Farklı Derin Öğrenme Metotları ile Kısa Dönem Elektrik Yükü Tahmin Karşılaştırması

Yıl 2021, Sayı: 31, 616 - 623, 31.12.2021

İpek Atik

https://doi.org/10.31590/ejosat.1017137

Öz

Elektrik üretim miktarının tahmin edilmesi iletim ve dağıtım sistemlerinin planlaması, üretim ekonomisi, ünite çalışma programları ve bakım onarım zamanlaması için önemli role sahiptir. Doğru tahmin modellemeleri ile kesintisiz ve güvenilir elektrik enerjisi üretimi sağlanabilir. Çalışmamızda Türkiye için saatlik elektrik üretim verileri kullanılarak farklı derin öğrenme algoritmaları ile 1saat, 2 saat ve 3 saat ilerisi için tahmin çalışmaları yapılmıştır. Modellere ait MAE, RMSE ve korelasyon katsayısı değerleri ile performansları karşılaştırılmıştır. Çalışmada gerçek değerlere en yakın tahmin yapan modelin belirlenmesi amaçlanmıştır. Çalışmanın bu bağlamda gelecek tahmin çalışmaları için faydalı olacağı öngörülmektedir.

Anahtar Kelimeler

Derin öğrenme, kısa dönem tahmin, LSTM, SVM, ANN

Comparison of Short-Term Electricity Load Forecasting Using Different Deep Learning Methods

Öz

Estimation of the amount of electricity generation plays an important role in the planning of transmission and distribution systems, generation economy, unit work schedules and maintenance repair timing. With accurate forecasting models, uninterrupted and reliable electrical energy production can be achieved. In our study, 1-hour, 2-hour and 3-hour ahead predictions were made with different deep learning algorithms using Turkey's hourly electricity generation data. With the MAE, RMSE and correlation coefficient values of the models, their performances were compared. The study aimed to determine the model that makes the closest estimation to the real values. In this context, it is anticipated that the study will be useful for future prediction studies.

Anahtar Kelimeler

Deep learning, short-term prediction, LSTM, SVM, ANN

Kaynakça

• Al Mamun, M., & Nagasaka, K. (2006). Implementation of an Intelligent Method to Forecast Long-term Electric Demand. Iranian Journal of Electrical and Computer Engineering (IJECE), 5(2), 75–82.
• Altan, G. (2019). DeepGraphNet: Grafiklerin Sınıflandırılmasında Derin Öğrenme Modelleri. Avrupa Bilim ve Teknoloji Dergisi, 319–327. doi:10.31590/ejosat.638256
• B. E. Türkay, & D. Demren. (2011). Electrical Load Forecasting Using Support Vector Machines (pp. 49–53). Presented at the International Conference on Electrical and Electronics Engineering, Nagpur.
• Božić, M., & Stojanović, M. (2011). Application of SVM Methods for Mid-Term Load Forecasting. Serbian Journal Of Electrical Engineering, 8(1), 73–83.
• Elattar, E. E., Goulermas, J., & Wu, Q. H. (2010). Electric Load Forecasting Based on Locally Weighted Support Vector Regression. IEEE Transactions on Systems, Man, And Cybernetics—Part C: Applications And Reviews, 40(4), 438–447.
• Ghanbari, A., Naghavi, A., Ghaderi, S. F., & Sabaghian, M. (2009). Artificial Neural Networks and Regression Approaches Comparison for Forecasting Iran’s Annual Electricity Load (pp. 675–679). Presented at the International Conference on Power Engineering, Energy and Electrical Drives. doi:10.1109/POWERENG.2009.4915245
• Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory (Vols. 1-8, Vol. 9). Neural Computation.
• Kaggle. (2021, December 6). Kaggle. Kaggle data set. dataset. Retrieved from https://www.kaggle.com/datasets
• Karaca, C., & Karacan, H. (2016). Investigation Of Factors Affecting Demand For Electricity Consumption With Multiple Regression Method. SUJEST, 4(3), 182–195.
• Khan, A. R., Razzaq, S., Alquthami, T., Moghal, M., Amin, A., & Mahmood, A. (2018). Day Ahead Load Forecasting for IESCO Using Artificial Neural Network and Bagged Regression Tree. Presented at the 1st International Conference on Power, Energy and Smart Grid, Mirpur.
• Kong, W., Dong, Z. Y., Jia, Y., Hill, D., Xu, Y., & Zhang, Y. (2017). Short-Term Residential Load Forecasting based on LSTM Recurrent Neural Network. IEEE Transactions on Smart Grid, PP, 1–1. doi:10.1109/TSG.2017.2753802
• Matijaš, M. (2013). Electric Load Forecasting Using Multivariate Meta-Learning (Ph.D Thesis). University of Zagreb, Zagreb.
• Nazarko, J., & Zalewski, W. (1999). The Fuzzy Regression Approach to Peak Load Estimation. IEEE Transactions on Power Systems, 14(3), 809–814.
• O. Kaynar, H. Ozekicioglu, & Demirkoparan, F. (2017). Forecasting of Turkey’s Electricity Consumption with Support Vector Regression and Chaotic Particle Swarn Algorithm. Journal of Administrative Sciences, 15(29), 2011–224.
• Omidi, A., Barakati, S., & Tavakoli, S. (2015). Application of nuSupport Vector Regression in Short-Term Load Forecasting (pp. 32–36). Presented at the The 20th Iranian Electrical Power Distribution Conference, Zahedan.
• Sarhani, M., & El Afia, A. (2015). Electric Load Forecasting Using Hybrid Machine Learning Approach Incorporating Feature Selection. Presented at the Proceedings of the International Conference on Big Data Cloud and Applications, Morocco.
• Sun, Z., Qingdang, L., & Wang, L. (2020). Deep Learning Based Visual Object Tracker With Template Update. University Politehnica of Bucharest Scientific Bulletin Series C-Electrical Engineering And Computer Science, 82(2), 65–76.
• Tosun, S., Ozturk, A., & Taspinar, F. (2019). Short Term Load Forecasting for Turkey Energy Distribution System with Artificial Neural Networks. Tehnicki Vjesnik-Technical Gazette, 26(6), 1545–1553.
• Wang, J., Zhu, S., Zhang, W., & Lu, H. (2010). Combined modeling for electric load forecasting with adaptive particle swarm optimization. Energy, 35(4), 1671–1678.
• Yuan, X., Li, L., & Wang, Y. (2019). Nonlinear Dynamic Soft Sensor Modeling With Supervised Long Short-Term Memory Network. IEEE Transactions on Industrial Informatics, 1–1. doi:10.1109/TII2019.2902129
• Zhang, G., Patuwo, B. E., & Hu, M. Y. (1998). Forecasting with Artificial Neural Networks:The State of the Art. International Journal of Forecasting, 14(1), 35–62.
• Zheng, S., Ristovski, K., Farahat, A., & Gupta, C. (2017). Long Short-Term Memory Network for Remaining Useful Life estimation.