Efficient anomaly detection method for offshore wind turbines

Abstract

Time-series anomaly detection plays a crucial role in the operation of offshore wind turbines. Various wind turbine monitoring systems rely on time-series data to monitor and identify anomalies in real-time, as well as to initiate early warning processes. However, for offshore wind turbines with a high data density, conventional methods have high computational overhead in detecting anomalies while failing to accurately detect anomalies due to variations in data scales. To address this challenge, we propose an efficient anomaly detection method with contrastive learning, called Hawkeye. Hawkeye is based on residual clustering, an unsupervised anomaly detection method for multivariate time-series data. To ensure accurate anomaly detection, a trend-capturing prediction module is also combined with an automatic labeling module. As a result, the most common information can be learned from multivariate time-series data to reconstruct data trends. By evaluating Hawkeye on public datasets and real-world offshore wind turbine operational datasets, the results show that Hawkeye’s F₁-score improves by an average of 14% compared with Isolation Forest, and its size shrinks by up to 11.5 times on the largest dataset compared with other methods. The proposed Hawkeye is potential to real-time monitoring and early warning systems for wind turbines, accelerating the development of intelligent operation and maintenance.

Keywords

Anomaly detection / Offshore wind turbines / Residual clustering / Time-series / Unsupervised learning

Funding

This work was supported by Shanghai Electric Power Energy Technology Co., Ltd. 2023 Science and Technology Project under Grant No. 33019006220801.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Yi-Feng Li, Zhi-Ang Hu, Jia-Wei Gao, and Yi-Sheng Zhang are currently employed by Shanghai Electric Power Energy Technology Co., Ltd., Shanghai, China, and also the research project is funded by Shanghai Electric Power Energy Technology Co., Ltd.

References

[1] L.-F. Shen, Z.-C. Li, J.T. Kwok, Timeseries anomaly detection using temporal hierarchical one-class network, in: Proc. of the 34th Conf. on Neural Information Processing Systems, Red Hook, USA, (2020), pp. 13016-13026.

[2] Y.-C. Zhang, X.-J. Su, K. Meng, Z.-Y. Dong, Robust fault detection approach for wind farms considering missing data tolerance and recovery, IET Renew. Power Gen. 14 (19) (Dec. 2020) 4150-4158.

[3] M. Thill, W. Konen, H. Wang, T. Bäck, Temporal convolutional autoencoder for unsupervised anomaly detection in time series, Appl. Soft Comput. 112 (Nov. 2021) 107751:1-22.

[4] K. Krishna, M.N. Murty, Genetic k-means algorithm, IEEE T. Syst. Man Cy. B 29 (3) (Jun. 1999) 433-439.

[5] M.M. Breunig, H.-P. Kriegel, R.T. Ng, J. Sander, LoF: identifying density-based local outliers, in: Proc. of ACM SIGMOD Intl. Conf. on Management of Data, Dallas, USA, (2000), pp. 93-104.

[6] L.A. Maglaras, J.-M. Jiang, T.J. Cruz, Combining ensemble methods and social network metrics for improving accuracy of OCSVM on intrusion detection in SCADA systems, J. Inf. Secur. Appl. 30 (Oct. 2016) 15-26.

[7] F.T. Liu, K.-M. Ting, Z.-H. Zhou, Isolation Forest, in: Proc. of the 8th IEEE Intl. Conf. on Data Mining, Pisa, Italy, (2008), pp. 413-422.

[8] S.R. Moreno, L. dos Santos Coelho, H.V.H. Ayala, V.C. Mariani, Wind turbines anomaly detection based on power curves and ensemble learning, IET Renew. Power Gen. 14 (19) (Dec. 2020) 4086-4093.

[9] H. Xu, C. Caramanis, S. Mannor, Outlier-robust PCA: the high-dimensional case, IEEE Trans. Inf. Theor. 59 (1) (Jan. 2012) 546-572.

[10] S. Guha, N. Mishra, G. Roy, O. Schrijvers, Robust random cut forest based anomaly detection on streams, in: Proc. of the 33rd Intl. Conf. on Machine Learning, New York, USA, (2016), pp. 2712-2721.

[11] J.-B. Li, W. Pedrycz, I. Jamal, Multivariate time series anomaly detection: A framework of hidden Markov models, Appl. Soft Comput. 60 (Nov. 2017) 229-240.

[12] A. Ryzhikov, M. Borisyak, A. Ustyuzhanin, D. Derkach, NFAD: Fixing anomaly detection using normalizing flows [Online]. Available, https://arxiv.org/abs/1912.09323, November 2021.

[13] P.-F. Marteau, S. Soheily-Khah, N. Béchet, Hybrid Isolation Forest―application to intrusion detection [Online]. Available, https://arxiv.org/abs/1705.03800, May 2017.

