Prediction of Solar Radiation with Artificial Neural Network for Samsun Region

Abstract

It is very important to predict solar radiation in the design of solar energy systems from renewable energy sources. Solar energy depends on global solar radiation. Various meteorological variables such as air temperature, sunshine duration, vapor pressure, cloudiness are used as input variables in estimating solar radiation with Artificial Neural Network Model (ANN). In this study, ANN was used to predict the daily solar radiation values of Samsun region between March 2017 and February 2019. Levenberg-Marquardt training algorithm, logarithmic sigmoid and linear transfer function were used for different input parameters in ANN method. The best model performance was obtained with 9-input meteorological data (average temperature, average wind speed, average vapor pressure, average cloudiness rate, sunshine duration, maximum temperature, minimum temperature, soil temperature at 5 cm). The correlation coefficient (R) for the test data was 0.9603 and the mean square error (MSE) was 0.3516. It has been observed that the feed forward ANN model provides a high performance for predicting solar radiation along with other meteorological parameters. In addition, when the sunshine duration was given as input to the ANN, R value was obtained as 0.9032.

Keywords

• Solar radiation
• Meteorological variables
• Feed-Forward neural network
• Levenberg-Marquardt

Samsun Bölgesi için Güneş Radyasyonunun Yapay Sinir Ağı ile Tahmini

Öz

Yenilenebilir enerji kaynaklarından güneş enerji sistemlerinin tasarımında güneş radyasyonunu tahmin etmek oldukça önemlidir. Güneş enerjisi, global güneş radyasyonuna bağlıdır. Güneş radyasyonunun Yapay Sinir ağı modeli (YSA) ile tahmin edilmesinde girdi değişkeni olarak, hava sıcaklığı, güneşlenme süresi, buhar basıncı, bulutluluk gibi çeşitli meteorolojik değişkenler kullanılmaktadır. Bu çalışmada, Samsun bölgesi için Mart 2017-Şubat 2019 tarihleri arasındaki günlük güneş radyasyonu YSA yöntemi ile tahmin edilmeye çalışılmıştır. YSA yönteminde farklı giriş değişkenleri için Levenberg-Marquardt eğitim algoritması, logaritmik sigmoid ve doğrusal transfer fonksiyonu kullanılmıştır. Model performansı en yüksek 9 girişli meterolojik veriler (ortalama sıcaklık, ortalama nispi nem, ortalama rüzgar hızı, ortalama buhar basıncı, ortalama bulutluluk oranı, güneşlenme süresi, maksimum sıcaklık, minimum sıcaklık, 5 cm’de toprak sıcaklığı) ile elde edilmiştir. Test verileri için R değeri 0.9603, MSE değeri 0.3516 olarak bulunmuştur. İleri beslemeli YSA modeli yaklaşımının diğer meteorolojik değişkenler ile birlikte güneş radyasyonunu tahmin etmek için yüksek bir performans sağladığı görülmüştür. Ayrıca, YSA’ya güneşlenme süresi giriş olarak verildiğinde, R değeri 0.9032 olarak elde edilmiştir.

Anahtar Kelimeler

• Güneş radyasyonu
• Meteorolojik değişkenler
• İleri beslemeli yapay sinir ağı
• Levenberg-Marquardt

Kaynakça

1. Alva, G., Lin, Y., Fang, G. (2018). An overview of thermal energy storage systems. Energy, 144, 341–378.
2. Amanollahi, J., Kaboodvandpour, S., Majidi, H., (2017). Evaluating the accuracy of ANN and LR models to estimate the water quality in Zarivar International Wetland, Iran. Natural Hazards, 85,1511–1527.
3. Atik, K., Deniz, E., Yıldız, E. (2007). Meteorolojik Verilerin Yapay Sinir Ağları ile Modellenmesi. KSÜ Fen ve Mühendislik Dergisi, 10 (1), 148-152.
4. Atkins, M.J., Walmsley, M.R.W., Morrison, A.S. (2010). Integration of solar thermal for improved energy efficiency in low-temperature-pinch industrial processes. Energy 35, 1867–1873.
5. Azadeh, A., Maghsoudi, A. and Sohrabkhani, S. (2009). An integrated artifcial neural networks approach for predicting global radiation. Energ. Convers. Manag. 50, 1497–1505.
6. Bayat, K. and Mirlatifi, S.M. (2009). Estimation of Global Solar Radiation using Regression and Artificial Neural Networks Models. Bimonthly Journal of Agricultural Sciences and Natural Resources, 16, 3, 270-280.
7. Behrang, M. A, Assareh, E., Ghanbarzadeh, A. and Noghrehabadi, A., (2010). The potential of different artiBcial neural network (ANN) techniques in daily global solar radiation modelling based on meteorological data. Sol. Energy 84; 1468–1480.
8. Bojanowski, J. S,. Donatelli, M., Skidmore, A. K. and Vrieling, A. (2013). An auto-calibration procedure for empirical solar radiation models; Environ. Model. Softw. 49, 118–128.

