A new optimization method for parameter determination in modeling solid solubility in supercritical CO2

Abstract

For reliable design of supercritical extraction processes, it is necessary to have a sound knowledge of the solute solubilities and their representations by mathematical models so that the solute-supercritical fluid behavior can be predicted accurately. In this work, a new optimization method, the pattern search method, was used to optimize binary interaction parameters for correlating solute solubilities in supercritical CO2 with cubic EOS models. For this purpose, experimentally measured solubility data collected for 10 binary systems were used to examine this parameter optimization method. The results show that the calculated solubility values agree very well with the corresponding experimental data. The average absolute relative deviations between calculated and experimental data are mostly found to be less than 10% and some even less than 1%. Meanwhile, a comparison between the pattern search method and genetic algorithm reflects that the former gives better correlation performance and more significant reduction of computing times than the latter.