期刊
EXTREME MECHANICS LETTERS
卷 52, 期 -, 页码 -出版社
ELSEVIER
DOI: 10.1016/j.eml.2022.101645
关键词
Machine learning; Neural networks; Plasticity; Finite element method
资金
- Laboratory Directed Re-search & Development Program at Idaho National Laboratory un-der the Department of Energy (DOE) Idaho Operations Office (an agency of the U.S. Government) [DE-AC07-05ID145142]
- National Science Foundation [DMR-1810720]
This article introduces a hybrid finite element method based on neural networks, which learns the constitutive relations of materials from full-field data. By using the full-field data of non-uniform displacement fields, neural networks can be efficiently trained to accurately predict the constitutive relations of materials.
Neural networks (NNs) have demonstrated strong capabilities of learning constitutive relations from big data. However, most NN-based constitutive models require experimental data from a considerable number of stress-strain paths that are expensive to collect. Here, we develop a hybrid finite element method - NN (FEM-NN) framework for learning the constitutive relations from full-field data. As a result, the non-uniform displacement field from a deformed sample with geometrical inhomogeneities can be used for training NNs. Such full-field data have the advantage of providing many different stress-strain paths at different locations in the sample by a single test, thereby enabling the highly efficient training of NNs. We apply FEM-NN simulations to learn the constitutive relations of several model materials characterized by rate-independent J2 plasticity. These FEM-NN studies demonstrate that the trained NNs produce the constitutive relations of plasticity with high accuracy and efficiency.(c) 2022 Elsevier Ltd. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据