SOLARGIS: SOLAR DATA AND ONLINE APPLICATIONS ......Finding site-specific high-quality data, such as solar radiation and air temperature is not straightforward task; it often demands

Šúri M., Cebecauer T., Skoczek A., 2011. SolarGIS: Solar Data And Online Applications For PV Planning And Performance Assessment.. 26th European Photovoltaics Solar Energy Conference, September 2011, Hamburg, Germany.

1

SOLARGIS: SOLAR DATA AND ONLINE APPLICATIONS FOR PV PLANNING AND PERFORMANCE ASSESSMENT

Marcel Šúri, Tomáš Cebecauer, Artur Skoczek GeoModel Solar s.r.o., Pionierska 15, 831 02 Bratislava, Slovakia, tel: +421 2 492 12 422, fax: +421 2 492 12 423;

website: http//geomodelsolar.eu; corresponding author: [email protected]

ABSTRACT: SolarGIS is a new generation web service (http://solargis.info) aimed to increase efficiency and reduce uncertainty in planning and performance assessment of PV systems. SolarGIS is based on more than 10 years of R&D, international collaboration, and experience with previous projects, such as PVGIS. The system consists of high-resolution database of solar radiation and air temperature (which are operationally calculated for Europe, Africa, Asia and Brazil) and some other auxiliary parameters. Solar radiation is calculated from Meteosat Prime and IODC satellites data covering a period from 1994 (2000) to the present. Derived solar parameters are calculated for any fixed-mounted or sun-tracking PV. Air temperature is derived from atmospheric models from ECMWF and NCEP. SolarGIS integrates previously published and validated PV performance models to simulate module surface reflectivity, non-linearities in the conversion efficiency due to irradiance/temperature outdoor conditions for c-Si, CdTe, and CIS/CIGS modules and shading. Four applications are implemented in the system: (i) iMaps - high-resolution global interactive maps and data, (ii) climData - interactive and automated access to solar radiation and air temperature; (iii) pvPlanner - PV performance simulator with a new concept of simulation algorithms and data formats and (iv) pvSpot – a tool for performance evaluation and monitoring of PV systems. The key innovation of SolarGIS involve: (i) high-resolution and detailed solar radiation database, extensively validated and documented, (ii) Implementation of a new aerosol database from ECMWF into solar radiation model, (iii) Increased accuracy of PV simulation, (iv) Possibilities of near-real time data delivery, (v) Detailed interactive maps, precise geo-positioning and availability of other support data. Keywords: SolarGIS online system, solar radiation data, air temperature data, site assessment, PV simulation, PV performance evaluation, Meteosat

1 INTRODUCTION A number of data and software options are available in Europe and worldwide [15, 16]. However, there are still options to improve efficiency of site selection, planning, and design of solar energy projects. SolarGIS (http://solargis.info/) is a new generation web service offering a range of unique features. The core of the system is based on (i) a new high-resolution climate database, systematically build from satellite and meteorological sources, and with extensively validated accuracy. The innovations of the SolarGIS web system improve the grounds introduced to solar community by PVGIS web prototype [1]: (ii) fast and easy geographical navigation using interactive maps, (iii) easy to understand tools for accessing and display of data, simulation software and tools for performance evaluation and monitoring for solar energy systems. 2 PRIMARY DATABASE

At present the data geographically cover Africa,

major parts of Europe, Asia and eastern part of Latin America (fig. 1). Besides solar radiation and air temperature parameters (which are particularly described in chapters 2.1 and 2.2) the database includes also high resolution Digital Elevation Model (source SRTM-3, spatial resolution 80 metres in Central Europe), data about land cover and population. Wind speed and wind direction at 10 m height and relative humidity parameters has been recently implemented, too. In order to provide fast access to these resources, data are organised in data structures that are optimised for a point access.

Figure 1: Present coverage of SolarGIS database

represents the areas in the orange colour. 2.1. Solar radiation A new model for high-performance calculation of

Global Horizontal Irradiance (GHI) and Direct Normal Irradiance (DNI) from satellite was implemented [9]. The model is designed for operational processing of Meteosat MSG and MFG satellite data at full spatial and temporal resolution. The algorithms are based on the Heliosat-2 calculation scheme [17] and the approach of Perez et al. operational satellite-to-irradiance model [18].

