Prediction of solubility of N-alkanes in supercritical CO2 using RBF-ANN and MLP-ANN

Mahdi Abdi-Khanghah, Amin Bemani, Zahra Naserzadeh, Zhien Zhang

Index: 10.1016/j.jcou.2018.03.008

Full Text: HTML

Abstract

Recently, due to declination of oil production the importance of enhancement of oil recovery becomes highlighted. CO2 injection as one of popular approaches because of economically and environmental friendly has wide applications in enhancement of oil recovery. Supercritical carbon dioxide is defined as CO2 which is placed at the pressure and temperature above the critical pressure and temperature which is used widely in petroleum industry. After CO2 injection to the reservoir, the light hydrocarbons of crude oil can be extracted by liquid CO2 and these processes are affected by different parameters such as solubility, so this study was performed to investigate solubility of alkanes in supercritical CO2. Two types of artificial neural networks, i.e., Radial Basis Function (RBF) and Networks Multi-layer Perceptron (MLP) were applied for this investigation. Results show that the MLP-ANN (artificial neural network) has better performance than RBF-ANN to predict solubility of n-alkane in supercritical carbon dioxide.