Desenvolvimento de modelo computacional para transposição de dados fluviométricos para aplicação no cálculo da garantia física de centrais geradoras hidrelétricas

Abstract

The present study displays the state-of-the-art methodologies for estimation of fluviometric data in ungauged catchments besides discussing the uncertainties associated with data collection and streamflow estimation methods. In order to evaluate the uncertainties achieved with this type of methods, it was proposed the implementation and test of a simplified model for streamflow data estimation in ungauged catchments. The proposed model was developed on MATLAB® platform and seeks for complete automation of the process. In order to verify the uncertainty level of the results obtained with the software, an analysis of 271 gauges distributed around the Brazilian territory was conducted, concluding that the model is applicable to 40% of the sites, presenting errors below 30%. However, for the remaining 60% of the sites the model was unable to explain the discharges with great accuracy, showing errors greater than 30%. Therefore, it is recommended an increase in the number of variables as input data, seeking to minimize errors.