Hava durumu verilerinin kısa dönem elektrik yük tahmini üzerine etkileri

Öz

Elektrik tüketim talebinin doğru bir şekilde tahmin edilmesi, enerji şirketleri için planlama ve kaynak yönetimi açısından büyük bir öneme sahiptir ve bu nedenle, hava durumunun elektrik yük tahminine olan etkilerinin tam olarak belirlenmesi gerekmektedir. Bu çalışmada, hava durumunun kısa dönem elektrik yük tahmini üzerindeki etkileri incelenmiştir. Catboost, LightGBM ve XGBoost gibi güçlü makine öğrenmesi modelleri ile geçmiş yük verileri, gün özellikleri ve tatil günleri gibi bağımsız değişkenlerle birleştirilerek elektrik yük tahmini yapılmıştır. İlk olarak hava durumu verileri hariç tutularak tahmin yapılmış, ardından hava durumu verileri eklenerek modellerin performansları karşılaştırılmıştır. Modellerin performans değerlendirmesi için ortalama mutlak yüzdesel hata (MAPE), ortalama karesel hata (MSE) ve ortalama mutlak hata (MAE) kullanılmıştır. Bir yıl boyunca modellerin günlük performansı ölçülmüştür. Hava durumu verilerinin veri setine eklenmesiyle birlikte MAPE değerlerinde LightGBM modelinde %3.213, XGBoost modelinde %3.404 ve CatBoost modelinde %6.671 performans iyileşmesi olduğu tespit edilmiştir.

Anahtar Kelimeler

• Kısa dönem yük tahmini
• Hava durumu
• Makine öğrenmesi
• MAPE

The effects of weather data on short-term electricity load forecasting

Öz

Accurate estimation of electricity consumption demand is of great importance for energy companies in terms of planning and resource management, and therefore, it is necessary to fully determine the effects of weather conditions on electricity load estimation. In this study, the effects of weather conditions on short-term electricity load estimation were investigated. Electric load estimation was performed by combining historical load data with independent variables such as day characteristics and holidays with a powerful machine learning model such as CatBoost, LightGBM and XGBoost. First, estimation was performed by excluding weather data, and then the performances of the models were compared by adding weather data. Mean absolute percentage error (MAPE), mean square error (MSE) and mean absolute error (MAE) were used to evaluate the performance of the models. Daily performance of the models was measured for a year. With the addition of weather data to the dataset, MAPE values were found to improve by 3.213% for the LightGBM model, 3.404% for the XGBoost model and 6.671% for the CatBoost model.

Anahtar Kelimeler

• Short-term load estimation
• Weather
• Machine learning
• MAPE

Kaynakça

1. [1] Ibrahim IA, Hossain MJ. “Short-term multivariate time series load data forecasting at low-voltage level using optimised deep-ensemble learning-based models”. Energy Conversion and Management, 296, 117663, 2023.
2. [2] Hafiz F, Queiroz ARD, Husain I. “Solar generation, storage, and electric vehicles in power grids: Challenges and solutions with coordinated control at the residential level”. IEEE Electrification Magazine, 6(4), 83-90, 2018.
3. [3] Pallonetto F, Jin C, Mangina E. “Forecast electricity demand in commercial building with machine learning models to enable demand response programs”. Energy and AI, 7, 100121, 2022.
4. [4] Mehigan L, Deane JP, Gallachóir BPÓ, Bertsch V. “A review of the role of distributed generation (DG) in future electricity systems”. Energy, 163, 822-836, 2018.
5. [5] Hong T, Fan S. “Probabilistic electric load forecasting: A tutorial review”. International Journal of Forecast, 32(3), 914-938, 2016.
6. [6] Şekerci H. “Load demand forecast of organized ındustrial zone and ımbalance cost analysis”. Innovations in Intelligent Systems and Applications Conference, Ankara, Turkey, 16-18 October 2024.
7. [7] Chandrasekaran R, Paramasivan SK. “Advances in deep learning techniques for short-term energy load forecasting applications: a review,” Archives of Computational Methods in Engineering, 32, 663-692, 2024.
8. [8] Ekonomou L, Christodoulou CA, Mladenov V. “A short-term load forecasting method using artificial neural networks and wavelet analysis”. International Journal of Power Systems, 1, 64-68, 2016.

