Simulation and Estimation of Vapor-Liquid Equilibrium for Asymmetric Binary Systems (CO2-Alcohols) Using Artificial Neural Network

Since it is not always possible to carry out experiments at all desired temperatures and pressures, generally thermodynamic models based on equations of state are used for estimation of vapor-liquid equilibrium. In this work, a method using artificial neural network (ANN) was designed and applied to simulate and estimate the VLE for the binary asymmetric systems containing CO₂ and Alcohols. The vapor-liquid equilibrium data of six systems include (CO₂-methanol), (CO₂-ethanol), (CO₂-1-butanol), (CO₂-2-butanol), (CO₂-1-pentanol) and (CO₂-2-pentanol) were used to developed the ANN for simulation of VLE. The results when using a developed ANN model or other methods such as SRK equations of state with LCVM, PSRK, WS, were compared with experimental data at various temperatures and pressures. Finally, it was observed that there is more qualitative and quantitative conformity between ANN model results and experimental data. Furthermore, the developed ANN model showed more accurate estimation over wide range of experimental conditions.