A Web-based Interactive Structural Pattern Generation Tool with Graphic Statics and Machine Learning of Dragonfly Wings

Designers, engineers, and scientists have always drawn inspiration from nature. The dragonfly wing is among the many natural structures that have intrigued many researchers to study its geometry and performance as a bioinspired design. In previous research, we developed a workflow to use Graphic Statics to analyze the dragonfly wing structure and machine learning models to generate the topology and geometry of a dragonfly wing structure. However, the current workflow involves multiple geometric algorithms and the implementation of complex machine learning models, making it is difficult for designers to follow and use. Therefore, in this paper, we introduce a web-based tool that implements the workflow. It includes (1) the input control panel from the user to define the constraints of the structure; (2) the backend server that proceeds the generation of the structure with Graphic Statics and machine learning models; (3) the output control panel to allow the interaction of the result for frontend display; and (4) the file manager to store and restore the generated result. On the web page, designers can easily input the boundary and parameters of a wing/cantilever structure and generate a funicular structural form in our cloud server. Analytical results such as Minkowski Sum and FEM analysis are shown to the user. Finally, the user can export the STL model for other purposes such as aerodynamic analysis or digital fabrication. © Copyright 2023 Association for Computer Aided Design in Architecture (ACADIA).