9. [9] Tsalikidis N, Mystakidis A, Tjortjis C, Koukaras P, Ioannidis D. “Energy load forecasting: one-step ahead hybrid model utilizing ensembling”. Computing, 106, 241-273, 2024.
10. [10] Kuo PH, Huang CJ. “A high precision artificial neural networks model for short-Term energy load forecasting”. Energies, 11(1), 213, 2018.
11. [11] Farsi B, Amayri M, Bouguila N, Eicker U. “On short-term load forecasting using machine learning techniques and a novel parallel deep LSTM-CNN approach”. IEEE Access, 9, 31191-31212, 2021.
12. [12] Enerji Piyasası Düzenleme Kurumu (EPDK), “Elektrik Piyasası Dengeleme ve Uzlaştırma Yönetmeliği”. https://www.epdk.gov.tr/Detay/Icerik/3-0-159-3/yonetmelikler (01.02.2024).
13. [13] Iwafune Y, Yagita Y, Ikegami T, Ogimoto K. “Short-term forecasting of residential building load for distributed energy management”. IEEE International Energy Conference (ENERGYCON), Cavtat, Croatia, 13-16 May 2014.
14. [14] Ekinci F. “YSA ve ANFIS tekniklerine dayalı enerji tüketim tahmin yöntemlerinin karşılaştırılması”. Düzce Üniversitesi Bilim ve Teknoloji Dergisi, 7(3), 1029-1044, 2019.
15. [15] Kotecha R, Ukarande S, Hosalikar K, Maniar P, Maru S, Pallikuth D, Biswas T, Rane V. “Short-Term load demand forecasting based on weather and ınfluencing factors using deep neural network experts for sustainable development goal 7”. SN Computer Science, 5(2), 253, 2024.
16. [16] Nguyen NA, Dang TD, Verdú E, Solanki EK. “Short-term forecasting electricity load by long short-term memory and reinforcement learning for optimization of hyper-parameters”. Evolutionary Intelligence, 16(5), 1729-1746, 2023.
17. [17] Işık G, Öğüt H, Mutlu M. “Deep learning based electricity demand forecasting to minimize the cost of energy imbalance: A real case application with some fortune 500 companies in Türkiye”. Engineering Applications of Artificial Intelligence, 118, 105664, 2023.
18. [18] Son J, Cha J, Kim H, Wi YM. “Day-Ahead short-term load forecasting for holidays based on modification of similar days’ load profiles”. IEEE Access, 10, 17864-17880, 2022.
19. [19] Sultana N, Hossain SMZ, Almuhaini S, Düştegör D. “Bayesian optimization algorithm-based statistical and machine learning approaches for forecasting short-term electricity demand”. Energies, 15(9), 3425, 2022.
20. [20] Porteiro R, Callejo LH, Nesmachnow S. “Electricity demand forecasting in industrial and residential facilities using ensemble machine learning”. Universidad de Antioquia, 102, 9-25, 2022.
21. [21] Barman M, Choudhury NBD, Sutradhar S. “A regional hybrid GOA-SVM model based on similar day approach for short-term load forecasting in Assam, India”. Energy, 145, 710-720, 2018.
22. [22] Başoǧlu B, Bulut M. “Kisa dönem elektrik talep tahminleri için yapay sinir aǧlari ve uzman sistemler tabanli hibrit sistem geliştirilmesi”. Journal of the Faculty of Engineering and Architecture of Gazi University, 32(2), 575-583, 2017.
23. [23] Zhang L, Jánošík D. “Enhanced short-term load forecasting with hybrid machine learning models: CatBoost and XGBoost approaches”. Expert Systems with Applications, 241, 122686, 2024.
24. [24] Madrid EA, Antonio N. “Short-term electricity load forecasting with machine learning”. Information, 12(2), 1-21, 2021.
