Intelligent design of engineered metabarrier based on deep learning. (15th January 2022)

• Record Type: Journal Article
• Authors: Liu, Chen-Xu; Yu, Gui-Lan
• Is Part Of: Composite structures. Volume 280(2022)
• Publication Date: 2022-01-15
• Subjects: Metabarrier -- Intelligent design -- Deep learning -- Local resonance -- Bandgap; Composite construction -- Periodicals; Composites -- Périodiques; 624.18
• DOI: 10.1016/j.compstruct.2021.114911
• Languages: English
• ISSNs: 0263-8223

Abstract

Highlights: An approach based on deep learning is presented to achieve the design of metabarriers rapidly and accurately. "One-to-many" design by deep learning method is realized. A new activation function is developed to improve design accuracy. A new loss function is used to realize the design for mixed waves. P wave, S wave, and mixed waves are considered, respectively. Geometric parameters and material (including soil and rubber) parameters are taken into account simultaneously. Abstract: This study presents an approach based on deep learning to design engineered metabarriers for all body waves. Ten variables, containing geometric and material parameters, are taken into account. Two design cases are considered, and three different wave modes are discussed in each case. In order to increase design accuracy and realize the design for mixed waves, a new activation function and a new loss function are proposed, respectively. The designed results are highly consistent with expectations. It takes a very short time to complete a design and many different results meeting the same target can be given by our method. The deep learning model has great universality, feasibility, rapidity, and accuracy on designing the engineered metabarriers.