Very-short term forecasting of photovoltaic plants generation based on meteorological data from open sources using machine learning

Abstract

This paper is devoted to the problem of predicting electrical energy generation by photovoltaic power plants based on meteorological data from open sources using machine learning methods. The paper presents an overview of existing sources of meteorological data for solving the presented problem and possible methods for their processing, including a comparative analysis of the two most effective methods of machine learning application: ensemble algorithms and neural networks for generation forecasting of a real power plant in the Russian Federation.