Search academic texts

Information × Registration Number 0225U003684, (0124U003219) , R & D reports Title Smart Grid model for operational management of distribution networks based on artificial intelligence methods popup.stage_title Використання прогнозів вузлового навантаження для розрахунку режимів електричної мережі Head Blinov Ihor V., Доктор технічних наук Registration Date 31-07-2025 Organization Institute of Electrodynamics of the National Academy of Sciences of Ukraine popup.description1 The main goal of the work is to create a test Smart Grid model for the operational planning and management of distribution networks (Lithuanian partners) and to estimate the value of electrical energy losses based on the forecasting of nodal loads and the generation of renewable energy sources (RES) using artificial intelligence methods (Ukrainian partners) popup.description2  As part of the third stage of the research project, a new deep learning architecture based on the Transformer neural network was developed and tested to forecast node loads and electricity losses in distribution grids. Special attention was given to improving prediction accuracy by detecting and correcting anomalous values in the historical data. The study used retrospective measurements from a Ukrainian distribution system operator, enabling realistic modeling of power system behavior. The results show that combining node load forecasting with subsequent loss calculation yields better accuracy than direct loss prediction. The best performance (MAPE = 8.4%) was achieved using this two-step approach. The proposed model demonstrates strong potential for deployment in modern Smart Grid systems, particularly in scenarios involving frequent topological changes—an increasingly relevant issue in Ukraine due to ongoing security threats. This research confirms the effectiveness of deep neural networks in energy loss management, reducing the need for physical measurements and improving forecast reliability. The findings have been published in a peer-reviewed journal indexed by Scopus, highlighting their scientific significance and practical relevance. Product Description popup.authors Miroshnyk Volodymyr O. Parus Yevhenii V. Sychova Viktoriia V. Shymaniuk Pavlo V. popup.nrat_date 2025-07-31 Close