DOI: https://doi.org/10.15407/publishing2018.50.005

SHORT-TERM ELECTRICAL LOAD FORECASTING FOR THE ELECTRICAL SUPPLY COMPANY WITH DEEP NEURAL NETWORK

P. Chernenko, V. Miroshnyk
Institute of Electrodynamics of the National Academy of Sciences of Ukraine,
Peremohy, 56, Kyiv-57, 03680, Ukraine,
е-mail: This e-mail address is being protected from spambots. You need JavaScript enabled to view it

With liberalization of the electricity market of Ukraine, electricity supply companies will have direct economic incentives to increase the accuracy of hourly load forecasts. Over the past 10 years, significant results achieved in the areas of computer vision, automated control, text and sound processing which outperform human-level. The basis for the breakthrough was a significant increase in computing capabilities, due to modern graphics processors (GPUs), increased the availability of data, the development of more sophisticated machine learning algorithms. We present a new deep learning architecture eResNet for short-term forecasting of the hourly electrical load of the electrical supply company. Basic blocks of this architecture are the layers of the autoencoder type with the shortcut connections. The first layer of the block reduces the dimension of the data, to select the most informative signals, the second layer restores the dimension. Each layer includes a non-linear SELU (scaled exponential linear unit) function. Shortcut connections simplify the error gradient flow, which allows to effectively train all layers of the neural network. The study of the influence of the size of the training set on the accuracy of forecasting conducted. The MAPE of the eResNet is 3.69 % (when training set includes information for 11 years), the error of the multilayer perceptron is 3.85 % (using information for 8 years). References 13, figures 3, table 1.
Key words: electrical load, short-term forecasting, artificial neural network, deep learning.

1. Pro rynok elektrychnoi enerhii: Zakon Ukrainy vid 13.04.2017 r. № 2019-VIII. URL: http://zakon3.rada.gov.ua/laws/show/2019-19 (data zatverdzhennia 21.07.2017). (Ukr)
2. Postanova NKREKP Pro zatverdzhennia Pravyl rynku №307. URL: http://zakon2.rada.gov.ua/laws/show/v0307874-18#n9 (data zatverdzhennia 14.03.2018). (Ukr)
3. Chernenko P.O., Martyniuk O.V. Dekompozytsiia dobovoho hrafika elektrychnoho navantazhennia enerhosystemy ta modeliuvannia yoho skladovykh pid chas korotkostrokovoho prohnozuvannia. Visnyk Vinnytskoho politekhnichnoho instytutu. Vinnytsia, 2017. № 6. Vyp. 135. S. 86–93. (Ukr).
4. LeCunn Y., Bengio Y., Hinton G. Deep learning. Nature. 2015. Vol. 521. P. 436–444. DOI: https://doi.org/10.1038/nature14539
5. Amarasinghe K., Marino D.L., Manic M. Deep Neural Networks for Energy Load Forecasting. Proc. 2017 IEEE 26TH International Symposium on Industrial Electronics (ISIE). 2017. P. 1483–1488. DOI: https://doi.org/10.1109/ISIE.2017.8001465
6. Kuo P.H., Huang, C.J. A High Precision Artificial Neural Networks Model for Short-Term Energy Load Forecasting. Energies. 2018. Vol. 11. Issue 1. P. 213–226.  DOI: https://doi.org/10.3390/en11010213
7. He W. Load Forecasting via Deep Neural Networks. Proc. 5TH International Conference On Information Technology And Quantitative Management. 2017. Vol. 122. P. 308–314. DOI: https://doi.org/10.1016/j.procs.2017.11.374
8. Zhang B., Wu J.L., Chang P.C. A multiple time series-based recurrent neural network for short-term load forecasting. Soft Computing. 2018. Vol. 22. Issue 12. P. 4099–4112. DOI: https://doi.org/10.1007/s00500-017-2624-5
9. Mhaskar H.N., Poggio T. Deep vs. shallow networks: An approximation theory perspective. Analysis and Applications. 2016. Vol.14. Issue 6. P. 829–848. DOI: https://doi.org/10.1142/S0219530516400042
10. He K.M., Zhang X.Y., Ren S.Q., Sun J. Deep Residual Learning for Image Recognition. Proc. Conference on Computer Vision and Pattern Recognition. 2016. P. 770–778. DOI: https://doi.org/10.1109/CVPR.2016.90
11. Veit A., Wilber M., Velongie S. Residual networks behave like ensembles of relatively shallow networks. Proc. Advances in Neural Information Processing Systems 29. 2016. P. 550–558.
12. Klambauer G., Unterthiner T., Mayr A., Hochreiter S. Self-Normalizing Neural Networks. Proc. Advances in Neural Information Processing Systems 30. 2017. P. 971–980.
13. Reddi S.J., Kale S., Kumar S. On the Convergence of Adam and Beyond. ICLR 2018 Conference Submission. URL: https://openreview.net/pdf?id=ryQu7f-RZ