Day-ahead and Intraday wind speed forecasting method based on OS-ELM model with CEEMDAN

Nguyen Thi Hoai  Thu; Ngo Van  Khanh; Do Xuan  Bach

doi:10.6180/jase.2026.26030001

Day-ahead and Intraday wind speed forecasting method based on OS-ELM model with CEEMDAN

Nguyen Thi Hoai ThuThis email address is being protected from spambots. You need JavaScript enabled to view it., Ngo Van Khanh, and Do Xuan Bach

PGRE Lab., School of Electrical and Electronic Engineering, Hanoi University of Science and Technology

Received: April 29, 2025
Accepted: October 26, 2025
Publication Date: November 30, 2025

Copyright The Author(s). This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are cited.

Download Citation: ||https://doi.org/10.6180/jase.2026.26030001

This paper presents a model for day-ahead and intraday wind speed forecasting, integrating Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN)and the Online Sequential Extreme Learning Machine (OS-ELM). The CEEMDAN method decomposes the original wind speed data into various Intrinsic Mode Functions (IMFs). The predictive outputs of all IMFs are aggregated to generate the final forecast. To ensure a thorough evaluation, different machine learning models were analyzed at 1-hour interval for a 24-step forecast and at 15-minute interval for 1, 4, and 16-step forecasts. The models were compared using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Normalized Root Mean Squared Error (N-RMSE). The RMSE of the proposed model was 0.34 m/s, 0.64 m/s, 1.07 m/s, and 1.50 m/s for 1, 4, 16, and 24-step forecasting, respectively. Compared to other decomposition-based methods, the proposed model’s RMSE was approximately 3-10% lower. Additionally, compared to single models such as LSTM and ELM, the proposed model’s RMSE was reduced by approximately 30– 50%. The results indicate that the integration of the CEEMDAN method with the OS-ELM model yields superior performance in wind speed forecasting compared to other models.

Keywords: Windspeed forecasting; Online Sequential Extreme Learning Machine (OS-ELM); CEEMDAN decomposition; Day-ahead forecasting; Intraday forecasting

