Composite materials are valued in architecture for their remarkable strength-to-weight ratio and ability to shape intricate structures. However, conventional methods relying on single-use molds raise environmental concerns. Recent advancements in moldless fabrication, particularly self-morphing techniques, leverage geometric frustration— internal stresses generated by material architecture. Uniaxial shrinkage in composites, 

traditionally seen as distortions, can be harnessed to create a self-shaping mechanism, enabling the achievement of complex geometries by varying fiber orientations. This paper addresses the inverse problem of self-morphing composites, aiming at the generation of production plans from desired designs for morphing. We propose leveraging machine learning, notably Convolutional Neural Networks (CNNs), to predict fiber layouts using 2D data matrices. The paper outlines the use of simulations to construct a dataset for training CNN models to predict the fiber layouts required to achieve design geometry. The goal of this work is to advance the understanding of self-morphing techniques and work towards their implementation in architectural fabrication 

​

When: August 2024

Where: IASS annual conference, Zurich Switzerland

With: Gal Kapon, Guy Austern