Abstract
The generalized maximum likelihood principle and other statistical regression techniques were applied to 21 sets of ternary LLE, VLE and VLLE data for estimating the pair parameters of three local-composition models (NRTL, UNIQUAC and a modified Wilson equation) and the corresponding fitting accuracies were compared. Results show that the correlation of multicomponent fluid-phase equilibria may be greatly improved by applying the maximum likelihood principle. Unique identifiability of the pair parameters for the constituent partially miscible system was observed in ternary data reduction.
Furthermore, the reliability and limitation of cross-prediction among LLE, VLE and VLLE were investigated. The gap and interdependence between LLE and VLE are also discussed.
