Development of a methodology for estimating the daily generation of electricity from photovoltaic systems based on operating history and meteorological networks

• Autor(a) principal: Zandomenego, Raffaela
• Data de Publicação: 2022
• Outros Autores: Rampinelli, Giuliano Arns
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Revista de Engenharia Química e Química
• Texto Completo: https://periodicos.ufv.br/jcec/article/view/16265
• topic: Sistema fotovoltaico. Predição de geração de energia elétrica. Redes meteorológicas.; Photovoltaic system. Electricity generation prediction. weather networks.

There are many debates going on about energy transition and the development of new technologies, the generation of electricity from photovoltaic systems is becoming increasingly attractive and competitive, being one of the main agents of transformation for the energy transition. In this way, the prediction of electricity generation from photovoltaic systems becomes essential, as it contributes to mitigating the intermittency and uncertainty of the solar resource. Likewise, the prediction of electric power generation is important for the planning and modeling of future photovoltaic plants. In this way, the general objective of this dissertation was to develop, model and validate a methodology for predicting the daily generation of electricity from photovoltaic systems based on the operation history and meteorological networks for the horizons of 24, 48 and 72 hours. The period of analysis was 5 months, between August and November 2022. The weather forecast data were obtained from the EPAGRI platform and were divided into five forecast profiles: sunny, cloudy, cloudy + rain, rain and sun + rain. The reference system of the present study was a 17.6 kWp photovoltaic system installed on the roof of a consumer unit in the rural area of ??Tubarão (SC). To analyze and compare the performance of the methodology for predicting the generation of photovoltaic systems proposed in this dissertation, the persistence method was used as a reference model, in addition to the use of precision error indicators such as MAE, RMSE and MAPE. MAE, RMSE and MAPE values ??for the 24-hour horizon obtained the best results, with emphasis on the month of August, which presented values ??of 7.46 kWh, 10.83 kWh and 20.87% respectively. The presented methodology proved to be promising and with relevant information for further studies.