Wind Speed Forecasting Based On Candlestick Chart Representation, Minimum Redundancy Maximum Relevance Algorithm and XGBoost Model

Abstract

In today's electricity networks, the number of electricity generation facilities based on renewable energy sources is increasing in order to reduce the dependence on fossil energy sources. Thanks to the Wind Turbine (WT), the wind energy is converted into electrical energy and the integration of WT into the electricity network is ensured at the point of meeting the daily electricity needs. WT output power is affected by external factors such as wind speed and direction, as well as internal factors such as WT’s height, blade structure, generator output power, mechanical and electrical converter efficiencies. Forecasting wind speed allows wind farm operators to optimize electricity production. In this way, wind energy can be better integrated into the electricity network. Existing studies reveal that short-term forecasting approaches are inadequate in terms of accuracy and the nonlinear and stochastic nature of wind speed cannot be fully modeled. For this reason, the use of hybrid models, rather than standalone models, is becoming increasingly prevalent as they offer superior predictive performance. In this study, a novel approach based on candlestick representation, features evaluated with Minimum Redundancy Maximum Relevance (MRMR) approach and XGBoost model proposed for wind speed prediction. The dataset collected from the Supervisory Control and Data Acquisition (SCADA) system in WT for a sampling time of 10 minutes within a 1-year period is used. The data set is first passed through the preprocessing step and the wind direction, wind speed distribution and statistical values are examined. Then, the candlestick representation process step is applied to the time series. Afterwards, candlestick representation is utilized to derive trend and oscillator-based features, the feature group is evaluated with the MRMR approach. A wind speed prediction model is created with the XGBoost method and a situation with low model complexity and lowest prediction error is achieved. This proposed hybrid approach, especially based on candlestick chart representation, aims to increase the accuracy in short-term wind speed forecasting and overcome the limitations of traditional methods. The proposed method shows the best performance with the lowest error rate (RMSE: 0.0644) and the highest correlation coefficient (R: 0.8601) compared to all other models. This shows that the model is superior in terms of both accuracy and error rate.

Keywords

• Wind Speed Forecasting
• Candlestick Chart
• XGBoost
• MRMR
• Oscillators.

Mum Çubuğu Grafik Gösterimi, Minimum Artıklık Maksimum İlgililik Algoritması ve XGBoost Modeline Dayalı Rüzgâr Hızı Tahmini

Öz

Günümüz elektrik şebekelerinde fosil enerji kaynaklarına bağımlılığı azaltmak için yenilenebilir enerji kaynaklarına dayalı elektrik üretim tesislerinin sayısı giderek artmaktadır. Rüzgâr türbinleri (RT) sayesinde rüzgâr enerjisi elektrik enerjisine çevrilmekte ve RT’lerin günlük elektrik ihtiyacını karşılama noktasında elektrik şebekesine entegrasyonu sağlanmaktadır. RT’nin yüksekliği, rüzgâr türbininin kanat yapısı, jeneratör çıkış gücü, mekanik ve elektrik dönüştürücü verimliliği gibi iç faktörler ile birlikte rüzgâr hızı ve yönü gibi dış faktörlere bağlı olarak RT’nin çıkış gücü etkilenmektedir. Rüzgâr hızını tahmin etmek rüzgâr çiftliği operatörlerinin elektrik üretimini optimize etmesine olanak tanımaktadır. Bu sayede rüzgâr enerjisi elektrik şebekesine daha iyi entegre edilebilmektedir. Mevcut çalışmalar, kısa vadeli tahmin yaklaşımlarının doğruluk açısından yetersiz kaldığını ve rüzgâr hızının doğrusal olmayan ve stokastik doğasının tam anlamıyla modellenemediğini ortaya koymaktadır. Bu nedenle, tekil modeller yerine hibrit modellerin kullanımı giderek yaygınlaşmakta ve daha yüksek tahmin performansı sağlamak amacıyla tercih edilmektedir. Bu çalışmada, rüzgâr hızını tahmin etmek için mum çubuğu gösterimi, özniteliklerin Minimum Artıklık Maksimum Uygunluk (Minimum Redundancy Maximum Relevance-MRMR) yaklaşımı ile değerlendirildiği XGBoost modeline dayalı yeni bir yöntem önerilmektedir. RT’de bulunan Merkezi Denetleme Kontrol ve Veri Toplama (SCADA) sisteminden 10 dakikalık örnekleme zamanı için 1 yıllık zaman dilimi içerisinde toplanan veri seti kullanılmaktadır. Veri seti öncelikle önişleme adımından geçirilerek rüzgâr yönü, rüzgâr hızı dağılımı gibi değerler ile istatistiksel değerlere bakılmaktadır. Daha sonra zaman serisine mum çubuğu gösterimi işlem adımı uygulanmaktadır. Elde edilen mum çubuğu gösterimi için trend ve osilatör tabanlı öznitelikler uygulanarak MRMR yaklaşımı ile öznitelik grubu değerlendirilmiştir. XGBoost yöntemi ile rüzgâr hızı tahmin modeli oluşturulmakta ve model karmaşıklığının az ve tahmin hatasının en düşük olduğu durum elde edilmektedir. Özellikle mum çubuğu grafik gösterimine dayalı olarak önerilen bu hibrit yaklaşım, kısa vadeli rüzgâr hızı tahmininde doğruluğu artırmayı ve geleneksel yöntemlerin sınırlamalarını aşmayı hedeflemektedir. Önerilen yöntem, tüm diğer modellere göre en düşük hata oranı (RMSE: 0.0644) ve en yüksek korelasyon katsayısı (R: 0.8601) ile en iyi performansı göstermektedir. Bu, modelin hem doğruluk hem de hata oranı açısından üstün olduğunu göstermektedir.

