Excel library for process calc's including distillation

[cmarc]

thanks,
if I enter GPSA VLE data in data regression procedure I get a quite good agreement with calculated PR-Wilson parameters, I was not able to find a good agreement (large errors) with PR alone, I set the extended Alpha function as you suggested.

[rense]

can you share an example to plot a tpxy diagram ?

[carlo.stenali]

Rense,
there is a Excel page distributed with Properties which calculates TXY and PXY diagrams, look in the Excel directory, you should find a page named tpxy.xls.
I attach the images txy.jpg and pxy.jpg which show the diagrams for pentane-ethanol at 25 Bar and 440 K , the model is Peng Robinson-UNIQUAC with Wong Sandler mixing rules, it is one of the examples includes with Properties, see "www.prode.com" for additional information and examples.

[rense]

thanks Carlo,
the capability to print PXY , TXY diagrams from Excel is exactly what I need.
Nice.

[dbert]

someone has methyl acetate cyclohexane vle data ? I need to model a system at about 5 Bar

[carlo.stenali]

hi dbert,
I have some methyl acetate cyclohexane VLE points at 760 mmHg,

first column X(1)
second column Y(1)
third column temperature in degrees Celsius

0.033000 0.182000 74.300
0.085000 0.350000 68.400
0.142000 0.443000 64.900
0.283000 0.575000 59.700
0.313000 0.594000 59.000
0.373000 0.625000 57.900
0.478000 0.664000 56.800
0.507000 0.673000 56.700
0.616000 0.714000 56.000
0.688000 0.744000 55.800
0.722000 0.759000 55.700
0.781000 0.789000 55.500
0.835000 0.820000 55.550
0.940000 0.914000 55.800

Prode Properties is very useful for VLE/LLE data regression with nonstandard models as for example PR-WILSON or PR- NRTL (Wong-Sandler) which I suggest for this binary, see "http://www.prode.com" for additional information.

[dbert]

thank you very much Carlo,
I need also to model cyclohexane-toluene, actually I calculated the PR-Wilson parameters with the UNIFAC option included in the data regression procedure in Properties, do you think UNIFAC is reliable for this binary ?

[ariek]

such a great sharing! thank a lot

[carlo.stenali]

for cyclohexane - toluene I have (X1, Y1, T -C-)

0.000000 0.000000 110.560
0.063000 0.156000 107.040
0.142000 0.310000 102.780
0.178000 0.362000 101.250
0.290000 0.509000 96.620
0.313000 0.536000 96.000
0.393000 0.615000 93.390
0.444000 0.663000 91.810
0.534000 0.727000 89.610
0.597000 0.774000 88.050
0.754000 0.865000 85.000
0.890000 0.941000 82.470
0.938000 0.968000 81.730
1.000000 1.000000 80.750

with Properties I calculated the BIPs for PR-Wilson (Wong Sandler mixing rules) with both UNIFAC predictions and measured data then in Excel I generated the two TXY diagrams at 5 Bar, as you see in this (fortunate) case results are very similar (errors about 1-2 K)

[dbert]

thanks Carlo,
very useful data, btw I noticed that for this binary the Peng Robinson standard does a decent job and the selection of extended Alpha function can improve accuracy. I am curious if someone has compared the three parameters TWU formulation with PRSV which uses one empirical factor, Kappa, for fitting pure component vapour pressures.

[mouradbz]

Thank you

[carlo.stenali]

dbert,
I presume you refer to Stryjek and Vera correlation
alpha = ka + kb*(1-Tr^0.5)*(0.7-Tr)
ka = 0.378893 + 1.4897153*W -0.171318*W^2 + 0.0196554*W^3

generally referred as PRSV this correlation has been included in Hyprotech simulators, there is a parameter kb which is specific to the fluid of interest.
PRSV works well fro hydrocarbons and other fluids at low temperatures however , it doesn't extrapolate well at high temperatures and you should use another correlation for these cases.
Twu correlation produces equivalent (or better) results at low temperatures but works well even at high temperatures, so it should be preferred for all applications but specifically to mixtures with fluids in supercritical state.