Open AccessELECTRICAL LOAD FORECASTING ON HIERARCHICAL LEVELS OF IPS OF UKRAINE USING LSTM NEURAL NETWORK
Author(s) -
S.S. Loskutov,
Pavlo Shymaniuk
Publication year - 2021
Publication title -
pracì ìnstitutu elektrodinamìki nacìonalʹnoï akademìï nauk ukraïni/pracì ìnstitutu elektrodìnamiki nacìonalʹnoï akademìï nauk ukraïni
Language(s) - English
Resource type - Journals
eISSN - 1727-9895
pISSN - 2786-7064
DOI - 10.15407/publishing2021.59.081
Subject(s) - artificial neural network , stability (learning theory) , computer science , electric power system , artificial intelligence , table (database) , power quality , electrical load , power (physics) , electrical network , electric power , machine learning , data mining , engineering , electrical engineering , physics , quantum mechanics
The scientific research presents the results of a study of one-factor forecasting of the total electrical load at three hierarchical levels of the integrated power system (IPS) of Ukraine using artificial neural networks, such as LSTM. Based on research, forecasting errors at each hierarchical level of the power system were analyzed. Methods for improving the quality and stability of forecasts were proposed. The obtained results are the basis for the study of the assessment of the accuracy of forecasting the summary electrical load in the IPS of Ukraine. Ref. 9, fig. 4, table.