Makine Öğrenimi Tabanlı Kısa Vadeli Fotovoltaik Çıkış Gücü Tahminlemesi

Öz

Fosil yakıt kaynaklarının sınırlı olması ve çevreye zararlı etkilerinin olması nedeniyle fotovoltaik (PV) sistemlerinin kurulumuna olan ihtiyaç giderek artmaktadır. PV sistemlerinin hava koşularına bağımlılığı PV güç çıkışlarında kararsızlığa, gerilim, frekans dalgalanmaları ve kesintilere neden olmaktadır. Bu durum ise PV enerjisinin şebekelere entegrasyonunu zorlaştırmaktadır. Bu yüzden PV güç çıkışını önceden kısa süreli tahmin etmek karşılaşılan zorlukların üstesinden gelmek için çok önemlidir. Bu çalışmanın amacı, literatürde makine öğrenimi modellerinde yaygın olan aşırı öğrenme ve yavaş öğrenme dezavantajlarının üstesinden gelerek daha hızlı öğrenen ve yüksek doğrulukta performans gösteren Gürbüz Düzenlenmiş Rastgele Vektör Fonksiyon Bağlantı (GD-RVFL) ağı modelini kısa vadeli PV çıkış gücünü tahmin etmede kullanmak ve bu kapsamda önerilen modeli 10 farklı makine öğrenimi yöntemi olan Bayesian Ridge Regressor (BRR), Linear Regressor (LR), Gaussian Process Regressor (GPR), Support Vector Machine (SVM), Extreme Learning Machine (ELM), Yapay Sinir Ağı (YSA), Gradient Boosting Regressor (GBR), Random Forest Regressor (RFR), Lasso Regressor (LAR) ve Ridge Regressor (RR) yöntemleri ile karşılaştırılarak modellerinin performansını değerlendirmektir. Yapılan bu karşılaştırma sonucunda GD-RVFL’ nin etkinliği diğer 10 makine öğrenimi modeline göre önemli ölçüde daha iyi performans gösterdiği görülmüştür

Anahtar Kelimeler

• GD-RVFLN
• KKA
• Makine öğrenimi
• PV çıkış gücü tahmini

Kaynakça

1. [1] P. Markuse, Provisional Report on the State of the Global Climate 2020, World Meteorological Organization (WMO), 2021. Erişim Tarihi: 21 Ocak 2022. https://library.wmo.int/index.php?lvl=notice_display&id=21804#.YwThLEdBxPZ
2. [2] H.E. Murdock, D. Gibb, T. André vd. Renewables 2021 Global Status Report, World Energy Coucil, REN21, 2021. Erişim Tarihi: 21 Ocak 2022. https://www.ren21.net/reports/global-status-report/
3. [3] T. Gould, J. Coppel, T. Bienassis vd., World Energy Investment 2022, International Energy Agency, 2022. Erişim Tarihi: 21 Ocak 2022. https://www.iea.org/reports/world-energy-investment-2022
4. [4] F. Barbieri, S. Rajakaruna, ve A. Ghosh. "Very short-term photovoltaic power forecasting with cloud modeling: A review," Renewable and Sustainable Energy Reviews, 75, Aug. 2017, s. 242-263.
5. [5] G. Gowrisankaran, S. S. Reynolds ve M. Samano, “Intermittency and the value of renewable energy,” Journal of Political Economy, vol. 124, no.4, Aug 2016, s. 1187-1234.
6. [6] G. Stein ve T. M. Letcher, “Integration of PV generated electricity into national grids. In A comprehensive guide to solar energy systems” Academic Press, 2018, s.321-332.
7. [7] G. Cervone, L. Clemente-Harding, S. Alessandrini, ve L. Delle Monache, “Short-term photovoltaic power forecasting using Artificial Neural Networks and an Analog Ensemble,” Renewable Energy, vol.108, Aug. 2017, s. 274–286.
8. [8] Z.L. Yang, M. Mourshed, K. Liu, vd., “A novel competitive swarm optimized RBF neural network model for short-term solar power generation forecasting,” Neurocomputing, vol. 397, Jul. 2020, s.415-421.

