Date of Award
Master of Science in Chemical Engineering - (M.S.)
Chemical Engineering and Chemistry
Dimitrios P. Tassios
Mahmoud F. Abd-El-Bary
The NRTL, LEMF, and UNIQUAC models are used in the prediction and correlation of liquid-liquid equilibrium data for ten type I systems. The results indicate that best predictions are obtained with the LEMF equation, with an overall average absolute percentage error in distribution coefficients of 19.5 as compared to 23.0 and 24.7 with the NRTL and UNIQUAC equations, respectively. Good correlation of liquid-liquid equilibrium data alone is obtained with all three models; the LEMF equation is slightly better with an overall average absolute percentage error in distribution coefficients of 5.5 as compared to 6.7 and 5.9 for the NRTL and UNIQUAC equations, respectively. The same applies for the simultaneous correlation of liquid-liquid and binary vapor-liquid equilibrium data, where the combined errors for the distribution and activity coefficients are: LEMF, 6.5; NRTL, 8.8; and UNIQUAC, 8.2. In all cases best results are obtained for systems having large mutual solubilities, and become progressively worse as the mutual solubility decreases.
Simonetty, Jose Jr., "Prediction and correlation of liquid-liquid equilibria" (1989). Theses. 2050.