9. Changa, N.B., Daranpob, A., Yang, Y.J. and Jinc, K.R. (2009). Comparative Data Mining Analysis for Information Retrieval of MODIS Images: Monitoring Lake Turbidity Changes at Lake Okeechobee, Florida. Journal of Applied Remote Sensing, 3, 033549.
10. Culley, M.R., Carton, A.D., Weaver, S.R., Ogley-Oliver, E., StreetSun, J.C. (2011). Wind, Rock and Metal: Attitudes toward Renewable and Non-renewable Energy Sources in the Context of Climate Change and Current Energy Debates. Curr Psychol, 30, 215–233.
11. Gibb, D., Johnson, M., Romaní, J., Gasia, J., Cabeza, L.F., Seitz, A. (2018). Process integration of thermal energy storage systems – evaluation methodology and case studies. Applied Energy, 230, 750–760.
12. Heidari, E., Sobati, M.A., Movahedirad, S. (2016). Accurate prediction of nanofluid viscosity using a multilayer perceptron artificial neural network (MLP-ANN). Chemometrics and Intelligent Laboratory Systems, 155, 73–85.
13. IEA - International Energy Agency, (2018a). Key World energy statistics. Chemometrics and Intelligent Laboratory Systems 155 (2016) 73–85
14. IEA- International Energy Agency, (2018b). CO2 emissions from fuel combustion 2018 report.
15. Intergovernmental Panel on Climate Change. (2007). Climate change 2007: Synthesis report. New York: Cambridge University Press Koçak, B., Fernandez, A.I., Paksoy, H. (2020). Review on sensible thermal energy storage for industrial solar applications and sustainability aspects. Solar energy, 135-169.
16. Kumar, S., Kaur, T., ve Arora, M. (2015). Estimationof Solar Radıation Using Artifical Neural Network. International Journal of Sciencen Technology & Management, 4, 1, 658-662.
17. Lee, J.S., Choi, H. (2019). Influence of Construction Schemes for a Non-compensatory Composite Indicator on Flood Vulnerability Assessments in the Korean Peninsula. International Journal of Civil Engineering, 1317–1325.
18. Naderloo, L (2020). Prediction of solar radiation on the horizon using neural network methods, ANFIS and RSM (case study: Sarpol-e-Zahab Township, Iran). J. Earth Syst. Sci., 129, 148.
19. Moghadassi, A., Parvizian, F., Hosseini, S. (2009). A new approach based on artificial neural networks for prediction of high pressure vapor–liquid equilibrium Australian Journal of Basic and Applied Sciences, 3, 1851–1862.
20. Oğuz, K. ve Pekin, M.A. (2019). Predictability of Fog Visibility with Artificial Neural Network for Esenboga Airport. European Journal of Science and Technology, 15, 542-551.
21. Oreskes, N. (2004). The scientific consensus on climate change. Science, 306, 1686–1686.
22. Özcan, C ve Köprü, E.Y. (2020). Yapay Sinir Ağları ile Sıvı Ham Demir Tahmini ve 5.Yüksek Fırın Uygulaması. Avrupa Bilim ve Teknoloji Dergisi, Özel sayı, 155-162.
23. Strobl, R.O., Forte , F. (2007) Artificial neural network exploration of the influential factors in drainage network derivation. Hydrol Process 21(22):2965–2978.
24. Şahan, M., Okur, Y. (2016). Akdeniz bölgesine ait meteorolojik veriler kullanılarak yapay sinir ağları yardımıyla güneş enerjisinin tahmini. SDÜ Fen Edebiyat Fakültesi Fen Dergisi, 11, 1 .
25. Turan, M., Dengiz, O., Turan, İ.D. (2018). Samsun İlinin Newhall Modeline Göre Toprak Sıcaklık ve Nem Rejimlerinin Belirlenmesi. Türkiye Tarımsal Araştırmalar Dergisi , 5(2), 131-142.
26. UNEP (2015). Solar Heat for Industrial Processes.
27. Zhang, T., Liu, Y., Rao, Y., Li, X., Zhao, Q. (2020). Optimal design of building environment with hybrid genetic algorithm, artificial neural network, multivariate regression analysis and fuzzy logic controller. Building and Environment, 175, 106810.