Accurate prediction of CO2 solubility in eutectic mixture of levulinic acid (or furfuryl alcohol) and choline chloride

This research presents the application of two predictive models named adaptive neuro fuzzy inference system optimized by combination of hybrid and particle swarm optimization methods (CHPSO-ANFIS) and gene expression programming (GEP) for estimation of CO2 solubility in a deep eutectic solvent based on mixtures of levulinic acid or furfuryl alcohol and choline chloride. The input parameters of both models were temperature, pressure, and ratio of mole of levulinic acid and furfuryl alcohol to mole of choline chloride. The output parameter of model was the solubility of CO2. Results demonstrate that the predictions of developed models are in acceptable agreement with experimental data. However, the CHPSO-ANFIS model provides more accurate results compared to GEP model. The overall R2 and AARD% values of proposed CHPSO-ANFIS model were 0.999 and 1.07 and for GEP model were 0.995 and 3.63 respectively. The developed model and correlation are effective in providing quick and accurate predictions of CO2 solubility without conducting any time consuming and difficult experimental measurements.