processes. By addressing material efficiency and structural resilience, the study contributes to the development of sustainable and innovative solutions for high-rise construction. These findings not only inform updates to design standards but also pave the way for further advancements in applying artificial neural networks to structural engineering challenges.