Multi-Horizon Solar Radiation Estimation Using a Deep Learning-Based LSTM Model

Abstract

In this study, a multi-horizon time series forecasting approach is developed to estimate daily global solar radiation. The dataset is aligned according to daily frequency, and missing observations are handled with a two-stage assignment strategy combining time-based interpolation and seasonality-sensitive statistics based on day of the year (DOY). Input features are enriched with lagged values (1, 14, 30 days) and sliding window statistics (mean, standard deviation, minimum, and maximum). A stacked deep LSTM architecture with multiple output headers is used as the forecasting model and trained using the Log-Cosh loss function. Model performance is evaluated using RMSE, MAE, nRMSE, and R² metrics. The results show that forecasting accuracy improves significantly as the forecasting horizon increases. The MAE value is 1.166 for a 1-day horizon, while it decreases to 0.645 and 0.474 for 14-day and 30-day horizons, respectively. Similarly, R² increases from 0.642 for 1-day forecasts to 0.912 for 30-day forecasts.

Keywords

• Solar Radiation
• LSTM
• Deep Learning
• Time Series Analysis.

Derin Öğrenme Tabanlı LSTM Modeli Kullanarak Çok Ufuklu Güneş Radyasyonu Tahmini

Abstract

Bu çalışmada, günlük küresel güneş radyasyonunu tahmin etmek için çok ufuklu bir zaman serisi tahmin yaklaşımı geliştirilmiştir. Veri seti günlük frekansa göre hizalanmıştır ve eksik gözlemler, zamana dayalı enterpolasyon ve yılın gününe (DOY) dayalı mevsimsellik duyarlı istatistikleri birleştiren iki aşamalı bir atama stratejisi ile ele alınmıştır. Giriş özellikleri, gecikmeli değerler (1, 14, 30 gün) ve kayan pencere istatistikleri (ortalama, standart sapma, minimum ve maksimum) ile zenginleştirilmiştir. Çoklu çıktı başlıklarına sahip yığılmış derin LSTM mimarisi, tahmin modeli olarak kullanılmış ve Log-Cosh kayıp fonksiyonu kullanılarak eğitilmiştir. Model performansı, RMSE, MAE, nRMSE ve R² metrikleri kullanılarak değerlendirilmiştir. Sonuçlar, tahmin ufku arttıkça tahmin doğruluğunun önemli ölçüde iyileştiğini göstermektedir. 1 günlük ufuk için MAE değeri 1,166 iken, 14 günlük ve 30 günlük ufuklar için sırasıyla 0,645 ve 0,474'e düşmektedir. Benzer şekilde, R², 1 günlük tahminler için 0,642'den 30 günlük tahminler için 0,912'ye yükselmektedir.

Keywords

• Güneş Radyasyonu
• LSTM
• Derin Öğrenme
• Zaman Serisi Analizi.

Supporting Institution

Meteoroloji Bölge Müdürlüğü

Thanks

Bu çalışmada kullanılan meteorolojik verilerin temin edilmesi sürecinde sağladıkları destek ve katkılarından dolayı Van Meteoroloji Bölge Müdürlüğü’ne teşekkür ederiz.

References

1. Rahman S., Mohandes, M. Artificial neural network estimation of global solar radiation using air temperature and relative humidity, Energy Policy, vol. 36, no. 2, pp. 571–576, 2008.
2. Mellit, A., & Kalogirou, S. A., Artificial intelligence techniques for photovoltaic applications: A review,” Progress in Energy and Combustion Science, vol. 34, no. 5, pp. 574–632, 2008.
3. Perez R., Kivalov S., Schlemmer J., et al. Validation of short and medium term operational solar radiation forecasts in the US, Solar Energy, vol. 84, no. 12, pp. 2161–2172, 2010.
4. Angstrom, A. K. Solar and terrestrial radiation, Quarterly Journal of the Royal Meteorological Society, vol. 50, pp. 121–125, 1924.
5. Besharat F., Dehghan A. A., Faghih A. R. Empirical models for estimating global solar radiation: A review and case study, Renewable and Sustainable Energy Reviews, vol. 21, pp. 798–821, 2013.
6. De Souza J. L., Nicacio R. M., Moura, M. A. L. Global solar radiation measurements in Maceio, Brazil, Renewable Energy, vol. 30, pp. 1203–1220, 2005.
7. Menges H. O., Ertekin C., & Sonmete, M. H., Evaluation of global solar radiation models for Konya, Turkey, Energy Conversion and Management, vol. 47, pp. 3149–3173, 2006.
8. Yao W., Li Z., Wang Y., Jiang F., Hu, L. Evaluation of global solar radiation models for Shanghai, China, Energy Conversion and Management, vol. 84, pp. 597–612, 2014.

