Please use this identifier to cite or link to this item:
|Title:||Uncertainties in Solar Electricity Yield Prediction from Fluctuation of Solar Radiation|
|Authors:||SURI MARCEL; HULD THOMAS; DUNLOP EWAN; ALBUISSON Michel; LEFÈVRE Mireille; WALD LUCIEN|
|Citation:||Proceedings of the 22nd European Photovoltaic Solar Energy Conference p. 3547-3552|
|Type:||Articles in periodicals and books|
|Abstract:||Photovoltaic (PV) yield prediction is often based on estimates of long-term averages of solar radiation. However, uncertainty due to year-by-year variability of solar resource should also be considered especially for financing large PV projects and for management of electricity grids in regions with high concentration of PV. The required information includes the quantification of how much the yearly and/or seasonal values may vary from the long term averages. We present variability analysis of predicted PV yield by comparing yearly and monthly averages to long-term values. The 90% confidence interval has been calculated from the HelioClim-1 database. This gives the limit of deviation of yearly or monthly radiation sums from the long-term averages for which there is only a 10% probability of exceeding this deviation. Regionally, this study covers Central Europe and Mediterranean Basin. The HelioClim-1 database consists of daily global irradiation values which have been computed from Meteosat images1. The data series has primary resolution 15’ x 15’ and covers the period 1985-2005; some data are missing in 1986 and 1988. The database was integrated and processed within the PVGIS2. The monthly and yearly averages were calculated for each year as well as for the whole period. To enable calculation of global irradiation on inclined surfaces, the monthly and yearly long-term averages of the diffuse component were estimated using the Page model applied in the NASA SSE database3. A generic PV system was considered assuming a system performance factor of 0.75 and modules mounted at the optimum angle. An average yield was estimated for each month and year in the period 1985-2004. From the averages for each year the long-term monthly and yearly averages and standard deviations were calculated and the 90% probability distribution of values were estimated assuming a Gaussian distribution of inter-annual values. The variability of yearly sums of solar electricity is fairly low in the Mediterranean basin, while it increases further North in Europe, mainly in latitudes above 45°. Results indicate that due to year-by-year weather variations, in Southern Germany, there is a probability of 90% that the yearly PV yields are within the limits of +/- 8 to 10 % from the long-term average. In Southern Spain the weather regime is more stable, and the annual yields are within 4% from the long-term average, assuming 90% probability. Comparison of monthly average values with long-term averages shows higher stability of PV yields (in relative terms) in summer and the opposite trend (high variability) in winter. The solar resource has distinctive geographical patterns that might affect electricity flows in the transmission network. The variability estimates from HelioClim-1 database are compared to calculations based on the ground-measured global irradiation data available from the World Radiation Data Center. The sensitivity of results is tested against the inaccuracies of solar radiation estimates. We conclude that in some regions of Europe the inter-annual variation of predicted yearly PV yields may differ significantly (more than 10%) from the long-term average. Higher inter-annual variability in seasons may require enforcement of grid transmission system in/between some regions.|
|JRC Institute:||Sustainable Resources|
Files in This Item:
There are no files associated with this item.
Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.