TY - GEN
T1 - Efficient Structural Damage Detection Using Joint Vibration Signals and 1D-CNN Mode
AU - Kuo, Chien Chih
AU - Lee, Ching Hung
N1 - Publisher Copyright:
© 2023 IEEE.
PY - 2023
Y1 - 2023
N2 - In existing detection methodologies, identifying structural damage often necessitates the deployment of multiple sensors. This study introduces a novel approach to structural damage detection, focusing on the joints of planar steel frameworks. The proposed method aims to address the issue of excessive sensor requirements. Initially, we capture vibration signals from the joints using sensors. Subsequently, by analyzing the signal discrepancies between damaged and undamaged joints, we achieve the substitution of vibration signals from damaged joints. This method effectively reveals the characteristic features of structural damage. Furthermore, we explore various combinations of sensors and employ a one-dimensional Convolutional Neural Network model (1D-CNN) for detecting structural damage. In practical implementation, we validate the viability of our proposed approach using benchmark data from the Qatar University Grandstand Simulator (QUGS). Results demonstrate that our method significantly reduces sensor requirements, utilizing only 40% of the total count. Additionally, employing a single 1D-CNN model enhances detection accuracy to an impressive 96.53%. This straightforward yet powerful approach not only enhances detection performance but also optimizes sensor utilization, ultimately elevating detection efficiency. In conclusion, our proposed method not only embodies simplicity but also yields remarkable advancements in structural damage detection. It further optimizes sensor deployment, thus amplifying overall detection performance.
AB - In existing detection methodologies, identifying structural damage often necessitates the deployment of multiple sensors. This study introduces a novel approach to structural damage detection, focusing on the joints of planar steel frameworks. The proposed method aims to address the issue of excessive sensor requirements. Initially, we capture vibration signals from the joints using sensors. Subsequently, by analyzing the signal discrepancies between damaged and undamaged joints, we achieve the substitution of vibration signals from damaged joints. This method effectively reveals the characteristic features of structural damage. Furthermore, we explore various combinations of sensors and employ a one-dimensional Convolutional Neural Network model (1D-CNN) for detecting structural damage. In practical implementation, we validate the viability of our proposed approach using benchmark data from the Qatar University Grandstand Simulator (QUGS). Results demonstrate that our method significantly reduces sensor requirements, utilizing only 40% of the total count. Additionally, employing a single 1D-CNN model enhances detection accuracy to an impressive 96.53%. This straightforward yet powerful approach not only enhances detection performance but also optimizes sensor utilization, ultimately elevating detection efficiency. In conclusion, our proposed method not only embodies simplicity but also yields remarkable advancements in structural damage detection. It further optimizes sensor deployment, thus amplifying overall detection performance.
UR - http://www.scopus.com/inward/record.url?scp=85179838092&partnerID=8YFLogxK
U2 - 10.1109/CACS60074.2023.10326210
DO - 10.1109/CACS60074.2023.10326210
M3 - Conference contribution
AN - SCOPUS:85179838092
T3 - 2023 International Automatic Control Conference, CACS 2023
BT - 2023 International Automatic Control Conference, CACS 2023
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2023 International Automatic Control Conference, CACS 2023
Y2 - 26 October 2023 through 29 October 2023
ER -