Previsão de vazões utilizando redes neurais para Sobradinho

Abstract

Artificial Neural Networks are computational models based on the nervous system of living beings, thus possessing learning ability through the search of patterns. This paper aims to use artificial neural networks to flow forecasting of the Sobradinho reservoir, located in the São Francisco River at Bahia State. Using data natural flow and climatic variables provided by the National System Operator (NSO) and International Research Institute for Climate and Society (IRI), respectively. Artificial Neural networks was developed with four inputs, two hidden layers with three neurons and output natural flow variable. Satisfactory results and good performance were obtained in the forecasts. Artificial Neural Networks use is a good tool for forecasting, thus providing valuable information for making important decisions.