The key enhancements of the new SolarGIS model include: (1) multi-spectral satellite information to improve classification of snow/land/cloud signals (especially with MSG satellite data), (2) a new algorithm to more accurate calculation of lower bound preserving diurnal variability, (3) implementation of backscatter correction (4) variable upper bound for dynamic range and cloud index calculations, (5) new clear sky model, adapted DirIndex model for calculation of DNI from Global Horizontal Irradiance, (7) downscaling with high


2

resolution DEM to include local variability of solar irradiance (algorithm published recently by [14]), (8) implementation of a new aerosol database from ECMWF [13] resulted in a substantial improvement of especially Direct Normal Irradiance (fig. 2).

Two global atmospheric data sets are implemented in the clear-sky model: water vapour (WV) and Atmospheric Optical Depth (AOD), both received from ECMWF [10]. New AOD data set is one of the main outcomes of the European project GEMS. It better captures daily variability, especially events with extreme atmospheric load of aerosols and water vapour. Thus it reduces uncertainty of instantaneous GHI and especially DNI estimates. Main accuracy improvements were achieved in reduction of Root Mean Square Deviation (RMSD) and improved distribution functions.

Compared to other approaches, the advancements of the model improved results in mountains, coastal and arid zones, and for periods with snow coverage. IEA SHC Task 36 and MESOR [10] recommendations were used for database validation. Quality is compared with high-quality ground measurements from 60 stations over Europe and Africa, 5 stations being located in high mountains. The relative mean bias (rMB) for Global Horizontal Irradiance GHI is 1.1%, and root mean square difference (rRMSD) is 18.5%, 9.6% and 4.8% for hourly, daily and monthly data, respectively. Mean bias of GHI

a)

b)

Figure 2: Frequency distribution of DNI values [W/m2] for Valladolid (Spain), a) results based on the AOD long-term monthly averages, b) results based on daily values

from the new GEMS/ECMWF database in the Alps is between -8% and +4% and hourly rRMSD is around 30%, and the data show significant improvements in mountains, coastal zones, and arid

areas. The mean bias of Direct Normal Irradiance (DNI) is -

3.1 W/m2 (-0.9%), standard deviation of biases 6.4% and Root Mean Square Deviation (RMSD) is 123, 75, and 32 W/m2 (35.7%, 21.9% and 9.3%), for hourly, daily and monthly data, respectively. In Spain, the DNI model was validated against data from 14 ground stations. The mean bias of DNI is 16.4 W/m2 (4.2%), standard deviation of biases 3.9%, and hourly RMSD is 110 W/m2 (32.2%). For sites in desert (Tamanrasset and SedeBoqer) the quality validation shows higher deviation: mean bias is -26 and -71 W/m2 (-4.2 and -11.7%), respectively, and RMSD is 144 and 174 W/m2 (23.3 and 28.4%) respectively. DNI data for deAar (South Africa): bias -3.1%, RMSE 16.5%. This is consistent with other ground data used in the region.

Solar radiation data are available: i) in PRIME satellite region from April 2004 onwards at 15-minute time-step, for period 1994-2005 at 30-minute time-step; ii) in IODC satellite region from 1999 onwards at 30-minute time-step. Aggregated time domains are available in high spatial resolution, too.

2.2. Air temperature Air temperature is derived by disaggregation of ECMWF ERA Interim and/or NCEP GFS and CFSR data (depending on region) to a final grid resolution 1 km. In the disaggregation high-resolution digital terrain model SRTM-3 is used to adapt vertical gradient of temperature from low resolution grid cells to 1-km grid. Primary data include 6-hourly values covering a period from 1994 up to present. By temporal interpolation, 15-minute time steps are calculated from the primary data in SolarGIS applications. 3 ONLINE APPLICATIONS The volume of databases extends several terabytes, and the database is continuously updated. Therefore the only way to access them is via online applications. Four applications are implemented in the system: (i) iMaps - high-resolution global interactive maps, (ii) climData - interactive and automated access to solar radiation and air temperature; (iii) pvPlanner - PV performance simulator with a new concept of simulation algorithms and data formats and (iv) pvSpot – a tool for performance evaluation and monitoring of PV systems. All applications use an interactive map interface and geographical search utility (based on Google technology) and have extensive reporting capabilities (info delivered in standardized formats, such as csv, xls, and pdf). The interactive information is complemented by an e-shop with poster-size maps for solar energy at a professional cartographic quality. 3.1 SolarGIS iMaps The first map-based prototype for solar energy was made within PVGIS project [1]. The main task in SolarGIS was to adapt new geoinformation approaches


