IMPROVING THE ACCURACY OF THE FORECAST OF ELECTRICITY PRODUCTION BY PHOTOELECTRIC PLANTS BASED ON THE RANDOM FOREST METHOD

The paper considered the Random Forest method as a tool for forecasting the generation of photovoltaic plants, which is a key element for optimizing the structure of the electricity balance in the power system. Photoelectric stations are characterized by significant dependence on meteorological conditions, which creates difficulties in accurately forecasting their generation. The use of the Random Forest method allows taking into account numerous variables and complex interrelationships between them, which increases the accuracy of the forecast in conditions of significant uncertainty inherent in renewable energy. On the basis of the conducted research, an effective Random Forest application algorithm was proposed, which allows to reduce forecast errors. Particular attention is paid to how the model can adapt to changes in weather conditions and seasonal variations, which are important factors in maintaining the stable operation of the power system. The results of the study demonstrate that the application of the Random Forest method for forecasting the generation of photovoltaic plants provides an acceptable level of accuracy for planning power reserves in the power system. In particular, it allows power system operators to more effectively manage the balance of electricity, planning in advance the necessary power reserves depending on the forecasted volumes of renewable energy generation. This is especially important to ensure the stability of the power system during peak loads and in conditions of a growing share of renewable energy sources in the overall generation structure.