2019
DOI: 10.1016/j.actamat.2018.12.045
|View full text |Cite
|
Sign up to set email alerts
|

Establishing structure-property localization linkages for elastic deformation of three-dimensional high contrast composites using deep learning approaches

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1

Citation Types

2
66
0
1

Year Published

2019
2019
2024
2024

Publication Types

Select...
7

Relationship

1
6

Authors

Journals

citations
Cited by 146 publications
(69 citation statements)
references
References 47 publications
2
66
0
1
Order By: Relevance
“…Yang et al [30] present a novel feature-engineering-free approach for localization using deep learning. They used two datasets of contrast-10 (2500 MVEs) and contrast-50 (3000 MVEs) of size 21 × 21 × 21 with varying volume fraction and periodic boundary conditions, and were split into training, validation, and test sets.…”
Section: Multiscale Homogenization and Localization Linkages In High-mentioning
confidence: 99%
See 2 more Smart Citations
“…Yang et al [30] present a novel feature-engineering-free approach for localization using deep learning. They used two datasets of contrast-10 (2500 MVEs) and contrast-50 (3000 MVEs) of size 21 × 21 × 21 with varying volume fraction and periodic boundary conditions, and were split into training, validation, and test sets.…”
Section: Multiscale Homogenization and Localization Linkages In High-mentioning
confidence: 99%
“…Although 3-D CNNs could be used for this problem as well, the dataset here is almost four orders of magnitude larger, and since 3-D CNNs are much more computationally expensive, the authors in Ref. 30 designed a neat workaround to be able to use 2-D CNNs for localization. They accomplished this by treating the 3-D image of size 11 × 11 × 11 as 11 channels of a 2-D images of size 11 × 11, perpendicular to the maximum principal strain direction.…”
Section: Multiscale Homogenization and Localization Linkages In High-mentioning
confidence: 99%
See 1 more Smart Citation
“…D ata-driven methods such as machine learning (ML) and statistical analysis (SA) are efficient toolsets for extracting process-structure-property relation or for designsynthesis-characterization of materials [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18] . ML and SA are able to address large and complex tasks by focusing on the most relevant information in an overwhelming quantity of data while providing similar or better accuracy to the finite element analysis (FEA) and experiment [19][20][21][22] .…”
mentioning
confidence: 99%
“…Tiryaki and Aydin employed ML models to predict the compression strength of heat-treated woods 27 . Deep learning approaches were used for mining structure-property linkages in high contrast composites from simulation datasets, and to establish structure-property localization linkages for elastic deformation of three-dimensional high contrast composites 16,18 .…”
mentioning
confidence: 99%