3

Figure 3: Snapshot of SolarGIS iMaps in action

and web technologies and to make search of information and map exploration more efficient and understandable also for non experts. The iMaps application enables interactive browsing on maps (spatial resolution up to 250 meters) of solar radiation and temperature from a continental to local scale. Other support maps are also available: terrain, land cover, population and standard Google maps. For any point clicked on a map, the system interactively shows values of annual solar radiation (global horizontal, 2-axis tracking surfaces, diffuse and direct normal irradiation) and air temperature (yearly average, January and July). In addition, support information about terrain, land cover and population are also displayed. The information in the database is stored in optimized data formats so that the site-specific information is available on one click in full resolution for the whole database extent. On top of it, the parameters of long-term monthly averages can be downloaded for every selected point. The system is designed to provide users with a fast access to key information needed for a site selection or comparison. Map-based technology shows information in a spatial context, thus a user can better understand regional changes of climate resulting from proximity to a seaside, mountains, terrain shading, elevation, etc.

3.2. SolarGIS climData

In different stages of solar project development raw data are needed. Finding site-specific high-quality data, such as solar radiation and air temperature is not straightforward task; it often demands a lot of time, and results in approximate solutions. Still it is very difficult to access data in many countries or regions at the reasonable quality and accuracy [15, 16]. SolarGIS offers historical averaged data, as well as time series for a selected time period and temporal resolution up to the recent history for any selected location. A user can choose different periods and data parameters (global, direct diffuse, radiation, temperature, etc.) and time scales: § Long-term averages: monthly, § Typical meteorological year (TMY): hourly § Time series: monthly, daily, hourly and 15-minute (or

30-minute) values

Figure 4: SolarGIS climData – data selection dialogue

Data can be exported to XLS, CSV and PDF formats. Data types are compatible with most used design software. Operational services (e.g. FTP/email services) offer data in near-real time for sites with existing solar energy installations, enabling thus flexible performance monitoring and audit. 3.3 SolarGIS pvPlanner The pvPlanner application is a second generation online simulator for photovoltaic systems. It is aimed at providing fast results at any selected site online. It integrates previously published and validated PV performance models to simulate module surface reflectivity [2], non-linearities in the conversion efficiency due to irradiance/temperature outdoor conditions for c-Si, CdTe, and CIS/CIGS modules, and shading [3,4,5,6]. An improved approach for modeling a-Si modules is presented in a separate abstract [7]. The simulation methodology is described in [8]. Optimally, full time series of solar radiation and air temperature are needed for accurate simulation of non-linearities in PV output. Due to computational demand this is not possible in online applications, and therefore simplifications, such as use of monthly-averaged daily profiles of solar data are used, e.g. in [1]. This simplification leads to systematic errors up to more than 3% [4]. The SolarGIS methodology is based on using statistically aggregated solar and temperature data stored in the database with a time step of 15 minutes. Simplified input parameters enable to consider key characteristics of a PV system, such as its position, geometry, type and mounting of modules, efficiency of inverter, and assumed losses in DC and AC sections. The model calculates reflectance losses at the surface of PV modules and losses due irradiance and temperature characterising operating performance of modules in a site-specific climate conditions. The other system losses, mainly at the DC and AC section are to be set by a user. To consider the non-linearities in calculation of in-plane solar radiation and PV output, global and direct irradiance for each Meteosat grid cell (about 4x5 km in Central Europe) is categorised into seven percentiles. Thus, online calculation for each month reads averaged temperature profile and seven irradiance percentiles in


4

each step of 15 minutes (or 30 minutes). In total about 5500 model runs are needed for calculating PV potential output by pvPlanner for any change of input parameters, which is still acceptable and does not evoke long response time. This approach is still a simplification compared to calculating full 15-minute time series, however it reduces error to ±0.7%, and the error is not systematic (fig. 5).

