a review on grid-connected pv systems in brazil ...

29th European Photovoltaic Solar Energy Conference and Exhibition

A REVIEW ON GRID-CONNECTED PV SYSTEMS IN BRAZIL INCLUDING SYSTEM PERFORMANCE Andrigo Filippo Antoniolli1,#, André M. Nobre1,2,Thomas Reindl2 and Ricardo Rüther1 Federal de Santa Catarina (UFSC), Caixa Postal 476, 88040-900, Florianópolis – SC, Brazil 2Solar Energy Research Institute of Singapore (SERIS), National University of Singapore, 7 Engineering Drive 1, Block E3A, #06-01, Singapore 117574 # +55 48 9633 5056, [email protected]

1Universidade

ABSTRACT: Brazil is seen as a potential GW-scale photovoltaic (PV) market of the future. It possesses extensive land area (~8.5 million km2), excellent solar resource (annual irradiation values ranging from ~1,500 to ~2,200 kWh/m2), relatively high electricity tariffs and a booming economy, all ingredients which put the country on the path of massive PV energy implementation. This work presents a review of the previously installed grid-connected PV capacity up to 2013. The performance of some of the PV systems already connected to the grid is presented for ten sites. Performance metrics “specific energy yield” [kWh/kWp] as well as “performance ratio”, PR [%] are used, with both locally measured solar irradiation as well as other available databases working as a benchmark to the current system knowledge-base in the country. Keywords: PV System, Solar PV Potential, System Performance, Irradiation Databases

1

INTRODUCTION

ten of these existing systems taking into account figures of merit for the performance assessment of PV systems (specific yield and performance ratio). Additionally, the availability of the local solar resource through several different data sources namely SWERA (Solar Atlas) [8, 9], OLADE [10], METEONORM [11], NASA [12] e RORIZ [13]) will be explored, thus allowing for the decision on which values are more reliable to be used as reference in the simulation of PV systems, when local irradiance readings or in the vicinity are not available.

In terms of solar energy, Brazil has one of the best irradiation resources in the world (~1,500 to ~2,200 kWh/m2 per year of global horizontal irradiation) [1]. The combination of optimum natural characteristics favoring the adoption of solar renewable energy, coupled with the presence of investments in research and development in solar photovoltaics (PV), has not been sufficient to leverage the local market to considerable numbers. By the end of 2013, out of the ~139 GWp of PV systems deployed worldwide, only circa 10 MWp were installed in Brazil (less than 0.01% of the world market) [2]. With its great land extension (~8.5 million km2), Brazil experiences several climatic conditions. It is subdivided into eight different climatic zones [3]. In Florianópolis (Santa Catarina state), located in the south part of Brazil, the climate is subtropical. In this state capital, throughout the past 15 years, 1,527 kWh/m² in average was measured for global horizontal irradiation (GHI) at the campus site where Universidade Federal de Santa Catarina (UFSC) has been continuously operating the first grid-connected, building-integrated PV generator in the country [4]. The comparison of these values with the ones measured by the Brazilian Solar Atlas [1], which is based on satellite observations and models, have shown an overestimation of the latter by about 10%, already considering for horizon-dependent losses [5]. Brazil is targeted to ramp up its grid-connected PV systems installed capacity in the next years. In a recent solar PV-only energy auction for the sale of future capacities, a total of almost 10 GW were proposed [6]. Additionally, the residential rooftop distributed market is expected to boom due to high electricity tariffs, especially in major city centers. For example, in Belo Horizonte, in the Minas Gerais state, residential electricity prices as of September 2014 are ~0.58 BRL (Brazilian reais) per kWh (~0.26 USD/kWh) [7]. Nevertheless, there are questions about the solar irradiation database to be used for the planning and simulation of solar photovoltaic systems in the country. Trying to address that, as well as the status of PV in the country, this work focuses on two main areas: (1) cataloguing the current situation of grid-connected PV systems in Brazil (up to August 2014) and (2) evaluate

