Structural health monitoring (SHM) is essential for ensuring the safety and durability of civil infrastructure. Traditional SHM approaches, based on manual inspections or threshold-based analyses, often fail to detect early or subtle structural changes. In this work, we propose a data-driven framework for detecting structural regime transitions using long short-term memory (LSTM) networks trained on power spectral density data. This method does not require prior knowledge of the excitation sources or structural dynamics, enabling robust and interpretable transition detection under real-world conditions. A key component of the framework is the empirical transition point, $ T_a $, computed from prediction probabilities through persistence thresholding and entropy filtering. This allows for precise and automated detection of regime shifts, even in the absence of explicit ground truth. The model is validated on vibration data collected under ambient and train-induced excitations, achieving high accuracy in distinguishing pre- and post-retrofitting states. It demonstrates strong robustness across a range of operational and dynamic conditions. To enhance interpretability, we introduce the confidence variability index (CVI), which quantifies the temporal stability of the model's predictions and serves as an indicator of transition consistency. While the framework does not currently identify the physical causes of transitions, its sensitivity to dynamic changes makes it a valuable early-warning tool in SHM. Despite the inferential nature of transition detection due to the lack of ground truth, the approach offers a scalable, interpretable, and real-time solution for structural regime monitoring—contributing to the advancement of SHM systems through uncertainty quantification, causal inference, and intelligent infrastructure management.
Citation: A. Presno Vélez, M. Z. Fernández Muñiz, J. L. Fernández Martínez. Data-driven structural transition detection using vibration monitoring and LSTM networks[J]. AIMS Mathematics, 2025, 10(8): 18558-18585. doi: 10.3934/math.2025829
Structural health monitoring (SHM) is essential for ensuring the safety and durability of civil infrastructure. Traditional SHM approaches, based on manual inspections or threshold-based analyses, often fail to detect early or subtle structural changes. In this work, we propose a data-driven framework for detecting structural regime transitions using long short-term memory (LSTM) networks trained on power spectral density data. This method does not require prior knowledge of the excitation sources or structural dynamics, enabling robust and interpretable transition detection under real-world conditions. A key component of the framework is the empirical transition point, $ T_a $, computed from prediction probabilities through persistence thresholding and entropy filtering. This allows for precise and automated detection of regime shifts, even in the absence of explicit ground truth. The model is validated on vibration data collected under ambient and train-induced excitations, achieving high accuracy in distinguishing pre- and post-retrofitting states. It demonstrates strong robustness across a range of operational and dynamic conditions. To enhance interpretability, we introduce the confidence variability index (CVI), which quantifies the temporal stability of the model's predictions and serves as an indicator of transition consistency. While the framework does not currently identify the physical causes of transitions, its sensitivity to dynamic changes makes it a valuable early-warning tool in SHM. Despite the inferential nature of transition detection due to the lack of ground truth, the approach offers a scalable, interpretable, and real-time solution for structural regime monitoring—contributing to the advancement of SHM systems through uncertainty quantification, causal inference, and intelligent infrastructure management.
| [1] |
W. Fan, P. Qiao, Vibration-based damage identification methods: a review and comparative study, Struct. Health Monit., 10 (2010), 83–111. https://doi.org/10.1177/1475921710365419 doi: 10.1177/1475921710365419
|
