Fig. 4.6 Binary data fit for the six binary subsystems of the system methanol + acetic acid + methyl acetate + water at 1 bar with two variants of the UNIQUAC model. (McCabe Thiele Diagrams for p — 1 bar in mole fractions of the component appearing first in the system title)

with three binary parameters allowing to take into account more flexibly influences of the temperature. From Fig. 4.6 it can be seen that these two variants of UNIQUAC yield comparable correlations of the available experimental data for all six binary subsystems of the quaternary mixture methanol + acetic acid + methyl acetate + water. The predictions of ternary systems, however, appreciably differ. This can be seen from Fig. 4.7, in which distillation lines for the four ternary subsystems of the studied quaternary mixture are shown. Whereas for the three water-containing ternary systems, very similar predictions of the distillation lines are obtained with both models, the predictions for the system methanol + methyl acetate + acetic acid differ unexpectedly strong. These differences also lead to differing predictions of the RD process itself, as can be seen from Fig. 4.8. From Fig. 4.6 alone, one would not have expected such differences. Taking into account that in the example discussed here the data base for the binary data fit was identical for both models and that only different variants of the same thermodynamic model were used, clearly highlights the high sensitivity of RD process models to the thermodynamic fluid-properties models they rely on.

Fig. 4.7 Predictions of ternary distillation lines at p = 1 bar based on the binary data fit shown in Fig. 4.6. (Gibbs triangular diagrams in mole fractions)

acetic acid mole fraction (liq.)

water mole fraction (liq.)

temperature / C

0 0

Post a comment