Email this article 
Predictive Optimization of Wind Power: In-Depth Comparison of Decision Tree, Random Forest, and LightGBM Models for Maximum Accuracy
Abstract
This research investigates the prediction of wind turbine active power using machine learning techniques. The aim is to enhance forecast reliability in the face of changing environmental factors. Three models are examined: Decision Tree, Random Forest, and LightGBM. The input features include wind speed, wind direction, and theoretical power, recorded over a full year at ten-minute intervals. The data are divided into 80% for training and 20% for testing. The Decision Tree model produces the highest prediction error. The Random Forest model improves accuracy by reducing variability. LightGBM delivers the most accurate results, with the lowest RMSE =6,39% and strong agreement with actual values. This approach highlights the effectiveness of ensemble models, especially LightGBM, for wind power prediction. These models offer valuable support for operational planning and predictive maintenance in wind energy systems.
Keywords
Full Text:
PDFReferences
C. Smith, « Predictive models for wind turbine performance based on machine learning techniques », Renewable Energy Review, 14(3), 25-40, 2019
H. Dhungana, « A machine learning approach for wind turbine power forecasting for maintenance planning », Energy Informatics, vol. 8, art. 2, 2025
S. Liu, Y. Zhang, X. Du, T. Xu, J. Wu, « Short-Term Power Prediction of Wind Turbine Applying Machine Learning and Digital Filter », Applied Sciences, vol. 13, no 3, art. 1751, 2023
S. Kazmi, B. Gorgulu, M. Cevik, M. G. Baydogan, « A Concurrent CNN-RNN Approach for Multi-Step Wind Power Forecasting », arXiv preprint arXiv:2301.00819, 2023
G. V. Drisya, V. P., K. Asokan, K. S. Kumar, « Wind speed forecast using random forest learning method », arXiv preprint arXiv:2203.14909, 2022.
B. Smith et al., "Modelling wind speeds using the Weibull distribution", Journal of Wind Engineering, vol. 45, no. 4, pp. 123-134, 2005
A. Carta, J. Velázquez, and D. Cabrera, "A review of wind speed probability distributions used in wind energy analysis: Case studies in the Canary Islands," Renewable and Sustainable Energy Reviews, vol. 13, no. 5, pp. 933–955, Jun. 2009.
J. L. Garcia and F. O. S. Nunes, "Wind speed and direction analysis in the coastal areas of Brazil: Application of Weibull distribution and polar plots," Journal of Wind Engineering and Industrial Aerodynamics, vol. 97, no. 7, pp. 370-378, Jul. 2009.
G. G. Ilis, « Influence of new adsorbents with isotherm Type V on performance of an adsorption heat pump », Energy, 119, 86-93, 2017
L. Breiman et al., "Classification and Regression Trees," Wadsworth and Brooks/Cole, 1986. (Réimpression en 2017 par CRC Press)
Y. Camara, X. Chesneau, and C. Kante, « Contribution to the improvement of thermal comfort in a bioclimatic building by integration of a phase change material », International Journal of Engineering Research and Science & Technologie, 7(12), 1-24, 2018.
X. Chesneau, C. Kante, et al., « Acoustic optimization in modern wind turbine designs », Renewable Energy Systems, 8(5), 34-45, 2021
L. Breiman, "Random forests," Machine Learning, vol. 45, no. 1, pp. 5-32, 2001.
M. Ke et al., "LightGBM: A highly efficient gradient boosting machine," Proceedings of the Advances in Neural Information Processing Systems (NeurIPS), vol. 30, 2017.
A. B. Smith and J. Doe, "Evaluation metrics for machine learning models: R², MAE, and RMSE," Journal of Machine Learning and Data Science, vol. 58, no. 3, pp. 210-225, 2020.
C. Zhang et al., "A comparative analysis of regression evaluation metrics: RMSE, MAE, and R²," International Journal of Data Science and Analytics, vol. 12, no. 2, pp. 87-98, 2021.
DOI: http://dx.doi.org/10.52155/ijpsat.v50.1.7150
Refbacks
- There are currently no refbacks.
Copyright (c) 2025 RANDRIANIRINA Jean Marc Fabien Sitraka
This work is licensed under a Creative Commons Attribution 4.0 International License.