Anahtar Kelimeler

• Rüzgâr Hızı Tahmini
• Mum Çubuğu Grafiği
• XGBoost
• MRMR
• Osilatörler

Kaynakça

1. [1] Behera, S., Sahoo, S., & Pati, B. B. (2015). A review on optimization algorithms and application to wind energy integration to grid. Renewable and Sustainable Energy Reviews, 48, 214-227.
2. [2] Shafiullah, G. M., Oo, A. M., Ali, A. S., & Wolfs, P. (2013). Potential challenges of integrating large-scale wind energy into the power grid–A review. Renewable and sustainable energy reviews, 20, 306-321.
3. [3] MansourLakouraj, M., Shahabi, M., Shafie-khah, M., & Catalão, J. P. (2022). Optimal market-based operation of microgrid with the integration of wind turbines, energy storage system and demand response resources. Energy, 239, 122156.
4. [4] Msigwa, G., Ighalo, J. O., & Yap, P. S. (2022). Considerations on environmental, economic, and energy impacts of wind energy generation: Projections towards sustainability initiatives. Science of the Total Environment, 157755.
5. [5] Lin, Z., Liu, X., & Collu, M. (2020). Wind power prediction based on high-frequency SCADA data along with isolation forest and deep learning neural networks. International Journal of Electrical Power & Energy Systems, 118, 105835.
6. [6] Suo, L., Peng, T., Song, S., Zhang, C., Wang, Y., Fu, Y., & Nazir, M. S. (2023). Wind speed prediction by a swarm intelligence based deep learning model via signal decomposition and parameter optimization using improved chimp optimization algorithm. Energy, 276, 127526.
7. [7] Gong, Y., Jiang, Q., & Baldick, R. (2015). Ramp event forecast based wind power ramp control with energy storage system. IEEE Transactions on Power Systems, 31(3), 1831-1844.
8. [8] Neshat, M., Nezhad, M. M., Abbasnejad, E., Mirjalili, S., Groppi, D., Heydari, A., ... & Wagner, M. (2021). Wind turbine power output prediction using a new hybrid neuro-evolutionary method. Energy, 229, 120617.

