Journal of Electronic Science and Technology, Volume. 22, Issue 4, 100285(2024)
Efficient anomaly detection method for offshore wind turbines
Figures & Tables(11)
Fig. 1. Two examples on the real-world offshore wind turbine operational dataset: (a) contrastive of L1L2 line voltage and (b) contrastive of wind speed (mechanical).
Fig. 3. Difference between DSW embedding and conventional embedding approaches. DSW uses episodes of different features to predict different episodes of data with different features. While the conventional embedding approach uses all the features in the same time episode for embedding and does not take into account the correlation between the multivariate variables.
Fig. 5. Comparison of (a) multi-headed self-attention mechanism and (b) our proposed router approach.
Fig. 7. Anomaly detection results shown on two features: (a) L1L2 line voltage and (b) wind speed (mechanical).
Table 1. Datasets description.
View tableView in ArticleTable 1. Datasets description.
Dataset Dimension Train Test Contamination (%) PSM 27 70272 17568 30 KDD99 41 296413 98804 20 OWTD 61 60359 17279 7
Table 1. [in Chinese]
View tableView in ArticleTable 1. [in Chinese]
Algorithm 1: Hawkeye process Data: Data ${\mathbf{x}}$ (n objects × D attributes)Result: Anomaly labels 1: Function Predict( ${\mathbf{x}}$ )2: Initialization 3: FillMissingData( ${\mathbf{x}}$ )4: Standardization( ${\mathbf{x}}$ )5: PredictionModule(x) → Predicteddata 6: residual ← $|{\mathbf{x}} - {{\hat {\bf{x}}}}|$ 7: residual ← Standardization(residual) 8: return residual 9: Function Automatic labeling (residual data, max_iter, tol) 10: Initialize cluster centers C with 3 randomly selected points from residual 11: for iter ← 1 to max_iter do 12: for each point residuali ∈ residual do 13: Compute distance di,k from residuali to each cluster center 14: Assign residuali to the nearest cluster center, update labelsi 15: end 16: for each cluster j∈{1, 2, ···, k} do 17: Update cluster center ${C_j}$ as the mean of points within the cluster18: end 19: if change in cluster centers is less than tol then 20: break 21: end 22: end 23: return Anomaly labels
Table 2. Performance on labeled public datasets.
View tableView in ArticleTable 2. Performance on labeled public datasets.
Method KDD99 PSM P R ${F_1}$ P R ${F_1}$ LoF 0.47 0.23 0.26 0.58 0.54 0.56 OCSVM 0.64 0.75 0.69 0.67 0.59 0.61 Isolation Forest 0.49 0.30 0.37 0.72 0.65 0.66 USAD 0.69 0.72 0.66 0.73 0.62 0.62 Ours 0.89 0.77 0.79 0.77 0.77 0.77
Table 3. Anomaly detection results on an offshore wind turbine dataset.
View tableView in ArticleTable 3. Anomaly detection results on an offshore wind turbine dataset.
Ours USAD Isolation Forest OCSVM Amount 818 2803 3456 8639 Ratio (%) 4.7 10.1 16.0 20.0
Tools
Get Citation
Copy Citation Text
Yi-Feng Li, Zhi-Ang Hu, Jia-Wei Gao, Yi-Sheng Zhang, Peng-Fei Li, Hai-Zhou Du. Efficient anomaly detection method for offshore wind turbines[J]. Journal of Electronic Science and Technology, 2024, 22(4): 100285
Paper Information
Category:
Received: May. 12, 2024
Accepted: Oct. 16, 2024
Published Online: Jan. 23, 2025
The Author Email: Du Hai-Zhou (duhaizhou@shiep.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20