A Data-driven Framework for Mechanics Analysis of Origami Structures Using Machine Learning Methods

Origami-based metamaterials have been widely used in the field of energy-absorption structures. However, due to substantial computational costs and time consumption, designing or optimising origami-based metamaterials is a highly iterative process. Therefore, a new data-driven framework based on machine learning has been proposed in this thesis to assist in the mechanical analysis of origami structures. The feasibility and value of the framework were presented through the mechanical analysis of the Miura-origami-based structure in-plane quasi-static compression, and the framework applied to predict the mechanical behaviours of a non-rigid foldable square-twist origami pattern (Type 1), which lacks an analytical or empirical solution.