9. [9] D. Li and K. Sun, “Random Forest solar power forecast based on classification optimization,” Energy, vol. 187, Nov. 2019, s. 1-11.
10. [10] I. A. Ibrahim ve T. Khatib, “A novel hybrid model for hourly global solar radiation prediction using random forests technique and firefly algorithm,” Energy Convers Manag,
11. [11] H.Z. Wang, H.Y. Yi, J.C. Peng, vd., “Deterministic and probabilistic forecasting of photovoltaic power based on deep convolutional neural network,” Energy Convers Manag, vol. 153, Dec. 2017, s.409-422.
12. [12] H.Z. Wang, Z.X. Lei, X. Zhang, vd., “A review of deep learning for renewable energy forecasting,” Energy Convers Manag, vol. 198, Oct 2019, s.1-16.
13. [13] M.R. Douiri, “Particle swarm optimized neuro-fuzzy system for photovoltaic power forecasting model,” Sol Energy, vol. 184, s.91-104, 2019.
14. [14] Y.T. Li, Y. He, Y. Su, vd., “Forecasting the daily power output of a grid connected photovoltaic system based on multivariate adaptive regression splines,” Appl Energy, vol.180, 15 Oct. 2016, s.392-401.
15. [15] H. Wang ve J. Shen, "An Improved Model Combining Evolutionary Algorithm and Neural Networks for PV Maximum Power Point Tracking," in IEEE Access, vol. 7, 2019, s. 2823-2827.
16. [16] Duman Altan , B. Diken ve B. Kayişoğlu , “Prediction of Photovoltaic Panel Power Outputs using Time Series and Artificial Neural Network Methods”, Tekirdağ Ziraat Fakültesi Dergisi, c. 18, sayı. 3, ss. 457-469, Eyl. 2021, doi:10.33462/jotaf.837446
17. [17] K. Tümay Ateş , "Çok Katmanlı Yapay Sinir Ağı Modeli ve Kültürel Algoritma Modeli Kullanılarak Geliştirilen Melez Yöntem ile Kısa Vadeli Fotovoltaik Enerji Santrali Çıkış Gücü Tahmini", Osmaniye Korkut Ata Üniversitesi Fen Bilimleri Enstitüsü Dergisi, c. 5, sayı. 1, ss. 342-354, Mar. 2022, doi:10.47495/okufbed.1028813
18. [18] K.J. Wang, X.X. Qi, H.D. Liu, vd., “Deep belief network-based k-means cluster approach for short-term wind power forecasting,” Energy, vol. 168, Dec. 2018, s. 840-852.
19. [19] K.J Wang, X.X. Qi, ve H.D. Liu, “A comparison of day-ahead photovoltaic power forecasting models based on deep learning neural network,” Appl Energy, vol. 251, Oct. 2019.
20. [20] L.Y. Liu, Y. Zhao, D.L. Chang, vd., “Prediction of short-term PV power output and uncertainty analysis,” Appl Energy, vol. 228, Oct. 2018, s.700-711
21. [21] G. Cervone, L. Clemente-Harding, S. Alessandrini, vd., “Short-term photovoltaic power forecasting using artifificial neural networks and an analog ensemble,” Renew Energy, vol. 108, Aug. 2017, s.274-284.
22. [22] S. Sobri, S. Koohi-Kamali, ve N. Rahim Abd, “Solar photovoltaic generation forecasting methods: a review,” Energy Convers Manag, vol.156, Jan. 2018, s. 459-497. [23] M.Q. Raza, M. Nadarajah,ve C. Ekanayake, “Review on recent advances in PV output power forecast,” Sol Energy, vol. 136, Oct. 2016, s.125-144.
23. [24] L.W. Zheng, Z.K. Liu, J.N. Shen, vd., “Very short-term maximum Lyapunov exponent forecasting tool for distributed photovoltaic output,” Appl Energy, vol. 229, Nov. 2018, s.1128-1139.
24. [25] X. Zhao, H. Wei, H. Wang, vd., “3d-cnn-based feature extraction of ground-based cloud images for direct normal irradiance prediction,” Sol Energy, vol.181, 2019, s.510-518.
25. [26] Y. Zhou, N. Zhou, L. Gong, vd., “Prediction of photovoltaic power output based on similar day analysis, genetic algorithm and extreme learning machine,” Energy, vol. 204, 2020, s. 117894.
26. [27] R. Ahmed, V. Sreeram, Y. Mishra, vd., “A review and evaluation of the stateof-the-art in PV solar power forecasting: techniques and optimization,” Renew Sustain Energy Rev, vol. 24, May 2020, s.1-26.
27. [28] W. VanDeventer, E. Jamei, G. S. Thirunavukkarasu, vd., “Short-term PV power forecasting using hybrid GASVM technique,” Renewable energy, vol.140, 2019, s.367-379.
28. [29] M. Massaoudi, S. S. Refaat, H. Abu-Rub, vd., "A Hybrid Bayesian Ridge Regression-CWT-Catboost Model For PV Power Forecasting," 2020 IEEE Kansas Power and Energy Conference (KPEC), 2020, s. 1-5
29. [30] L. Gutiérrez, J. Patiño, ve E. Duque-Grisales, “A Comparison of the Performance of Supervised Learning Algorithms for Solar Power Prediction,” Energies, vol.14, no. 15, 2021, s. 4424.
