Estimation of a simple model of solar power generation using partial information

Abstract

This paper presents a heuristic method for the estimation of a model of a photovoltaic plant to be used for power generation forecasting purposes. The problem is addressed in a scenario where measurements of meteorological variables (i.e., irradiance and temperature) at the plant site are not available. The proposed approach efficiently exploits only power generation measurements and theoretical clear-sky irradiance, and is characterized by very low computational effort. A real-world application is presented to validate the procedure, which is also currently in pre-deployment phase at a Distribution System Operator control center in Northern Italy.