Forecasting solar energy generation using deep learning models

2025;
: pp. 669–681
https://doi.org/10.23939/mmc2025.02.669
Received: February 27, 2025
Revised: June 23, 2025
Accepted: June 25, 2025

Khasyshyn N., Liubinskyi B.  Forecasting solar energy generation using deep learning models.  Mathematical Modeling and Computing. Vol. 12, No. 2, pp. 669–681 (2025)

1
Lviv Polytechnic National University
2
Lviv Polytechnic National University

The application of deep learning models for forecasting solar energy generation is considered.  An analysis and comparison of the efficiency of recurrent (LSTM, GRU), convolutional (CNN), and temporal convolutional networks (TCN) for forecasting time series of solar energy generation were conducted.  The possibility of improving forecasting accuracy by constructing a hybrid model combining ARIMA and CNN was explored.  The results of experiments for different EU countries are presented, and a comparison of models in terms of forecasting accuracy and computational efficiency is performed as well.

  1. Chollet F.  Deep Learning with Python. Manning Publications (2018).
  2. Majka M.  ARIMAX: Time Series Forecasting with External Variables (2024). https://www.researchgate.net/publication/384196976_ARIMAX_Time_Series_Fo....
  3. Brockwell P. J., Davis R. A.  Introduction to Time Series and Forecasting.  Springer, New York (2002).
  4. Jebli I., Belouadha F.-Z., Kabbaj M. I., Tilioua A.  Deep Learning based Models for Solar Energy Prediction.  Advances in Science, Technology and Engineering Systems Journal.  6 (1), 349–355 (2021).
  5. Abumohsen M., Owda A. Y., Owda M.  Electrical Load Forecasting Using LSTM, GRU, and RNN Algorithms.  Energies.  16 (5), 2283 (2023).
  6. Arslan S.  A hybrid forecasting model using LSTM and Prophet for energy consumption with decomposition of time series data.  PeerJ Computer Science.  8, e1001 (2022).
  7. Senzeybek B.  Time-series analysis and forecasting with Meta's Prophet library.  Medium (2023).  https://medium.com/@barisenzeybek/time-series-analysis-and-forecasting-w....
  8. Verma J., Agarwal S.  A Hybrid Deep Neural Network Model for Time Series Forecasting.  International Journal of Advanced Trends in Computer Science and Engineering.  11 (1), 20–25 (2022).
  9. Scott G.  Box–Jenkins Model: Definition, Uses, Timeframes, and Forecasting (2022).  https://www.investopedia.com/terms/b/box-jenkins-model.asp.
  10. Kusnadi R.  Introduction to RNN and LSTM algorithm.  Medium (2020).  https://medium.com/@ryo.kusnadi/introduction-to-rnn-and-lstm-algorithm-5....
  11. Sagheer A., Kotb M.  Time series forecasting of petroleum production using deep LSTM recurrent networks.  Neurocomputing.  324, 56–66 (2018).
  12. Rink K.  Time Series Forecast Error Metrics You Should Know.  An overview and introduction to the most common error metrics (2021).  https://towardsdatascience.com/time-series-forecast-error-metrics-you-sh....
  13. Zhang A., Lipton Z. C., Li M., Smola A. J.  Gated recurrent units (GRU).  In: Dive into Deep Learning. Cambridge University Press (2021).  https://d2l.ai/chapter_recurrent-modern/gru.html.
  14. Hochreiter S., Schmidhuber J.  Long Short-Term Memory.  Neural Computation.  9 (8), 1735–1780 (1997).
  15. Liquet B., Moka S., Nazarathy Y.  The Mathematical Engineering of Deep Learning.  Ch. 5 Convolutional Neural Networks (2021).
  16. Solar hourly generation time series at country, NUTS 1, NUTS 2 level and bidding zones.  EMHIRES: European Meteorological derived High Resolution RES generation time series for present and future scenarios.  https://data.jrc.ec.europa.eu/dataset/jrc-emhires-solar-generation-time-....
  17. Photovoltaic Geographical Information System (PVGIS).  https://joint-research-centre.ec.europa.eu/photovoltaic-geographical-inf....
  18. Cho K., van Merriënboer B., Bahdanau D., Bougares F., Schwenk H., Bengio Y.  Learning phrase representations using RNN encoder-decoder for statistical machine translation.  Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP).  1724–1734 (2014).
  19. Khashei M., Bijari M.  A novel hybridization of artificial neural networks and ARIMA models for time series forecasting.  Applied Soft Computing.  11 (2), 2664–2675 (2011).
  20. Gers F. A., Schmidhuber J., Cummins F.  Learning to Forget: Continual Prediction with LSTM.  Neural Computation.  12 (6), 2451–2471 (2000).
  21. Ivashchenko H., Timoshenko D., Bliznyuk O, Kononenko O.  Deep learning models for time series forecasting. Control, Navigation and Communication Systems.  1, 82–87 (2024), (in Ukrainian).