國立虎尾科技大學 |

Accelerating Phase-Field Simulations of Materials Microstructures by Machine Learning.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Accelerating Phase-Field Simulations of Materials Microstructures by Machine Learning./
Author:	Dai, Minyi.
Description:	1 online resource (150 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-11, Section: B.
Contained By:	Dissertations Abstracts International85-11B.
Subject:	Materials science. -
Online resource:	click for full text (PQDT)
ISBN:	9798382720685

Accelerating Phase-Field Simulations of Materials Microstructures by Machine Learning.
Dai, Minyi.

Accelerating Phase-Field Simulations of Materials Microstructures by Machine Learning. - 1 online resource (150 pages)

Source: Dissertations Abstracts International, Volume: 85-11, Section: B.

Thesis (Ph.D.)--The University of Wisconsin - Madison, 2024.

Includes bibliographical references

Phase-field modeling is a mesoscale computational method that can simulate the spatial and temporal evolution of microstructures. With no need to explicit track the interface between different phases, this method has been successfully applied to a variety of material system and physical phenomena. In this thesis, its application to magnetization switching and effective properties calculation is discussed first.On the other hand, the small grid size required to spatially resolve the narrow (down to 1-2 nanometers) diffuse interfaces in heterogeneous materials results in a large computational load, especially for large-scale microstructures. Moreover, for complex physical phenomena, it is challenging to accurately determine all the thermodynamic and kinetic parameters involved in a phase-field model.To address these issues, machine learning models to accelerate the prediction of microstructure properties are discussed. Specifically, simple machine learning regression models are employed to predict the phase diagram of magnetization switching behaviors in magnetic nanodisks. In addition, graph neural networks are used to predict various properties of polycrystalline materials, including magnetostriction, electrical conductivity and Young's modulus. These machine learning approaches have demonstrated high accuracy and computational efficiency. Furthermore, the interpretability and the transferability of these models are also discussed.The final chapter outlines future directions to integrate phase-field modeling with machine learning, aiming to further enhance the fast and accurate prediction of material properties and dynamics.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798382720685Subjects--Topical Terms:

557839
Materials science.
Subjects--Index Terms:

Phase-field modelingIndex Terms--Genre/Form:

554714
Electronic books.

Accelerating Phase-Field Simulations of Materials Microstructures by Machine Learning.
LDR:03038ntm a22003977 4500 001 1151926
005 20241118135910.5
006 m o d
007 cr mn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798382720685
035 $a (MiAaPQ)AAI31299760
035 $a AAI31299760
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Dai, Minyi. $3 1478770
245 1 0 $a Accelerating Phase-Field Simulations of Materials Microstructures by Machine Learning.
264 0 $c 2024
300 $a 1 online resource (150 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-11, Section: B.
500 $a Advisor: Hu, Jiamian.
502 $a Thesis (Ph.D.)--The University of Wisconsin - Madison, 2024.
504 $a Includes bibliographical references
520 $a Phase-field modeling is a mesoscale computational method that can simulate the spatial and temporal evolution of microstructures. With no need to explicit track the interface between different phases, this method has been successfully applied to a variety of material system and physical phenomena. In this thesis, its application to magnetization switching and effective properties calculation is discussed first.On the other hand, the small grid size required to spatially resolve the narrow (down to 1-2 nanometers) diffuse interfaces in heterogeneous materials results in a large computational load, especially for large-scale microstructures. Moreover, for complex physical phenomena, it is challenging to accurately determine all the thermodynamic and kinetic parameters involved in a phase-field model.To address these issues, machine learning models to accelerate the prediction of microstructure properties are discussed. Specifically, simple machine learning regression models are employed to predict the phase diagram of magnetization switching behaviors in magnetic nanodisks. In addition, graph neural networks are used to predict various properties of polycrystalline materials, including magnetostriction, electrical conductivity and Young's modulus. These machine learning approaches have demonstrated high accuracy and computational efficiency. Furthermore, the interpretability and the transferability of these models are also discussed.The final chapter outlines future directions to integrate phase-field modeling with machine learning, aiming to further enhance the fast and accurate prediction of material properties and dynamics.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Materials science. $3 557839
650 4 $a Electromagnetics. $3 1178899
650 4 $a Nanoscience. $3 632473
653 $a Phase-field modeling
653 $a Microstructures
653 $a Nanodisks
653 $a Machine learning
653 $a Nanometers
655 7 $a Electronic books. $2 local $3 554714
690 $a 0794
690 $a 0565
690 $a 0607
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a The University of Wisconsin - Madison. $b Materials Science and Engineering. $3 1478683
773 0 $t Dissertations Abstracts International $g 85-11B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31299760 $z click for full text (PQDT)