DEVELOPMENT OF NEW MODELS FOR PREDICTING CRUDE OIL VISCOSITY USING GENETIC ALGORITHM

Abstract

Accurate determination of crude oil viscosity is necessary in multidiscipline such as reservoir simulation, enhanced oil recovery, evaluation of hydrocarbon reserves and designing production equipment and pipelines. Furthermore, viscosity data are comprised in dimensionless parameters specially in calculating flow regimes, friction factors and pressure gradients in multiphase flow problems. Direct measurements are usually expensive or unavailable. Thus, empirical correlations are used for predicting crude oil viscosity. A number of researchers developed viscosity correlations that are applicable of predicting crude oil viscosity as a function of production data, and/or composition of well stream fluids accompanied with appropriate accuracy. Therefore, searching for a better accuracy approach for predicting crude oil viscosity is vital.