Teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials

Abstract We propose a deep neural network (DNN) as a fast surrogate model for local stress calculations in inhomogeneous non-linear materials. We show that the DNN predicts the local stresses with 3.8% mean absolute percentage error (MAPE) for the case of heterogeneous elastic media and a mechanical...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Jaber Rezaei Mianroodi, Nima H. Siboni, Dierk Raabe
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2021
Materias:	Materials of engineering and construction. Mechanics of materials TA401-492 Computer software QA76.75-76.765
Acceso en línea:	https://doaj.org/article/7ffda2b990bd48ada3fb0971fd99c195
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:7ffda2b990bd48ada3fb0971fd99c195
record_format	dspace
spelling	oai:doaj.org-article:7ffda2b990bd48ada3fb0971fd99c1952021-12-02T18:18:32ZTeaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials10.1038/s41524-021-00571-z2057-3960https://doaj.org/article/7ffda2b990bd48ada3fb0971fd99c1952021-07-01T00:00:00Zhttps://doi.org/10.1038/s41524-021-00571-zhttps://doaj.org/toc/2057-3960Abstract We propose a deep neural network (DNN) as a fast surrogate model for local stress calculations in inhomogeneous non-linear materials. We show that the DNN predicts the local stresses with 3.8% mean absolute percentage error (MAPE) for the case of heterogeneous elastic media and a mechanical contrast of up to factor of 1.5 among neighboring domains, while performing 103 times faster than spectral solvers. The DNN model proves suited for reproducing the stress distribution in geometries different from those used for training. In the case of elasto-plastic materials with up to 4 times mechanical contrast in yield stress among adjacent regions, the trained model simulates the micromechanics with a MAPE of 6.4% in one single forward evaluation of the network, without any iteration. The results reveal an efficient approach to solve non-linear mechanical problems, with an acceleration up to a factor of 8300 for elastic-plastic materials compared to typical solvers.Jaber Rezaei MianroodiNima H. SiboniDierk RaabeNature PortfolioarticleMaterials of engineering and construction. Mechanics of materialsTA401-492Computer softwareQA76.75-76.765ENnpj Computational Materials, Vol 7, Iss 1, Pp 1-10 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Materials of engineering and construction. Mechanics of materials TA401-492 Computer software QA76.75-76.765
spellingShingle	Materials of engineering and construction. Mechanics of materials TA401-492 Computer software QA76.75-76.765 Jaber Rezaei Mianroodi Nima H. Siboni Dierk Raabe Teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials
description	Abstract We propose a deep neural network (DNN) as a fast surrogate model for local stress calculations in inhomogeneous non-linear materials. We show that the DNN predicts the local stresses with 3.8% mean absolute percentage error (MAPE) for the case of heterogeneous elastic media and a mechanical contrast of up to factor of 1.5 among neighboring domains, while performing 103 times faster than spectral solvers. The DNN model proves suited for reproducing the stress distribution in geometries different from those used for training. In the case of elasto-plastic materials with up to 4 times mechanical contrast in yield stress among adjacent regions, the trained model simulates the micromechanics with a MAPE of 6.4% in one single forward evaluation of the network, without any iteration. The results reveal an efficient approach to solve non-linear mechanical problems, with an acceleration up to a factor of 8300 for elastic-plastic materials compared to typical solvers.
format	article
author	Jaber Rezaei Mianroodi Nima H. Siboni Dierk Raabe
author_facet	Jaber Rezaei Mianroodi Nima H. Siboni Dierk Raabe
author_sort	Jaber Rezaei Mianroodi
title	Teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials
title_short	Teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials
title_full	Teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials
title_fullStr	Teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials
title_full_unstemmed	Teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials
title_sort	teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials
publisher	Nature Portfolio
publishDate	2021
url	https://doaj.org/article/7ffda2b990bd48ada3fb0971fd99c195
work_keys_str_mv	AT jaberrezaeimianroodi teachingsolidmechanicstoartificialintelligenceafastsolverforheterogeneousmaterials AT nimahsiboni teachingsolidmechanicstoartificialintelligenceafastsolverforheterogeneousmaterials AT dierkraabe teachingsolidmechanicstoartificialintelligenceafastsolverforheterogeneousmaterials
_version_	1718378284513755136

Teaching solid mechanics to artificial intelligence—a fast solver for heterogeneous materials

Ejemplares similares