Application of expert systems for accurate determination of dew-point pressure of gas condensate reservoirs

• A reliable model based on machine learning was developed for accurate prediction of dew-point pressure.
• A large database of more than 560 data points has been used to develop the dew-point pressure model.
• The reliability and accuracy of the proposed intelligent model was successfully examined.

Dew-point pressure is a parameter that has a key role in development of gas condensate reservoirs. Dropping of reservoir pressure below the dew-point pressure results in a decrease in production because of near wellbore blockage. In addition, due to separation of liquids, the produced gas has fewer valuable components. This study tries to develop a dependable method based on machine learning to adequately predict this important parameter. The intelligent system used in this work is Radial Basis Function (RBF) network that is a very flexible tool for pattern recognition. This model was developed and tested using a total set of 562 experimental data point acquired from different retrograde gas condensate fluids covering a wide range of variables. To optimize the tuning parameters of the proposed model, genetic algorithm was incorporated. This study also presents a detailed comparison between the results predicted by the proposed RBF model and those of other universal empirical correlations and intelligent systems for estimation dew-point pressure. The results showed that the presented model is superior to the pervious classic correlations and also intelligent systems.