Amirkabir Journal of Mechanical Engineering

Amirkabir Journal of Mechanical Engineering

A Data-Driven Framework for Wind Turbine Power Curve Cleaning and Abnormal Data detection Based on Binning and Quantiles

Document Type : Research Article

Authors
Alireza Aghajani Mobarakeh
Javad Poshtan
Control Group, Department of Electrical Engineering, Iran University of Science and Technology, Tehran, Iran
10.22060/mej.2026.24688.7894
Abstract
In this study, a preprocessing approach is proposed to improve training accuracy and computational efficiency by leveraging SCADA data and integrating machine learning algorithms with statistical techniques for unlabeled data. The method reduces the training dataset substantially by partitioning the data into equal points intervals and selecting representative samples based on quantiles, a procedure referred to as RD. Using this strategy, only 0.2% of the original dataset is required for model training in this study. Subsequently, abnormal data points are identified using a power curve model with quantile thresholds. The RD method is evaluated against DBSCAN and a KNN-based model. Experimental results obtained from real-world wind farm data indicate that RD combined with KNN outperforms DBSCAN. Specifically, both MAE and RMSE are reduced by approximately 15%, reflecting improved predictive accuracy. From a computational perspective, the execution time of RD is about 0.15 seconds, compared to 0.99 seconds for DBSCAN, corresponding to a reduction in runtime exceeding 50%. Moreover, unlike DBSCAN, which requires precise parameter tuning or additional constraints tailored to the power curve structure when dealing with dense or linear outliers, the proposed approach is capable of automatically eliminating outlier data points from the wind turbine power curve without the need for predefined filters or explicit boundary definitions prior to the cleaning process.
Keywords
Data Cleaning
Wind Turbine Power Curve (WTPC)
Machine Learning (ML)
Supervisory Control and Data Acquisition (SCADA)
Subjects
[1] C. Global Wind Energy, Global wind report 2023, Lisbon, 2023.
[2] C. Global Wind Energy, Global wind report 2024, Lisbon, 2024.
[3] F. Bilendo, H. Badihi, N. Lu, P. Cambron, B. Jiang, A normal behavior model based on power curve and stacked regressions for condition monitoring of wind turbines, IEEE Transactions on Instrumentation and Measurement, 71 (2022) 1–13.
[4] A. Llombart, J. Marques, J. Riera, Power curve characterization I: improving the bin method, in, International Conference on Renewable Energies and Power Quality (ICREPQ), 2005, pp. 367–371.
[5] R.K. Pandit, D. Infield, A. Kolios, Gaussian process power curve models incorporating wind turbine operational variables, Energy Reports, 6 (2020) 1658–1669.
[6] R.K. Pandit, D. Infield, Comparative assessments of binned and support vector regression-based blade pitch curve of a wind turbine for the purpose of condition monitoring, International Journal of Energy and Environmental Engineering, 10(2) (2019) 181–188.
[7] W. Hernandez, A. Méndez, J.L. Maldonado-Correa, F. Balleteros, Modeling of a Robust Confidence Band for the Power Curve of a Wind Turbine, Sensors, 16(12) (2016) 1–13.
[8] M. Ester, H.P. Kriegel, J. Sander, X. Xu, A density-based algorithm for discovering clusters a density-based algorithm for discovering clusters in large spatial databases with noise, in:  Proceedings of the Second International Conference on Knowledge Discovery and Data Mining, 1996, pp. 226 –231.
[9] H. Yang, J. Tang, W. Shao, J. Yin, B. Liu, Wind power data cleaning using RANSAC-based polynomial and linear regression with adaptive threshold, Scientific Reports, 15(1) (2025) 5105.
