Estimation of natural gas compressibility factors using artificial neural network approach

Prediction of compressibility factor of natural gas is an important key in many gas and petroleum engineering calculations. In this study compressibility factors of different compositions of natural gas are modeled by using an artificial neural network (ANN) based on back-propagation method. A reliable database including more than 5500 experimental data of compressibility factors is used for testing and training of ANN. The designed neural network can predict the natural gas compressibility factors using pseudo-reduced pressure and pseudo reduced temperature with average absolute relative deviation percent of 0.593. The accuracy of designed ANN has been compared to the mostly used empirical models as well as equations of state of Peng–Robinson and statistical association fluid theory. The comparison indicates that the proposed method provide more accurate results relative to other methods used in this work.

Figure optionsDownload high-quality image (111 K)Download as PowerPoint slideHighlights
► Natural gas z-factors are modeled by using a designed ANN.
► We compared the designed ANN model to the mostly used empirical models as well as EOSs.
► The designed ANN model can predict z-factors of natural gas mixtures precisely.