Adequate Knowledge of reservoir fluid characteristics (e.g., bubble point pressure) plays a crucial role while conducting modeling/simulation of production processes in petroleum reservoirs. Although many efforts have been made to obtain proper correlations for prediction of bubble point pressure (BPP) of reservoir fluids, there is still relatively high magnitude of error with the developed predictive tools available in the literature. To fill this lacuna, a robust and effective technique, called gene expression programming (GEP), is employed to determine BPP of crude oil samples as a function of temperature, oil composition, molecular weight of C7 þ , and specific gravity of C7 þ . The GEP method is built based on the experimental (or real) data used for training and testing phases in order to develop an appropriate correlation. The previous predictive methods are also reported in this study and employed to calculate BPP as a function of independent parameters when the same data bank is utilized. Comparing the outputs obtained from the previous models with the BPP values predicted by the GEP technique, it was found that the GEP approach exhibits higher accuracy and lower uncertainty on the basis of statistical analysis in terms of coefficient of determination (R2) and mean squared error (MSE). Great precision attained in this study through using GEP recommends linking reservoir simulator packages with the GEP tool when thermodynamic properties such as BPP are required for modeling and optimization purposes.
کلید واژگان :experimental study bubble point pressure gene expression programming crude oil predictive tools
ارزش ریالی : 600000 ریال
با پرداخت الکترونیک