| [2] | S. W. Doebling, C. R. Farrar, M. B. Prime, D. W. Shevitz, Damage identification and health monitoring of structural and mechanical systems from changes in their vibration characteristics: a literature review, Technical Report, Los Alamos National Laboratory Laboratory, New Mexico, 1996. https://doi.org/10.2172/249299 |
| [3] |
S. Hassani, U. Dackermann, A systematic review of optimization algorithms for structural health monitoring and optimal sensor placement, Sensors, 23 (2023), 3293. https://doi.org/10.3390/s23063293 doi: 10.3390/s23063293
|
| [4] |
T. Xin, Y. Yang, X. Zheng, J. Lin, S. Wang, P. Wang, Time series recovery using adjacent channel data based on LSTM: A case study of subway vibrations, Appl. Sci., 12 (2022), 11497. https://doi.org/10.3390/app122211497 doi: 10.3390/app122211497
|
| [5] | S. Hochreiter, J. Schmidhuber, Long short-term memory, Neural Comput., 9 (1997), 1735–1780. https://doi.org/10.1162/neco.1997.9.8.1735 |
| [6] |
M. Ahmadzadeh, S. M. Zahrai, M. Bitaraf, An integrated deep neural network model combining 1D CNN and LSTM for structural health monitoring utilizing multisensor time-series data, Struct. Health Monit., 24 (2025), 447–465. https://doi.org/10.1177/14759217241239041 doi: 10.1177/14759217241239041
|
| [7] |
V. Barzegar, S. Laflamme, C. Hu, J. Dodson, Multi-time resolution ensemble lstms for enhanced feature extraction in high-rate time series, Sensors, 21 (2021), 1954. https://doi.org/10.3390/s21061954 doi: 10.3390/s21061954
|
| [8] |
M. M. Eltahir, G. Aldehim, N. S. Almalki, M. M. Alnfiai, A. E. Osman, Reinforced concrete bridge damage detection using arithmetic optimization algorithm with deep feature fusion, AIMS Math., 8 (2023), 29290–29306. https://doi.org/10.3934/math.20231499 doi: 10.3934/math.20231499
|
| [9] |
M. G. Al-Thani, Z. Sheng, Y. Cao, Y. Yang, Traffic transformer: transformer-based framework for temporal traffic accident prediction, AIMS Math., 9 (2024), 12610–12629. https://doi.org/10.3934/math.2024617 doi: 10.3934/math.2024617
|
| [10] |
O. Hrizi, K. Gasmi, A. Alyami, A. Alkhalil, I. Alrashdi, A. Alqazzaz, et al., Federated and ensemble learning framework with optimized feature selection for heart disease detection, AIMS Math., 10 (2025), 7290–7318. https://doi.org/10.3934/math.2025334 doi: 10.3934/math.2025334
|
| [11] |
J. Hou, X. Lu, Y. Zhong, W. He, D. Zhao, F. Zhou, A comprehensive review of mechanical fault diagnosis methods based on convolutional neural network, J. Vibroeng., 26 (2023), 44–65. https://doi.org/10.21595/jve.2023.23391 doi: 10.21595/jve.2023.23391
|
| [12] |
E. L. Lydia, C. Santhaiah, M. Altafahmed, K. V. Kumar, G. P. Joshi, W. Cho, An equilibrium optimizer with deep recurrent neural networks enabled intrusion detection in secure cyber-physical systems, AIMS Math., 9 (2024), 11718–11734. https://doi.org/10.3934/math.2024574 doi: 10.3934/math.2024574
|
| [13] |
W. Liu, Z. Tang, F. Lv, X. Chen, An efficient approach for guided wave structural monitoring of switch rails via deep convolutional neural network-based transfer learning, Meas. Sci. Technol., 34 (2022), 024004. https://doi.org/10.1088/1361-6501/ac9ad3 doi: 10.1088/1361-6501/ac9ad3
|
| [14] |
W. Liu, S. Wang, Z. Yin, Z. Tang, Structural damage detection of switch rails using deep learning, NDT E Int., 147 (2024), 103205. https://doi.org/10.1016/j.ndteint.2024.103205 doi: 10.1016/j.ndteint.2024.103205
|
| [15] |
W. Liu, J. Hu, F. Lv, Z. Tang, A new method for long-term temperature compensation of structural health monitoring by ultrasonic guided wave, Measurement, 252 (2025), 117310. https://doi.org/10.1016/j.measurement.2025.117310 doi: 10.1016/j.measurement.2025.117310
|
| [16] |
W. Liu, Z. Tang, F. Lv, X. Chen, Multi-feature integration and machine learning for guided wave structural health monitoring: Application to switch rail foot, Struct. Health Monit., 20 (2021), 2013–2034. https://doi.org/10.1177/1475921721989577 doi: 10.1177/1475921721989577
|
| [17] |
F. Deng, X. Tao, P. Wei, S. Wei, A robust deep learning-based damage identification approach for SHM considering missing data, Appl. Sci., 13 (2023), 5421. https://doi.org/10.3390/app13095421 doi: 10.3390/app13095421
|
| [18] |
P. Ba, S. Zhu, H. Chai, C. Liu, P. Wu, L. Qi, Structural monitoring data repair based on a long short-term memory neural network, Sci. Rep., 14 (2024), 9974. https://doi.org/10.1038/s41598-024-60196-2 doi: 10.1038/s41598-024-60196-2
|
| [19] |
