Transfer Learning-Enhanced Residual Attention Temporal Network for Structural Damage Identification
Buildings · 13.07.2026 · 16 (14) · 2779
DOI: 10.3390/buildings16142779 ↗Kurz gefasst
Fachartikel in der Zeitschrift Buildings, online erschienen am 13.07.2026. Verfasst von Xinwei Wang und 4 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 41 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Key Scientific Research Projects of Colleges and Universities, Key R&D and Promotion Projects in Henan Province.
Abstract (Originalsprache)
Accurate structural damage identification under limited data availability and measurement noise remains a persistent challenge in structural health monitoring (SHM). This study proposes TLCA-RATNet, a transfer learning-enhanced residual attention temporal network for vibration-based damage state classification under noisy and small sample conditions. RATNet integrates adaptive threshold residual denoising, residual attention, and bidirectional gated recurrent unit (BiGRU)-based temporal modeling to suppress noise, emphasize damage-sensitive features, and capture global temporal dependencies. The current implementation is formulated as a single-task classifier, in which local feature refinement and global temporal representation are jointly optimized end-to-end through a unified damage classification objective. Transfer learning further initializes the target domain model using knowledge learned from a data-rich source structure, while regularized fine-tuning reduces overfitting on limited target samples. Experiments were conducted on a six-story lumped-mass shear structure, a three-story physical frame, and the IASC-ASCE SHM benchmark structure, using 10–145 training samples per damage class and additive-noise conditions ranging from 0 to 30 dB signal-to-noise ratio. On Dataset 2 at 5 dB, TLCA-RATNet achieved an accuracy of 89.86%, exceeding LSTM and CNN-BiGRU by 9.94 and 13.10 percentage points, respectively. On Dataset 3 at 0 dB, it achieved 86.00% accuracy, outperforming CNN-BiGRU by 10.72 percentage points. In the limited sample transfer experiment on Dataset 2, transfer learning increased the accuracy from 93.05% to 100.00%, representing a gain of 6.95 percentage points over training from scratch. These results indicate that TLCA-RATNet provides a data-efficient and noise-robust approach for damage state screening and rapid model adaptation in SHM applications with scarce labeled data and noisy measurements.
Förderung
- National Natural Science Foundation of China
- Key Scientific Research Projects of Colleges and Universities
- Key R&D and Promotion Projects in Henan Province
- Henan Out-standing Youth Science Foundation
- Science and Technology Research and Development Project
Zitieren
APA
Wang, X., Shahzad, M. M., Wei, Z., Yang, S., & Wang, T. (2026). Transfer Learning-Enhanced Residual Attention Temporal Network for Structural Damage Identification. Buildings, 16(14), 2779. https://doi.org/10.3390/buildings16142779
Vancouver
Wang X, Shahzad MM, Wei Z, Yang S, Wang T. Transfer Learning-Enhanced Residual Attention Temporal Network for Structural Damage Identification. Buildings. 2026;16(14):2779. doi:10.3390/buildings16142779
BibTeX
@article{Wang2026,
author = {Wang, Xinwei and Shahzad, Muhammad Moman and Wei, Zheng and Yang, Shixuan and Wang, Tianlong},
title = {Transfer Learning-Enhanced Residual Attention Temporal Network for Structural Damage Identification},
journal = {Buildings},
year = {2026},
volume = {16},
number = {14},
pages = {2779},
doi = {10.3390/buildings16142779}
}
EUVIMED