SAULO CUSTODIO DE AQUINO FERREIRA
Data: 09/05/2024, 14:30 h
Sala Zoom: https://puc-rio.zoom.us/j/92668200136?pwd=T0trcUszNDhRVjlaczVkY05jblNGQT09
Orientadores: Fernando Luiz Cyrino Oliveira & Paula Medina Maçaira Louro
Resumo: Brazil has always been a country that had its electricity matrix based mostly on renewable sources, more specifically on water. Over the years, this has diversified and demonstrated a greater participation of the wind source. To better explore it, research aimed at modeling its behavior is essential. However, it is not always possible to have wind speed and wind generation data available in quantity and in the locations of interest. This data is essential to identify potential installation sites for wind farms, improve the performance of existing ones and encourage forecasting research and simulation of wind generation, which are inputs to help improve the performance of the planning and operation of the Brazilian electricity sector. In the absence of wind speed data, an alternative is to use data from reanalyses. They provide long historical data of climate and atmospheric variables for various parts of the globe and free of charge. Thus, the first contribution of this work focused on verifying the representativeness of wind speed data, made available by MERRA-2, in the Brazilian territory. Following the recommendations in the literature, interpolation, extrapolation and bias correction techniques were used to improve adequacy to the speeds provided by the reanalysis base, those that occur at the height of the turbine rotors of the wind farms. In a second contribution, MERRA-2 data were combined with power measured in wind farms in northeastern Brazil to model stochastically and non-parametrically the relationship between speed and power in wind turbines. For this, clustering techniques, estimation of density curves and simulation were used. Finally, in a third contribution, it is expected to improve the methodology developed in the second contribution, in order to develop a simulation methodology based on the Markov Chain Monte Carlo method, still using MERRA-2 wind speed and measured wind power.