Dragonfly-Inspired Wing Design Enabled by Machine Learning and Graphic Statics

Abstract

This research relates the morphology of the structural network of a dragonfly wing to the static equilibrium of forces using a geometry-based equilibrium method known as graphic statics. It then develops machine learning algorithms to generate similar structural networks for designing airplane wings using the in-plane equilibrium information of the dragonfly wing. This approach can generate a structural network in equilibrium with no prior information related to the topology or geometry of the network just by receiving the boundary geometry of the wing. This research shows that the internal network of the dragonfly wing can be assumed to be compression or tension-only system on a 2D plane. Consequently, another diagram is constructed to represent the geometric equilibrium of the forces in the initial network. This diagram is called the force diagram in the context of graphic statics. A new wing geometry is reconstructed from the force diagram, with its members sized according to the force magnitude. Both form of the network and its force diagram are used to train machine learning models for the generation of the structural network of the wing. The developed methodology is also capitalized to generate microstructural patterns inspired by other species with networks of convex polygons. We numerically and experimentally examine one application of this method in designing the cellular core, 3D printed by fused deposition modeling, of the airfoil wing, which suggests up to 25% improvement in the out-of-plane stiffness. Our findings suggest this ML-assisted approach enables leveraging million years of evolution in nature to develop the next generation of high-performance, lightweight structures.