Forecasting the electricity generation of solar power plants

Abstract

Modern methods of forecasting and the actual basis for its implementation are analyzed. To predict the production of electricity using solar power plants (SES), data on long periods of operation of stations in the past are used. This data is processed using statistical methods and machine learning methods to obtain forecasts of the stations' performance in the future. The currently accepted methods of forecasts are given, which allow us to evaluate errors and compare the effectiveness of various forecasting methods. An example of the SES performance forecast using the PV System software product is given.