Monthly Electricity Generatıon Forecast in Solar Power Plants with LSTM
Abstract
Today, with the intensive use of electrical devices, the need for electricity has increased. Fossil fuels are generally used to meet this need. However, considering the damage caused by fossil fuels to the environment, governments make various incentives for renewable energy sources. The incentives of countries for solar power plants are quite large. Recently, there are many investors who want to build solar power plants. The sunshine duration of our country is quite high. And the fact that the climatic conditions are efficient for the generation of electricity attracts many investors. However, the installation of these power plants is quite costly. It is possible to predict the amortization periods of these costs with the ever-developing artificial intelligence technology. In this study, the energy data to be produced in the future is estimated by using real solar power plant data with machine learning algorithms. Data, take from solar power plants owned by Humartaş Energy company. In the study, predictions and analyses were made using the LSTM (Long Short-Term Memory) method, which is one of the artificial neural networks. The error rate of the study between 1% and 15%. It is foreseen that studies will also be implemented with other renewable energy sources such as wind, geothermal, hydraulic energy data in the coming stages.
Keywords
LSTM, Artificial intelligence, Solar power plants, Energy forecast
References
- [1] Word Energy Outlook 2013, (2013). World Energy Outlook 2013 [Online]. Available: www.iea.org/reports/world-energy-outlook-2013.
- [2] T.C. Enerji ve Tabii Kaynaklar Bakanlığı, (2020). Güneş [Çevrimiçi]. Erişim: www.enerji.gov.tr/bilgi-merkezi-enerji-gunes.
- [3] F. Özen, (2019, Haziran). Yenilenebilir Enerjide Yapay Zeka Uygulamaları [Çevrimiçi]. Erişim: www.researchgate.net/publication/333816757_YENILENEBILIR_ENERJIDE_YAPAY_ZEKA_UYGULAMALARI.
- [4] T. Boyekin, I. Kıyak, “Rooftop solar power plant based electric vehicle charging station,” in 6. European Conference On Renewable Energy Systems (ECRES), 2018, pp. 959-966.
- [5] H. Hemza, C. Abdeslam, M. P. Rachid and D. Barakel, “Tracing current-voltage curve of solar panel based on LabVIEW Arduino interfacing,” Bilişim Teknolojileri Dergisi, vol. 8, no. 3, pp. 117–123, 2015.
- [6] V. Ciocia, A. Boicea, A. Dematteis, P. D. Leo, F. Giordano and F. Spertino, “PV system integration in buildings: an energy and economic case study,” in 2017 10th International Symposium on Advanced Topics in Electrical Engineering (ATEE), 2017, pp. 786-790.
- [7] P. Grunow, A. Preiss, S. Koch and S. Krauter “Yield and Spectral Effects of A-Si Modules,” in Proceedings of the 24th European Photovoltaic Solar Energy Conference, 2009, pp. 2846-2829.
- [8] B. Herteleer, J. Cappelle and J. Driesen, “Quantifying low-light behaviour of photovoltaic modules by identifying their irradiance-dependent efficiency from data sheets,” in European Photovoltaic Solar Energy Conference, pp. 1-8, 2012.
- [9] Solar PV Word Expo, (2021, August). Solar Cells [Online], Available: https://www.pvresources.com/en/solarcells/solarcells.php.
- [10] E. Koutroulis, F. Blaabjerg, “Methods for the Optimal Design of Grid-Connected PV Inverters, International Journal Of Renewable Energy Research,” IJRER, vol. 1, no. 2, pp. 54-64, 2011.