9. [9] Oh, S. Y., Joung, C., Lee, S., Shim, Y. B., Lee, D., Cho, G. E., ... & Park, Y. B. (2024). Condition-based maintenance of wind turbine structures: A state-of-the-art review. Renewable and Sustainable Energy Reviews, 204, 114799.
10. [10] Kisvari, A., Lin, Z., & Liu, X. (2021). Wind power forecasting–A data-driven method along with gated recurrent neural network. Renewable Energy, 163, 1895-1909.
11. [11] Zheng, Y., Ge, Y., Muhsen, S., Wang, S., Elkamchouchi, D. H., Ali, E., & Ali, H. E. (2023). New ridge regression, artificial neural networks and support vector machine for wind speed prediction. Advances in Engineering Software, 179, 103426.
12. [12] Yang, Q., Huang, G., Li, T., Xu, Y., & Pan, J. (2023). A novel short-term wind speed prediction method based on hybrid statistical-artificial intelligence model with empirical wavelet transform and hyperparameter optimization. Journal of Wind Engineering and Industrial Aerodynamics, 240, 105499.
13. [13] Li, Y., Shen, X., & Zhou, C. (2023). Dynamic multi-turbines spatiotemporal correlation model enabled digital twin technology for real-time wind speed prediction. Renewable Energy, 203, 841-853.
14. [14] Malakouti, S. M. (2023). Improving the prediction of wind speed and power production of SCADA system with ensemble method and 10-fold cross-validation. Case Studies in Chemical and Environmental Engineering, 8, 100351.
15. [15] Zhang, Z., Wang, J., Wei, D., Luo, T., & Xia, Y. (2023). A novel ensemble system for short-term wind speed forecasting based on Two-stage Attention-Based Recurrent Neural Network. Renewable Energy, 204, 11-23.
16. [16] Wang, X., Wang, J., Niu, X., & Wu, C. (2024). Novel wind-speed prediction system based on dimensionality reduction and nonlinear weighting strategy for point-interval prediction. Expert Systems with Applications, 241, 122477.
17. [17] Zhu, A., Zhao, Q., Yang, T., Zhou, L., & Zeng, B. (2024). Wind speed prediction and reconstruction based on improved grey wolf optimization algorithm and deep learning networks. Computers and Electrical Engineering, 114, 109074.
18. [18] Wang, M., & Tian, Z. (2024). Ultra-short-term wind speed prediction based on empirical wavelet transform and combined model. Earth Science Informatics, 17(1), 539-560.
19. [19] Wang, J. W., Yang, H. J., & Kim, J. J. (2020). Wind speed estimation in urban areas based on the relationships between background wind speeds and morphological parameters. Journal of Wind Engineering and Industrial Aerodynamics, 205, 104324.
20. [20] Zhou, S., Gao, C. Y., Duan, Z., Xi, X., & Li, Y. (2023). A robust error correction method for numerical weather prediction wind speed based on Bayesian optimization, variational mode decomposition, principal component analysis, and random forest: VMD-PCA-RF (version 1.0. 0). Geoscientific Model Development, 16(21), 6247-6266.
21. [21] Malik, P., Gehlot, A., Malik, P. K., & Singh, R. (2023, April). Global horizontal irradiance and wind speed prediction using ANN: Comprehensive Study. In 2023 IEEE Devices for Integrated Circuit (DevIC) (pp. 190-193). IEEE.
22. [22] Li, Y., Feng, Z., & Feng, L. (2015). Using candlestick charts to predict adolescent stress trend on micro-blog. Procedia Computer Science, 63, 221-228.
23. [23] Xu, R., Liu, X., Wan, H., Pan, X., & Li, J. (2021). A Feature Extraction and Classification Method to Forecast the PM2. 5 Variation Trend Using Candlestick and Visual Geometry Group Model. Atmosphere, 12(5), 570.
24. [24] Hsu, Y. C. (2020). Using machine learning and candlestick patterns to predict the outcomes of American football games. Applied Sciences, 10(13), 4484.
25. [25] Guilizzoni, M., & Maldonado Eizaguirre, P. (2022). Trend lines and Japanese candlesticks applied to the forecasting of wind speed data series. Forecasting, 4(1), 165-181.
26. [26] Erisen, B. Wind Turbine Scada Dataset. 2018. Available online: http//www.kaggle.com/berkerisen/wind-turbine-scada-dataset (accessed on 10 May 2024).
27. [27] Chande, T. S., & Kroll, S. (1994). The new technical trader: boost your profit by plugging into the latest indicators.
28. [28] Johnston, F. R., Boyland, J. E., Meadows, M., & Shale, E. (1999). Some properties of a simple moving average when applied to forecasting a time series. Journal of the Operational Research Society, 50(12), 1267-1271.
29. [29] Klinker, F. (2011). Exponential moving average versus moving exponential average. Mathematische Semesterberichte, 58, 97-107.
30. [30] Lane, G. C. (1984). Lane's stochastics. Technical Analysis of Stocks & Commodities, 2(3), 87-90.
31. [31] Gumparthi, S. (2017). Relative strength index for developing effective trading strategies in constructing optimal portfolio. International Journal of Applied Engineering Research, 12(19), 8926-8936.
32. [32] Chande, T. S., & Kroll, S. (1994). The new technical trader: boost your profit by plugging into the latest indicators.
33. [33] De Ville, B. (2013). Decision trees. Wiley Interdisciplinary Reviews: Computational Statistics, 5(6), 448-455.
34. [34] Yeşilyurt, Sefa, & Dalkılıç, Hüseyin. (2021). Xgboost ve gradient boost machine ile günlük nehir akımı tahmini. In 3rd International Symposium of III Engineering Applications on Civil Engineering and Earth Sciences, Karabük, Türkiye.
35. [35] Gülgezen, G., Çataltepe, Z., & Yu, L. MRMR Algoritması Kullanılarak Kararlı Öznitelik Seçimi Stable Feature Selection Using MRMR Algorithm.