A Sustainable Anomaly Detection Framework for Autonomous Surface Ship: Adaptive Subsystem-Level Anomaly Detection Algorithm via MLOps
##plugins.themes.bootstrap3.article.main##
##plugins.themes.bootstrap3.article.sidebar##
Abstract
For the stable operation of Maritime Autonomous Surface Ships (MASS), this study proposes a sustainable anomaly detection framework that integrates a subsystem-level Condition-Based Maintenance (CBM) model with an adaptive MLOps pipeline. The main engine is decomposed into 14 functional units, each monitored by a hybrid algorithm that combines an Attention-LSTM-AutoEncoder and an Isolation Forest to detect subtle anomalies. To address model performance degradation caused by gradual data drift in maritime environments, an Autonomous Maintenance Mechanism is developed. This mechanism utilizes state severity (Z-Score) and drift velocity (ΔZ) indicators to algorithmically distinguish between sudden physical faults and gradual sensor drift. Based on this distinction, the MLOps pipeline accumulates confirmed drift in a buffer and selectively retrains and redeploys models using local onboard data once sufficient evidence has been gathered, while bypassing suspected fault conditions to avoid learning anomalous patterns. Experiments on an engine testbed indicate that the proposed system can suppress the Anomaly Rate (AR_t) during data drift and help restore diagnostic reliability, suggesting a practical basis toward self-sustaining condition monitoring for MASS.
How to Cite
##plugins.themes.bootstrap3.article.details##
MLOps, Sustainable, Data Drift, New Normal, Attention-LSTM-AE
Ashraf, W. M., Ansar, T., Ahmed, F., Hussain, J., Abbas, M. M., & Dua, V. (2026). From drift to adaptation to the failed ML model: Transfer learning in industrial MLOps. arXiv preprint arXiv:2602.00957.
Haque, A., & Soliman, H. (2025). A transformer-based autoencoder with isolation forest and XGBoost for malfunction and intrusion detection. Future Internet, 17(4), 164. doi: 10.3390/fi17040164
Kodakandla, N. (2024). Data drift detection and mitigation: A comprehensive MLOps approach for real-time systems. International Journal of Science and Research Archive, 12(1), 3127–3139.
Li, H., Meng, X., Liu, J., Zhang, W., Zhou, X., & Yang, X. (2025). A framework of predictive maintenance for maritime autonomous surface ships considering component degradation. Journal of Marine Science and Engineering, 13(3), 512. doi: 10.3390/jmse13030512
Liang, Q., Knutsen, K. E., Vanem, E., Zhang, H., & Æsøy, V. (2024). Unsupervised anomaly detection in marine diesel engines using transformer neural networks and residual analysis. International Journal of Prognostics and Health Management, 15(1), 1–18.
Lu, J., Liu, A., Dong, F., Gu, F., Gama, J., & Zhang, G. (2019). Learning under concept drift: A review. IEEE Transactions on Knowledge and Data Engineering, 31(12), 2346–2363. doi: 10.1109/TKDE.2018.2876857
Malhotra, P., Ramakrishnan, A., Anand, G., Goyal, L., Lodha, P., & Singh, P. (2016). LSTM-based encoder-decoder for multi-sensor anomaly detection. arXiv preprint arXiv:1607.00148.
Proulx, C., & Reichard, K. (2021). Automated condition-based maintenance (CBM) using artificial intelligence (AI) and machine learning (ML) for unmanned systems. Annual Conference of the PHM Society, 13(1).
Vanem, E., & Storvik, G. O. (2017). Anomaly detection using dynamical linear models and sequential testing on a marine engine system. International Journal of Prognostics and Health Management, 8(1).
Xu, H., Pang, G., Wang, Y., & Wang, Y. (2023). Deep isolation forest for anomaly detection. IEEE Transactions on Knowledge and Data Engineering, 35(12), 12591–12604. doi: 10.1109/TKDE.2023.3270299
Xu, J., Wu, H., Wang, J., & Long, M. (2022). Anomaly transformer: Time series anomaly detection with association discrepancy. In International Conference on Learning Representations (ICLR).
Xu, X., Yan, X., Yang, K., Zhao, J., Sheng, C., & Yuan, C. (2021). Review of condition monitoring and fault diagnosis for marine power systems. Transportation Safety and Environment, 3(2), 85–102. doi: 10.1093/tse/tdab004
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.