Graph neural networks for mechanical property prediction of 2D fiber composites

Erdem Caliskan; Reza Abedi; Massimiliano Lupo Pasini

doi:10.1016/j.matdes.2025.114500

Graph neural networks for mechanical property prediction of 2D fiber composites

Erdem Caliskan, Reza Abedi, Massimiliano Lupo Pasini

Computational Coupled Physics

Research output: Contribution to journal › Article › peer-review

Abstract

This work investigates the ability of graph neural networks (GNNs) to homogenize 2D fiber composite microstructures. We use different inhomogeneity and anisotropy indices to motivate and show that the Volume Elements (VEs) used in ML methods should ideally be far from their Representative Volume Element (RVE) size limit and, consequently, are notably anisotropic. Hence, training only the isotropic limit properties may not be acceptable. Another aspect is the need to normalize elastic stiffness values for ML, especially when high elastic contrast ratios are encountered between composite phases or in the material set. We introduce a normalization technique based on the mean-field method (MFM) to handle such high contrast ratios and train for the entire stiffness tensor. We show that the proposed GNN approaches exhibit high accuracy and efficiency compared to traditional methods and convolutional neural networks, utilizing unstructured graphs constructed from microstructure topology. Our model successfully predicts the stiffness tensor, peak strength under bulk damage, and brittle fracture initiation strength across diverse microstructure configurations while maintaining high accuracy even for extreme material contrasts and volume fractions. We also present a method to improve prediction accuracy for small dataset sizes using Voronoi partitioning.

Original language	English
Article number	114500
Journal	Materials and Design
Volume	257
DOIs	https://doi.org/10.1016/j.matdes.2025.114500
State	Published - Sep 2025

Funding

This research is sponsored by the Artificial Intelligence Initiative as part of the Laboratory Directed Research and Development (LDRD) Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the US Department of Energy under contract DE-AC05-00OR22725 . This work used resources of the Oak Ridge Leadership Computing Facility, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725 , under INCITE award CPH161 . Reza Abedi acknowledges the partial support of his time from Air Force Office of Scientific Research (AFOSR) under grant number FA9550-22-1-0359 . The authors would also like to thank the University of Tennessee Infrastructure for Scientific Applications and Advanced Computing (ISAAC). A portion of the computation for this work was performed with ISAAC computational resources. This manuscript has been authored in part by UT-Battelle, LLC, under contract DE-AC05-00OR22725 with the US Department of Energy (DOE). The US government retains and the publisher, by accepting the article for publication, acknowledges that the US government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for US government purposes. DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan).The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Erdem Caliskan reports financial support was provided by Oak Ridge Institute for Science and Education (ORISE). If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Massimiliano Lupo Pasini thanks Dr. Vladimir Protopopescu for his valuable feedback in the preparation of the manuscript. Erdem Caliskan would like to thank Professor Katherine Acton for providing the Voronoi dataset. Erdem Caliskan also appreciates the help from Professor Ravindra Duddu for providing the phase-field damage dataset. This research is sponsored by the Artificial Intelligence Initiative as part of the Laboratory Directed Research and Development (LDRD) Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the US Department of Energy under contract DE-AC05-00OR22725. This work used resources of the Oak Ridge Leadership Computing Facility, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725, under INCITE award CPH161. Reza Abedi acknowledges the partial support of his time from Air Force Office of Scientific Research (AFOSR) under grant number FA9550-22-1-0359. The authors would also like to thank the University of Tennessee Infrastructure for Scientific Applications and Advanced Computing (ISAAC). A portion of the computation for this work was performed with ISAAC computational resources.

Keywords

Brittle strength
Fiber composites
Graph neural network
Homogenization
Machine learning

