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

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.

Keywords

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

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)

Funding

National Natural Science Foundation of China(52361165658)

National Natural Science Foundation of China(U24A20169)