25. [25] Wang Y, Sun S, Chen X, Zeng X, Kong Y, Chen Y, Wang T. “Short-term load forecasting based on SVMD and XGBoost”. International Journal of Electrical Power & Energy Systems, 129, 106830, 2021.
26. [26] Zhang XL, Song Z. “A comparative study of the data-driven day-ahead hourly provincial load forecasting methods: From classical data mining to deep learning”. Renewable and Sustainable Energy Reviews, 119, 109632, 2020.
27. [27] Morais LBS, Aquila G, Faria VAD, Lima LMM, Lima JWM, Queiroz AR. “Short-term load forecasting using neural networks and global climate models: An application to a large-scale electrical power system”. Applied Energy, 348, 121439, 2023.
28. [28] Feinberg EA, Genethliou D. Load Forecasting. Editors: Chow JH, Wu FF, Momoh J. Applied Mathematics for Restructured Electric Power Systems, 269-285, New York, USA, Springer Press, 2005.
29. [29] Sobhani M, Hong T, Marti C. “Temperature anomaly detection for electric load forecasting”. International Journal of Forecasting, 36(2), 324-333, 2020.
30. [30] Fahad MU, Arbab N. “Factor affecting short term load forecasting”. Journal of Clean Energy Technologies, 2(4), 305-309, 2014.
31. [31] Erişen E, Iyigun C, Tanrısever F. “Short-term electricity load forecasting with special days: an analysis on parametric and non-parametric methods”. Annals of Operations Research, 2017. https://doi.org/10.1016/j.soildyn.2019.106004.
32. [32] Friedman JH. “Greedy function appraximation: a gradient boosting machine”. Annals of Statistics, 29(5), 1189-1232, 2001.
33. [33] Prokhorenkova L, Gusev G, Vorobev A, Dorogush AV, Gulin A. “CatBoost: unbiased boosting with categorical features”. 32nd International Conference on Neural Information Processing Systems, Montréal, Canada, 3-8 December 2018.
34. [34] Aydın ZE, Erdem Bİ, Çiçek ZİE. “Prediction bike-sharing demand with gradient boosting methods”. Pamukkale University Journal of Engineering Sciences, 29(8), 824-832, 2023.
35. [35] Ke G, Meng Q, Finley T. Wang T, Chen W, Ma W, Ye Q, Liu TY. “LightGBM: A highly efficient gradient boosting decision tree”. Advances in Neural Information Processing Systems Conference (NIPS), Long Beach CA, USA, 4-9 December 2017.
36. [36] Chen T, Guestrin C. “XGBoost: A scalable tree boosting system”. in Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Association for Computing Machinery”. Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, USA, 13-17 August, 2016.
37. [37] Lewis CD. “Industrial and Business Forecasting Methods: a practical guide to exponential smoothing and curve fitting”. Journal of Forecasting, 2(2), 194-196, 1982.
38. [38] Weisstein EW. “Correlation”. https://mathworld.wolfram.com/Correlation.html (08.02.2024).
39. [39] Rodgers JL, Nicewander WA. “Thirteen Ways to Look at the Correlation Coefficient”. The American Statistician, 42, 186-195, 1988.
40. [40] Emhamed AA, J. Shrivastava J. “Electrical load distribution forecasting utilizing support vector model (SVM)”. Materials Today: Proceedings, 47(1), 41-46, 2021.
41. [41] Microsoft. “microsoft.com/tr-tr”. https://learn.microsoft.com/tr-tr/azure/cosmos-db/notebooks-overview (25.02.2024).
42. [42] EPİAŞ. “Seffaflik”. https://seffaflik.epias.com.tr/home (20.01.2024).
43. [43] OpenMeteo, “OpenMeteo”. https://open-meteo.com/en/docs/historical-weather-api (15.01.2024).