Article

Received 30.03.2025

Revised 04.04.2025

Accepted 28.06.2025

Retrieved from Vol. 29, No. 1, 2025

Pages 98 -104

Myskiv, S., & Viter, M. (2025). Analysis of approaches to computer modelling and forecasting of electricity consumption data. The National Transport University Bulletin, 29(1), 98-104. https://doi.org/10.33744/2308-6645-2025-1-60-098-104

Analysis of approaches to computer modelling and forecasting of electricity consumption data

Serhii Myskiv Mykhailo Viter

This article provides a comprehensive analysis of methods for computer modeling and forecasting of electricity consumption, which is critically important for ensuring the stability and efficiency of modern energy systems, especially in conditions of economic and geopolitical instability. The object of the study is the processes and methods for forecasting time series of electricity consumption. 104 The purpose of the study is to systematize, analyze, and compare traditional statistical approaches (ARIMA, SARIMA) and modern machine learning methods (LSTM, GRU, ensemble models) for electricity consumption forecasting, as well as to identify promising research directions. Methods of the study – to achieve this goal, methods of system analysis, comparative analysis, generalization, and synthesis of scientific publications were used. The paper details the advantages and disadvantages of different model classes. Statistical methods, while simple to implement, have limitations when dealing with non-linear dependencies. In contrast, machine learning models, particularly recurrent neural networks, demonstrate significantly higher accuracy due to their ability to capture complex temporal patterns. Special attention is given to hybrid models that combine the strengths of both approaches and to the contribution of Ukrainian scientific schools in developing adaptive models for crisis conditions. The results of the article can be used by energy companies to improve planning accuracy, optimize network operations, and reduce costs. The results can also be implemented in the educational process for training specialists in the fields of energy and computer science. Forecast assumptions regarding the development of the research object include the further development of multimodal and interpretable models (Explainable AI), the implementation of federated learning for working with confidential data, and the creation of systems adaptive to the transition to renewable energy sources

electricity consumption forecasting; computer modeling; machine learning; statistical models; hybrid models; energy systems
  1. Hippert, H. S., Pedreira, C. E., & Souza, R. C. (2011). Neural networks for short-term load forecasting: A review and evaluation. IEEE Transactions on Power Systems, 26(1), 44–55.
  2. Hong, T., Pinson, P., & Fan, S. (2020). Global energy forecasting competition 2014. International Journal of Forecasting, 36(3), 820–838.
  3. Hyndman, R. J., Liu, X. A., & Koehler, A. B. (2018). Dynamic time warping averaging of time series allows faster and more accurate classification. IEEE International Conference on Data Mining, 470–479.
  4. Rangapuram, S. S., Seeger, M. W., Gasthaus, J., Stella, L., Wang, Y., & Januschowski, T. (2021). Deep state space models for time series forecasting. Advances in Neural Information Processing Systems, 31, 7785–7794.
  5. Salinas, D., Flunkert, V., Gasthaus, J., & Januschowski, T. (2022). DeepAR: Probabilistic forecasting with autoregressive recurrent networks. International Journal of Forecasting, 38(3), 1181–1196.
  6. Makridakis, S., Spiliotis, E., & Assimakopoulos, V. (2023). Statistical and machine learning forecasting methods: Concerns and ways forward. PLoS ONE, 13, 22–30.
  7. Koenker, R., & Bassett, G. (2019). Quantile regression. Econometrica, 46(1), 33–50.
  8. Bottieau, J., Hubert, L., Mercier, Z., Sebastien, D., & Vallee, F. (2020). Metaheuristic scheduling of the last mile in smart cities. IEEE Transactions on Smart Grid, 11(1), 16–26.
  9. Yelisieieva, O. K., Hrynko, A. P., & Velyhura, A. V. (2022). Prohnozuvannia enerhospozhyvannia promyslovykh pidpryiemstv za dopomohoiu hlybynnykh neironnykh merezh [Forecasting of energy consumption of industrial enterprises using deep neural networks]. Visnyk KhNURE - Bulletin of KNURE, 4, 56–64. [in Ukrainian].
  10. Kyrylenko, O. V., Blinov, I. V., & Parus, Ye. V. (2022). Metody prohnozuvannia rezhymiv elektrospozhyvannia v umovakh enerhorynku Ukrainy [Methods for forecasting electricity consumption modes in the conditions of the energy market of Ukraine]. Tekhnichna elektrodynamika - Technical Electrodynamics, 3, 59–67. [in Ukrainian].
  11. Blinov, I. V., & Parus, Ye. V. (2022). Zastosuvannia ansamblevykh metodiv prohnozuvannia elektrospozhyvannia v umovakh nevyznachenosti [Application of ensemble methods for electricity consumption forecasting under uncertainty]. Pratsi Instytutu elektrodynamiky NAN Ukrainy - Proceedings of the Institute of Electrodynamics of the National Academy of Sciences of Ukraine, 60, 5–13. [in Ukrainian].
  12. Tereshchenko, T. O., Yamnenko, Yu. S., & Klepach, L. Ye. (2021). Intehrovana systema prohnozuvannia ta upravlinnia rezhymamy elektrospozhyvannia [Integrated system for forecasting and managing electricity consumption modes]. Elektronika ta zviazok - Electronics and Communication, 95(4), 17–25. [in Ukrainian].
  13. Mysliuk, Yu. V., & Panchenko, V. M. (2021). Prohnozuvannia navantazhennia dlia malykh enerhetychnykh system z vidnovliuvanymy dzherelamy enerhii [Load forecasting for small energy systems with renewable energy sources]. Vidnovliuvana enerhetyka - Renewable Energy, 1, 26–34. [in Ukrainian].

https://doi.org/10.33744/2308-6645-2025-1-60-098-104