Structural stability investigation in bioinspired metamaterials based on glass sponge microstructures

This work presents an innovative design for a lattice microstructure, drawing inspiration from the deep-sea glass sponge Euplectella aspergillum. The computational framework developed enables real-time synchronization between COMSOL Multiphysics and MATLAB, utilizing a genetic algorithm and artificial neural networks to optimize the microstructure. During the optimization procedure, key geometric parameters were adjusted while keeping the representative volume element's fraction constant to maximize the critical load factor under uniaxial vertical compression. The genetic algorithm evaluated numerous parameter combinations, while neural networks predicted the occurrence of either local or global instability for each configuration. Designs lying to global instability were penalized, favoring local instability in the final optimized structure. The results indicated a 65% enhancement in ultimate buckling deformation and a 34% increase in buckling load compared to the unoptimized design.