Prediction of moisture content of natural gases using a GA-RBF model

Abstract

It is of great importance to know the phase behavior of systems containing water and hydrocarbon in petroleum industry because reservoir fluids are usually saturated with water at pressure and temperature condition of reservoir. There are various correlation/charts and thermodynamic models for prediction of water content of natural gases. The existing approaches usually do not present accurate predictions at low temperature conditions. In addition, since during the development of these correlations, there were not reported experimental data dealing with phase equilibrium in system containing water and hydrocarbons at low temperature conditions, these correlations and models need to be further verified. Present study aims to develop a method based on radial basis function networks optimized by genetic algorithm, namely GA-RBF for estimation of water content of natural gases. The model was proposed based on reported experimental data. Most of utilized experimental data were at low temperature conditions. The performance of the proposed model in prediction of experimental data was compared with literature correlations. Results show that the developed model exhibits accurate results, which are more accurate and superior to literature correlations.