[1] DUAN Yuanfeng, DING Pengyao, DUAN Zhengteng, CHENG J. J. Roger, et al. Reconstruction of bridge-sensor data and detection of structural damage based on gradient-coupled autoencoder and fully connected network [J]. Journal of Southeast University (English Edition), 2026, 42 (1): 1-11. [doi:10.3969/j.issn.1003-7985.2026.01.001]

Copy

Reconstruction of bridge-sensor data and detection of structural damage based on gradient-coupled autoencoder and fully connected network()

基于梯度耦合自编码器与全连接网络的桥梁传感器数据重构与结构损伤识别

Journal of Southeast University (English Edition)[ISSN:1003-7985/CN:32-1325/N]

Volumn:

42

Issue:

2026 1

Page:

1-11

Research Field:

Civil Engineering

Publishing date:

2026-03-20

Info

Title:

Reconstruction of bridge-sensor data and detection of structural damage based on gradient-coupled autoencoder and fully connected network

基于梯度耦合自编码器与全连接网络的桥梁传感器数据重构与结构损伤识别

Author(s):

DUAN Yuanfeng, DING Pengyao, DUAN Zhengteng, CHENG J. J. Roger

College of Civil Engineering and Architecture, Zhejiang University, Hangzhou 310058, China

段元锋, 丁芃尧, 段政腾, 郑荣俊

浙江大学建筑工程学院, 杭州 310058

Keywords:

structural health monitoring;  machine learning;  data compression;  damage identification;  convolutional neural network;  fully connected neural network;  gradient-coupled mechanism

结构健康监测;  机器学习;  数据压缩;  损伤识别;  卷积神经网络;  全连接神经网络;  梯度耦合机制

PACS:

TU317.1

DOI:

10.3969/j.issn.1003-7985.2026.01.001

Abstract:

A dual-task parallel machine learning framework was developed by integrating a convolutional autoencoder (CAE) and a fully connected neural network (FCNN) via the gradient-coupled mechanism, enabling simultaneous data compression-reconstruction and structural damage identification. Under the condition where 40% of the sensor nodes are missing, the model successfully reconstructs the full sensor network with an R² of 0.916 and normalized root mean square error (NRMSE) of 0.028 8. Even under significant noise contamination with an SNR of 12 dB, the model maintains strong reconstruction performance, achieving a R² of 0.910 and NRMSE of 0.025 3. Forty-six structural damage scenarios were simulated using the scaled bridge model. The accuracy of spatial localization and quantification of the damage severity using the framework exceeds 99.3%. The proposed framework reduces the training time by 54.4% and iteration counts by 45.5% compared to conventional two-stage machine learning approaches, demonstrating the efficiency of gradient-coupled optimization.

为了提升桥梁传感器数据压缩编码的效率与任务适配性,提出了一种双任务并行机器学习框架。该框架由一个卷积自编码器(CAE)和一个全连接神经网络(FCNN)组成,利用梯度耦合机制(GCM)实现高效并行的数据压缩重构与损伤识别。该框架在一座三跨连续梁桥缩尺模型上进行了实验验证。实验结果表明,在传感器网络40%节点缺失的情况下,模型仍能取得R²为0.916、NRMSE为0.028 8的全传感器网络重构效果;在SNR为12 dB的高噪声水平影响下,模型仍取得R²为0.910、NRMSE为0.025 3的重构效果;在结构损伤空间定位与损伤程度识别的46个工况中,模型均取得了99.3%以上的识别精度。此外,模型训练时长与训练迭代次数相较于传统两阶段机器学习框架降幅分别达54.4%和45.5%。

References:

[1]GHAREHBAGHI V R, NOROOZINEJAD F E, NOORI M, et al. A critical review on structural health monitoring: Definitions, methods, and perspectives[J]. Archives of Computational Methods in Engineering, 2022, 29(4): 2209-2235.
[2]SUN L M, SHANG Z Q, XIA Y, et al. Review of bridge structural health monitoring aided by big data and artificial intelligence: From condition assessment to damage detection[J]. Journal of Structural Engineering, 2020, 146(5): 04020073.
[3]SHAN J Z, ZHANG X, LOONG C N, et al. Predictive maintenance and its applications in civil engineering structures: A review[J]. Journal of Southeast University (English Edition), 2024, 40(3): 245-256.
[4]HU S Y, XIE Z N, YANG Y. Interference wind pressure prediction of high-rise buildings with square section based on machine learning[J]. Journal of Southeast University (Natural Science Edition), 2024, 54(6): 1425-1437. (in Chinese)
[5]CHEN H, ZHU Y K, LEI B, et al. Sensor fault self-detection based on the mean shift method [J]. Journal of Southeast University (English Edition), 2024, 40(2): 140-147.
[6]LIU F P. Study on the application of structural health monitoring system for Dongying Yellow River Bridge [J]. Qinghai Science and Technology of Transportation, 2024, 36(2): 141-146. (in Chinese)
[7]YU B, QIU H X, WANG H, et al. Health monitoring system for Sutong Yangtze River Bridge [J]. Journal of Earthquake Earthquake Engineering and Engineering Vibration, 2009, 29(4): 170-177. (in Chinese)
[8]LIU Z Q, LI N, GUO J, et al. Design and implementation of structural monitoring systems for Xihoumen Bridge(Ⅱ): Implementations[J]. Engineering Sciences, 2010, 12(7): 101-106. (in Chinese)
[9]ZHAO Z Q, ZHANG Y, HU J M, et al. Comparative study of PCA and ICA based traffic flow compression [J]. Journal of Highway and Transportation Research and Development, 2008, 25(11): 109-113, 118. (in Chinese)
[10]ZHOU S W, LIN Y P, YE S T, et al. A wavelet data compression algorithm with memory-efficiency for wireless sensor network[J]. Journal of Computer Research and Development, 2009, 46(12): 2085-2092. (in Chinese)
[11]QUER G, MASIERO R, PILLONETTO G, et al. Sensing, compression, and recovery for WSNs: Sparse signal modeling and monitoring framework[J]. IEEE Transactions on Wireless Communications, 2012, 11(10): 3447-3461.
[12]DUAN Y F, DUAN Z T, ZHANG H M, et al. Bridge damage identification based on convolutional autoencoders and extreme gradient boosting trees [J]. Journal of Southeast University (English Edition), 2024, 40(3): 221-229.
[13]LIU Z J, JIN M R, ZHOU L C, et al. Bridge damage identification method based on structural response vectors and support vector machine algorithms [J]. Journal of University of Jinan (Science and Technology), 2020, 34(2): 106-112. (in Chinese)
[14]DAI L C, CAO W, YI S C, et al. Damage identification of concrete structure based on WPT-SVD and GA-BPNN [J]. Journal of Zhejiang University (Engineering Science), 2023, 57: 100-110, 132. (in Chinese)
[15]YANG D H, SUN J Z, YI T H, et al. Early warning technology of long-span bridge bearing deterioration considering time lag effects of thermal-induced displacement [J]. Journal of Southeast University (Natural Science Edition), 2024, 55(2): 268-274. (in Chinese)
[16]SHAN D S, SHI L, TAN K X. Bridge damage identification based on CNN and LSTM deep network [J]. Bridge Construction, 2023, 53(4): 41-46. (in Chinese)
[17]ZHANG C W, CHUN Q, MA Y K, et al. Research on damage detection of ancient stone arch bridges based on spatio-temporal difference graph convolutional neural network [J]. Journal of Southeast University (Natural Science Edition), 2025, 55(2): 370-379. (in Chinese)
[18]ZHANG H M, HU F, DUAN Y F, et al. A vision-based deformation tracking for self-centering structures during shaking table tests[J]. Engineering Structures, 2025, 330: 119800.
[19]CHEN H, LI J B, YIN X G. A new experimental method for structural damage identification [J]. Journal of Experimental Mechanics, 2011, 26(1): 96-102. (in Chinese)

Memo

Memo:

Received: 2025-04-11; Revised: 2025-06-19.
Biography: DUAN Yuanfeng (1977—), male, doctor, professor, ceyfduan@zju.edu.cn.
Foundation items: The National Natural Science Foundation of China (No.52361165658, U24A20169).
Citation: DUAN Yuanfeng, DING Pengyao, DUAN Zhengteng, et al. Reconstruction of bridge-sensor data and detection of structural damage based on gradient-coupled autoencoder and fully connected network[J]. Journal of Southeast University (English Edition), 2026, 42(1): 1-11. DOI: 10. 3969/j. issn. 1003-7985. 2026. 01. 001.

Common functions