Bubble point pressure is the most crucial Pressure-Volume-Temperature property of reservoir fluid, which plays a critical role in almost all tasks related to reservoir and production engineering. There are numerous approaches for predicting various Bubble point pressure properties, namely, empirical correlations and few computational intelligence schemes. The achievements of Neural Networks (NN), Fuzzy Logic (FL) Genetic Algorithm (GA), and Expert System (ES) alone open the door to the Hybrid Systems; a genetically optimized neural network (GA-ANN) and Neuro-Fuzzy (NF) modeling techniques to play a major role in petroleum industry.
In this paper, a novel comprehensive approach to the prediction of the bubble point pressure (Pb) using two hybrid systems (GA-ANN and NF) and Expert System is introduced. A total of about 160 data points from Middle East oil samples were used. Twenty three correlations of Pb are integrated to develop Expert System. The performance of the proposed techniques is compared against performance of the most accurate general correlations for Pb calculation. Statistical error analysis was also used to check the validation of the proposed techniques. From the results of this study, it can be pointed out that these methods are more accurate and reliable.