2

METHOD

In this work, grid-connected solar PV systems in Brazil will be the focus of the study. The systems to be included are both building applied photovoltaics (BAPV) or building integrated photovoltaics (BIPV), and include systems that have been connected to the grid in the country up to August 2014. Off-grid solar systems are not part of the scope of this work. The first goal is to catalogue PV systems in operation or about to be connected to the grid, trying to identify their technical characteristics such as installed capacity, module technology, tilt, azimuth, date of commissioning and also system geographic location. The federal government agency ANEEL (Associação Nacional de Energia Elétrica, Brazilian Electrical Energy Agency [14]), which currently enrolls new PV systems being connected to the grid was consulted for that purpose. Local utility companies, universities, technical institutes, PV system integrators and in some cases given the availability of the information, the system owners, were also consulted. In a second effort, ten case studies using PV systems were evaluated, which have already being installed in many locations in the country. Their operational behavior was simulated and compared to energy generation data recorded on site. 2.1 Repository of grid-connected PV systems in Brazil In order to create a database of grid-connected PV systems in Brazil, a form was created for each system where basic information was sought after such as location

2747


(city, state and coordinates), installed capacity (in kWp), system current status (deployed, under commissioning or planned), start of operations (for systems already deployed), module information (technology and brand), brand of inverters, system tilt and azimuth, shading level (heavy, partial or no shading), mode of installation (BAPV or BIPV), and finally the sector responsible for the given installation (residential, commercial, industrial, government, educational or R&D). Interviews with local and state utilities, the relevant federal government agency (ANEEL), several integrators and system owners took place, in order to identify whether a given PV system was already in operation or in the planning stage. Through the consultation, it was observed that some of the information requested was not available from a particular site. At least four data sets were available for all sites – installed capacity, system location, year of commissioning and sector. 2.2 Case studies Figure 1: Brazilian Federal States where the ten PV systems under evaluation in this study are located.

Ten PV systems were added to this study, being mostly from R&D projects with the utility company Tractebel Energia (SC-2, SC-3 and SC-4 in this paper) and some with direct request to system owners, whose data could be downloaded from their monitoring systems. These systems are distributed throughout the following states: Santa Catarina (SC), Paraná (PR), Rio de Janeiro (RJ), São Paulo (SP), Minas Gerais (MG), Bahia (BA) and Pará (PA). Table I shows basic characteristics for these ten installations, and Figure 1 shows the Federal States in which they are installed.

For each grid-connected PV system, the monthly solar electricity generation was requested (kWh/month) for the years of 2012 and 2013. For some systems, data for the entire two years were not available due to the PV system recent installation date. Plans of array PV system solar irradiation data were sourced, when available. For some cases, due to the absence of a local sensor, irradiation was obtained from alternative sources such as nearby meteorological stations. For the systems located inside the University campus in Florianópolis-SC (SC-2 and SC-4), data from a global horizontal irradiance pyranometer installed at the site of system SC-11 were utilized. Software RADIASOL, using the Perez model [10], was used for the calculation of the in-plane irradiation. For system SC-3, which is located at the Florianópolis airport, approximately 9 km from SC-1, -2 and -4, irradiation was obtained from a reference solar cell installed on the plane of array for that system. For PV system BA-1, integrated to the rooftop of the Pituaçu Stadium in Salvador, Bahia (BA), the irradiation was captured by a reference cell tilted at 5 degrees, on the same plane of the photovoltaic array. For PV system PR-1 located in Curitiba, Paraná, irradiation was acquired by the Instituto Nacional de Meteorologia (INMET, National Meteorology Institute) for the station A807, located in the downtown Curitiba area (INMET, 2013). As the pyranometer is also installed horizontally, software RADIASOL was again used for the determination of the irradiation on the plane of the array. For the system located in Belém, Pará (PA-1), the irradiation data was acquired from a station 5 km away from the PV system. The values are also presented by Blasques et al. [15], who analyzed the first 20 months of operation of the PV system. For the grid-connected PV system located in São Bernardo do Campo, São Paulo (SP-1), irradiation values were obtained from the meteorological station at Instituto

Table I: Basic system characteristics and location of ten grid-connected PV systems evaluated in this study.

1

For system SC-1, where the mentioned pyranometer is installed, data from a second pyranometer installed on the plane of the array is used for the needed analysis for SC-1.

