Optimization of mechanical properties of bio-inspired Voronoi structures by genetic algorithm

Numerous patterns in nature comprise aesthetics structures and intriguing geometries. The geometric factors behind these patterns include area, aspect ratio, circularity, orientation, and regularity. In this study, the quasi-3D porous structures with varying regularities were systematically described and generated by the Voronoi segmentation. Then, genetic algorithms were applied to optimize the mechanical properties of these structures under tension condition. The image-based finite element simulation could calculate each structure's fitness value (strength). Afterward, structures with higher fitness values were selected and underwent the replication process, including crossover and mutation. New structures were generated based on the input structures. The above steps driven by the genetic algorithms were repeated 30 iterations until the tensile strength of the structures was improved and converged. Both simulation and tensile testing results as well as digital image correlation of specimens fabricated by stereolithography 3D printing showed that the Voronoi structures optimized by the genetic algorithm could enhance their stiffness, strength, and toughness values by ∼30%. This research has the potential to be applied in the fields of structural materials and biomimetic micro-aerial vehicles, which require lightweight, strength, and toughness simultaneously.