Solar irradiance forecasting at one-minute intervals for different sky conditions using sky camera images. (15th November 2015)

• Record Type: Journal Article
• Authors: Alonso-Montesinos, J.; Batlles, F.J.; Portillo, C.
• Is Part Of: Energy conversion and management. Volume 105(2016)
• Page Start: 1166
• Page End: 1177
• Publication Date: 2015-11-15
• Subjects: Solar irradiance forecasting -- Sky camera -- Image processing -- CSP plants -- PV plants; Direct energy conversion -- Periodicals; Energy storage -- Periodicals; Energy transfer -- Periodicals; Énergie -- Conversion directe -- Périodiques; Direct energy conversion; Periodicals; 621.3105
• DOI: 10.1016/j.enconman.2015.09.001
• Languages: English
• ISSNs: 0196-8904

Abstract

Highlights: The solar resource has been predicted for three hours at 1-min intervals. Digital image levels and cloud motion vectors are joint for irradiance forecasting. The three radiation components have been predicted under different sky conditions. Diffuse and global radiation has an nRMSE value around 10% in all sky conditions. Beam irradiance is predicted with an nRMSE value of about 15% in overcast skies. Abstract: In the search for new techniques to predict atmospheric features that might be useful to solar power plant operators, we have carried out solar irradiance forecasting using emerging sky camera technology. Digital image levels are converted into irradiances and then the maximum cross-correlation method is applied to obtain future predictions. This methodology is a step forward in the study of the solar resource, essential to solar plant operators in adapting a plant's operating procedures to atmospheric conditions and to improve electricity generation. The results are set out using different statistical parameters, in which beam, diffuse and global irradiances give a constant normalized root-mean-square error value over the time interval for all sky conditions. The average measure is 25.44% for beam irradiance; 11.60% for diffuse irradiance and 11.17% for global irradiance.