Page 1 of 1

property package for condensable polar gases

PostPosted: 19 August 2020, 15:03
by Stephan

I want to model heating and cooling of methanol and water. Temperatures are between 20-240°C, pressures between 2 bar and 20 bar abs.
Cofe/Coco is used for this.

When usinig the TEA property package I obtain significant enthalpy deviations within liquid and vapor phase compared to
enthalpy values from tables such as within VDI heat atlas.
I tried all enthalypy calculation types within TEA and found that the ideal-enthalpy setting produces most accurate enthalpy values.
However, I would like to have it more accurate with deviations less than 10%.
The water property package does work very well for water enthalpies, however, no other compounds and no methanol can be calculated within the same stream when using that package.

Is there another model/option available that is more suitable for the compounds and conditions mentioned above?
Eventually from the Chemsep property package?

Re: property package for condensable polar gases

PostPosted: 20 August 2020, 09:47
by jasper
I would expect any of the activity models to do a reasonable job at moderate conditions.

At 20 bar, you'd probably need an equation of state. But then you should put a binary interaction parameter in for the water-methanol pair.

Re: property package for condensable polar gases

PostPosted: 01 October 2020, 11:11
by Stephan
We tested meanwhile a couple of models in a mixture of H2O, CH3OH, air components and H2.
We used Chmsep property package and testet within Cofe.

1) We used Prausnitz for k-value, Hayden O'Connel as EOS and NRTL for activity coefficient.
Enthalpy values were calculated via Excess enthalpy. That model setup up gave quite acurate enthalpy values for water and methanol, both liquid and gaseous,
for pressures bewtween 2-30 bar.
However, VLE-equilibria methanol/water were not so accurate.
We didn't test gases with that model yet.

2) We tested EOS for K-value, Predictive SRK as EOS.
Here we got very accurate VLE-quilibria, but not so accurate enthalpy values.

Is there any suggestion how to get both enthalpy and VLE calculations close to empirical values for the given system?

Re: property package for condensable polar gases

PostPosted: 01 October 2020, 15:37
by jasper
Quoting Prof. Ross Taylor:
The test is not precisely described. Were all those compounds in the same test? Part of the message suggests yes, part of the message suggests not.

In any case, this is an odd choice of model for VLE of methanol water given that any NRTL BIPs for that pair come from DECHEMA and they do not use HOC for the vapor phase. Also, the Prausnitz K-value model was created to use UNIQUAC. In principle, you can do what you are trying, but all of the BIPs would need to be refit.

If that combination and excess enthalpy gives reasonable enthalpy then I would expect a DECHEMA style model to do the same. I am quire sure that can give good VLE for MeOH water although not perhaps at high pressure. For that mixture at high pressure then something like PSRK is your only choice. (High pressure is not defined here so it it hard to be definitive.

If the light gases are in the same system as MeOH and Water then you will need to find Henry coefficients and model them that way. ChemSep does not come with a datbank of Henry coefficients.

Re: property package for condensable polar gases

PostPosted: 05 October 2020, 13:09
by Stephan
We did two kind of tests.

pure compound tests for enthalpy (for all compounds of concern)
binary test with methanol and water for VLE (mole fractions of water an methanol in liquid and vapor phase for given temperature and pressure).

Solution of gases within water and methanol is not of concern for our calculations and therefore has not been tested.

What does "DECHEMA style model" mean? Which setting for K-value, activity coefficient and EOS?

Re: property package for condensable polar gases

PostPosted: 06 October 2020, 14:03
by hkooijman
Why don't you share the collected validation data on the forum? Then there is more information to help you.