Estudo comparativo entre simulações e o desempenho real de um sistema de geração fotovoltaica

Abstract

Brazil has taken on a significant role in the global energy transition, driven by the substantial growth in energy generation from renewable sources, particularly solar energy. In this scenario, Distributed Micro and Mini Generation (DMMG), led by photovoltaic solar energy, has played a central role in the expansion of installed capacity in the country. This progress is supported by favorable public policies, the reduction in photovoltaic system costs, and the high availability of solar resources. In this context, analyzing the accuracy of performance predictions for photovoltaic systems is essential to optimize both return on investment and operational efficiency. This study presents a comparative analysis between simulated energy generation forecasts and the actual performance of a Grid-Connected Photovoltaic System (GCPS) in the city of Garanhuns, Pernambuco. The photovoltaic software PVsyst and PVSOL premium were used, along with meteorological data from Meteonorm versions 8.1 and 8.2, to conduct three simulations that were compared with real data collected over a one-year period through the Solar Portal monitoring platform. The methodology combined a literature review and a case study with a quantitative approach, focusing on the analysis of the percentage error between simulated and actual values, as well as evaluating the performance of the photovoltaic system using figures of merit such as capacity factor (CF), final yield (YF), and overall performance ratio (PR). The obtained data reveal that the real system's annual generation was 3872.4 kWh, while the simulations predicted results ranging from 3730.3 kWh to 4008.9 kWh, with annual percentage errors varying between -3.52% and 3.67%, all within a margin of uncertainty considered acceptable. The figures of merit confirm the operational efficiency of the real system and the accuracy of the simulations, with an annual overall performance ratio (PR) exceeding 80%. This study emphasizes the importance of selecting the appropriate meteorological database to achieve accurate energy generation estimates. Both PVsyst and PVSOL premium software proved to be effective tools; however, for this work, PV*SOL premium using the Meteonorm 8.2 database presented the lowest percentage error, closely matching the real data. It is concluded that this research contributes to electrical engineering by validating the reliability of simulation software, assisting engineers and designers in choosing tools and making decisions in photovoltaic solar generation projects