Tampere University of Technology

TUTCRIS Research Portal

Knowledge-based optimization of artificial neural network topology for additive manufacturing process modeling: a case study for fused deposition modeling: A new approach and case study for fused deposition modeling

Research output: Contribution to journalArticleScientificpeer-review

Standard

Knowledge-based optimization of artificial neural network topology for additive manufacturing process modeling: a case study for fused deposition modeling : A new approach and case study for fused deposition modeling. / Nagarajan, Hari P. N.; Mokhtarian, Hossein; Jafarian, Hesam; Dimassi, Saoussen ; Bakrani Balani, Shahriar; Hamedi, Azarakhsh; Coatanea, Eric; Wang, G. Gary; Kari R., Haapala.

In: Journal of Mechanical Design, Vol. 141, No. 2, 02.2019.

Research output: Contribution to journalArticleScientificpeer-review

Harvard

APA

Vancouver

Author

Nagarajan, Hari P. N. ; Mokhtarian, Hossein ; Jafarian, Hesam ; Dimassi, Saoussen ; Bakrani Balani, Shahriar ; Hamedi, Azarakhsh ; Coatanea, Eric ; Wang, G. Gary ; Kari R., Haapala. / Knowledge-based optimization of artificial neural network topology for additive manufacturing process modeling: a case study for fused deposition modeling : A new approach and case study for fused deposition modeling. In: Journal of Mechanical Design. 2019 ; Vol. 141, No. 2.

Bibtex - Download

@article{e3b8429335a043e0867a82bdf88e58c3,
title = "Knowledge-based optimization of artificial neural network topology for additive manufacturing process modeling: a case study for fused deposition modeling: A new approach and case study for fused deposition modeling",
abstract = "Additive manufacturing (AM) continues to rise in popularity due to its various advantages over traditional manufacturing processes. AM interests industry, but achieving repeatable production quality remains problematic for many AM technologies. Thus, modeling the different process variables in AM using modeling techniques, such as, machine learning, can be highly beneficial in creating useful knowledge of the process. Such developed artificial neural network models would aid designers and manufacturers to make informed decisions about their products and processes. However, accurately defining an artificial neural network topology is challenging due to the need to integrate AM system behavior during modeling. Towards that goal, an approach combining dimensional analysis conceptual modeling (DACM), experimental sampling, factors selection, and modeling based on Knowledge-Based Artificial Neural Network (KB-ANN) is proposed. This approach integrates existing literature and expert knowledge of the AM process to implement system behavior centered topology optimization of the knowledge-based artificial neural network model. The usefulness of the method is demonstrated using a case study to model wall thickness, height of part, and total mass of the part in a Fused Deposition Modeling (FDM) process. The KB-ANN based model for FDM has better performance and generalization model with low mean squared error in comparison to a conventional ANN.",
author = "Nagarajan, {Hari P. N.} and Hossein Mokhtarian and Hesam Jafarian and Saoussen Dimassi and {Bakrani Balani}, Shahriar and Azarakhsh Hamedi and Eric Coatanea and Wang, {G. Gary} and {Kari R.}, Haapala",
year = "2019",
month = "2",
doi = "10.1115/1.4042084",
language = "English",
volume = "141",
journal = "Journal of Mechanical Design",
issn = "1050-0472",
publisher = "American Society of Mechanical Engineers",
number = "2",

}

RIS (suitable for import to EndNote) - Download

TY - JOUR

T1 - Knowledge-based optimization of artificial neural network topology for additive manufacturing process modeling: a case study for fused deposition modeling

T2 - A new approach and case study for fused deposition modeling

AU - Nagarajan, Hari P. N.

AU - Mokhtarian, Hossein

AU - Jafarian, Hesam

AU - Dimassi, Saoussen

AU - Bakrani Balani, Shahriar

AU - Hamedi, Azarakhsh

AU - Coatanea, Eric

AU - Wang, G. Gary

AU - Kari R., Haapala

PY - 2019/2

Y1 - 2019/2

N2 - Additive manufacturing (AM) continues to rise in popularity due to its various advantages over traditional manufacturing processes. AM interests industry, but achieving repeatable production quality remains problematic for many AM technologies. Thus, modeling the different process variables in AM using modeling techniques, such as, machine learning, can be highly beneficial in creating useful knowledge of the process. Such developed artificial neural network models would aid designers and manufacturers to make informed decisions about their products and processes. However, accurately defining an artificial neural network topology is challenging due to the need to integrate AM system behavior during modeling. Towards that goal, an approach combining dimensional analysis conceptual modeling (DACM), experimental sampling, factors selection, and modeling based on Knowledge-Based Artificial Neural Network (KB-ANN) is proposed. This approach integrates existing literature and expert knowledge of the AM process to implement system behavior centered topology optimization of the knowledge-based artificial neural network model. The usefulness of the method is demonstrated using a case study to model wall thickness, height of part, and total mass of the part in a Fused Deposition Modeling (FDM) process. The KB-ANN based model for FDM has better performance and generalization model with low mean squared error in comparison to a conventional ANN.

AB - Additive manufacturing (AM) continues to rise in popularity due to its various advantages over traditional manufacturing processes. AM interests industry, but achieving repeatable production quality remains problematic for many AM technologies. Thus, modeling the different process variables in AM using modeling techniques, such as, machine learning, can be highly beneficial in creating useful knowledge of the process. Such developed artificial neural network models would aid designers and manufacturers to make informed decisions about their products and processes. However, accurately defining an artificial neural network topology is challenging due to the need to integrate AM system behavior during modeling. Towards that goal, an approach combining dimensional analysis conceptual modeling (DACM), experimental sampling, factors selection, and modeling based on Knowledge-Based Artificial Neural Network (KB-ANN) is proposed. This approach integrates existing literature and expert knowledge of the AM process to implement system behavior centered topology optimization of the knowledge-based artificial neural network model. The usefulness of the method is demonstrated using a case study to model wall thickness, height of part, and total mass of the part in a Fused Deposition Modeling (FDM) process. The KB-ANN based model for FDM has better performance and generalization model with low mean squared error in comparison to a conventional ANN.

U2 - 10.1115/1.4042084

DO - 10.1115/1.4042084

M3 - Article

VL - 141

JO - Journal of Mechanical Design

JF - Journal of Mechanical Design

SN - 1050-0472

IS - 2

ER -