Uma estratégia baseada em rede neural de base radial aplicada ao gerenciamento da produção de petróleo e gás natural

Abstract

This work consists of a Neural Network for modeling the relationship between the well head surface variables and the multiphase flows of the test separator vessels in the East Urucu Field (LUC). In practice, this relationship is obtained when each producing unit is aligned to a test separator for production estimation. Data were obtained when the well is aligned to a test separator. Those data are used for training a Neural Network of Radial Basis Function (NN - RBF). The goal is to make the NN - RBF to recognize the patterns of the wel l head variables (input RBF) and the phases flow in the separator (out RBF), through using, as the training set, the patterns obtained during the production test previously executed.