Attention Based Energy Demand Forecasting in Smart Grid Environments

Öz

The smart grid is a crucial aspect of the modern energy landscape, providing a reliable, efficient, and sustainable way of meeting the growing energy demands. However, the vast amounts of data generated by smart grid technology necessitate the development of advanced data processing and analysis techniques. In this paper, we propose an attention-based time series workflow that combines dilated convolution and attention mechanisms for time series forecasting in smart grid applications. This workflow extracts temporal features from time series data using dilated convolutions and emphasizes significant temporal points in the hidden states using attention mechanisms. Experimental evaluations showed up to an 8% better performance for energy demand forecasting compared to commonly used deep learning-based methods. Our workflow achieved this gain by requiring 1/3 of the training time other models took. We also improved performance by 42% in various domains, demonstrating the adaptability of our approach across different areas. This study may assist researchers in constructing accurate forecasting models for smart grid environments. Furthermore, it highlights that the attention-based approach can be employed to promote sustainable energy and optimize smart grid environments.

Anahtar Kelimeler

• Smart grid
• Time series forecasting
• Attention
• Transformers
• Energy demand forecasting

Akıllı Şebeke Ortamlarında Dikkat Tabanlı Enerji Talep Tahmini

Öz

Akıllı şebeke, modern enerji peyzajının kritik bir unsuru olup, artan enerji taleplerini karşılamak için güvenilir, verimli ve sürdürülebilir bir yol sağlamaktadır. Bununla birlikte, akıllı şebeke teknolojisi tarafından üretilen büyük miktardaki veri, gelişmiş veri işleme ve analiz tekniklerinin geliştirilmesini gerektirmektedir. Bu makalede, akıllı şebeke uygulamalarında zaman serisi tahmininde kullanılmak üzere, dilatasyonlu konvolüsyon ve dikkat mekanizmalarını birleştiren bir dikkat tabanlı zaman serisi iş akışı öneriyoruz. Bu akış, dilatasyonlu konvolüsyonları kullanarak zaman serisi verilerinden zamansal özellikler çıkarır ve dikkat mekanizmalarını kullanarak gizli durumlardaki önemli zaman noktalarını vurgular. Deneysel değerlendirmeler sonucunda, enerji talebi tahmininde, yaygın olarak kullanılan derin öğrenme tabanlı yöntemlere göre %8'e kadar daha iyi bir performans gösterdiği gözlemlendi. Bu kazancı diğer modellerin aldığı eğitim süresinin yalnızca 1/3'ü kadar bir sürede elde edilmiştir. Ayrıca, tamamen farklı bir alanda %42'lik bir kazanç elde edilmiştir ve akışın diğer alanlara uyarlanabileceği gösterilmiştir. Bu çalışma, araştırmacılara akıllı şebeke uygulamaları için daha doğru ve verimli tahmin modelleri geliştirmelerine yardımcı olabilir, ayrıca enerji sistemlerinin sürdürülebilir yönetimi ve akıllı şebeke operasyonlarının optimizasyonu için yapay zeka ve dikkat tabanlı tahmin tekniklerinin potansiyeli hakkında değerli bilgiler sunabilir.

Anahtar Kelimeler

• Akıllı şebeke
• Zaman serisi tahmini
• Attention mekanizması
• Dönüştürücüler
• Enerji talebi tahmini

Kaynakça

1. V. Kulkarni, S. K. Sahoo, S. B. Thanikanti, S. Velpula, and D. I. Rathod, "Power systems automation, communication, and information technologies for smart grid: A technical aspects review," TELKOMNIKA (Telecommunication Computing Electronics and Control), vol. 19, no. 3, pp. 1017-1029, 2021.
2. A. H. Bagdadee, M. Aurangzeb, S. Ali, and L. Zhang, "Energy management for the industrial sector in smart grid system," Ener. Repor, vol. 6, pp. 1432-1442, 2020.
3. M. S. Hossain, N. A. Madlool, N. A. Rahim, J. Selvaraj, A. K. Pandey, and A. F. Khan, "Role of smart grid in renewable energy: An overview," Ren. and Sust. Ener. Rev., vol. 60, pp. 1168-1184, 2016.
4. S. Raghavendra, S. Neelakandan, M. Prakash, B. T. Geetha, S. M. R. Asha, and M. K. Roberts, "Artificial humming bird with data science enabled stability prediction model for smart grids," Sust. Comp.: Infor. and Syst., vol. 36, p. 100821, 2022.
5. D. F. Costa Silva, A. R. Galvão Filho, R. V. Carvalho, F. de Souza L. Ribeiro, and C. J. Coelho, "Water flow forecasting based on river tributaries using long short-term memory ensemble model," Ener., vol. 14, no. 22, pp. 7707, 2021.
6. Z. Chen, F. Xiao, F. Guo, and J. Yan, "Interpretable machine learning for building energy management: A state-of-the-art review," Adv. in Appl. Ener., vol. 100123, 2023.
7. S. Dong, P. Wang, and K. Abbas, "A survey on deep learning and its applications," Comp. Scie. Rev., vol. 40, p. 100379, 2021.
8. Y. Xu and L. Zhao, "Inception-LSTM Human Motion Recognition with Channel Attention Mechanism," Comput. and Math. Meth. in Med., 2022.

