LL equilibrium

[andrejkucma]

Hello, everybody, Hello Jasper
I solve LL equilibrium in 2 liquids flash. Beetwen H2O (cca1800 kg/h) and 1,2 EDC (8100 kg/h) (equimoar composition) does it go very well. Than I try add new component chloral (cca170 kg/h). Simulation was done, but with not realistic equilibrium, chloral went in phases 1,2 EDC:H20 (4,8:1) of chloral weightflow. In mine laboratory tests, its been 1:4,8. In real mixture chloral in water doesnt exists, so I create new compound- chloralhydrate. After create new compound simulation has been done very similar according to chloral behavior in this problem.

Can anybody help me with - how to obtain results as in my labor- and in literature too?
Thank you to all.
Andrej

[andrejkucma]

and my flowsheet

[jasper]

You are using ChemSep's internal thermo, and for some reason the new compounds shows up as n-Eicosane inside ChemSep instead of your added compound. I am getting the ChemSep authors to look at this as well.

[jasper]

Harry will look at this somewhere next week - he is away from his desk right now; in the mean time, try to add your custom made compound to ChemSep's configuration as well, and then compare the fugacities of the two product streams to check that the phase equilibrium has actually be obtained?

[andrejkucma]

Ok, I try it. I made the same problem again. and i start with only EDC and water in equimolar composition. I use EoS UNIFAC VLE,chemsep internal thermodynamic: liquid liquid gama K value and UNIFAC activity coeff. Then I look beter on physical properties and watch an error in phase density:

Light liquid: Light liquid
Mole weight (kg/kmol) 98.2100
Density (kg/m3) 1225.59
Std.density (kg/m3) 1258.58
Viscosity (N/m2.s) 6.7406E-04
Heat capacity (J/kmol/K) 129724
Thermal cond. (J/s/m/K) 0.129670

Liquid: Feed Liquid
Mole weight (kg/kmol) 58.4870 18.1628
Density (kg/m3) 1345.06 1263.32
Std.density (kg/m3) 1210.13 999.058
Viscosity (N/m2.s) 6.7714E-04 6.9431E-04
Heat capacity (J/kmol/K) 102889 75648.1
Thermal cond. (J/s/m/K) 0.203702 0.609538
Surface tension (N/m) * 0.0186595
Upthere is high difference of densities against to the Std. density. can it be any form of problems, that chloral/chloralhydrate has reverse ratio of divided coeff beetwen light liquid/liquid ?
BR
A.

[jasper]

From what I could see in your previous document a completely different compound was used. It is hard to say something about your current setup without being able to inspect it.

But what I can say is that there is no kind of dividing coefficient involved; you were using an activity model, and the phase equilibrium is calculated such that the fugacities in both phases are equal. For an activity model and two liquid phases, this works out to x * gamma for each compound in the two phases being equal at equilibrium. If this condition is not satisfied, the phase equilibrium has not actually been obtained.

[andrejkucma]

Please, excuse me, I forgot add a file.Here the FSD with water and edc.
Exists any other method in COFE how to obtain LL equilibrium as using UNIFAC LL aktivity model?

[jasper]

You have selected EOS for the liquid density model in ChemSep. This is not the best choice. You may try Rackett, COSTALD, or even ideal (Amagat) for the liquid density.

What bothers me though is that TEA reports the fugacities of streams 5 and 6 to differ, with UNIFAC. ChemSep uses also UNIFAC. The fugacities should be consistent. This is being looked into.

TEA does not support calculation of L-L equilibria at this point. However, as COCO uses CAPE-OPEN thermodynamics, you can use any external package that does do LL equilibria. Unfortunately ChemSep thermo also does not yet provide LL equilibria (yet).

[jasper]

in CHEMSEP if you use UNIFAC to calculate the acitivties, UNIFAC-LLE is selected for LLE. However, TEA is setup up to use UNIFAC-VLE. If you go into the TEA configuration and change the activity method to UNIFAC-LLE you will find that streams 5 and 6 are indeed in equilibrium (equal fugacities).