[1] M.Q.Raza, M. Nadarajah, and C. Ekanayake, (2016) “On recent advances in PV output power forecast" Solar Energy 136: 125–144. DOI: 10.1016/j.solener.2016.06.073.
[2] A. Dolara, S. Leva, M. Mussetta, and E. Ogliari. “PV hourly day-ahead power forecasting in a micro grid context”. In: 2016 IEEE 16th International Conference on Environment and Electrical Engineering (EEEIC). 2016, 1–5. DOI: 10.1109/EEEIC.2016.7555636.
[3] A. Gupta, A. Kumar, and B. Kadhirvel, (2021) “Day ahead and intra-day wind power forecasting based on feed back error correction" IET Renewable Power Generation 15: DOI: 10.1049/rpg2.12211.
[4] J. Cunha and C. Pereira, (2024) “A hybrid model based on STL with simple exponential smoothing and ARMA for wind forecast in a Brazilian nuclear power plant site" Nuclear Engineering and Design 421: 113026. DOI: 10.1016/j.nucengdes.2024.113026.
[5] A. ., S. Singh, and A. Mohapatra, (2019) “Repeated wavelet transform based ARIMA model for very short term wind speed forecasting" Renewable Energy 136: DOI: 10.1016/j.renene.2019.01.031.
[6] O. Shukur and M. H. Lee, (2015) “Daily wind speed forecasting through hybrid KF-ANN model based on ARIMA" Renewable Energy 76: 637–647. DOI: 10.1016/j.renene.2014.11.084.
[7] B. Gu, H. Hu, J. Zhao, H. Zhang, and X. Liu, (2023) “Short-term wind power forecasting and uncertainty analysis based on FCM–WOA–ELM–GMM" Energy Re ports 9: 807–819. DOI: 10.1016/j.egyr.2022.11.202.
[8] T. H. T. Nguyen, N. V. Pham, V. N. N. Nguyen, H. M. Pham, and Q. B. Phan, (2022) “Forecasting Wind Speed Using A Hybrid Model Of Convolutional Neural Network And Long-Short Term Memory With Boruta Algorithm Based Feature Selection" Journal of Applied Science and Engineering 26: 1053–1060. DOI: 10.6180/jase. 202308_26(8).0001.
[9] N. T. H. Thu, P. Q. Bao, and P. N. Van, (2023) “A Hybrid Model of Decomposition, Extended Kalman Filter and Autoregressive- Long Short- Term Memory Network for Hourly Day Ahead Wind Speed Forecasting" Journal of Applied Science and Engineering 27: 3133–3141. DOI: 10.6180/jase.202409_27(9).0004.
[10] U. Yuzgec, E. Dokur, and M. Balci, (2024) “A novel hybrid model based on Empirical Mode Decomposition and Echo State Network for wind power forecasting" Energy 300: 131546. DOI: 10.1016/j.energy.2024.131546.
[11] N.N.V.Nhat,D.N.Huu,andT.T.H.Nguyen, (2024) “Short-term multi-step forecasting of rooftop solar power generation using a combined data decomposition and deep learning model of EEMD-GRU" Journal of Renew able and Sustainable Energy 16(1): DOI: 10.1063/5.0176951.
[12] Z. Gong, A. Wan, Y. Ji, K. AL-Bukhaiti, and Z. Yao, (2024) “Improving short-term offshore wind speed forecast accuracy using a VMD-PE-FCGRU hybrid model" Energy 295: 131016. DOI: 10.1016/j.energy.2024.131016.
[13] Z.-H. Liu, C.-T. Wang, H.-L. Wei, B. Zeng, M. Li, and X.-P. Song, (2024) “A wavelet-LSTM model for short term wind power forecasting using wind farm SCADA data" Expert Systems with Applications 247: 123237. DOI: 10.1016/j.eswa.2024.123237.
[14] Z. Wu and N. Huang, (2009) “Ensemble Empirical Mode Decomposition: a Noise-Assisted Data Analysis Method" Advances in Adaptive Data Analysis 1: 1 41. DOI: 10.1142/S1793536909000047.
[15] N. T. Hoai Thu, P. Quoc Bao, and N. V. Nhat Nam. “Multiple Step Ahead Forecasting of Rooftop Solar Power Based on a Novel Hybrid Model of CEEM DAN-Bidirectional LSTM Network with Structure Optimized by PSO Method”. In: 2022 11th International Conference on Control, Automation and Information Sciences (ICCAIS). 2022, 522–528. DOI: 10.1109/ICCAIS56082.2022.9990323.
[16] X. Zhang, H. Ren, J. Liu, Y. Zhang, and W. Cheng, (2024) “A monthly temperature prediction based on the CEEMDAN–BO–Bi LSTM coupled model" Scientific Reports 14(1): 808. DOI: 10.1038/s41598-024-51524-7.
[17] Tomas Johannesson & Halldor Bjornsson, Icelandic Met. Office; Gabor Grothendieck. stinepack: Stineman, a Consistently Well Behaved Method of Interpolation. en. Institution: Comprehensive R Archive Network Pages: 1.5. 2006. DOI: 10.32614/CRAN.package. stinepack.
[18] A. Smiti, (2020) “A critical overview of outlier detection methods" Computer Science Review 38: 100306. DOI: 10.1016/j.cosrev.2020.100306.
[19] M.E. Torres, M. A. Colominas, G. Schlotthauer, and P. Flandrin. “A complete ensemble empirical mode decomposition with adaptive noise”. In: 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2011, 4144–4147. DOI: 10.1109/ ICASSP.2011.5947265.
[20] N.-y. Liang, G.-b. Huang, P. Saratchandran, and N. Sundararajan, (2006) “A Fast and Accurate Online Sequential Learning Algorithm for Feedforward Networks" IEEE Transactions on Neural Networks 17(6): 1411 1423. DOI: 10.1109/TNN.2006.880583.