2748


de Astronomia, Geofísica e Ciências Atmosféricas of the Universidade de São Paulo (IAG-USP), some 7 km (27.65° S; 46.62° W) away from the PV system. Similarly to the case of PR-1, SC-2 and SC-4, in-plane irradiation was derived from the use of software RADIASOL. For the system in Macaé, Rio de Janeiro (RJ-1), irradiation data were acquired from a meteorological station A608 from INMENT (22.38° S; 41.81° W), 11 km away from the system. Finally, for the PV system located in Belo Horizonte, Minas Gerais (MG-1), the irradiation data were obtained from meteorological station A521 (19.88° S; 43.97° W) also from INMET. Firstly, with installed system power capacity and monthly electricity generation data at hand, specific yield was calculated. Using the irradiation data, an estimated value of performance ratio (PR) is calculated. Both yield and performance ratio values were calculated on a monthly and yearly basis. As mentioned previously, different irradiation databases were taken into account in this study. These values were benchmarked with the goal as acting as guidance for researchers and engineers. For the comparison, the measured data from 2012 and 2013 (either in loco or near the PV system) were used as baseline and compared against the other five data sources, which are based on averages ranging from 10 to 22 years. After finding the percentage variations among the irradiation databases, the influence of the use of each particular source of data was discussed in terms of how it affects the performance calculation for the PV systems (taking the energy readings of 2012 and 2013 into account). An analysis of the irradiation databases with a baseline performance ratio of 80% was conducted in order to benchmark these several [16-21].

Figure 2: Annual and cumulative installed PV capacities in Brazil until Aug/2014.

Figure 3: Pictorial view of the existing installed gridconnected solar PV capacity distribution in Brazil per state, with the darkest color states of Minas Gerais (MG), Santa Catarina (SC) and Bahia (BA) leading the way.

RESULTS Data gathered as described in this work (Figure 2) show the strong growth over the past 4 years of gridconnected PV systems in Brazil. Although the absolute values are negligible when compared with leading photovoltaic markets in the world, it does show a compound annual growth rate of more than 100%. Figure 3 shows the states with the majority of the solar PV adoption thus far are Minas Gerais (MG), Santa Catarina (SC) and Bahia (BA), although it is expected that the solar-rich North-eastern region of Brazil will see many MW-scale PV systems in the near future [22]. Figure 4 exhibits the types of the existing system installations, with R&D/educational sites leading the first wave of the country’s solar implementation roadmap. Performance results obtained thus far indicate that the ten PV systems under investigation here show average specific yields of ~1,230 kWh/kWp (Figure 5). Year-onyear variations were observed, with a few cases suffering from detectable system downtimes. Measured performance ratios have ranged from 60-80%, in line with PRs found elsewhere in the tropics [21]. The investigation displays monthly and yearly performance metrics, which further assist in determining seasonal trends in the solar resource throughout Brazil.

Figure 4: In A, the percentage of installed solar PV capacity in Brazil per sector. It can be observed that the majority of systems in the country originated from R&D and educational efforts. In B, taking absolute number of PV installations into account, it can be seen that most of the systems are residential units.

2749


Table 2: Measured annual irradiation records, used as baseline to compare with other irradiation databases.

Figure 5: Ten PV systems under investigation in this work and their annual yields in 2012 and 2013. Annual yield averages point out to a sunnier 2012 in Brazil in comparison with 2013. Several established irradiation databases were used for benchmarking of PV system performance (Figure 6) and compared with calculated PR results based on local irradiation measured.

Figure 6: Performance ratios (PR) in 2013 of ten sites based on several irradiation databases of the study.

2750


Figure 7 shows an analysis of the measured specific yield of PV system MG-1. Taking years 2012 and 2013 into account, a yield of 1,340 kWh/kWp was achieved. Using the other irradiation sources, with a theoretical performance ratio of 80%, the following specific yields are achieved for each database: 1,321 kWh/kWp for Meteornorm, 1,415 kWh/kWp for OLADE, 1,650 kWh/kWp for SWERA, 1,830 kWh/kWp for RORIZ and 1,545 kWh/kWp for NASA.

