Abstract

SELLITTO, Miguel Afonso; RIBEIRO, José Luis Duarte  and  PETRILLO, Fabiano. Long-term non-linear model for the power requirement of the brazilian electricity system. Inv. Op. [online]. 2007, vol.27, n.1, pp.1-20. ISSN 0874-5161.

The formerly closed Brazilian electrical energy market has been disclosed for investors and domestic consumers, which now have choices for theirs decisions concerning energy sales and purchasing, in long-term contracts. To establish minimal planning condition, suppliers need to forecast the required load in order to project and operate power plants. With the aid of a method that integrates non-linear regression and soft-system methodology, we depict a non-linear model for the gross total electric power requirement, counting on domestic activity, population and human developing index. A goodness-of-fit better then 96% was achieved. The final analysis focused on the strong contingency period occurred in Brazil in 2001.

Keywords : electrical power; multiple nonlinear regression; multivariate data analysis; SSM; long-term load forecasting.

        · abstract in Portuguese     · text in Portuguese     · Portuguese ( pdf )