Short-time forecasting of renewable production energy in solar photovoltaic installations

In this paper we describe an automatic system that is able to perform real-time monitoring of a photovoltaic system and short-time forecasting of the production energy. By comparing the theoretical production energy with the real production energy one we can easily detect losses in efficiency. The proposed system was tested on data collected from a photovoltaic installation with two fixed arrays (each connected to an inverter) of solar panels. We made use of two types of least squares regression, the linear regression (LR) and the quadratic regression (QR). The best results were obtained by the QR algorithm using one week as training set for each inverter.