Exploring the structure-property relations of thin-walled, 2D extruded lattices using neural networks

This paper investigates the structure–property relations of thin-walled lattices, characterized by their cross-sections and heights, under dynamic longitudinal compression. These elucidate interactions different geometric features a design on mechanical response, including energy absorption. We proposed combinatorial, key-based system to generate lattice designs used finite element method simulate response with Johnson–Cook material model. Using an autoencoder, we encoded cross-sectional images lattices into latent feature vectors, which were supplied neural network model predictions. The trained models can accurately predict absorption curves in be extended new outside via transfer learning.