Magnetic Properties Identification by Using a Bi-Objective Optimal Multi-Fidelity Neural Network

Research output: Contribution to journalArticleResearchpeer review

Authors

External Research Organisations

  • University of Pavia
View graph of relations

Details

Original languageEnglish
JournalIEEE transactions on magnetics
Volume57
Issue number6
Early online date24 Mar 2021
Publication statusPublished - 17 May 2021

Abstract

In order to identify the magnetic properties of magnetic steel, the synergy between the data arising from the experimental activity, an FE model, and the use of a multi-fidelity surrogate could relieve the burden of the total cost. A neural network, with as many outputs as fidelity levels, is adopted in quality of metamodel to describe the forward problem [forward neural network (FNN)]. FNN is trained using multiple losses aiming at getting a robust surrogate that is poorly sensitive to the chosen norm. This makes it bi-objective optimal since several error metrics are simultaneously minimized. In addition, a conjugate, inverse net (INNCJ) is built, which is a ready-to-use tool for inverse properties identification, since no optimization runs are required. Its performances are compared to those obtained with a transfer learning-based approach (INNTR) and a single-fidelity inverse neural network (INNSF). Finally, a real $B - H$ curve identification task has been solved, thereby validating the conjugate inverse net.

Keywords

    Bi-objective optimality, inverse problem, magnetic permeability, multi-fidelity, neural network

ASJC Scopus subject areas

Cite this

Magnetic Properties Identification by Using a Bi-Objective Optimal Multi-Fidelity Neural Network. / Baldan, Marco; Barba, Paolo Di; Nacke, Bernard.
In: IEEE transactions on magnetics, Vol. 57, No. 6, 17.05.2021.

Research output: Contribution to journalArticleResearchpeer review

Baldan M, Barba PD, Nacke B. Magnetic Properties Identification by Using a Bi-Objective Optimal Multi-Fidelity Neural Network. IEEE transactions on magnetics. 2021 May 17;57(6). Epub 2021 Mar 24. doi: 10.1109/TMAG.2021.3068705
Download
@article{8e1540c5b6804423a139fa218725a65d,
title = "Magnetic Properties Identification by Using a Bi-Objective Optimal Multi-Fidelity Neural Network",
abstract = "In order to identify the magnetic properties of magnetic steel, the synergy between the data arising from the experimental activity, an FE model, and the use of a multi-fidelity surrogate could relieve the burden of the total cost. A neural network, with as many outputs as fidelity levels, is adopted in quality of metamodel to describe the forward problem [forward neural network (FNN)]. FNN is trained using multiple losses aiming at getting a robust surrogate that is poorly sensitive to the chosen norm. This makes it bi-objective optimal since several error metrics are simultaneously minimized. In addition, a conjugate, inverse net (INNCJ) is built, which is a ready-to-use tool for inverse properties identification, since no optimization runs are required. Its performances are compared to those obtained with a transfer learning-based approach (INNTR) and a single-fidelity inverse neural network (INNSF). Finally, a real $B - H$ curve identification task has been solved, thereby validating the conjugate inverse net.",
keywords = "Bi-objective optimality, inverse problem, magnetic permeability, multi-fidelity, neural network",
author = "Marco Baldan and Barba, {Paolo Di} and Bernard Nacke",
year = "2021",
month = may,
day = "17",
doi = "10.1109/TMAG.2021.3068705",
language = "English",
volume = "57",
journal = "IEEE transactions on magnetics",
issn = "0018-9464",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "6",

}

Download

TY - JOUR

T1 - Magnetic Properties Identification by Using a Bi-Objective Optimal Multi-Fidelity Neural Network

AU - Baldan, Marco

AU - Barba, Paolo Di

AU - Nacke, Bernard

PY - 2021/5/17

Y1 - 2021/5/17

N2 - In order to identify the magnetic properties of magnetic steel, the synergy between the data arising from the experimental activity, an FE model, and the use of a multi-fidelity surrogate could relieve the burden of the total cost. A neural network, with as many outputs as fidelity levels, is adopted in quality of metamodel to describe the forward problem [forward neural network (FNN)]. FNN is trained using multiple losses aiming at getting a robust surrogate that is poorly sensitive to the chosen norm. This makes it bi-objective optimal since several error metrics are simultaneously minimized. In addition, a conjugate, inverse net (INNCJ) is built, which is a ready-to-use tool for inverse properties identification, since no optimization runs are required. Its performances are compared to those obtained with a transfer learning-based approach (INNTR) and a single-fidelity inverse neural network (INNSF). Finally, a real $B - H$ curve identification task has been solved, thereby validating the conjugate inverse net.

AB - In order to identify the magnetic properties of magnetic steel, the synergy between the data arising from the experimental activity, an FE model, and the use of a multi-fidelity surrogate could relieve the burden of the total cost. A neural network, with as many outputs as fidelity levels, is adopted in quality of metamodel to describe the forward problem [forward neural network (FNN)]. FNN is trained using multiple losses aiming at getting a robust surrogate that is poorly sensitive to the chosen norm. This makes it bi-objective optimal since several error metrics are simultaneously minimized. In addition, a conjugate, inverse net (INNCJ) is built, which is a ready-to-use tool for inverse properties identification, since no optimization runs are required. Its performances are compared to those obtained with a transfer learning-based approach (INNTR) and a single-fidelity inverse neural network (INNSF). Finally, a real $B - H$ curve identification task has been solved, thereby validating the conjugate inverse net.

KW - Bi-objective optimality

KW - inverse problem

KW - magnetic permeability

KW - multi-fidelity

KW - neural network

UR - http://www.scopus.com/inward/record.url?scp=85103254054&partnerID=8YFLogxK

U2 - 10.1109/TMAG.2021.3068705

DO - 10.1109/TMAG.2021.3068705

M3 - Article

AN - SCOPUS:85103254054

VL - 57

JO - IEEE transactions on magnetics

JF - IEEE transactions on magnetics

SN - 0018-9464

IS - 6

ER -

By the same author(s)