FORECASTING VOLUMES AND PRICES OF BALANCING SERVICES OF IPS OF UKRAINE

Abstract

The new electricity market model in Ukraine aims to enhance and optimize market dynamics, particularly through the transition from a "single buyer" model to a decentralized system. One of the main segments of the wholesale market is the balancing market, which operates in near real-time to improve the stability and efficiency of the power system. This paper aims to analyze the use of probabilistic neural networks (PNNs), specifically Bayesian networks, for forecasting the volumes of balancing services purchased by the transmission system operator, and to investigate classical models for forecasting the price of balancing services. The study included an analysis of demand volumes for balancing services in the upward (loading) and downward (unloading) directions for the periods from March 1, 2022, to June 20, 2023. Overall, the forecasting results for the demand volumes of balancing services are satisfactory but require further improvement. ARIMA and VARMA models were used for price forecasting. Price forecasting for balancing services indicated that the ARIMA model better replicates actual data; however, the accuracy of the forecasts remains low, particularly for the price series of unloading services. To improve forecasting results, it is necessary to optimize the models and use longer data histories. References 7,  table 1, figures 4.