Physics-Driven Data Collection in 3-D Printing: Traversing the Realm of Social Manufacturing

Tariku Sinshaw Tamir; Gang Xiong; Zhen Shen; Jiewu Leng

doi:10.1109/TCSS.2024.3407823

Physics-Driven Data Collection in 3-D Printing: Traversing the Realm of Social Manufacturing

Tariku Sinshaw Tamir
, Gang Xiong
, Zhen Shen
, Jiewu Leng^*

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

5 Scopus citations

Abstract

Additive manufacturing (AM), also called 3-D printing, is a supporting technology in social manufacturing that has gained significant attention recently. As the AM industry grows, collecting and analyzing data are essential to ensure product quality, process efficiency, and cost-effectiveness. However, obtaining experimental data is challenging owing to cost and time constraints. Therefore, cost-effective and time-efficient strategies for collecting AM data are urgently required. This study proposes a novel data-collection approach that integrates the concept of finite element analysis (FEA) and physics-informed machine learning (PIML). We begin by discussing the importance of data collection in AM and the associated challenges. We then present various types of data that can be collected in AM, including the 3-D models and end-to-end data. End-to-end data comprise experimental data (i.e., sensors and images) and simulation data. Moreover, we present a case study that demonstrates the generation of simulation data and provides a detailed analysis of warpage. The STereoLithography (STL) file format of the BeltClip object from the Thingiverse possesses slicing through the Ultimaker.

Original language	English
Pages (from-to)	7909-7928
Number of pages	20
Journal	IEEE Transactions on Computational Social Systems
Volume	11
Issue number	6
DOIs	https://doi.org/10.1109/TCSS.2024.3407823
State	Published - 2024
Externally published	Yes

Bibliographical note

Publisher Copyright:
© Cura software. The resulting G-code file is input to the Digimat-AM platform for virtual simulation of the BeltClip printing process. Digimat-AM, as a FEA simulation tool, then generates observational sample data. These data function as a roadmap for understanding the application of physical information for learning, which constitutes the observational bias aspect of PIML. The observational data obtained from the Digimat-AM is suggested for building a machine-learning model. Finally, we conclude with a discussion of inductive and learning biases in the prediction, control, and optimization aspects of AM.

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 9 Industry, Innovation, and Infrastructure

Keywords

3-D printing
experimental data
physics-informed machine learning (PIML), simulation data
social manufacturing
warpage analysis

ASJC Scopus subject areas

Modeling and Simulation
Social Sciences (miscellaneous)
Human-Computer Interaction

Access to Document

10.1109/TCSS.2024.3407823

Cite this

@article{7139a369a2f74cf08f847fe8a38ea555,

title = "Physics-Driven Data Collection in 3-D Printing: Traversing the Realm of Social Manufacturing",

abstract = "Additive manufacturing (AM), also called 3-D printing, is a supporting technology in social manufacturing that has gained significant attention recently. As the AM industry grows, collecting and analyzing data are essential to ensure product quality, process efficiency, and cost-effectiveness. However, obtaining experimental data is challenging owing to cost and time constraints. Therefore, cost-effective and time-efficient strategies for collecting AM data are urgently required. This study proposes a novel data-collection approach that integrates the concept of finite element analysis (FEA) and physics-informed machine learning (PIML). We begin by discussing the importance of data collection in AM and the associated challenges. We then present various types of data that can be collected in AM, including the 3-D models and end-to-end data. End-to-end data comprise experimental data (i.e., sensors and images) and simulation data. Moreover, we present a case study that demonstrates the generation of simulation data and provides a detailed analysis of warpage. The STereoLithography (STL) file format of the BeltClip object from the Thingiverse possesses slicing through the Ultimaker.",

keywords = "3-D printing, experimental data, physics-informed machine learning (PIML), simulation data, social manufacturing, warpage analysis",

author = "Tamir, \{Tariku Sinshaw\} and Gang Xiong and Zhen Shen and Jiewu Leng",

