Development of a Novel Empirical Method for Prediction of Gas Viscosity

Gas viscosity is one of the most important parameters in petroleum engineering affecting fluid flow in porous media, well, and pipelines. Therefore, it is important to use an accurate value in any ranges of operational pressure and temperature. Gas viscosity is measurable in laboratories but it is costly and time consuming. Also, in Iran, there is no apparatus to measure gas viscosity accurately in laboratories; thus engineers rely on empirical correlations to estimate gas viscosity. In this study, a novel method is used to predict hydrocarbon gas viscosity. This new gas viscosity correlation is developed using artificial neural network, statistical techniques, and a non-linear optimization. Moreover, the validation of this correlation has been approved. The results show that this model has more accuracy compared to other ones for a massive data set.