An improved method for predicting CO2 minimum miscibility pressure based on artificial neural network
Corresponding Author and Email:Xinwei Liao, xinwei@cup.edu.cn; ORCID: https://orcid.org/0000-0002-7820-6821
Citation:Dong, P., Liao, X., Chen, Z., Chu, H. An improved method for predicting CO2 minimum miscibility pressure based on artificial neural network. Advances in Geo-Energy Research, 2019, 3(4): 355-364, doi: 10.26804/ager.2019.04.02.
Article Type:Original article
Abstract:
The CO2 enhanced oil recovery (EOR) method is widely used in actual oilfields. It is extremely important to accurately predict the CO2 minimum miscibility pressure (MMP) for CO2-EOR. At present, many studies about MMP prediction are based on empirical, experimental, or numerical simulation methods, but these methods have limitations in accuracy or computation efficiency. Therefore, more work needs to be done. In this work, with the results of the slim-tube experiment and the data expansion of the multiple mixing cell methods, an improved artificial neural network (ANN) model that predicts CO2 MMP by the full composition of the crude oil and temperature is trained. To stabilize the neural network training process, L2 regularization and Dropout are used to address the issue of over-fitting in neural networks. Predicting results show that the ANN model with Dropout possesses higher prediction accuracy and stronger generalization ability. Then, based on the validation sample evaluation, the mean absolute percentage error and R-square of the ANN model are 6.99 and 0.948, respectively. Finally, the improved ANN model is tested by six samples obtained from slim-tube experiment results. The results indicate that the improved ANN model has extremely low time cost and high accuracy to predict CO2 MMP, which is of great significance for CO2-EOR.
Keywords:Artificial neural network, L2 regularization, dropout, CO2 minimum miscibility pressure, multiple mixing cell methods.
Download this Manuscript:PDF