Figure 5: Estimate of the annual sum of global tilted

irradiation as a function of PV module inclination, using three approaches. The error of the estimate

increases with increasing module inclination when monthly-averaged daily profiles are used (full line).

When using aggregated statistics (percentiles, implemented in SolarGIS), the error remains within

±0.7% (dotted line). Results are compared to the calculation based on full 15-minute time series.

The spatial resolution of data is equivalent to the Meteosat grid cell size; however the data are on the fly disaggregated to about 80 meters using terrain elevation and horizon angle values, which are pre-calculated from SRTM-3 DEM data. As described above, implementation of new AOD data and disaggregation algorithms make possible to evaluate more precisely terrain effects. Roof-mounted systems are supported by tools, which help to calculate roof length, area and roof azimuth. The pvPlanner application brings fast and easy to understand approach for site selection, prefeasibility, and planning of PV systems. Sites can be compared and different technology options tested. New software tools are available for a comprehensive PDF reporting and data export to common formats.

Figure 6: SolarGIS pvPlanner -

- snapshot of the main screen showing the simulation of PV electricity potential.

3.4 SolarGIS pvSpot SolarGIS pvSpot is the comparative tool for performance appraisal of a PV system in operation. The comparison is based on a user-specified real production data with independent simulations applicable for any site or roof. pvSpot integrates operational data services (near-real time data) with pvPlanner functionality. Based on aggregated 15-minute solar radiation data, air temperature data, basic PV system definition and geographical position, pvSpot provides estimated electricity production of any defined PV system for certain period of time (e.g. for yesterday, last week, month, year or user specified period of time). This is extremely useful for the evaluation of small PV systems, where no control sensor is installed or for comparison with control systems on PV installation.

Figure 7: SolarGIS pvSpot – snapshot of the screen

showing daily overview of potential electricity production for defined PV system

4 CONLUSION The key innovation of SolarGIS database and numerical models: § High-resolution and detailed solar radiation database,

extensively validated and documented [9,10,11]; increased accuracy, especially in the mountains, arid and tropical climate, coastal zones, and for events with snow and ice; the database provides a new insight in solar climate of many regions in Asia and Africa. Data are extensively validated using IEA SHC Task 36 and MESoR standards [12].

§ Implementation of a new aerosol database from ECMWF [13] and data disaggregation model [14] resulted in a substantial improvement of especially Direct Normal Irradiance, thus reducing uncertainty in CPV (and also CSP) applications;

§ Increased accuracy of PV simulation, based on the use of statistically aggregated data, enabling calculation at high spatial and temporal data resolution with improved accuracy and fast model response. Compared to simplified approaches, based on the assumption of a typical daily profile, and still used in some PV simulators, this implementation resolves issues related to the non-linear behaviour of PV components.


5

§ Detailed interactive maps, precise geo-positioning and availability of other support data and maps with a global coverage (terrain, land cover and population density) allow exploration, effective site identification and preliminary assessment;

§ The operational models deliver solar and temperature data in near-real time for performance assessment and monitoring of PV.

SolarGIS combines scientific knowledge with modern data and IT technology. It aims to reduce uncertainty and time needed for accessing the key information, and to increase accuracy and reliability of the data services. SolarGIS provides standardized reporting and transactions to facilitate the effective and transparent communication between financing institutions, project developers, operators and owners. REFERENCES [1] Šúri M., Huld T., Cebecauer T., Dunlop E.D., 2008.

Geographic Aspects of Photovoltaics in Europe: Contribution of the PVGIS Web Site. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 1, 1, 34-41.

[2] Martín N., Ruiz J. M., 2005, Annual angular reflection losses in PV modules, Progress in Photovoltaics, 13, 1, pages 75–84

[3] Huld T., Šúri M., Dunlop E.D., 2008. Geographical Variation of the Conversion Efficiency of Crystalline Silicon Photovoltaic Modules in Europe, Progress in Photovoltaics: Research and Applications, 16, 585-607.

