Wind Energy Forecasting Based on Grammatical Evolution

Abstract

The energy generated by wind turbines exhibits a continually fluctuating structure due to the dynamic variations in wind speed. In addition, in the context of seasonal transitions, increasing energy demand, and national/international energy policies, the necessity arises for short and long-term forecasting of wind energy. The use of machine learning algorithms is prevalent in the prediction of energy generated from wind. However, in machine learning algorithms such as deep learning, complex and lengthy equations emerge. In this study, the grammatical evolution algorithm, a type of symbolic regression method, is proposed to obtain equations with fewer parameters instead of complex and lengthy equations. This algorithm has been developed to derive a suitable equation based on data. In the study, through the use of grammatical evolution (GE), it has been possible to obtain a formula that is both simple and capable of easy computation, with a limited number of parameters. The equations obtained as a result of the conducted analyses have achieved a performance value of approximately 0.91. The equations obtained have been compared with methods derived using the genetic expression programming (GEP) approach. In conclusion, it has been ascertained that the grammatical evolution method can be effectively employed in the forecasting of wind energy.

Keywords

• Wind Energy
• Energy Forecasting
• Symbolic Regression
• Grammatical Evolution
• Genetic Expression Programming

References

1. [1] Ş. Fidan, M. Cebeci, ve A. Gündoğdu, “Extreme Learning Machine Based Control of Grid Side Inverter for Wind Turbines”, Teh. Vjesn., c. 26, sy 5, ss. 1492-1498, Eki. 2019, doi: 10.17559/TV-20180730143757.
2. [2] R. Saidur, N. A. Rahim, M. R. Islam, ve K. H. Solangi, “Environmental impact of wind energy”, Renew. Sustain. Energy Rev., c. 15, sy 5, ss. 2423-2430, Haz. 2011, doi: 10.1016/j.rser.2011.02.024.
3. [3] M. Schmidt ve H. Lipson, “Symbolic Regression of Implicit Equations”, içinde Genetic Programming Theory and Practice VII, R. Riolo, U.-M. O’Reilly, ve T. McConaghy, Ed., içinde Genetic and Evolutionary Computation. , Boston, MA: Springer US, 2010, ss. 73-85. doi: 10.1007/978-1-4419-1626-6_5.
4. [4] Y. H. Çelik ve Ş. Fidan, “Analysis of cutting parameters on tool wear in turning of Ti-6Al-4V alloy by multiple linear regression and genetic expression programming methods”, Measurement, c. 200, s. 111638, Ağu. 2022, doi: 10.1016/j.measurement.2022.111638.
5. [5] C. Ferreira, “Gene Expression Programming: a New Adaptive Algorithm for Solving Problems”. arXiv, 30 Aralık 2001. doi: 10.48550/arXiv.cs/0102027.
6. [6] N. Lourenço, F. Assunção, F. B. Pereira, E. Costa, ve P. Machado, “Structured Grammatical Evolution: A Dynamic Approach”, içinde Handbook of Grammatical Evolution, C. Ryan, M. O’Neill, ve J. Collins, Ed., Cham: Springer International Publishing, 2018, ss. 137-161. doi: 10.1007/978-3-319-78717-6_6.
7. [7] M. O’Neill ve C. Ryan, “Grammatical evolution”, IEEE Trans. Evol. Comput., c. 5, sy 4, ss. 349-358, Ağu. 2001, doi: 10.1109/4235.942529.
8. [8] A. S. Dufek, D. A. Augusto, P. L. S. Dias, ve H. J. C. Barbosa, “Data-driven symbolic ensemble models for wind speed forecasting through evolutionary algorithms”, Appl. Soft Comput., c. 87, s. 105976, Şub. 2020, doi: 10.1016/j.asoc.2019.105976.