CONCLUSIONS Through the analysis proposed in this work, it was possible to highlight the growth potential of gridconnected PV systems in Brazil for the past 5 years. The R&D sector had a crucial role in deploying the majority of the installed capacity of the systems (~43%) with the goal of leveraging this nascent photovoltaic market. Although residential systems are typically of a smaller size, this sector has led the number of installations in the country. The presence of home solar PV systems tends to increase due to high costs of electricity, especially in major urban areas. The existence of system database in Brazil helps disseminating the knowledge base of solar PV in the country, helping utilities, governments, industry and R&D groups alike. The close observation of the market growth, as well as best practices of integrating solar systems, how they behave in the several different climatic zones in Brazil, among other technology-related knowhow, tend to work as a tool for the years ahead with major deployment of solar capacities in the South American country Trying to source for ten case studies in Brazil presented challenges, among them were the identification of relevant systems in different climatic zones, lack of irradiation databases near the systems, and ultimately reliability on the data recorded by utilities or system owners. Irradiation data used varied in sensor type (pyranometer and reference cells) and distance to site (from in loco to up to 30 km of separation). Also, due to the infant stage of the technology in Brazil, not many systems are under operation for a relevant number of years, which constrained the analysis to two full years in this study. Comparing two years of recorded data (both on energy generation and local irradiation resource), versus irradiation databases that ranged from 10 to 22 years of averages, it was possible to see considerable deviation among them, especially between sources that are based on ground measurements and others from databases. One of the goals of this work was to act as a guide to engineers, practitioners, investors and academic segment when dealing with PV systems in Brazil. This study aimed at acting as a stepping stone in areas of PV database management and performance assessment. Present and future R&D efforts are in place or under deployment for the improvement of such analyses.

Figure 7: Grid-connected PV system MG-1: measured specific yield for 2012 (green bar) and 2013 (blue bar) averaged 1,340 kWh/kWp. The orange bars represent the specific yield for the PV system taking into account the use of the given irradiation database and a performance ratio of 80%. DISCUSSIONS The database of grid-connected PV systems proposed in this study aimed at covering the maximum number of PV systems installed in Brazil as possible. Brazilian regulatory agency ANEEL has published in its website the database “BIG” (Banco de Informações de Geração, Generation Information Database), with a list of all gridconnected PV systems being added to the Brazilian network under resolution 482, which dates only from April 17th, 2012 [14]. However, there are many PV systems in operation that have not been authorized by ANEEL or have preceded the creation of the regulatory law. The creation of a repository containing system characteristics would work well in promoting the nascent photovoltaic market in the country among governments, institutions and the public. The database could also play the role of tracking national market evolution on an official basis for international reporting purposes. A future work proposal would be to develop an online map where new system owners could enroll their PV installations. In such a map, similar to Figure 3 presented in this work, a color scale for each state, or even specific cities could be deployed. The methodology and results arising from the use of several irradiation databases showed considerable variation. The availability of irradiation data at PV system sites is limited and comparability among many systems has much room for improvement. For a more accurate exercise on the assessment of PV system performance in Brazil, dedicated irradiance sensors need to be installed on the plane of the array, preferably of research-grade make or calibrated by reputable institutions. Future R&D work aims at addressing performance of PV systems in a comparable basis in several different Brazilian climates

ACKNOWLEDGEMENTS The authors would like to acknowledge the financial support of Tractebel Energia under the ANEEL R&D program and the Brazilian Scientific Research Council (CNPQ). The Alexander Von Humboldt foundation is acknowledged for supporting the installation of the first grid-connected BIPV system in Brazil (SC-1 in this paper). The authors further acknowledge CEMIG and ANEEL, for the information about grid connected PV systems in Brazil. A special acknowledgement also goes to INMET and IAG-USP for the irradiation data from some of the analyzed PV systems.

