Rapid inverse design of metamaterials based on prescribed mechanical behavior through machine learning

Designing and printing metamaterials with customizable architectures enables the realization of unprecedented mechanical behaviors. Existing inverse design approaches struggle with multiple design objectives, nonlinear behavior, and manufacturing errors. This paper reports a rapid inverse design methodology using generative machine learning and desktop additive manufacturing to create metamaterials with nearly any desired uniaxial compressive stress-strain curve, accounting for printing errors. Results show 90% fidelity between target and measured behaviors, offering a potential bypass for iterative design-manufacturing cycles.

Chan Soo Ha, Desheng Yao, Zhenpeng Xu, Chenang Liu, Han Liu, Daniel Elkins, Matthew Kile, Vikram Deshpande, Zhenyu Kong, Mathieu Bauchy, Xiaoyu (Rayne) Zheng