Bidirectional Pattern Recognition and Prediction of Bending-Active Thin Sheets via Artificial Neural Networks

• DOI: 10.3390/electronics14030503
• Publication Date: 2025-01-27T14:42:23Z
• Authors: Yuxin Xie; Xiang Wang; Xinjie Zhou; Qiang Zhou

Currently, active-bending structures and their shape optimization techniques have become a hot topic in the design of spatial freeform buildings. However, form-finding process is usually time-consuming, application finite element methods (FEM) requires huge computational effort. In face these challenges, artificial intelligence great potential for bring many important advantages to this field. paper, we propose novel, data-driven, bidirectional prediction method based on neural networks. It can both forward infer bending deformation shapes thin plate under specific complex conditions reverse boundary necessary given shape. comparison traditional simulation, proposed quicker simpler utilize during facilitates predictions. Communication between construction be facilitated ensure quality efficiency relevant bent structural components. experimentally demonstrated that network control mean value deviation below 40 mm 4 m × 0.5 aluminum plate.