Previsão de vazão utilizando o modelo autorregressivo integrado média móvel: reservatório de itutinga/mg

Abstract: The hydrologic modeling is an important tool for planning and management of river basins water resources, mainly when this tool intended to provide hydroelectric power stations. In this work the hydrological stochastic model was used, ARMA (p,q), to adjust the Itutinga reservoir annual average flows in the state of Minas Gerais/Brazil which, together with the Camargos reservoir contributes to the Furnas hydroelectric power station that is located in the municipality of Lavras/MG/Brazil. The objective of this study was to simulate the reservoir annual month average flow using flow data from the ONS database (02144067, Agência Nacional de Águas-ANA), for the historical series of 1931-2010, and there foreto verify the existence of similarity between the observed data and the ones simulated-adjusted by the model. The stochastic model ARMA (p,q), has shown its capacity to capture the time series dynamics, presenting satisfying results on the simulation of annual average flows to the observed data in the Itutinga reservoir.