note = "Publisher Copyright: {\textcopyright} Cura software. The resulting G-code file is input to the Digimat-AM platform for virtual simulation of the BeltClip printing process. Digimat-AM, as a FEA simulation tool, then generates observational sample data. These data function as a roadmap for understanding the application of physical information for learning, which constitutes the observational bias aspect of PIML. The observational data obtained from the Digimat-AM is suggested for building a machine-learning model. Finally, we conclude with a discussion of inductive and learning biases in the prediction, control, and optimization aspects of AM. ",

year = "2024",

doi = "10.1109/TCSS.2024.3407823",

language = "English",

volume = "11",

pages = "7909--7928",

journal = "IEEE Transactions on Computational Social Systems",

issn = "2329-924X",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "6",

}

TY - JOUR

T1 - Physics-Driven Data Collection in 3-D Printing

T2 - Traversing the Realm of Social Manufacturing

AU - Tamir, Tariku Sinshaw

AU - Xiong, Gang

AU - Shen, Zhen

AU - Leng, Jiewu

N1 - Publisher Copyright: © Cura software. The resulting G-code file is input to the Digimat-AM platform for virtual simulation of the BeltClip printing process. Digimat-AM, as a FEA simulation tool, then generates observational sample data. These data function as a roadmap for understanding the application of physical information for learning, which constitutes the observational bias aspect of PIML. The observational data obtained from the Digimat-AM is suggested for building a machine-learning model. Finally, we conclude with a discussion of inductive and learning biases in the prediction, control, and optimization aspects of AM.

PY - 2024

Y1 - 2024

N2 - Additive manufacturing (AM), also called 3-D printing, is a supporting technology in social manufacturing that has gained significant attention recently. As the AM industry grows, collecting and analyzing data are essential to ensure product quality, process efficiency, and cost-effectiveness. However, obtaining experimental data is challenging owing to cost and time constraints. Therefore, cost-effective and time-efficient strategies for collecting AM data are urgently required. This study proposes a novel data-collection approach that integrates the concept of finite element analysis (FEA) and physics-informed machine learning (PIML). We begin by discussing the importance of data collection in AM and the associated challenges. We then present various types of data that can be collected in AM, including the 3-D models and end-to-end data. End-to-end data comprise experimental data (i.e., sensors and images) and simulation data. Moreover, we present a case study that demonstrates the generation of simulation data and provides a detailed analysis of warpage. The STereoLithography (STL) file format of the BeltClip object from the Thingiverse possesses slicing through the Ultimaker.

AB - Additive manufacturing (AM), also called 3-D printing, is a supporting technology in social manufacturing that has gained significant attention recently. As the AM industry grows, collecting and analyzing data are essential to ensure product quality, process efficiency, and cost-effectiveness. However, obtaining experimental data is challenging owing to cost and time constraints. Therefore, cost-effective and time-efficient strategies for collecting AM data are urgently required. This study proposes a novel data-collection approach that integrates the concept of finite element analysis (FEA) and physics-informed machine learning (PIML). We begin by discussing the importance of data collection in AM and the associated challenges. We then present various types of data that can be collected in AM, including the 3-D models and end-to-end data. End-to-end data comprise experimental data (i.e., sensors and images) and simulation data. Moreover, we present a case study that demonstrates the generation of simulation data and provides a detailed analysis of warpage. The STereoLithography (STL) file format of the BeltClip object from the Thingiverse possesses slicing through the Ultimaker.

KW - 3-D printing

KW - experimental data

KW - physics-informed machine learning (PIML), simulation data

KW - social manufacturing

KW - warpage analysis

UR - https://www.scopus.com/pages/publications/85197085227

U2 - 10.1109/TCSS.2024.3407823

DO - 10.1109/TCSS.2024.3407823

M3 - Article

AN - SCOPUS:85197085227

SN - 2329-924X

VL - 11

SP - 7909

EP - 7928

JO - IEEE Transactions on Computational Social Systems

JF - IEEE Transactions on Computational Social Systems

IS - 6

ER -

Physics-Driven Data Collection in 3-D Printing: Traversing the Realm of Social Manufacturing

Abstract

Bibliographical note

UN SDGs

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this