Access to Document

10.1016/j.matdes.2025.114500

Cite this

@article{f1c374e1d44b44f985239a9a79def94c,

title = "Graph neural networks for mechanical property prediction of 2D fiber composites",

abstract = "This work investigates the ability of graph neural networks (GNNs) to homogenize 2D fiber composite microstructures. We use different inhomogeneity and anisotropy indices to motivate and show that the Volume Elements (VEs) used in ML methods should ideally be far from their Representative Volume Element (RVE) size limit and, consequently, are notably anisotropic. Hence, training only the isotropic limit properties may not be acceptable. Another aspect is the need to normalize elastic stiffness values for ML, especially when high elastic contrast ratios are encountered between composite phases or in the material set. We introduce a normalization technique based on the mean-field method (MFM) to handle such high contrast ratios and train for the entire stiffness tensor. We show that the proposed GNN approaches exhibit high accuracy and efficiency compared to traditional methods and convolutional neural networks, utilizing unstructured graphs constructed from microstructure topology. Our model successfully predicts the stiffness tensor, peak strength under bulk damage, and brittle fracture initiation strength across diverse microstructure configurations while maintaining high accuracy even for extreme material contrasts and volume fractions. We also present a method to improve prediction accuracy for small dataset sizes using Voronoi partitioning.",

keywords = "Brittle strength, Fiber composites, Graph neural network, Homogenization, Machine learning",

author = "Erdem Caliskan and Reza Abedi and \{Lupo Pasini\}, Massimiliano",

note = "Publisher Copyright: {\textcopyright} 2025 The Author(s)",

year = "2025",

month = sep,

doi = "10.1016/j.matdes.2025.114500",

language = "English",

volume = "257",

journal = "Materials and Design",

issn = "0264-1275",

publisher = "Elsevier",

}

TY - JOUR

T1 - Graph neural networks for mechanical property prediction of 2D fiber composites

AU - Caliskan, Erdem

AU - Abedi, Reza

AU - Lupo Pasini, Massimiliano

PY - 2025/9

Y1 - 2025/9

N2 - This work investigates the ability of graph neural networks (GNNs) to homogenize 2D fiber composite microstructures. We use different inhomogeneity and anisotropy indices to motivate and show that the Volume Elements (VEs) used in ML methods should ideally be far from their Representative Volume Element (RVE) size limit and, consequently, are notably anisotropic. Hence, training only the isotropic limit properties may not be acceptable. Another aspect is the need to normalize elastic stiffness values for ML, especially when high elastic contrast ratios are encountered between composite phases or in the material set. We introduce a normalization technique based on the mean-field method (MFM) to handle such high contrast ratios and train for the entire stiffness tensor. We show that the proposed GNN approaches exhibit high accuracy and efficiency compared to traditional methods and convolutional neural networks, utilizing unstructured graphs constructed from microstructure topology. Our model successfully predicts the stiffness tensor, peak strength under bulk damage, and brittle fracture initiation strength across diverse microstructure configurations while maintaining high accuracy even for extreme material contrasts and volume fractions. We also present a method to improve prediction accuracy for small dataset sizes using Voronoi partitioning.

AB - This work investigates the ability of graph neural networks (GNNs) to homogenize 2D fiber composite microstructures. We use different inhomogeneity and anisotropy indices to motivate and show that the Volume Elements (VEs) used in ML methods should ideally be far from their Representative Volume Element (RVE) size limit and, consequently, are notably anisotropic. Hence, training only the isotropic limit properties may not be acceptable. Another aspect is the need to normalize elastic stiffness values for ML, especially when high elastic contrast ratios are encountered between composite phases or in the material set. We introduce a normalization technique based on the mean-field method (MFM) to handle such high contrast ratios and train for the entire stiffness tensor. We show that the proposed GNN approaches exhibit high accuracy and efficiency compared to traditional methods and convolutional neural networks, utilizing unstructured graphs constructed from microstructure topology. Our model successfully predicts the stiffness tensor, peak strength under bulk damage, and brittle fracture initiation strength across diverse microstructure configurations while maintaining high accuracy even for extreme material contrasts and volume fractions. We also present a method to improve prediction accuracy for small dataset sizes using Voronoi partitioning.

KW - Brittle strength

KW - Fiber composites

KW - Graph neural network

KW - Homogenization

KW - Machine learning

UR - https://www.scopus.com/pages/publications/105012771843

U2 - 10.1016/j.matdes.2025.114500

DO - 10.1016/j.matdes.2025.114500

M3 - Article

AN - SCOPUS:105012771843

SN - 0264-1275

VL - 257

JO - Materials and Design

JF - Materials and Design

M1 - 114500

ER -

Graph neural networks for mechanical property prediction of 2D fiber composites

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this