X. Wang, J. Wu, C. Liu, H. Wang, Y. Du, W. Niu, Exploring LSTM based recurrent neural network for failure time series prediction, J. Beijing Univ. Aeronaut. Astron., 44 (2018), 772–784. https://doi.org/10.13700/J.BH.1001-5965.2017.0285 doi: 10.13700/J.BH.1001-5965.2017.0285
|
| [20] |
A. Vaishnavi, A. Sharma, V. P. S. Naidu, Fault detection in machine bearings using deep learning, SAE Tech. Pap., 26 (2024), 0473. https://doi.org/10.4271/2024-26-0473 doi: 10.4271/2024-26-0473
|
| [21] |
A. P. Vélez, M. Z. F. Muñiz, J. L. F. Martínez, Enhancing structural health monitoring with machine learning for accurate prediction of retrofitting effects, AIMS Math., 9 (2024), 30493–30514. https://doi.org/10.3934/math.20241472 doi: 10.3934/math.20241472
|
| [22] |
A. P. Vélez, M. Z. F. Muñiz, J. L. F. Martínez, Advancing structural health monitoring with deep belief network-based classification, Mathematics, 13 (2025), 1435. https://doi.org/10.3390/math13091435 doi: 10.3390/math13091435
|
| [23] |
B. Peeters, G. De Roeck, Stochastic system identification for operational modal analysis: a review, J. Dyn. Syst. Meas. Control, 123 (2001), 659–667. https://doi.org/10.1115/1.1410370 doi: 10.1115/1.1410370
|
| [24] |
J. M. W. Brownjohn, Structural health monitoring of civil infrastructure, Phil. Trans. R. Soc. A, 365 (2007), 589–622. https://doi.org/10.1098/rsta.2006.1925 doi: 10.1098/rsta.2006.1925
|
| [25] | C. R. Farrar, K. Worden, Structural health monitoring: a machine learning perspective, John Wiley & Sons, Ltd, 2012. https://doi.org/10.1002/9781118443118 |
| [26] |
H. Li, J. Ou, The state of the art in structural health monitoring of cable-stayed bridges, J. Civil. Struct. Health Monit., 6 (2016), 43–67. https://doi.org/10.1007/s13349-015-0115-x doi: 10.1007/s13349-015-0115-x
|
| [27] | T. Poggio, H. Mhaskar, L. Rosasco, B. Miranda, Q. Liao, Why and when can deep-but not shallow-networks avoid the curse of dimensionality: a review, Int. J. Autom. Comput., 14 (2017) 503–519. https://doi.org/10.1007/s11633-017-1054-2 |
| [28] | A. Jentzen, B. Kuckuck, P. von Wurstemberger, Mathematical introduction to deep learning: methods, implementations and theory, arXiv, 2023. https://doi.org/10.48550/arXiv.2310.20360 |
| [29] | P. L. Bartlett, S. Mendelson, Rademacher and gaussian complexities: risk bounds and structural results, J. Mach. Learn. Res., 3 (2002), 463–482. |
| [30] | M. Anthony, P. L. Bartlett, Neural network learning: theoretical foundations, Cambridge University Press, 2009. |
| [31] | D. P. Kingma, J. L. Ba, Adam: a method for stochastic optimization, arXiv, 2017. https://doi.org/10.48550/arXiv.1412.6980 |
| [32] | D. Siegmund, Sequential analysis: tests and confidence intervals, Springer, 1985. https://doi.org/10.1007/978-1-4757-1862-1 |
| [33] |
D. V. Hinkley, Inference about the change-point from cumulative sum tests, Biometrika, 58 (1971), 509–523. https://doi.org/10.1093/BIOMET/58.3.509 doi: 10.1093/BIOMET/58.3.509
|
| [34] | R. P. Adams, D. J. C. MacKay, Bayesian online changepoint detection, arXiv, 2007. https://doi.org/10.48550/arXiv.0710.3742 |
| [35] | M. Altamirano, F. X. Briol, J. Knoblauch, Robust and scalable Bayesian online changepoint detection, arXiv, 2023. https://doi.org/10.48550/arXiv.2302.04759 |
| [36] |
M. Z. Yousaf, S. Khalid, M. F. Tahir, A. Tzes, A. Raza, A novel dc fault protection scheme based on intelligent network for meshed dc grids, Int. J. Electr. Power Energy Syst., 154 (2023), 109423. https://doi.org/10.1016/j.ijepes.2023.109423. doi: 10.1016/j.ijepes.2023.109423
|
| [37] |
M. Z. Yousaf, H. Liu, A. Raza, A. Mustafa, Deep learning-based robust dc fault protection scheme for meshed HVDC grids, CSEE J. Power Energy Syst., 9 (2023), 2423–2434. https://doi.org/10.17775/CSEEJPES.2021.03550 doi: 10.17775/CSEEJPES.2021.03550
|
| [38] |
M. Z. Yousaf, M. F. Tahir, A. Raza, M. A. Khan, F. Badshah, Intelligent sensors for dc fault location scheme based on optimized intelligent architecture for HVdc systems, Sensors, 22 (2022), 9936. https://doi.org/10.3390/s22249936 doi: 10.3390/s22249936
|
Figures(14) / Tables(5)
A. Presno Vélez, M. Z. Fernández Muñiz, J. L. Fernández Martínez. Data-driven structural transition detection using vibration monitoring and LSTM networks[J]. AIMS Mathematics, 2025, 10(8): 18558-18585. doi: 10.3934/math.2025829
DownLoad: