2. Home
  3. ARCHIVE
  4. 2025
  5. VOL 47, NO 5 (2025)
  6. pp. 87-104

FORECASTING ELECTRICITY CONSUMPTION UNDER CONDITIONS OF ROCKET AND DRONE ATTACKS ON THE POWER SYSTEM

S. Saukh, O. Kliuzko

Èlektron. model. 2025, 47(5):87-104

https://doi.org/10.15407/emodel.47.05.087

ABSTRACT

A methodology is proposed to enhance the accuracy of short-term forecasts of hourly electricity consumption in conditions of destructive rocket and drone attacks on the power system. To quantitatively assess this impact, a special load-shedding factor is introduced, which combines binary, rank, and scale characteristics of emergency shutdowns. Based on the proposed forecasting methodology, the Energy AI software complex has been developed, which implements a full cycle of data processing — from determining key influencing factors to generating a forecast and assessing its adequacy. It uses the Random Forest machine learning algorithm with automatic optimization of hyperparameters using the intelligent algorithms of the Optima framework. The Random Forest algorithm uses data from open operational reports on electricity consumption and data from news reports on rocket and drone attacks on critical infrastructure, and the introduction of emergency power outage schedules for electricity consumers as input data. The results of the computational experiments confirm the need to use the load-shedding factor to improve the accuracy of forecasts. The proposed methodology for forecasting electricity consumption is recommended for use in energy information and analytical systems to improve the accuracy of forecast calculations and minimize losses for electricity suppliers.

KEYWORDS

electricity supply, forecasting, Random Forest algorithm, missile and drone attacks, emergency consumer shutdowns.

REFERENCES

  1. Information on massive attacks on Ukraine's critical infrastructure. (Ukr). URL:  https:// map.ua-energy.org/uk/resources/12f3148d-841a-478d-b9ed-72bf0764b286/ (accessed at 11.06.2025).
  2. Saukh, S. & Borysenko, A. (2025) Modeling of ukraine electric power system and its resi­lience assessment in the conditions of systematic terrorist attacks. Technical electrodyna­mics, 2, 57-70. DOI: https://doi.org/10.15407/techned2025.02.057.
  3. Information on the application of measures to limit electricity consumption. (Ukr). URL: https://map.ua-energy.org/uk/resources/0f8f9882-1fb2-47c6-81dc-31fbad914f16/ (accessed at 11.06.2025)
  4. Rangelov, D., Boerger, M., Tcholtchev, N., Lämmel, P. and Hauswirth, M. (2023) Design and Development of a Short-Term Photovoltaic Power Output Forecasting Method Based on RF, Deep Neural Network and LSTM Using Readily Available Weather Features in IEEE Access, 11, 41578-4159. DOI: 10.1109/ACCESS.2023.3270714.
  5. Xuan, Y. et al. (2021) Multi-Model Fusion Short-Term Load Forecasting Based on RF Feature Selection and Hybrid Neural Network, in IEEE Access. 9, 69002-69009. DOI: 10.1109/ACCESS.2021.3051337.
  6. Sartini, Luthfia Rohimah, Yana Iqbal Maulana, Supriatin Supriatin, Dewi Yuliandari (2023) Optimization of RF Prediction for Industrial Energy Consumption Using Genetic Algorithms PIKSEL Penelitian Ilmu Komputer Sistem Embedded and Logic, 11(1), 35-44, DOI: 10.33558/piksel.v11i1.5886
  7. Wang W., Keen J., Bank J., Giraldez J. and Montano-Martinez K. (2023) An Automated Approach for Screening Residential PV Applications Using a RF Model, in IEEE Open Access Journal of Power and Energy, 10, 327-334. DOI: 10.1109/OAJPE.2023.3270223.
  8. Liu F., Dong T., Hou T., Liu Y. (2021) A Hybrid Short-Term Load Forecasting Model Based on Improved Fuzzy C-Means Clustering, Random Forest and Deep Neural Networks IEEE Access 9, 59754-59768. DOI: 10.1109/ACCESS.2021.3063123.
  9. Olcay K., Tunca S.G., Özgür M.A. (2024) Forecasting and Performance Analysis of Energy Production in Solar Power Plants Using Long Short-Term Memory (LSTM) and Random Forest Models IEEE Access, 12, 103299-103312. DOI: 10.1109/ACCESS.2024.3432574.
  10. Contreras J., Espínola R., Nogales F.J., Conejo A.J. (2003) Day-Ahead Electricity Price Forecasting Using the Wavelet Transform and ARIMA Models IEEE Transactions on Power Systems, 18 (2), 1014-1020. DOI: 10.1109/TPWRS.2003.811182.
  11. Mohammadi, Sahar and Nabavi, Seyed Azad and Motlagh, Naser Hossein and Anvari-Moghaddam, Amjad and Tarkoma, Sasu and Geyer, Philipp (2022)A Comparative Study of Machine Learning Modeling in Electricity Load Forecasting. SSRN Electronic Journal Available at SSRN:  http://dx.doi.org/10.2139/ssrn.4200148.
  12. Planned and emergency shutdowns. Official website of DTEK Kyiv Electricity Network. https://www.dtek-kem.com.ua/ua/shutdowns (accessed at 15.06.2025).
  13. Official website of National Power Company “UKRENERGO”. https://ua.energy/uchasnikam_rinku/ (accessed at 15.06.2025).
  14. Passing of the Autumn-Winter Periods 2022-2024: State of the Energy System. DiXi Group. Kyiv. URL: https://dixigroup.org/wp-content/uploads/2024/04/2024_winterseasons_analysis_dixi_group_final.pdf  (accessed at 10.06.2025).

Full text: PDF