[4] Huld T, Gottschalg R, Beyer H G and Topič M, 2010. Mapping the performance of PV modules, effects of module type and data averaging, Solar Energy, 84, 324-338.

[5] Huld T., Cebecauer T., Šúri M., Dunlop E. D., 2010. Analysis of one-axis tracking strategies for PV systems in Europe. Progress in Photovoltaics: Research and Applications, 18, 3, 183-194.

[6] Huld T, Friesen G, Skoczek A, Kenny R, Sample T, Field M, Dunlop D, 2011. A power-rating model for crystalline silicon PV modules”, submitted to Solar Energy Materials and Solar Cells.

[7] Skoczek A., Virtuani A., Cebecauer T., 2011. Energy yield prediction of amorphous silicon modules using full time data series of irradiance and temperature for different geographical locations. Submitted to EUPVSEC, Hamburg.

[8] Description of methods implemented in SolarGIS pvPlanner. http://solargis.info/doc/_docs/SolarGIS_pvPlanner_methods.pdf

[9] Cebecauer T., Šúri M., Perez R., 2010. High performance MSG satellite model for operational solar energy applications, ASES National Solar Conference, Phoenix, USA.

[10] Cebecauer, T., Šúri M., 2010. Accuracy Improvements of Satellite-Derived Solar Resource Based on Gems Re-analysis Aerosols. Proceeding of the SolarPACES 2010, Perpignan, France.

[11] Šúri M., Cebecauer T., Perez R. 2010. Quality Procedures of SolarGIS for Provision Site-speficic Solar Resource Information. Proceeding of the SolarPACES 2010, Perpignan, France.

[12] Hoyer-Klick C., Beyer H.G., Dumortier D., Schroedter-Homscheidt M., Wald L., Martionli M., Schillings C., Gschwind B. Menard L., Gaboardi E., Ramirez-Santigosa L., Polo J., Cebecauer T., Huld T., Šúri M., De Blas M., Lorenz E., Kurz C., Remund J., Ineichen P., Tsvetkov A., Hofierka J., 2010. Management and Exploitation of Solar Resource Knowledge. Eurosun 2010, 29. 9. - 01.10. 2010, Graz, Austria.

[13] Morcrette J., Boucher O., Jones L., Salmond D., Bechtold P., Beljaars A., Benedetti A., Bonet A., Kaiser J.W., Razinger M., Schulz M., Serrar S., Simmons A.J., Sofiev M., Suttie M., Tompkins A., Uncht A., 2009, GEMS-AER team. Journal of Geophysical Research, 114.

[14] Ruiz-Arias J.A., Cebecauer T., Tovar-Pescador J., Šúri M., 2010. Spatial disaggregation of satellite-derived irradiance using a high-resolution digital elevation model. Solar Energy, 84, 9, 1644-1657.

[15] Šúri M., Remund J., Cebecauer T., Dumortier D., Wald L., Huld T., Blanc F., 2008. First steps in the cross-comparison of solar resource spatial products in Europe. Proceeding of the EUROSUN 2008, 1st International Conference on Solar Heating, Cooling and Buildings, Lisbon.

[16] Šúri M., Remund J., Cebecauer T., Hoyer-Click C., Dumortier D., Huld T., Stackhouse P.W., Ineichen P., 2009. Comparison of Direct Normal Irradiation Maps for Europe. Proceeding of the SolarPACES 2009, Berlin, Germany.

[17] Hammer A., Heinemann D., Hoyer C., Kuhlemann R., Lorenz E., Müller R., Beyer H.G., 2003. Solar Energy Assessment Using Remote Sensing Technologies. Remote Sensing of Environment, 86, 423–432.

[18] R. Perez, P. Ineichen, K. Moore, M. Kmiecik, C. Chain, R. George, F. Vignola, 2002. A New Operational Satellite-to-Irradiance Model. Solar Energy 73, 307-317.

Documents

SOLARGIS: SOLAR DATA AND ONLINE APPLICATIONS ......Finding site-specific high-quality data, such as solar radiation and air temperature is not straightforward task; it often demands