Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal

Jongyeop Kim; Jinki Kim; Matthew Sands; Seongsoo Kim

doi:10.1109/SERA57763.2023.10197684

Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal

Jongyeop Kim
, Jinki Kim
, Matthew Sands
, Seongsoo Kim

Georgia Southern University
North Carolina State University

Research output: Contribution to book or proceeding › Conference article › peer-review

1 Scopus citations

Abstract

This study utilizes vibration-based signal analysis as a non-destructive testing technique that involves analyzing the vibration signals produced by a structure to detect possible defects or damage. The study aims to employ deep learning models to identify defects in a 3D-printed cantilever beam by analyzing the beam's tip displacement given a random input signal generated by an electromagnetic shaker. This study is focused on the output signal without any information of the random input, which is common for structural health monitoring applications in practice. Additionally, the study has revealed that the number of times the test set is applied to the trained model significantly impacts the accuracy of the model's consistent predictions.

Original language	English
Title of host publication	Proceedings - 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications, SERA 2023
Editors	Yeong-Tae Song, Junghwan Rhee, Yuseok Jeon
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	201-209
Number of pages	9
ISBN (Electronic)	9798350345889
ISBN (Print)	9798350345889
DOIs	https://doi.org/10.1109/SERA57763.2023.10197684
State	Published - May 23 2023
Event	21st IEEE/ACIS International Conference on Software Engineering Research, Management and Applications, SERA 2023 - Orlando, United States Duration: May 23 2023 → May 25 2023

Publication series

Name	2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications (SERA)

Conference

Conference	21st IEEE/ACIS International Conference on Software Engineering Research, Management and Applications, SERA 2023
Country/Territory	United States
City	Orlando
Period	05/23/23 → 05/25/23

UN SDGs

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

SDG 7 Affordable and Clean Energy

Scopus Subject Areas

Energy Engineering and Power Technology
Renewable Energy, Sustainability and the Environment
Electrical and Electronic Engineering
Industrial and Manufacturing Engineering
Safety, Risk, Reliability and Quality
Environmental Engineering

Keywords

Deep learning
GRU
LSTM
Model Stability
Signal processing
Structural Health Monitoring
Vibrations

Access to Document

10.1109/SERA57763.2023.10197684

Cite this

Kim, J., Kim, J., Sands, M., & Kim, S. (2023). Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal. In Y.-T. Song, J. Rhee, & Y. Jeon (Eds.), Proceedings - 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications, SERA 2023 (pp. 201-209). (2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications (SERA)). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SERA57763.2023.10197684

Kim, Jongyeop ; Kim, Jinki ; Sands, Matthew et al. / Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal. Proceedings - 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications, SERA 2023. editor / Yeong-Tae Song ; Junghwan Rhee ; Yuseok Jeon. Institute of Electrical and Electronics Engineers Inc., 2023. pp. 201-209 (2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications (SERA)).

@inproceedings{61ee29896c204a87b162c1b90320dfb7,

title = "Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal",

abstract = "This study utilizes vibration-based signal analysis as a non-destructive testing technique that involves analyzing the vibration signals produced by a structure to detect possible defects or damage. The study aims to employ deep learning models to identify defects in a 3D-printed cantilever beam by analyzing the beam's tip displacement given a random input signal generated by an electromagnetic shaker. This study is focused on the output signal without any information of the random input, which is common for structural health monitoring applications in practice. Additionally, the study has revealed that the number of times the test set is applied to the trained model significantly impacts the accuracy of the model's consistent predictions.",

keywords = "Deep learning, GRU, LSTM, Model Stability, Signal processing, Structural Health Monitoring, Vibrations",

author = "Jongyeop Kim and Jinki Kim and Matthew Sands and Seongsoo Kim",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 21st IEEE/ACIS International Conference on Software Engineering Research, Management and Applications, SERA 2023 ; Conference date: 23-05-2023 Through 25-05-2023",

year = "2023",

month = may,

day = "23",

doi = "10.1109/SERA57763.2023.10197684",

language = "English",

isbn = "9798350345889",

series = "2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications (SERA)",

