Modelling of petroleum multiphase fluids in ESPs an intelligent approach

This paper proposes an artificial neural network (ANN) to estimate head of two-phase petroleum fluids in electrical submersible pumps (ESPs) as an alternative to existing widely used empirical models. Analytical models have been also developed for this purpose which are still unattractive due to their complexity, reliance on over-simplified assumptions, need to excessive extent of information or lack of accuracy. The proposed ANN is trained with the same data used in developing a number of empirical models; however, the ANN presents evidently higher accuracy in the entire operating area with a low computation time in the order of milliseconds to be run on an office personal computer.