A influência da resolução espacial em modelos de previsão em mesoescala na programação de despacho da produção de usinas eólicas

Abstract

The use of wind energy for electricity production has increased worldwide. According to the PDE 2022 (Ten Year Plan for Energy Expansion) provided by EPE (Energy Research Company) installed capacity of wind power stemmed in 2013 was 2,2 GW, about 1.8% of the Brazilian. The energy matrix, with perspective increase to 22.4 GW, which in 2023 represent about 11.8% of the Brazilian energy matrix. Thus the need to have a schedule of availability of power wind farms is extremely important in the operation of the Brazilian electrical system. Using the wind speed data generated by the WRF model with spatial resolution of 15 km and 5 km 100 m long made up an estimated power generation for wind complex of Cerro Chato and Osorio from 9 August 2014 to 29 August 2014. Then compared with the daily schedule of data for energy production made by plants, by ETA model and the actual generation of the same period, provided by the National System Operator in its weekly newsletters. During the study period of this study, the absolute percentage mean errors of prediction in wind generation complex of Cerro Chato were 18.42% of the Plant PDP, 15.97% of the ETA model, 20.96% and 20.30% of the WRF model with spatial resolution of 15 km disregarding and considering the influence of temperature variation, respectively, and 7.55% and 6.89% of the WRF model with spatial resolution of 5 km disregarding and considering the influence of temperature variation, respectively. The absolute percentage mean errors schedules of the wind farm of Osório power generation were 16.29% of the plant's PDP, 19.19% of the ETA model, 9.38% and 8.78% of the WRF model with spatial resolution 15 km disregarding and considering the influence of temperature variation, respectively, and 5.94% and 5.09% of the WRF model with spatial resolution of 5 km disregarding and considering the influence of temperature variation, respectively.