Tampere University of Technology

TUTCRIS Research Portal

Domain decomposition with subdomain pre-processing for finite element modelling of transformers with stranded conductors

Research output: Contribution to journalArticleScientificpeer-review

Standard

Domain decomposition with subdomain pre-processing for finite element modelling of transformers with stranded conductors. / Gürbüz, Ismet Tuna; Lehikoinen, Antti; Aydin, Ugur; Rasilo, Paavo; Takala, Eelis; Belahcen, Anouar.

In: AIP Advances, Vol. 10, 015133, 2020.

Research output: Contribution to journalArticleScientificpeer-review

Harvard

APA

Vancouver

Author

Gürbüz, Ismet Tuna ; Lehikoinen, Antti ; Aydin, Ugur ; Rasilo, Paavo ; Takala, Eelis ; Belahcen, Anouar. / Domain decomposition with subdomain pre-processing for finite element modelling of transformers with stranded conductors. In: AIP Advances. 2020 ; Vol. 10.

Bibtex - Download

@article{48e4ff630cb547a884800f83eeee8811,
title = "Domain decomposition with subdomain pre-processing for finite element modelling of transformers with stranded conductors",
abstract = "Finite element analysis of transformers with stranded conductors can be computationally heavy due to the fine mesh required for winding domains. In this paper, a domain decomposition method with subdomain pre-processing is proposed to overcome with this problem. In order to solve the field-circuit problem with finite element analysis, the winding domain is modelled initially. Next, the reduced matrix system is obtained from the winding domain formulation. The solution of the entire problem is obtained by solving the winding domain formulation and the reduced matrix system separately. The method is tested on a transformer model in no load operation. Accurate solutions are obtained in both frequency and time domains 7-9 times faster than when using brute-force method.",
author = "G{\"u}rb{\"u}z, {Ismet Tuna} and Antti Lehikoinen and Ugur Aydin and Paavo Rasilo and Eelis Takala and Anouar Belahcen",
note = "EXT={"}Lehikoinen, Antti{"}",
year = "2020",
doi = "10.1063/1.5130076",
language = "English",
volume = "10",
journal = "AIP Advances",
issn = "2158-3226",
publisher = "American Institute of Physics",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - Domain decomposition with subdomain pre-processing for finite element modelling of transformers with stranded conductors

AU - Gürbüz, Ismet Tuna

AU - Lehikoinen, Antti

AU - Aydin, Ugur

AU - Rasilo, Paavo

AU - Takala, Eelis

AU - Belahcen, Anouar

N1 - EXT="Lehikoinen, Antti"

PY - 2020

Y1 - 2020

N2 - Finite element analysis of transformers with stranded conductors can be computationally heavy due to the fine mesh required for winding domains. In this paper, a domain decomposition method with subdomain pre-processing is proposed to overcome with this problem. In order to solve the field-circuit problem with finite element analysis, the winding domain is modelled initially. Next, the reduced matrix system is obtained from the winding domain formulation. The solution of the entire problem is obtained by solving the winding domain formulation and the reduced matrix system separately. The method is tested on a transformer model in no load operation. Accurate solutions are obtained in both frequency and time domains 7-9 times faster than when using brute-force method.

AB - Finite element analysis of transformers with stranded conductors can be computationally heavy due to the fine mesh required for winding domains. In this paper, a domain decomposition method with subdomain pre-processing is proposed to overcome with this problem. In order to solve the field-circuit problem with finite element analysis, the winding domain is modelled initially. Next, the reduced matrix system is obtained from the winding domain formulation. The solution of the entire problem is obtained by solving the winding domain formulation and the reduced matrix system separately. The method is tested on a transformer model in no load operation. Accurate solutions are obtained in both frequency and time domains 7-9 times faster than when using brute-force method.

U2 - 10.1063/1.5130076

DO - 10.1063/1.5130076

M3 - Article

VL - 10

JO - AIP Advances

JF - AIP Advances

SN - 2158-3226

M1 - 015133

ER -