9. [9] P. Valsaraj, D. A. Thumba, K. Asokan, ve K. S. Kumar, “Symbolic regression-based improved method for wind speed extrapolation from lower to higher altitudes for wind energy applications”, Appl. Energy, c. 260, s. 114270, Şub. 2020, doi: 10.1016/j.apenergy.2019.114270.
10. [10] D. Vázquez, R. Guimerà, M. Sales-Pardo, ve G. Guillén-Gosálbez, “Automatic modeling of socioeconomic drivers of energy consumption and pollution using Bayesian symbolic regression”, Sustain. Prod. Consum., c. 30, ss. 596-607, Mar. 2022, doi: 10.1016/j.spc.2021.12.025.
11. [11] R. Rueda, L. G. B. Ruiz, M. P. Cuéllar, ve M. C. Pegalajar, “An Ant Colony Optimization approach for symbolic regression using Straight Line Programs. Application to energy consumption modelling”, Int. J. Approx. Reason., c. 121, ss. 23-38, Haz. 2020, doi: 10.1016/j.ijar.2020.03.005.
12. [12] M. Trabelsi vd., “An Effective Hybrid Symbolic Regression–Deep Multilayer Perceptron Technique for PV Power Forecasting”, Energies, c. 15, sy 23, Art. sy 23, Oca. 2022, doi: 10.3390/en15239008.
13. [13] S. Porras, E. Jove, B. Baruque, ve J. L. Calvo-Rolle, “A comparative analysis of intelligent techniques to predict energy generated by a small wind turbine from atmospheric variables”, Log. J. IGPL, c. 31, sy 4, ss. 648-663, Tem. 2023, doi: 10.1093/jigpal/jzac031.
14. [14] R. Rueda, M. P. Cuéllar, M. Molina-Solana, Y. Guo, ve M. C. Pegalajar, “Generalised Regression Hypothesis Induction for Energy Consumption Forecasting”, Energies, c. 12, sy 6, Art. sy 6, Oca. 2019, doi: 10.3390/en12061069.
15. [15] D. Criado-Ramón, L. G. B. Ruiz, ve M. C. Pegalajar, “Electric demand forecasting with neural networks and symbolic time series representations”, Appl. Soft Comput., c. 122, s. 108871, Haz. 2022, doi: 10.1016/j.asoc.2022.108871.
16. [16] O. Kochueva ve K. Nikolskii, “Data Analysis and Symbolic Regression Models for Predicting CO and NOx Emissions from Gas Turbines”, Computation, c. 9, sy 12, Art. sy 12, Ara. 2021, doi: 10.3390/computation9120139.
17. [17] P. Li, C. Tian, Z. Zhang, M. Li, ve Y. Zheng, “Analysis of influencing factors of energy consumption in rural Henan based on symbolic regression method and Tapio model”, Energy Sources Part Recovery Util. Environ. Eff., c. 43, sy 2, ss. 160-171, Oca. 2021, doi: 10.1080/15567036.2019.1623951.
18. [18] K. Kefer vd., “Simulation-Based Optimization of Residential Energy Flows Using White Box Modeling by Genetic Programming”, Energy Build., c. 258, s. 111829, Mar. 2022, doi: 10.1016/j.enbuild.2021.111829.
19. [19] D. Martínez-Rodríguez, J. M. Colmenar, J. I. Hidalgo, R.-J. Villanueva Micó, ve S. Salcedo-Sanz, “Particle swarm grammatical evolution for energy demand estimation”, Energy Sci. Eng., c. 8, sy 4, ss. 1068-1079, 2020, doi: 10.1002/ese3.568.
20. [20] J. M. Colmenar, J. I. Hidalgo, ve S. Salcedo-Sanz, “Automatic generation of models for energy demand estimation using Grammatical Evolution”, Energy, c. 164, ss. 183-193, Ara. 2018, doi: 10.1016/j.energy.2018.08.199.
21. [21] I. A. Aditya, A. A. Simaremare, J. Raharjo, Suyanto, ve I. Wijayanto, “Daily Power Plant Load Prediction using Grammatical Evolution”, içinde 2022 International Conference on Electrical Engineering, Computer and Information Technology (ICEECIT), Kas. 2022, ss. 122-126. doi: 10.1109/ICEECIT55908.2022.10030558.
