Estudo Comparativo da Eficácia de Autocodificadores na Identificação de Danos Estruturais

Marcos Spínola Neto; Luiz Tadeu Dias Júnior; Rafelle Piazzaroli Finotti; Flávio de Souza Barbosa; Alexandre Abrahão Cury

doi:10.63595/vetor.v35i2.18346

Authors

Marcos Spínola Neto Universidade Federal de Juiz de Fora, Faculdade de Engenharia - Juiz de Fora, MG, Brasil
Luiz Tadeu Dias Júnior Universidade Federal de Juiz de Fora, Faculdade de Engenharia - Juiz de Fora, MG, Brasil
Rafelle Piazzaroli Finotti Universidade Federal de Juiz de Fora, Faculdade de Engenharia - Juiz de Fora, MG, Brasil https://orcid.org/0000-0002-8569-0041
Flávio de Souza Barbosa Universidade Federal de Juiz de Fora, Faculdade de Engenharia - Juiz de Fora, MG, Brasil
Alexandre Abrahão Cury Universidade Federal de Juiz de Fora, Faculdade de Engenharia - Juiz de Fora, MG, Brasil https://orcid.org/0000-0002-8860-1286

DOI:

https://doi.org/10.63595/vetor.v35i2.18346

Keywords:

Structural Health Monitoring, Damage detection, Variational Autoencoder, autoencoders

Abstract

Structural Health Monitoring (SHM) aims to ensure the safety and functionality of structures. In recent years, various machine learning techniques have been employed for this purpose. Among them, autoencoders (AE) stand out as models capable of extracting features from vibration data, reducing dimensionality and proving to be effective tools for SHM applications. This work investigates the effectiveness of four methodologies based on autoencoders, combined with a statistical tool to detect and quantify structural changes in three different structures. The vibration signals from the structures are used as input data, and the values from the latent layer of the autoencoders are used as parameters in the Hotelling T² test to evaluate structural changes. The results obtained indicate that the autoencoder model with the best performance, Variational AE-T², outperforms the others in identifying and quantifying structural changes. Although the AE, Sparse AE, and Convolutional AE models exhibited limitations regarding the quantification of alterations, they showed relevant performance for anomaly detection.

Downloads

Download data is not yet available.

References

[1] H. Sousa e W. Courage, “Value of information of a pro-active SHM-based tool supported by advanced FE models and Bayesian statistics towards efficiency in bridge management,” Structure and Infrastructure Engineering, vol. 18, pp. 554–572, 2021. Disponível em: https://doi.org/10.1080/15732479.2021.1978507

[2] X. Wang, Y. Q. Ni, e K. C. Lin, “Comparison of statistical counting methods in SHM-based reliability assessment of bridges,” Journal of Civil Structural Health Monitoring , vol. 5, pp. 275–286, 2015. Disponível em: https://doi.org/10.1007/s13349-015-0103-1

[3] A. Cury, D. Ribeiro, F. Ubertini, e M. D. Todd, Structural Health Monitoring Based on Data Science Techniques, 1a ed., Springer Cham, 2022.

[4] F. J. Carrión, J. A. Quintana, e S. E. Crespo, “Techno-economical and practical considerations for SHM systems,” Journal of Civil Structural Health Monitoring, vol. 7, pp. 207-215, 2017. Disponível em: https://doi.org/10.1007/s13349-017-0215-x

[5] V. Alves e A. Cury, “An automated vibration-based structural damage localization strategy using filter-type feature selection,” Mechanical Systems and Signal Processing, vol. 190, p. 110145, 2023. Disponível em: https://doi.org/10.1016/j.ymssp.2023.110145

[6] M. Spínola Neto, R. Finotti, F. Barbosa, e A. Cury, “Structural Damage Identification Using Autoencoders: A Comparative Study,” Buildings, vol. 14, no. 7, p. 2014, 2024. Disponível em: https://doi.org/10.3390/buildings14072014

[7] R. P. Finotti, C. F. Silva, P. H. E. Oliveira, F. S. Barbosa, A. A. Cury, e R. C. Silva, “Novelty detection on a laboratory benchmark slender structure using an unsupervised deep learning algorithm,” Latin American Journal of Solids and Structures, vol. 20, no. 9, p. e512, 2023. Disponível em: https://doi.org/10.1590/1679-78257591

[8] M. Mou, X. Zhao, K. Liu, e Y. Hui, “Database for Structural Control and Monitoring Benchmark Problems,” 2015. Disponível em: https://datacenterhub.org/resources/257

[9] B. Peeters e G. De Roeck, “One year monitoring of the Z24-bridge: Environmental influences versus damage events,” Earthquake Engineering Structural Dynamics, vol. 30, no. 2, pp. 149-171, 2001. Disponível em: https://doi.org/10.1002/1096-9845(200102)30:2<149::AID-EQE1>3.0.CO;2-Z.

[10] D. Montgomery, Introduction to Statistical Quality Control, 4a ed., Hoboken, NJ, EUA: John Wiley & Sons, 2007.

[11] L. Resende, R. P. Finotti, F. S. Barbosa, e A. A. Cury, “Structural damage detection with autoencoding neural networks,” em Proceedings of the XLIII Ibero-Latin American Congress on Computational Methods in Engineering, Foz do Iguaçu, Brasil, 2022.

[12] T. Akiba, S. Sano, T. Yanase, T. Ohta, e M. Koyama, “Optuna: A Next-generation Hyperparameter Optimization Framework,” em Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD ’19), Anchorage, AK, EUA, 2019, pp. 2623–2631. Disponível em: https://doi.org/10.1145/3292500.3330701

[13] L. Resende, R. Finotti, F. Barbosa, H. Garrido, A. Cury, e M. Domizio, “Damage identification using convolutional neural networks from instantaneous displacement measurements via image processing,” Structural Health Monitoring , vol. 23, no. 3, pp. 1627-1640, 2024. Disponível em: https://doi.org/10.1177/14759217231193102

[14] R. P. Finotti, C. Gentile, F. Barbosa, e A. Cury, “Structural novelty detection based on sparse autoencoders and control charts,” Structural Engineering and Mechanics, vol. 81, no. 5, pp. 647-664, 2022. Disponível em: https://doi.org/10.12989/sem.2022.81.5.647.

[15] R. P. Finotti, F. S. Barbosa, A. A. Cury, e R. L. Pimentel, “Numerical and experimental evaluation of structural changes using sparse auto-encoders and SVM applied to dynamic responses,” Applied Sciences, vol. 11, no. 24, p. 11965, 2021. Disponível em: https://doi.org/10.3390/app112411965