Logo image
Bio-inspired structural optimization of three-dimensional Voronoi structures using genetic algorithms: Inspirations from avian wing bones
期刊文章   開放取用(OA)

Bio-inspired structural optimization of three-dimensional Voronoi structures using genetic algorithms: Inspirations from avian wing bones

C. C Lin, C. C. Tung, Y. Y. Chuang柏宇 陳
Materials & Design, 卷.248
2024

摘要

Avian wing bones;Genetic Algorithm;Mechanical Properties;Optimization;Struts;Three-dimensional Voronoi

Birds have evolved lightweight yet strong wing bones. Inside the wing bones, struts formed three-dimensional Voronoi structures, representing adaptations toward lightness. Inspired by this natural design, we proposed structural optimization method to optimize the mechanical properties of three-dimensional Voronoi structures using genetic algorithms(GA). The optimization process begins with the generation of three-dimensional porous Voronoi structures. Then, beam elements were extracted for finite element simulations as performance indicators. Finally, genes of structures with superior mechanical properties were retained and inherited through iterative evolution until optimal solutions were achieved. Samples were fabricated using additive manufacturing techniques, followed by compression testing to assess their mechanical properties. Experimental results showed that optimized Voronoi structures increased peak load capacity by 25.9% for Voronoi structures with eight Voronoi seeds. With regularity constraint, the peak load increased by 58.6% in lower regularity samples and the energy absorption increased by 39.3% in higher regularity ones. Normal vector projections revealed the principles behind the optimizations, and the Hausdorff distance measured structural similarity, validating the effectiveness of the genetic algorithm. Overall, we developed a novel method for optimizing 3D Voronoi structures using genetic algorithms, offering significant potential for designing porous, lightweight structures applicable across various fields. 

檔案與連結 (1)

url
https://doi.org/10.1016/j.matdes.2024.113501檢視
已出版(紀錄版本) 開放

相關連結

指標

1 檢視次數

詳細資料

Logo image