| Abstract Scope |
Recent advancements in architected materials enable their extensive applications in energy absorption and impact dissipation, etc. However, traditional design methodologies, including analytical prediction or numerical optimization, are not capable of capturing and replicating the full dynamic responses, attributed to limitations in multiple design objectives, nonlinear behaviors, and intrinsic trial-and-error design process. Herein, we exploited the artificial intelligence approach and established a machine learning based design framework to inversely create metamaterials that achieve target dynamic responses across a wide range of strain rates. Additionally, we developed a revised Molecular Dynamic simulation procedure to replace precisely evaluate the dynamic response of the lattice, which significantly reduces the time span of simulating each design from hours to minutes. Our work provides a rapid architected material design approach for future product design, targeting optimal performance in diverse application scenarios. |