Planejamento de mercado de energia elétrica no novo modelo do Setor Elétrico Brasileiro pela aplicação de análise de componente principal como método de combinação de previsões

Abstract

This work describes the application of Principal Component Analysis as an alternative technique of Forecast Combination in the market planning of a power distribution company. The Forecast Combination method is proposed as a solution to the decision problem faced by the business managers in choosing among several forecasting scenarios proposed by analysts in the preparation of the ten-year plan of EnerSul Inc. The residential, commercial, and industrial consumption classes were chosen to be studied because they compose, together, 74% of EnerSul market share, and so, they mean the principal adjustment challenge. The option made by the application of the present technique shows to be the most useful option, among all the predictions, to hammer out the baseline that will be applied for the power purchase agreements. Besides, it has the better performance in error evaluation, the technique previously ensures a scenario able to perform acceptable errors because it gathers the features of all the other predictions made.