Contrastive Metric Learning Loss-Enhanced Multi-Layer Perceptron for Sequentially Appearing Clusters in Acoustic Emission Data Streams

##plugins.themes.bootstrap3.article.main##

##plugins.themes.bootstrap3.article.sidebar##

Published Jun 27, 2024
Oualid Laiadi Ikram Remadna El yamine Dris Redouane Drai Sadek Labib Terrissa Noureddine Zerhouni

Abstract

Conventional structural health monitoring methods for interpreting unlabeled acoustic emission (AE) data typically rely on generic clustering approaches. This work introduces a novel approach for analyzing sequential and temporal acoustic emission (AE) data streams by enhancing a Multi-Layer Perceptron (MLP) with a contrastive metric learning loss function (MLP-CMLL)and Time Series K-means (TSKmeans) clustering. This dual approach, MLP-CMLL with TSKmeans, is crafted to refine cluster differentiation significantly. This method is designed to optimize cluster differentiation, bringing similar acoustic patterns closer and distancing divergent ones, thereby improving the MLP's ability to classify acoustic events over time. Addressing the limitations of traditional clustering algorithms in handling the temporal dynamics of AE data, MLP-CMLL with TSKmeans approach provides deeper insights into cluster formation and evolution. It promises enhanced monitoring and predictive maintenance capabilities in engineering applications by capturing the complex dynamics of AE data more effectively, offering a significant advancement in the field of structural health monitoring. Through experimental validation, we apply this method to characterize the loosening phenomenon in bolted structures under vibrations. Comparative analysis with two standard clustering methods using raw streaming data from three experimental campaigns demonstrates that our proposed method not only delivers valuable qualitative information concerning the timeline of clusters but also showcases superior performance in terms of cluster characterization.

How to Cite

Laiadi, O., Remadna, I., Dris, E. yamine ., Drai, R. ., Terrissa, S. L. ., & Zerhouni, N. . (2024). Contrastive Metric Learning Loss-Enhanced Multi-Layer Perceptron for Sequentially Appearing Clusters in Acoustic Emission Data Streams. PHM Society European Conference, 8(1), 10. https://doi.org/10.36001/phme.2024.v8i1.4134
Abstract 236 | PDF Downloads 145

##plugins.themes.bootstrap3.article.details##

Keywords

acoustic emission (AE), sequentially appearing clusters, data streams, structural health monitoring, contrastive metric learning, multi-layer perceptron (MLP)

References
Bolognani, D., Verzobio, A., Tonelli, D., Cappello, C., Glisic, B., Zonta, D., & Quigley, J. (2018). Iwshm 2017: Quantifying the benefit of structural health monitoring: what if the manager is not the owner? Structural Health Monitoring, 17(6), 1393–1409. Fu, W., Zhou, R., & Guo, Z. (2023). Automatic bolt tightness detection using acoustic emission and deep learning. In Structures (Vol. 55, pp. 1774–1782). Furnas, G. W., Deerwester, S., Durnais, S. T., Landauer,

T. K., Harshman, R. A., Streeter, L. A., & Lochbaum,

K. E. (2017). Information retrieval using a singular value decomposition model of latent semantic structure. In Acm sigir forum (Vol. 51, pp. 90–105). Gonzalez Andino, S., Grave de Peralta Menendez, R., Thut, G., Spinelli, L., Blanke, O., Michel, C., & Landis,

T. (2000). Measuring the complexity of time series: an application to neurophysiological signals. Human brain mapping, 11(1), 46–57.

Hassani, K., & Khasahmadi, A. H. (2020). Contrastive multi-view representation learning on graphs. In International conference on machine learning (pp. 41164126).

Hoła, J., & Sadowski, Ł. (2022). Non-destructive testing in civil engineering (Vol. 12) (No. 14). MDPI.

Huang, X., Ye, Y., Xiong, L., Lau, R. Y., Jiang, N., & Wang, S. (2016). Time series k-means: A new kmeans type smooth subspace clustering for time series data. Information Sciences, 367-368, 1-13. doi: https://doi.org/10.1016/j.ins.2016.05.040 Kharrat, M., Ramasso, E., Placet, V., & Boubakar, M. (2016).

A signal processing approach for enhanced acoustic emission data analysis in high activity systems: Application to organic matrix composites. Mechanical Systems and Signal Processing, 70, 1038–1055.

Kherif, F., & Latypova, A. (2020). Principal component analysis. In Machine learning (pp. 209–225). Elsevier. Liu, G., Lin, Z., & Yu, Y. (2010). Robust subspace segmentation by low-rank representation. In Proceedings of the 27th international conference on machine learning (icml-10) (pp. 663–670).

Maulik, U., & Bandyopadhyay, S. (2002). Performance evaluation of some clustering algorithms and validity indices. IEEE Transactions on pattern analysis and machine intelligence, 24(12), 1650–1654.

Ramasso, E., Denoeux, T., & Chevallier, G. (2022). Clustering acoustic emission data streams with sequentially appearing clusters using mixture models. Mechanical Systems and Signal Processing, 181, 109504. Ramasso, E., Placet, V., & Boubakar, M. L. (2015). Unsupervised consensus clustering of acoustic emission time-series for robust damage sequence estimation in composites. IEEE Transactions on Instrumentation and Measurement, 64(12), 3297–3307.

Ramasso, E., Verdin, B., & Chevallier, G. (2022). Monitoring a bolted vibrating structure using multiple acoustic emission sensors: A benchmark. Data, 7(3), 31. Saunshi, N., Plevrakis, O., Arora, S., Khodak, M., & Khandeparkar, H. (2019). A theoretical analysis of contrastive unsupervised representation learning. In International conference on machine learning (pp. 5628–5637). Sause, M. G., Gribov, A., Unwin, A. R., & Horn, S. (2012).

Pattern recognition approach to identify natural clusters of acoustic emission signals. Pattern Recognition Letters, 33(1), 17–23.

Sun, J., Yang, H., Li, D., & Xu, C. (2023). Experimental investigation on acoustic emission in fretting friction and wear of bolted joints. Journal of Sound and Vibration, 558, 117773. doi: https://doi.org/10.1016/j.jsv.2023.117773
Swedish Accident Investigation Authority. (2017). Windactionjvestas wind turbine collapse in lemnhult. Wall, M. E., Rechtsteiner, A., & Rocha, L. M. (2003). Singular value decomposition and principal component analysis. In A practical approach to microarray data analysis (pp. 91–109). Springer. Wang, T., Song, G., Wang, Z., & Li, Y. (2013). Proof-ofconcept study of monitoring bolt connection status using a piezoelectric based active sensing method. Smart Materials and Structures, 22(8), 087001. Xu, D., Liu, P., Li, J., & Chen, Z. (2019). Damage mode identification of adhesive composite joints under hygrothermal environment using acoustic emission and machine

learning. Composite structures, 211, 351–363. Xu, P., Zhou, Z., Liu, T., & Mal, A. (2021). Determination of geometric role and damage assessment in hybrid fiber metal laminate (fml) joints based on acoustic emission. Composite Structures, 270, 114068. doi: https://doi.org/10.1016/j.compstruct.2021.114068 Zhu, H., & Koniusz, P. (2022a). Ease: Unsupervised discriminant subspace learning for transductive few-shot learning. In Proceedings of the ieee/cvf conference on computer vision and pattern recognition (pp. 9078–9088). Zhu, H., & Koniusz, P. (2022b). Generalized laplacian eigenmaps. Advances in Neural Information Processing Systems, 35, 30783–30797.
Section
Technical Papers