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

pages = "201--209",

editor = "Yeong-Tae Song and Junghwan Rhee and Yuseok Jeon",

booktitle = "Proceedings - 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications, SERA 2023",

address = "United States",

}

Kim, J , Kim, J, Sands, M & Kim, S 2023, Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal. in Y-T Song, J Rhee & Y Jeon (eds), Proceedings - 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications, SERA 2023. 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications (SERA), Institute of Electrical and Electronics Engineers Inc., pp. 201-209, 21st IEEE/ACIS International Conference on Software Engineering Research, Management and Applications, SERA 2023, Orlando, United States, 05/23/23. https://doi.org/10.1109/SERA57763.2023.10197684

Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal. / Kim, Jongyeop ; Kim, Jinki; Sands, Matthew et al.
Proceedings - 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications, SERA 2023. ed. / Yeong-Tae Song; Junghwan Rhee; Yuseok Jeon. Institute of Electrical and Electronics Engineers Inc., 2023. p. 201-209 (2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications (SERA)).

Research output: Contribution to book or proceeding › Conference article › peer-review

TY - GEN

T1 - Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal

AU - Kim, Jongyeop

AU - Kim, Jinki

AU - Sands, Matthew

AU - Kim, Seongsoo

PY - 2023/5/23

Y1 - 2023/5/23

N2 - This study utilizes vibration-based signal analysis as a non-destructive testing technique that involves analyzing the vibration signals produced by a structure to detect possible defects or damage. The study aims to employ deep learning models to identify defects in a 3D-printed cantilever beam by analyzing the beam's tip displacement given a random input signal generated by an electromagnetic shaker. This study is focused on the output signal without any information of the random input, which is common for structural health monitoring applications in practice. Additionally, the study has revealed that the number of times the test set is applied to the trained model significantly impacts the accuracy of the model's consistent predictions.

AB - This study utilizes vibration-based signal analysis as a non-destructive testing technique that involves analyzing the vibration signals produced by a structure to detect possible defects or damage. The study aims to employ deep learning models to identify defects in a 3D-printed cantilever beam by analyzing the beam's tip displacement given a random input signal generated by an electromagnetic shaker. This study is focused on the output signal without any information of the random input, which is common for structural health monitoring applications in practice. Additionally, the study has revealed that the number of times the test set is applied to the trained model significantly impacts the accuracy of the model's consistent predictions.

KW - Deep learning

KW - GRU

KW - LSTM

KW - Model Stability

KW - Signal processing

KW - Structural Health Monitoring

KW - Vibrations

UR - https://www.scopus.com/pages/publications/85168769024

U2 - 10.1109/SERA57763.2023.10197684

DO - 10.1109/SERA57763.2023.10197684

M3 - Conference article

AN - SCOPUS:85168769024

SN - 9798350345889

T3 - 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications (SERA)

SP - 201

EP - 209

BT - Proceedings - 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications, SERA 2023

A2 - Song, Yeong-Tae

A2 - Rhee, Junghwan

A2 - Jeon, Yuseok

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 21st IEEE/ACIS International Conference on Software Engineering Research, Management and Applications, SERA 2023

Y2 - 23 May 2023 through 25 May 2023

ER -

Kim J , Kim J, Sands M, Kim S. Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal. In Song YT, Rhee J, Jeon Y, editors, Proceedings - 2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications, SERA 2023. Institute of Electrical and Electronics Engineers Inc. 2023. p. 201-209. (2023 IEEE/ACIS 21st International Conference on Software Engineering Research, Management and Applications (SERA)). doi: 10.1109/SERA57763.2023.10197684

Deep Learning Model to Improve the Stability of Damage Identification via Output-only Signal

Abstract

Publication series

Conference

UN SDGs

Scopus Subject Areas

Keywords

Access to Document

Other files and links

Fingerprint

Cite this