2751


REFERENCES

18. MONDOL, J. D.; YOHANIS, Y.; SMYTH, M.; NORTON, B. Long term performance analysis of a grid connected photovoltaic system in Northern Ireland. Energy Conversion and Management, 2006. 19. REICH, N. H.; MUELLER, B.; ARMBRUSTER, A.; VAN SARK, W. G. J. H. M.; KIEFER, K.; REISE, C. Performance ratio revisited: is PR > 90% realistic Progress in Photovoltaics: Research and Applications, 2012. 20. RUTHER, R.; NASCIMENTO, L.; URBANETZ, J.; PFITSCHER, P.; VIANA, T. Long-term performance of the first grid-connected, building-integrated, thin-film amorphous silicon PV installation in Brazil. EUA: Proceedings of the 35th IEEE Photovoltaic Specialists Conference. New York, 2010. 21. ZOMER, C., et al. - The balance between aesthetics and performance in building-integrated photovoltaics in the tropics. Progress in Photovoltaics, 2013. 22. PERNAMBUCO GOVERNO DO ESTADO. Pernambuco promove primeiro leilão de energia solar do País, que atrai investimentos de R$ 597 milhões. December of 2013. Available in [http://www.pe.gov.br/b/7038]. 23. NOBRE, A., YE, Z., CHEETAMUN, H., REINDL, T., LUTHER, J., REISE, C. High performing PV systems for tropical regions – optimization of system performance. Oral presentation and conference proceedings at the 27th European Photovoltaic Solar Energy Conference & Exhibition (PVSEC), Frankfurt, Germany, 24-28 September 2012, p. 3763 3769, http://dx.doi.org/10.4229/27thEUPVSEC201 2-5CO.6.2

1. PEREIRA, E. B. et al. Atlas Brasileiro de Energia Solar. [S.l.], 2006. [www.ccst.inpe.br/wpcontent/themes/ccst-2.0/pdf/atlas_solarreduced.pdf]. 2. EPIA - MARKET REPORT 2013. A: European Photovoltaic Industry Association, 2013. 3. NBR 15220 - Desempenho térmico de edificações parte 3: Zoneamento bioclimático brasileiro e diretrizes construtivas para habitações unifamiliares de interesse social, 2003. 4. RÜTHER, R., 1998. Experiences and operational results of the first grid-connected, buildingintegrated, thin-film photovoltaic installation in Brazil. WCPEC2 - 2nd World Conference on Photovoltaic Solar Energy Conversion, Vienna, Austria. 5. NASCIMENTO, L. R. A avaliação de longo prazo de um sistema fotovoltaico integrado à edificação urbana e conectado à rede elétrica pública. Departamento de Engenharia Civil, Universidade Federal de Santa Catarina - UFSC, Florianópolis - SC, 2013. 6. Solarbuzz - Solar Market Research and Analysis. Posted by Michael Barker on August 1, 2014 7. CEMIG - Conta de Energia Elétrica de BrumadinhoMG, Classe Residencial Trifásica, referente a Março/2014, obtida pelo autor diretamente com o consumidor. 8. SWERA - Solar and Wind Energy Resource Assessment. [http://en.openei.org/apps/SWERA]. Acessado em 2014. 9. PEREIRA, E. B. et al. Atlas Brasileiro de Energia Solar. [S.l.], 2006. [www.ccst.inpe.br/wpcontent/themes/ccst-2.0/pdf/atlas_solarreduced.pdf] 10. KRENZINGER, A., 1998. RADIASOL – Programa para determinação da radiação solar sob superficies com diferentes ângulos de inclinação. Laboratório de Energia Solar, Universidade Federal do Rio Grande do Sul, Porto Alegre. 11. Meteonorm 7 – Edition 2012 – [www.meteonorm.com] 12. SolRad-Net, http://solrad-net.gsfc.nasa.gov/ 13.RORIZ Bioclimática, 2012. [http://www.roriz.eng.br/epw_9.html] 14. ANEEL, Banco de Informações de Geração (BIG) [www.aneel.gov.br/area.cfm?idArea=15] 15. BLASQUES, L. C. M. et al. Avaliação de desempenho do primeiro SFCR instalado em edificação comercial na cidade de Belém, estado do Pará, após os primeiros 20 meses de operação monitorada. In: V CBENS V CONGRESSO BRASILEIRO DE ENERGIA SOLAR. Anais... Recife - PE: [s.n.], 2014. 16. MARION et al. - Performance Parameters for GridConnected PV Systems. Prepared for the 31st IEEE Photovoltaics Specialists Conference and Exhibition 2005, Florida, USA. 17. DECKER, B.; JAHN, U. Performance of 170 grid connected PV plants in Northern Germanyâ€”Analysis of yields and optimization potentials. Solar Energy, 1997.

2752