[14] S. Lee, T. Park, K. Lee, Soft contrastive learning for time series, in: Proc. of the 12th Intl. Conf. on Learning Representations, Vienna, Austria, (2024), pp. 1-25.

[15] H.-J. Li, H.-Z. Xu, W. Peng, C.-R. Shen, X.-W. Qiu, Multi-scale sampling based MLP networks for anomaly detection in multivariate time series, in: Proc. of the IEEE 29th Intl. Conf. on Parallel and Distributed Systems, Danzhou, China, (2023), pp. 1421-1428.

[16] C. Landin, J. Liu, K. Katsarou, S. Tahvili, Time series anomaly detection using convolutional neural networks in the manufacturing process of RAN, in: Proc. of IEEE Intl. Conf. on Artificial Intelligence Testing, Athens, Greece, (2023), pp. 90-98.

[17] Y. Su, Y.-J. Zhao, C.-H. Niu, R. Liu, W. Sun, D. Pei, Robust anomaly detection for multivariate time series through stochastic recurrent neural network, in: Proc. of the 25th ACM SIGKDD Intl. Conf. on Knowledge Discovery & Data Mining, Anchorage, USA, (2019), pp. 2828-2837.

[18] M.-L. Shyu, S.-C. Chen, K. Sarinnapakorn, L.-W. Chang, A novel anomaly detection scheme based on principal component classifier, in: Proc. of IEEE Foundations and New Directions of Data Mining Workshop, Piscataway, USA, (2003), pp. 172-179.

[19] W.-Q. Jiang, Y. Hong, B.T. Zhou, X. He, C. Cheng, A GAN-based anomaly detection approach for imbalanced industrial time series, IEEE Access 7 (Sept. 2019) 143608-143619.

[20] J.-H. Xu, H.-X. Wu, J.-M. Wang, M.-S. Long, Anomaly transformer: time series anomaly detection with association discrepancy, in: Proc. of the 10th Intl. Conf. on Learning Representations, Virtual Event, (2022), pp. 1-20.

[21] Z.A. Khan, T. Hussain, A. Ullah, S. Rho, M. Lee, S.W. Baik, Towards efficient electricity forecasting in residential and commercial buildings: A novel hybrid CNN with a LSTM-AE based framework, Sensors 20 (5) (Mar. 2020) 1399:1-16.

[22] Z.-X. Wang, C.-H. Pei, M.-H. Ma, et al., Revisiting VAE for unsupervised time series anomaly detection: A frequency perspective, in: Proc. of ACM on Web Conf., Singapore, Singapore, (2024), pp. 3096-3105.

[23] H. Cheng, Q.-S. Wen, Y. Liu, L. Sun, RobustTSF: Towards theory and design of robust time series forecasting with anomalies, in: Proc. of the Twelfth Intl. Conf. on Learning Representations, Vienna, Austria, (2024), pp. 1-24.

[24] H.-Z. Xu, G.-S. Pang, Y.-J. Wang, Y.-J. Wang, Deep isolation forest for anomaly detection, IEEE T. Knowl. Data Eng. 35 (12) (Dec. 2023) 12591-12604.

[25] H.-Z. Xu, Y.-J. Wang, S.-L. Jian, Q. Liao, Y.-J. Wang, G.-S. Pang, Calibrated one-class classification for unsupervised time series anomaly detection, IEEE T. Knowl. Data Eng., (2024),10.1109/TKDE.2024.3393996.

[26] Y.-H. Zhang, J.-C. Yan, Crossformer: Transformer utilizing cross-dimension dependency for multivariate time series forecasting, in: Proc. of the 11th Intl. Conf. on Learning Representations, Kigali, Rwanda, (2023), pp. 1-21.

[27] L. Al Shalabi, Z. Shaaban, B. Kasasbeh, Data mining: A preprocessing engine, J. Comput. Sci. 2 (9) (Sept. 2006) 735-739.

[28] A. Vaswani, N. Shazeer, N. Parmar, et al., Attention is all you need, in: Proc. of the 31st Intl. Conf. on Neural Information Processing Systems, Long Beach, USA, (2017), pp. 6000-6010.

[29] UCI Machine Learning Repository, KDD Cup, 1999. Data [Online]. Available, https://www.kaggle.com/datasets/galaxyh/kdd-cup-1999-data, October 1999

[30] J. Audibert, P. Michiardi, F. Guyard, S. Marti, M.A. Zuluaga, USAD: Unsupervised anomaly detection on multivariate time series, in: Proc. of the 26th ACM SIGKDD Intl. Conf. on Knowledge Discovery & Data Mining, Virtual Event, (2020), pp. 3395-3404.