Spatio-temporal models for photovoltaic power short-term forecasting

Xwégnon Ghislain Agoua (1) , Robin Girard (1) , Georges Kariniotakis (1)

1 PERSEE - Centre Procédés, Énergies Renouvelables, Systèmes Énergétiques

The interest for photovoltaic (PV) generation has grown in recent years, while some areas start to witness significant penetration of PV production in the grid. However, the power output of PV plants is characterized by an important variability since it depends on meteorological conditions. Accurate forecasts of the power output of PV plants is recognized today as a necessary tool to facilitate large scale PV penetration. In this paper, we propose a statistical method for very short-term forecasting (0-6 hours) of PV plants power output. The proposed method uses distributed power plants as sensors and exploits their spatio-temporal dependencies to improve the forecasts. It uses as input only production data of the geographically distributed power plants, while its computational requirements are small making it appropriate for large-scale application.

Mots clés: photovoltaic; forecasting; models

Domaines: Energétique

Xwégnon Ghislain Agoua, Robin Girard, Georges Kariniotakis. Spatio-temporal models for photovoltaic power short-term forecasting. Solar Integration workshop 2015, Oct 2015, Brussels, Belgium. ⟨hal-01220321⟩