Irradiance forecasting for the photovoltaic systems

Abstract

Global warming has emerged as a key environmental issue, with one of the most exciting approaches to greenhouse gas reductions being the use renewable energy. Photovoltaic (PV) generation of electricity is an important renewable energy source, especially at small scale, such as in homes. To increase the value of small-scale renewable generators, one efficient way is to aggregate and control their output and group them in zones. One challenge of doing this is the precise forecasting of each PV output. This paper introduces our developed irradiance forecasting algorithm based on the data from a sky camera. A SVM regression technique is used to generate regress curves between future irradiance and other available information. A series of experimental results are presented to evaluate and demonstrate our forecasting accuracy.