A Gan-Based Machine Learning Surrogate Model for the Stepwise Prediction of Full-Field Mechanical Behavior of Kirigami-Inspired Metamaterials

Abstract

The exploration of mechanical metamaterials, characterized by unique unit cells with significant macroscopic mechanical properties, is of crucial importance. Traditional methods relying on bioinspired structures often fall short, especially when practical requirements lack biological counterparts. Hence, there is a pressing need to advance methodologies for the expansive design of unit cells, catering to diverse practical demands. Inspired by kirigami structures, a design paradigm gains prominence for its extensive possibilities and unique deformations. However, kirigami-inspired design faces challenges in accurately assessing the deformation history and stress states of diverse units, requiring validation through experiments or simulations, and imposing efficiency constraints on design. Using data generated from a mimetic corrosion algorithm and computational simulations, a GAN-based machine learning surrogate model is built to predict the history of mechanical deformation and stress fields of kirigami-based metamaterials with great accuracy and incredible efficiency, in which a stepwise mapping strategy is applied to augment the comprehension of mechanical information in image processing. In distinct cases covering both explored domains and several unexplored domains, the model achieves effective predictions, demonstrating the robustness and adaptability of the model in extrapolating predictions of both time domain and design space.

Keywords: mechanical metamaterials, Machine learning, kirigami-inspired design, stepwise mapping strategy, mechanical prediction