Nature-inspired Design of High-performance Building Cantilever Structures

The research proposal aims to revolutionize structural design by drawing inspiration from the geometric and mechanical properties of natural structures. Natural structures,such as insect wings and plant rhizomes, exhibit lightweight yet highly efficient structural networks, characterized by primary and secondary veins in a statically determinate system optimized for their in-plane and out-of-plane stiffness. This study proposes to adapt these natural principles to the design of cantilevered building structures, a critical architectural element that often faces challenges in balancing weight, stiffness, and load-bearing capacity.Building upon recent advancements in machine learning and graphic statics, this research will employ Maxwell's reciprocal diagrams to analyze and replicate the force equilibrium of natural structural networks. These diagrams will be used to train artificial intelligence models capable of generating structural networks for arbitrary boundary geometries, specifically, cantilevered structures. The machine learning framework will integrate image-text-based Diffusion Models to predict form and force diagrams, ensuring that the generated designs maintain static equilibrium and geometric efficiency.The generated structural patterns will be evaluated for their mechanical performance under point loads and distributed loads, simulating real-world conditions such as wind and seismic forces. Our preliminary studies on scaled-down 3D-printed prototypes of dragonfly-inspired wing cores have demonstrated up to a 25% improvement in out-of-plane stiffness compared to traditional designs, with a significant reduction in material usage. The proposed research seeks to extend these findings, focusing on developing lightweight, high-performance cantilever designs that optimize stiffness-to-weight ratios while maintaining architectural flexibility.This work has broader implications for sustainable construction practices, as it leverages bio-inspired principles to minimize material usage without compromising structural integrity. The methodology is also adaptable to other natural structures, such as the vein networks of grasshopper wings or the rhizomes of Amazon water lilies, opening avenues for further exploration. The project will culminate in the development of a computational design platform, enabling architects and engineers to generate efficient cantilever structures based on user-defined boundary geometries.⁠⁠⁠⁠⁠⁠⁠By bridging biology, computational design, and structural engineering, this research aims to redefine how we design lightweight and efficient cantilever systems, contributing to more sustainable and resilient architectural innovations in the built environment.