Excess Free Energy-Equation Of State Mixing Rule And Combined Chemical And Phase Equilibrium

This work consists of two parts representing separate topics.In the first part, the excess free energy - equation of state (G~E—EOS)mixing rule has been reviewed. since the G~E — EOS mixing rule is capable of correlating, partial predicting and complete predicting vapor-liquid equilibrium from low to high pressures, it has been studied extensively. In the derivations of mostG~E—EOS mixing rules, the pressure limit conditions, i. e. either at infinite or zero pressure, are adopted. As a result, the third parameter is needed to compensate the postulation of infinite pressure in Wong-Sandier (WS) mixing rule.With no pressure prerequisite here, through rigorous derivations, a modified WS type — undefined pressure limit (UPL) mixing rule has beendeveloped. The UPL-WS G~E—EOS mixing rule remains the merits of the WS one, while eliminating the third parameter. The parameters in the UPL-WS mixing rule have been generalized and correlated with the dipole moments for all components. The correlation takes into account the difference between nonpolar and polar molecules.A comprehensive test on the complete capabilities using the UNIFAC method and SRK, PR EOSs, in particular, has been conducted. Compared with the WS mixing rule for 55 binary and ternary systems (3341 data points), including nonpolar-nonpolar, nonpolar-polar, polar-polar etc., the UPL-WS is , in general, superior to the WS mixing rules. In the complete predictions, the grand absolute average absolute pressure deviations for the UPL-WS mixing rule are 3.80 (SRK EOS), and 3.81 (PR EOS) per cent against that for the WS 5.61 and 5.64 per cent, respectively. Meanwhile , the grand absolute average absolute composition deviations are 1.86 (SRK EOS), and 2.04 (PR EOS) against 2.39 and 2.61 per cent, respectively. In addition, the partially predictive and correlative accuracy for the UPL-WS mixing rule is better than or comparable to the WS one. More importantly, the mixing rule proposed needs no the third parameter, and shows no divergence problem in calculations.