Title: High-Resolution Digital Elevation Model for Improved PV Yield Estimates
Authors: CEBECAUER TomasHULD THOMASSURI MARCEL
Citation: Proceedings of the 22nd European Photovoltaic Solar Energy Conference p. 3553-3557
Publisher: WIP-Renewable Energies
Publication Year: 2007
JRC N°: JRC36425
URI: http://publications.jrc.ec.europa.eu/repository/handle/JRC36425
Type: Contributions to Conferences
Abstract: The amount of solar radiation arriving at a PV system is often influenced by the local terrain conditions. The intensity of the direct and diffuse components of radiation are determined by the elevation above sea level (air mass), by shadows cast by neighbouring terrain features (hills and mountains) and by the sky fraction visible by the PV module. Reliable solar radiation data are needed for system siting and also for performance assessment. Terrain effects are not always considered in project planning, which may lead to yield overestimations. In this paper we analyse the attenuation of solar radiation caused by shadowing from local terrain features and the implications of this effect on modelled performance of fixed-plate and 2-axis tracking systems. High (100-m) and low (1-km and 3-km) resolution digital elevation models (DEM) are used to demonstrate the influence of DEM resolution on the model estimates. For demonstration we have chosen the highly-populated Alpine region in Southern Switzerland and Northern Italy. To calculate solar radiation maps, we use the solar radiation model r.sun and a set of other programs, all integrated in a geographical information system (GIS) GRASS. The model allows the local terrain and shadowing analysis. The tools are optimized for working with extensive GIS datasets and for satellite data processing. The developed tools were validated against global horizontal radiation measured at 10-min intervals in ground stations that are part of ASRB network. In this case elevation data with a grid resolution of 100-m were used together with the r.sun model. Fig. 1 compares clear-sky global irradiance values modelled by r.sun against the measured ones for Davos station. When assuming local terrain horizon (shadowing the direct component) the instantaneous calculation significantly improves, decreasing the estimate error (relative RMSE) of 10-min irradiance from 11.8% to 8.6% for all clear sky days in 2005. Ignoring shadowing by terrain (as is the case of most current software) leads to significant errors in mountains as demonstrated by Fig. 2. Similarly, the use of low-resolution DEM (1-km, or 3-km as is the case of Meteosat satellite) can underestimate the attenuation of the global irradiation in many locations. Terrain shadowing derived from different resolution DEMs has a pronounced effect on 2-axis tracking systems compared to fixed-angle systems, as demonstrated on the monthly average of PV output (averaged over the region) in Fig 3. The total annual overestimation in the Northern Lombardy (Italy) region due to low resolution terrain used in modelling is 6.1% for fixed optimum-angle systems and 10.5% for 2-axis tracking system as this system is more sensitive to the accuracy of the shadowing representation. The described modelling tools and high-resolution digital elevation data have been integrated into the map-based Internet application PVGIS that is designed for improved performance assessment of PV systems. The new version of PVGIS uses a DEM resolution of 100-meters, thus the accuracy of estimates (see an example in Fig. 4) can be compared with the previous version, where 1-km DEM resolution was used. Such improvement is beneficial mainly for those that are responsible for siting and planning a PV system in regions with complex terrain.
JRC Institute:Institute for Environment and Sustainability

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.