30. [31] A. Afzal, S. Alshahrani, A. Alrobaian, vd., “Power plant energy predictions based on thermal factors using ridge and support vector regressor algorithms” Energies, vol.14, no. 21, 2021, 7254.
31. [32] A. I. Khalyasmaa, S. A. Eroshenko, V. A. Tashchilin, vd., “Industry experience of developing day-ahead photovoltaic plant forecasting system based on machine learning,” Remote Sensing, vol. 12, no.20, 2020, 3420.
32. [33] S. Chahboun ve M. Maaroufi, “Novel comparison of machine learning techniques for predicting photovoltaic output power,” International Journal of Renewable Energy Research (IJRER), vol.11, no.3, 2021, s. 1205-1214.
33. [34] A. Aggarwal ve M. M. Tripathi, “Short-term solar power forecasting using Random Vector Functional Link (RVFL) network,” In Ambient Communications and Computer Systems, Springer, Singapore, 2018, s.29-39.
34. [35] S. P. Mishra, P. P. Padhi, J. Naik, vd., "An efficient Robust Random Vector Functional Link network for Solar Irradiance, Power and Wind speed prediction," 2021 1st Odisha International Conference on Electrical Power Engineering, Communication and Computing Technology(ODICON), 2021, s. 1-7.
35. [36] C. Haydaroğlu, “Dicle Üniversitesi Güneş Enerjisi Santralinin Performans Analizi,” (Yayımlanmamış Yüksek Lisans Tezi), Dicle Üniversitesi, Fen Bilimler Enstitüsü, 2017.
36. [37] C. Haydaroğlu ve B. Gümüş, “Dicle Üniversitesi güneş enerjisi santralinin PVsyst ile simülasyonu ve performans parametrelerinin değerlendirilmesi,” Dicle Üniversitesi Mühendislik Fakültesi Mühendislik Dergisi, vol. 7, no. 3, 2016. s. 491-500.
37. [38] C. Haydaroğlu ve B. Gümüş, “Investigation of the effect of short-term environmental contamination on energy production in photovoltaic panels: Dicle University solar power plant example,” Applied Solar Energy, vol.53, no.1, 2017, s.31-34.
38. [39] H. Kılıç, “Güneş Enerjisi ile İlgili Meteorolojik Verilerin Tahmini İçin Yöntem Geliştirilmesi,” (Yayımlanmamış Yüksek Lisans Tezi), Dicle Üniversitesi, Fen Bilimler Enstitüsü, 2016.
39. [40] B. Gumus ve H. Kilic, “Time dependent prediction of monthly global solar radiation and sunshine duration using exponentially weighted moving average in southeastern of Turkey,” Thermal Science, vol.22, no.2, 2018, s. 943-951.
40. [41] H. Kılıç, B. Gümüş ve M. Yılmaz, “Diyarbakır İli İçin Güneş Enerjisi Verilerinin Meteorolojik Standartlarda Ölçülmesi ve Analizi,” EMO Bilimsel Dergi, vol.5, no.10, 2016, s.15-19.
41. [42] M. Yılmaz, B. Gümüş, H. Kılıç ve M. E. Asker, "Chaotic analysis of the gloabal solar irradiance," 2017 IEEE 6th International Conference on Renewable Energy Research and Applications (ICRERA), 2017, s. 1058-1066.
42. [43] C. Shen, M. Sun, M., Tang, ve C. E. Priebe, “Generalized canonical correlation analysis for classification,” Journal of Multivariate Analysis, vol.130, 2014, s. 310-322.
43. [44] W. Zuobin, M. Kezhi ve G.-W. Ng, "Feature Regrouping for CCA- Based Feature Fusion and Extraction Through Normalized Cut," 2018 21st International Conference on Information Fusion (FUSION), 2018, s. 2275-2282.
44. [45] A. Afifi, V. A. Clark ve Susanne May, Canonical Correlation Analysis | R Data Analysıs Examples, UCLA: Statistical Consulting Group,2004, Erişim tarihi: 12 Temmuz 2022. https://stats.oarc.ucla.edu/r/dae/canonical-correlation-analysis/
45. [46] Y. H. Pao ve Y. Takefuji, "Functional-link net computing: theory, system architecture, and functionalities," in Computer, vol. 25, no. 5, May 1992, s. 76-79.
46. [47] Y.H. Pao, G.H. Park, ve D.J. Sobajic, “Learning and generalization characteristics of the random vector functional-link net,” Neurocomputing , vol.6, no.2, 1994, s.163–180.
47. [48] Q. Shi, R. Katuwal, P. N. Suganthan, ve M. Tanveer, “Random vector functional link neural network based ensemble deep learning” Pattern Recognition, vol.117, 2021, s.107978.
48. [49] A. E. Hoerl ve R. W. Kennard, ‘‘Ridge regression: Applications to nonorthogonal problems,’’ Technometrics, vol. 12, no. 1, 1970, s. 69–82.
49. [50] H. T. Huynh ve Y. Won, ‘‘Regularized online sequential learning algorithm for single-hidden layer feedforward neural networks,’’ Pattern Recognit. Lett., vol. 32, no. 14, Oct. 2011, s. 1930–1935.
50. [51] W. Dai, Q. Chen, F. Chu, X. Ma ve T. Chai, "Robust Regularized Random Vector Functional Link Network and Its Industrial Application," in IEEE Access, vol. 5, 2017, s. 16162-16172.