Search academic texts

Information × Registration Number 0226U000065, (0125U002919) , R & D reports Title Short-term forecasting of wholesale electricity market prices for microgrid energy resource management popup.stage_title Розробка методу короткострокового прогнозування цін на оптовому ринку електроененргії України Head Miroshnyk Volodymyr O., Registration Date 02-01-2026 Organization Institute of Electrodynamics of the National Academy of Sciences of Ukraine popup.description1 Development of a new method for short-term price forecasting in the wholesale electricity market of Ukraine in the "day-ahead" segment based on the analysis of supply and demand curves using deep learning artificial neural networks. popup.description2 The research developed a comprehensive approach to price forecasting that combines preliminary data cleaning (removal or correction of anomalous values) with the use of modern deep learning models. A comparative analysis of the accuracy of different ANN architectures in the daily price forecasting problem was conducted. A separate methodology based on two recurrent GRU networks was proposed for forecasting aggregate demand and supply curves, with subsequent determination of the equilibrium price. Main results: the developed approach provides improved price forecasting accuracy. In particular, the GRU-based forecasting model, which takes into account the full demand-supply curves, achieved the lowest values ​​of the average error less on test data, significantly outperforming classical point models. In general, the results of the work confirm the effectiveness of using artificial neural networks (especially convolutional and recurrent) for short-term forecasting of market prices, provided that the data is carefully prepared, allowing for more reliable and accurate forecasts without describing the prerequisites or motivation of the study. Product Description popup.authors popup.nrat_date 2026-01-02 Close