22. [22] B. Jamil, L. Serrano-Luján, ve J. Colmenar, “On the Prediction of One-Year Ahead Energy Demand in Turkey using Metaheuristic Algorithms”, c. 7, ss. 79-91, Ağu. 2022, doi: 10.25046/aj070411.
23. [23] N. Lourenço, J. M. Colmenar, J. I. Hidalgo, ve S. Salcedo-Sanz, “Evolving energy demand estimation models over macroeconomic indicators”, içinde Proceedings of the 2020 Genetic and Evolutionary Computation Conference, içinde GECCO ’20. New York, NY, USA: Association for Computing Machinery, Haz. 2020, ss. 1143-1149. doi: 10.1145/3377930.3390153.
24. [24] L. Serrano-Luján, C. Toledo, J. M. Colmenar, J. Abad, ve A. Urbina, “Accurate thermal prediction model for building-integrated photovoltaics systems using guided artificial intelligence algorithms”, Appl. Energy, c. 315, s. 119015, Haz. 2022, doi: 10.1016/j.apenergy.2022.119015.
25. [25] J. Jeschke, D. Sun, A. Jamshidnejad, ve B. De Schutter, “Grammatical-Evolution-based parameterized Model Predictive Control for urban traffic networks”, Control Eng. Pract., c. 132, s. 105431, Mar. 2023, doi: 10.1016/j.conengprac.2022.105431.
26. [26] V. Christou vd., “Grammatical Evolution-Based Feature Extraction for Hemiplegia Type Detection”, Signals, c. 3, sy 4, Art. sy 4, Ara. 2022, doi: 10.3390/signals3040044.
27. [27] Ş. Fi̇dan ve H. Çi̇men, “Rüzgâr Türbinlerinde Tork ve Kanat Eğim Açısı Kontrolü”, Batman Üniversitesi Yaşam Bilim. Derg., c. 11, sy 1, Art. sy 1, Haz. 2021.
28. [28] F. Noorian, A. M. de Silva, ve P. H. W. Leong, “gramEvol: Grammatical Evolution in R”, J. Stat. Softw., c. 71, ss. 1-26, Tem. 2016, doi: 10.18637/jss.v071.i01.
29. [29] B. Peng, S. Wan, Y. Bi, B. Xue, ve M. Zhang, “Automatic Feature Extraction and Construction Using Genetic Programming for Rotating Machinery Fault Diagnosis”, IEEE Trans. Cybern., c. 51, sy 10, ss. 4909-4923, Eki. 2021, doi: 10.1109/TCYB.2020.3032945.
30. [30] I. Arnaldo, K. Krawiec, ve U.-M. O’Reilly, “Multiple regression genetic programming”, içinde Proceedings of the 2014 Annual Conference on Genetic and Evolutionary Computation, içinde GECCO ’14. New York, NY, USA: Association for Computing Machinery, Tem. 2014, ss. 879-886. doi: 10.1145/2576768.2598291.
31. [31] S. Wagner vd., “Architecture and Design of the HeuristicLab Optimization Environment”, içinde Advanced Methods and Applications in Computational Intelligence, R. Klempous, J. Nikodem, W. Jacak, ve Z. Chaczko, Ed., içinde Topics in Intelligent Engineering and Informatics. , Heidelberg: Springer International Publishing, 2014, ss. 197-261. doi: 10.1007/978-3-319-01436-4_10.
32. [32] J. H. Steiger, “Tests for comparing elements of a correlation matrix”, Psychol. Bull., c. 87, sy 2, ss. 245-251, 1980, doi: 10.1037/0033-2909.87.2.245.
33. [33] Çelebi, S. B., & Fidan, Ş. RNN-Based Time Series Analysis for Wind Turbine Energy Forecasting. International Journal of Engineering and Innovative Research, 6(1), 15-28.
34. [34] Çelebi, S. B., & Karaman, Ö. A. Multilayer LSTM Model for Wind Power Estimation in the Scada System. European Journal of Technique (EJT), 13(2), 116-122.