Recovery rate of vapor extraction in heavy oil reservoirs - experimental, statistical, and modeling studies

The VAPor EXtraction process (also known as VAPEX) is a solvent-based enhanced oil recovery (EOR) technology that has great potential for the recovery of heavy oil and bitumen through mass transfer and gravity drainage mechanisms. In this study, laboratory tests, the multivariable regression technique, and the connectionist model optimized by a Genetic Algorithm (GA) were used to determine the oil production rate during the VAPEX process in homogeneous and fractured porous media. The smart technique and statistical models describe the VAPEX production rate in terms of three dimensionless numbers, namely the Schmidt number (Sc), the Peclet number (Pe), and a dimensionless parameter (N_S) referred to as the VAPEX number. The developed smart model was constructed based on a large number of experimental data conducted under various process conditions in both training and testing phases. A comparison of results obtained from connectionist modeling, the regressive model, and the experimental VAPEX data exhibits an average absolute error lower than 7% between the predicted and actual values. Using both experimental and modeling results, the statistical analysis suggests that the Peclet number is the most important parameter affecting the oil production rate in the VAPEX, and also the smart technique is superior to the regression model developed. This study shows the effectiveness of connectionist model in prediction of VAPEX production in the absence of sufficient laboratory and/or field data, which may lead to a proper design of heavy oil recovery schemes.