ACED: Redefining the Design of Mechanical Metamaterials via Generative Models

Mechanical metamaterials (MMs), with their unique structural configurations rather than their chemical compositions, offer a wide range of unprecedented mechanical properties, from enhanced toughness and energy absorption to advanced vibration damping and soundproofing capabilities. Recent advancements in additive manufacturing have enabled the creation of these intricate geometries, leading to materials that excel in diverse applications, including safety gear, aerospace, noise-canceling technologies, and impact-resistant devices. Designing MMs today involves a complex, time-consuming iterative process where experts intuitively define designs, validate them with physics-based simulations, and refine them through trial and error. Recent advances in generative AI and machine learning offer the potential to disrupt this design cycle by automating and accelerating the process. This research proposes the development of a computational pipeline that integrates AI-driven optimization techniques with advanced simulations to streamline MM design. Using a graph-based representation of MM structures, the system effectively reduces the design space of general 3D MMs from hundreds of thousands of dimensions to a compact space with 10–100 dimensions. The system will incorporate advanced optimization techniques that learn efficient design solutions and enable the transfer of insights across different material properties, for example, from stiffness to shear strength or impact resistance. Furthermore, the approach facilitates a more comprehensive understanding of material performance, enabling designs that balance multiple desired properties. Tested primarily on MMs, the approach will also be applicable to a broad range of scientific and engineering problems requiring the design of complex, high-performance materials. If successful, this project will establish a new class of AI-driven design tools that enable faster, more flexible, and sustainable material development. The project will also provide open-source tools, encouraging further research and collaboration in the field. The educational goal of this project is to create tools that inspire K–12 students to consider STEM pathways and to disseminate new curricula adapted to interdisciplinary research. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.