Detection of damage in steel truss bridges using a hybrid 1DCNN–BIGRU model and time-series data augmentation techniques
Transport and Communications Science Journal · 06.09.2025 · 76 (9) · 1281-1295
DOI: 10.47869/tcsj.76.9.11 ↗Kurz gefasst
Fachartikel in der Zeitschrift Transport and Communications Science Journal, online erschienen am 06.09.2025. Verfasst von Đăng Nguyễn Lê Minh und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Structural health monitoring (SHM) systems based on time-series data are pivotal tools for assessing the safety and serviceability of bridges. However, the lack of large-scale and diverse datasets often leads deep learning (DL) models to overfitting and poor generalization. This study proposes a comprehensive damage detection framework that simultaneously addresses both data and model aspects. On the data side, we employ data augmentation (AUG) techniques—including noise injection, sequence reversal, flipping, and window slicing—to expand and diversify the training set. Regarding the proposed methodology, we established a hybrid deep learning framework that integrates One-dimensional Convolutional Neural Networks (1DCNN) with Bidirectional Gated Recurrent Units (BiGRU). In this architecture, the 1DCNN layers are responsible for mining local features from raw time-series inputs, while the BiGRU layers specialize in capturing long-term temporal dependencies in both directions. The efficacy of this approach was verified using the Cua Rao steel truss bridge dataset. The experimental outcomes reveal that the data-augmented hybrid model significantly outperforms standalone baselines, achieving an accuracy of 93.9%. This is a substantial improvement over the individual 1DCNN (84.2%) and BiGRU (83.8%) models, demonstrating faster convergence and high stability for SHM applications
Zitieren
APA
Minh, Đ. N. L., Xuân, N. H., Mạnh, T. V., & Ngọc, B. N. K. (2025). Detection of damage in steel truss bridges using a hybrid 1DCNN–BIGRU model and time-series data augmentation techniques. Transport and Communications Science Journal, 76(9), 1281-1295. https://doi.org/10.47869/tcsj.76.9.11
Vancouver
Minh ĐNL, Xuân NH, Mạnh TV, Ngọc BNK. Detection of damage in steel truss bridges using a hybrid 1DCNN–BIGRU model and time-series data augmentation techniques. Transport and Communications Science Journal. 2025;76(9):1281-1295. doi:10.47869/tcsj.76.9.11
BibTeX
@article{Minh2025,
author = {Minh, Đăng Nguyễn Lê and Xuân, Nam Hồ and Mạnh, Trung Vũ and Ngọc, Bảo Nguyễn Kiều},
title = {Detection of damage in steel truss bridges using a hybrid 1DCNN–BIGRU model and time-series data augmentation techniques},
journal = {Transport and Communications Science Journal},
year = {2025},
volume = {76},
number = {9},
pages = {1281--1295},
doi = {10.47869/tcsj.76.9.11}
}
EUVIMED