Heat Transfer Fluids

[wmconlon]

I would like to model some processes using some of the commercial Heat Transfer Fluids (HTFs) such as the Dowtherm HTFs. Does anyone have suggestions about how to build up a model fluid using publicly available that includes specific heat, density, thermal conductivity, viscosity and vapor pressure as a function of temperature?

[jasper]

All the properties you need are simply on the spec sheets of the Dowtherm fluids. Here is that for Dowtherm A.

http://msdssearch.dow.com/PublishedLite ... age=GetDoc

I presume you are not mixing the fluids, so you can treat the fluids as a pure compound with the specified data. You can enter the data in PCD man, or take a similar compound into TEA and configure the data in TEA.

Do you have any specific question about how to go about it?

[wmconlon]

Thanks.

[wmconlon]

I am still uncertain about the process for entering the tabular data into PCD Manager. Does one use it to generate a correlation and then enter the coefficients?

[hkooijman]

ChemSep/COCO doesn't have a lot of heat transfer fluids in its library. We will take this under consideration for new releases but this will take some time.

To model heat transfer of fluids, it is important that you get the right models to describe their behavior. For a DOWtherm that would be to describe its ENTHALPY properly, as that determines the heat it can transport, as well as its DENSITY and VISCOSITY as that will determine the work pumps need to do and the sizing of your heat exchangers. Therefore, the physical property temperature correlations you setup for the heating fluids should have (accurate) correlations for these! For the enthalpy to be accurate, you must look up correlations in literature, or else measured points and try to fit correlations through them. In the case of enthalpy, it is composed from contributions of the ideal gas (IG) heat capacity and the heat of vaporization. So you will need good temperature correlations for these. The IG contributions can be predicted if you know the groups, but it is better fitted to data. Please look at the manual of PCDman at http://chemsep.org/downloads/docs/ChemS ... anager.pdf and the online technical book http://chemsep.org/book/contents.html However, to understand how all these models work and what is important, you need to get a understanding in the thermodynamics of your fluids. If you don't understand, it is quicker to ask for advise of collegues/fellow students or an expert/professor. Hope this helps

H

[wmconlon]

Thank you. I asked one of the HTF vendors if they had CAPE-OPEN property files. I think it would be very useful if the vendors provided this data, although there certainly might be some proprietary data that they might choose to restrict.

For my immediate needs, I am still uncertain as to how to enter data into PCD Manager.

For my HTF, I have developed a correlation of the liquid density (kg/m3) as a function of temperature (K) of the form d= mT +b. On the temperature correlations tab, I choose Liquid Density, Equation 2, enter the temperature limits of the correlation and the two fitting parameters A,B. When I press the Apply button, an Information Window pops up saying 'Selection not supported'

Is there a cookbook that shows how to use this software for a new compound?

[hkooijman]

The apply button is for something different, namely to apply the selected Estimate from the pull down box. All you need to do is enter the Eq.no. 2 and the A, B plus the min and max temperatures. Then a plot / table should show with which you can check your entries. If all is ok, you can directly press File-SaveAs. I always suggest people to save their compounds to a different file than the original ChemSep1.pcd file to ensure integrity of that library

H

[hkooijman]

A cookbook for hypotheticals (or pseudo's) is http://www.chemsep.com/downloads/docs/C ... ticals.pdf
Documentation of the PCD manager is http://www.chemsep.com/downloads/docs/C ... anager.pdf
H

[wmconlon]

Ok. I see the problem.

I entered the coefficient but checked 'Mass Densities'. This resulted in a straight line plot but, because I had not entered the Molecular Weight, the densities were order E-35. I assumed, that the check box would apply to the coefficients, not merely the plot.

So I see that the correlation coefficients for density (and presumably other properties) must be estimated on a molar basis.

I will have to ask the vendor of this HTF if they will provide the molecular weight, since it was not provided with the thermophysical data.

--bill

[jasper]

If no properties are on a molar basis, and all properties are on a mass basis, you should not need molecular weight - except that the simulation environment may require properties on a molar basis. In this case, a dummy molecular weight (e.g. 1 kg/mol) should work. Conversion from mass to molar basis and conversion from molar to mass basis with both use the same dummy weight. When using for example 1e3 gr/mol (1 kg/mol), properties per mole equal properties per kg.