9. J. Shi and Z. Wang, "A hybrid forecast model for household electric power by fusing landmark-based spectral clustering and deep learning," Sust., vol. 14, no. 15, pp. 9255, 2022.
10. M. M. Keskin, F. Irım, O. Karaahmetoğlu, and E. Kaya, "Time series prediction with hierarchical recurrent model," Sig., Ima. and Vid. Proce., vol. 17, no. 5, pp. 2121-2127, 2023.
11. I. S. Oh and J. S. Lee, "Dense sampling of time series for forecasting," IEEE Acc., vol. 10, pp. 75571-75580, 2022.
12. G. Hebrail and A. Berard, "UCI machine learning repository: Individual household electric power consumption dataset," EDF R&D, vol. 30, no. 08, 2012.
13. G. Garnero and D. Godone, "Comparisons between different interpolation techniques," The International Archives of the Photogr., Remote Sens. and Spatial Infor. Scien., vol. 40, pp. 139-144, 2014.
14. D. Garcia, "Robust smoothing of gridded data in one and higher dimensions with missing values," Comp. Stat. & Data Analy., vol. 54, no. 4, pp. 1167-1178, 2010.
15. S. Hochreiter and J. Schmidhuber, "Long short-term memory," Neural Comp., vol. 9, no. 8, pp. 1735-1780, 1997.
16. F. Yu and V. Koltun, "Multi-scale context aggregation by dilated convolutions," arXiv preprint arXiv:1511.07122, 2015.
17. R. Xi, M. Hou, M. Fu, H. Qu, and D. Liu, "Deep dilated convolution on multimodality time series for human activity recognition," in 2018 International Joint Conference on Neural Networks (IJCNN), July 2018, pp. 1-8.
18. A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, and I. Polosukhin, "Attention is all you need," in Adv. in Neural Infor. Proce. Syst., vol. 30, 2017.
19. I. Sutskever, O. Vinyals, and Q. V. Le, "Sequence to sequence learning with neural networks," in Adv. in Neural Infor. Proce. Syst., vol. 27, 2014.
20. M. S. Ko, K. Lee, J. K. Kim, C. W. Hong, Z. Y. Dong, and K. Hur, "Deep concatenated residual network with bidirectional LSTM for one-hour-ahead wind power forecasting," IEEE Transac. on Sustain. Ener., vol. 12, no. 2, pp. 1321-1335, 2020.
21. A. V. Seliverstova, D. A. Pavlova, S. A. Tonoyan, and Y. E. Gapanyuk, "The time series forecasting of the company’s electric power consumption," in Advances in Neural Computation, Machine Learning, and Cognitive Research II: Selected Papers from the XX International Conference on Neuroinformatics, October 8-12, 2018, Moscow, Russia, pp. 210-215, Springer International Publishing, 2019.
22. A. Shobol, M. H. Ali, M. Wadi, and M. R. TüR, ‘‘Overview of big data in smart grid,’’ in Proc. 8th Intl. Conf. Renew. Energy Res. Appl. (ICRERA), 2019, pp. 1022–1025.
23. D. C. Sekhar, P. R. Rao, R. Kiranmayi, "Large Scale Predictive Analytics based Real-Time Energy Management and Enhance Power Quality in Smart Grid," in Proceedings of the 2022 IEEE 7th International Conference for Convergence in Technology (I2CT), Apr. 1-5, 2022, pp. 1-5.
24. F. Yu and V. Koltun, "Multi-scale context aggregation by dilated convolutions," arXiv preprint arXiv:1511.07122, 2015.
25. R. Xi, M. Hou, M. Fu, H. Qu, and D. Liu, "Deep dilated convolution on multimodality time series for human activity recognition," in 2018 International Joint Conference on Neural Networks (IJCNN), July 2018, pp. 1-8.
26. A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, and I. Polosukhin, "Attention is all you need," in Adv. in Neural Inform. Proce. Syst., vol. 30, 2017.
27. I. Sutskever, O. Vinyals, and Q. V. Le, "Sequence to sequence learning with neural networks," in Adv. in Neural Infor. Proce. Syst., vol. 27, 2014.
28. M. S. Ko, K. Lee, J. K. Kim, C. W. Hong, Z. Y. Dong, and K. Hur, "Deep concatenated residual network with bidirectional LSTM for one-hour-ahead wind power forecasting," IEEE Transac.on Sustain. Ener., vol. 12, no. 2, pp. 1321-1335, 2020.