9. Wang H., Liu Y., Zhou B., Li C., Cao G., Voropai N., Barakhtenko E. Taxonomy research of artificial intelligence for deterministic solar power forecasting, Energy Conversion and Management, vol. 214, p. 112909, 2020.
10. Kumari P., Toshniwal, D. Deep learning models for solar irradiance forecasting: A comprehensive review, Journal of Cleaner Production, vol. 318, p. 128566, 2021
11. Al-Jaafreh T. M. L., Al-Odienat A., Altaharwah Y. A., The solar energy forecasting using LSTM deep learning technique, Proceedings of the 2022 International Conference on Emerging Trends in Computer Engineering Applications (ETCEA 2022), pp. 1–6, 2022.
12. Bellido-Jiménez J. A., Estévez J., García-Marín A. P. Assessing neural network approaches for solar radiation estimates using limited climatic data in the Mediterranean Sea, Proceedings of the 3rd International Electronic Conference on Atmospheric Sciences, p. 19, 2020.
13. Yu Y., Cao J., Zhu J. An LSTM short-term solar irradiance forecasting under complicated weather conditions, IEEE Access, vol. 7, pp. 145651–145666, 2019.
14. Mehdizadeh S., Behmanesh J., Khalili, K. Comparison of artificial intelligence methods and empirical equations to estimate daily solar radiation, Journal of Atmospheric and Solar-Terrestrial Physics, vol. 146, pp. 215–227, 2016.
15. Heng S. Y., Ridwan W. M., Kumar P., Ahmed A. N., Fai C. M,. Birima A. H., El-Shafie A. Artificial neural network model with different backpropagation algorithms and meteorological data for solar radiation prediction, Scientific Reports, vol. 12, no. 1, pp. 1–18, 2022.
16. Cornejo-Bueno L., Casanova-Mateo C., Sanz-Justo J., Salcedo-Sanz S. Machine learning regressors for solar radiation estimation from satellite data, Solar Energy, vol. 183, pp. 768–775, 2019.
17. Kotsiantis S. B., Decision trees: A recent overview, Artificial Intelligence Review, vol. 39, pp. 261–283, 2013.
18. Ikram R. M. A., Dai H., Ewees A. A., Shiri J., Kisi, O., Zounemat-Kermani M. Application of improved version of multi verse optimizer algorithm for modeling solar radiation, Energy Reports, vol. 8, pp. 12063–12080, 2022.
19. Azimi R., Ghayekhloo M., Ghofrani M. A hybrid method based on a new clustering technique and multilayer perceptron neural networks for hourly solar radiation forecasting, Energy Conversion and Management, vol. 118, pp. 331–344, 2016.
20. Natekin A., Knoll A. Gradient boosting machines, a tutorial, Frontiers in Neurorobotics, vol. 7, p. 21, 2013.
21. Güneş H., Hark C., Akkaya A. E. Comparison of Deep Learning Models and Optimization Algorithms in the Detection of Scoliosis and Spondylolisthesis from X-Ray Images, Sakarya University Journal of Science, vol. 28, no. 2, pp. 438–451, 2024.
22. Hark C. Sahte Haber Tespiti için Derin Bağlamsal Kelime Gömülmeleri ve Sinirsel Ağların Performans Değerlendirmesi, Fırat Üniversitesi Mühendislik Bilimleri Dergisi, vol. 34, no. 2, pp. 733–742, 2022.
23. Hark C., Ucxkan T., Karcı A. A new multi-document summarisation approach using saplings growing-up optimisation algorithms: Simultaneously optimised coverage and diversity, Journal of Information Science, vol. 50, no. 3, pp. 635–650, 2024.