Neuro-Fuzzy Approach for Gas Compressibility Factor Prediction

Good prediction of gas compressibility factor (z - factor) of natural gas is an essential key in numerous petroleum engineering calculations. Currently, many empirical correlations have been utilized to predict gas compressibility factor. However, the precision of these empirical correlations proved to be inadequate for optimal prediction due to their limitations. The wide range of applications of artificial intelligence (AI) techniques in petroleum engineering affirms the capability of AI techniques to handle such issue effectively. In this work, a published data set consisting of 6000 experimental data points of gas compressibility factors was used to develop a neuro-fuzzy model for gas deviation factor prediction. Neuro-Fuzzy technique was tested to develop a new approach. The derived Neuro-Fuzzy based model was compared to test its performance and reliability to predict gas compressibility factor. The result of this study showed the superiority of Neuro-Fuzzy models in predicting gas compressibility factor. It was observed that the proposed method provide more accurate results and more clear visualize the z-factor in 3D dimensional relative to other method used in this work.