[10] R. Morrison, X. Liu, Z. Lin, Anomaly detection in wind turbine SCADA data for power curve cleaning, Renewable Energy, 184 (2022) 473–486.
[11] Z. Mehmood, Z. Wang, Hybrid iForest-DBSCAN for anomaly detection and wind power curve modelling, Expert Systems with Applications, 289 (2025) 128381–128381.
[12] B. Jing, Z. Qian, H. Zareipour, Y. Pei, A. Wang, Wind turbine power curve modelling with logistic functions based on quantile regression, Applied Sciences, 11(7) (2021) 3048–3048.
[13] R. Koenker, G. Bassett Jr, Regression quantiles, Econometrica, 46(1) (1978) 33–50.
[14] C. Wan, J. Lin, J. Wang, Y. Song, Z.Y. Dong, Direct quantile regression for nonparametric probabilistic forecasting of wind power generation, IEEE Transactions on Power Systems, 32(4) (2017) 2767–2778.
[15] K. Xu, J. Yan, H. Zhang, H. Zhang, S. Han, Y. Liu, Quantile based probabilistic wind turbine power curve model, Applied Energy, 296 (2021) 116913–116913.
[16] G. Liang, Y. Su, X. Wu, J. Ma, H. Long, Z. Song, Abnormal data cleaning for wind turbines by image segmentation based on active shape model and class uncertainty, Renewable Energy, 216 (2023) 118965.
[17] Z. Wang, L. Wang, C. Huang, A fast abnormal data cleaning algorithm for performance evaluation of wind turbine, IEEE Transactions on Instrumentation and Measurement, 70 (2021) 1–12.
[18] Y. Su, F. Chen, G. Liang, X. Wu, Y. Gan, Wind power curve data cleaning algorithm via image thresholding, in:  2019 IEEE International Conference on Robotics and Biomimetics (ROBIO), 2019, pp. 1198–1203.
[19] C. Zhang, D. Hu, T. Yang, Anomaly detection and diagnosis for wind turbines using long short-term memory-based stacked denoising autoencoders and XGBoost, Reliability Engineering and System Safety, 222 (2022) 108445–108445.
[20] G. Jiang, P. Xie, H. He, J. Yan, Wind turbine fault detection using a denoising autoencoder with temporal information, IEEE Transactions on Mechatronics, 23(1) (2018) 89–100.
[21] J. Wang, M. Kou, R. Li, Y. Qian, Z. Li, Ultra-short-term wind power forecasting jointly driven by anomaly detection, clustering and graph convolutional recurrent neural networks, Advanced engineering informatics, 65 (2025) 103137–103137.
[22] A. Meyer, Multi-target normal behaviour models for wind farm condition monitoring, Applied Energy, 300 (2021) 117342–117342.
[23] K. Leahy, R.L. Hu, I.C. Konstantakopoulos, C.J. Spanos, A.M. Agogino, D.T.J. O’Sullivan, Diagnosing and predicting wind turbine faults from scada data using support vector machines, International Journal of Prognostics and Health Management, 9(1) (2018) 1–11.
[24] A. Encalada-Davila, L. Moyon, C. Tutiven, B. Puruncajas, Y. Vidal, Early fault detection in the main bearing of wind turbines based on Gated Recurrent Unit (GRU) neural networks and SCADA data, IEEE/ASME Transactions on Mechatronics, 27(6) (2022) 5583 – 5593.
[25] L. Xiang, X. Yang, A. Hu, H. Su, P. Wang, Condition monitoring and anomaly detection of wind turbine based on cascaded and bidirectional deep learning networks, Applied Energy, 305 (2022) 117925–117925.
[26] S. Paudel, S. Faulstich, S. Sheng, Recommended key performance indicators for operational management of wind turbines, in, Journal of Physics: Conference Series, 2019, pp. 1–24.
[27] A. Aghajani Mobarakeh, A Review of Intelligent Fault Detection Methods for Wind Turbines, Iran University of Science and Technology (IUST), Tehran, Iran, 2024.
[28] E. Rashidian, Wind turbine fault diagnosis using SCADA data and normal behavior modeling, Iran University of Science and Technology (IUST), Tehran, Iran, 2023.
[29] T. Ouyang, A. Kusiak, Y. He, Modeling wind-turbine power curve: A data partitioning and mining approach, Renewable Energy, 102 (2017) 1–8.
Amirkabir Journal of Mechanical Engineering
Volume 57, Issue 10
January 2026
Pages 1263-1286