Structural optimisation of cross-chiral metamaterial structures via genetic algorithm

Auxetic metamaterials have garnered the attention from researchers in the past decades due to their advanced mechanical performance unparalleled by naturally occurring materials. Cross-chiral auxetics, as a class of novel auxetic metamaterials, have become the centre of focus in this field of research. In this study, we combined the optimisation prowess of an ANN-aided GA with the modelling capability of FEA, to tackle the puzzling issue of accurately replicating the complex anisotropicity of CFRP material fabricated by FDM 3D printing. In this work, we present a novel approach that utilises a local periodic cell with porosity for mechanical homogenisation of CFRP materials, as well as a curvilinear system established to capture the locally varying orientation of the 3D printed structure. An ANN surrogate model was then trained based on FEA results which was utilised for a GA optimisation to improve the performance of cross-chiral structures printed with CFRP filaments. The optimised structure displays 13.2% greater auxeticity, as well as superior stress distribution, energy absorption and buckling resistance. DIC technology was implemented to verify FEA results. Good coherence was observed between the experimental and simulation data, indicating accurate prediction from the novel model.