Accelerated discoveries of mechanical properties of graphene using machine learning and high-throughput computation

Zesheng Zhang, Yang Hong, Bo Hou, Zhongtao Zhang, Mehrdad Negahban, Jingchao Zhang

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Machine learning (ML) has been vastly used in various fields, but its application in engineering science remains in infancy. In this work, for the first time, different machine learning algorithms and artificial neural network (ANN) structures are used to predict the mechanical properties of single-layer graphene under various impact factors of system temperature, strain rate, vacancy defect and chirality. The predictions include fracture strain, fracture strength and Young's modulus. High throughput computation (HTC) combined with classical molecular dynamics (MD) simulation is used to generate the training dataset for the ML models. It was discovered that both temperature and vacancy defect have negative effects on the predicted properties while strain rate has positive correlations with the prediction results. The stochastic gradient descent (SGD) method could not properly capture the effects of the different impact factors on the mechanical properties of graphene, while k-nearest neighbors (KNN), support vector machine (SVM), decision tree (DT) and ANN provided desirable prediction results. Discoveries in this work provide new perspectives on the study of mechanical properties using state-of-the-art computational methods.

Original languageEnglish (US)
Pages (from-to)115-123
Number of pages9
JournalCarbon
Volume148
DOIs
StatePublished - Jul 1 2019

Fingerprint

Graphite
Graphene
Learning systems
Throughput
Mechanical properties
Vacancies
Strain rate
Neural networks
Defects
Chirality
Decision trees
Computational methods
Learning algorithms
Support vector machines
Molecular dynamics
Fracture toughness
Elastic moduli
Temperature
Computer simulation

ASJC Scopus subject areas

  • Chemistry(all)
  • Materials Science(all)

Cite this

Accelerated discoveries of mechanical properties of graphene using machine learning and high-throughput computation. / Zhang, Zesheng; Hong, Yang; Hou, Bo; Zhang, Zhongtao; Negahban, Mehrdad; Zhang, Jingchao.

In: Carbon, Vol. 148, 01.07.2019, p. 115-123.

Research output: Contribution to journalArticle

Zhang, Zesheng ; Hong, Yang ; Hou, Bo ; Zhang, Zhongtao ; Negahban, Mehrdad ; Zhang, Jingchao. / Accelerated discoveries of mechanical properties of graphene using machine learning and high-throughput computation. In: Carbon. 2019 ; Vol. 148. pp. 115-123.
@article{04da19ec086747c1a1aefc752773fa25,
title = "Accelerated discoveries of mechanical properties of graphene using machine learning and high-throughput computation",
abstract = "Machine learning (ML) has been vastly used in various fields, but its application in engineering science remains in infancy. In this work, for the first time, different machine learning algorithms and artificial neural network (ANN) structures are used to predict the mechanical properties of single-layer graphene under various impact factors of system temperature, strain rate, vacancy defect and chirality. The predictions include fracture strain, fracture strength and Young's modulus. High throughput computation (HTC) combined with classical molecular dynamics (MD) simulation is used to generate the training dataset for the ML models. It was discovered that both temperature and vacancy defect have negative effects on the predicted properties while strain rate has positive correlations with the prediction results. The stochastic gradient descent (SGD) method could not properly capture the effects of the different impact factors on the mechanical properties of graphene, while k-nearest neighbors (KNN), support vector machine (SVM), decision tree (DT) and ANN provided desirable prediction results. Discoveries in this work provide new perspectives on the study of mechanical properties using state-of-the-art computational methods.",
author = "Zesheng Zhang and Yang Hong and Bo Hou and Zhongtao Zhang and Mehrdad Negahban and Jingchao Zhang",
year = "2019",
month = "7",
day = "1",
doi = "10.1016/j.carbon.2019.03.046",
language = "English (US)",
volume = "148",
pages = "115--123",
journal = "Carbon",
issn = "0008-6223",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Accelerated discoveries of mechanical properties of graphene using machine learning and high-throughput computation

AU - Zhang, Zesheng

AU - Hong, Yang

AU - Hou, Bo

AU - Zhang, Zhongtao

AU - Negahban, Mehrdad

AU - Zhang, Jingchao

PY - 2019/7/1

Y1 - 2019/7/1

N2 - Machine learning (ML) has been vastly used in various fields, but its application in engineering science remains in infancy. In this work, for the first time, different machine learning algorithms and artificial neural network (ANN) structures are used to predict the mechanical properties of single-layer graphene under various impact factors of system temperature, strain rate, vacancy defect and chirality. The predictions include fracture strain, fracture strength and Young's modulus. High throughput computation (HTC) combined with classical molecular dynamics (MD) simulation is used to generate the training dataset for the ML models. It was discovered that both temperature and vacancy defect have negative effects on the predicted properties while strain rate has positive correlations with the prediction results. The stochastic gradient descent (SGD) method could not properly capture the effects of the different impact factors on the mechanical properties of graphene, while k-nearest neighbors (KNN), support vector machine (SVM), decision tree (DT) and ANN provided desirable prediction results. Discoveries in this work provide new perspectives on the study of mechanical properties using state-of-the-art computational methods.

AB - Machine learning (ML) has been vastly used in various fields, but its application in engineering science remains in infancy. In this work, for the first time, different machine learning algorithms and artificial neural network (ANN) structures are used to predict the mechanical properties of single-layer graphene under various impact factors of system temperature, strain rate, vacancy defect and chirality. The predictions include fracture strain, fracture strength and Young's modulus. High throughput computation (HTC) combined with classical molecular dynamics (MD) simulation is used to generate the training dataset for the ML models. It was discovered that both temperature and vacancy defect have negative effects on the predicted properties while strain rate has positive correlations with the prediction results. The stochastic gradient descent (SGD) method could not properly capture the effects of the different impact factors on the mechanical properties of graphene, while k-nearest neighbors (KNN), support vector machine (SVM), decision tree (DT) and ANN provided desirable prediction results. Discoveries in this work provide new perspectives on the study of mechanical properties using state-of-the-art computational methods.

UR - http://www.scopus.com/inward/record.url?scp=85063544575&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85063544575&partnerID=8YFLogxK

U2 - 10.1016/j.carbon.2019.03.046

DO - 10.1016/j.carbon.2019.03.046

M3 - Article

VL - 148

SP - 115

EP - 123

JO - Carbon

JF - Carbon

SN - 0008-6223

ER -