Photovoltaic and Solar Forecasting: State of the Art

Download Full Document from the International Energy Agency Website (PDF, 1.5 MB)

Authors: Sophie Pelland, Natural Resources Canada
Jan Remund, Meteotest
Jan Kleissl, University of California
Takashi Oozeki, National Institute of Advanced Industrial Science and Technology
Karel De Brabandere, 3E

CETC number: 2013-119

Publication date: 2013-10-21

Abstract:

As costs come down for photovoltaic (PV) systems, which convert solar energy into electricity, concerns are being raised about integrating large numbers of these systems into electricity grids. One way to facilitate this integration is to generate forecasts of how much electricity PV systems will generate in the hours or days ahead, which are used by grid operators to decide which generators to use to meet electricity demand. Such forecasts are for instance being implemented by the Independent Electricity System Operator (IESO) in Ontario.

This report summarizes the state of the art in solar and PV forecasting, and is meant as a guide for operators like the IESO, and for researchers and forecast providers. It discusses the resources used to generate forecasts: measured weather and PV data, satellite and sky images of clouds and numerical weather prediction (NWP) models which form the basis of modern weather forecasting. It shows that measured data is key for forecasts 0 to 6 hours ahead, while NWP models are essential beyond that. The report discusses how to evaluate and compare the accuracy of different forecasts, and examines how accuracy depends on local climate, forecast horizon and the size of the geographic area of interest.

For more information about CanmetENERGY’s activities related to solar photovoltaic energy, visit the Solar Photovoltaic Energy section of the website.