Library "eos" for thermophysical properties of fluids

[Antech]

Hello.

I performed the basic CodeSaturne testing to learn the program and confirm it's applicability for our needs (industrial aerodynamic tasks). Now I'd like to add a library "eos" (listed in User's guide) to calculate thermophysical properties of fluids depending on pressure and (mainly) temperature. But I cannot find any information about "eos" in the manual or on Internet. Where can I find this library and is it free for commersial use or is it proprietary?

Thanks.

[Yvan Fournier]

Hello,

Unfortunately, the EOS library (used mostly in the context of the NEPTUNE_CFD code) is not Open Source.

Code_Saturne can use the freesteam (http://freesteam.sourceforge.net/) library, and we will add support for the CoolProp (http://www.coolprop.org/) library soon (actually, detecting/linking with is is available in the trunk, but using its calls is not, so the automation is minimalist at this point).

Regards,

Yvan

[Antech]

Yvan Fournier
Hello. Sorry for late reply, was busy on monday morning and forget about the topic...
Thanks for the info about fluid properties routines. If I get it right, the most promising option is a CoolProp intrgration in new Saturne versions. I've browsed CoolProp website and it seems to be capable enough for our purposes (we need mostly gaseous mixtures containing H2O, CO2, N2 and O2 for aerodynamic modeling).

At the moment one should use separate formulas to express properties of fluids in Code Saturne. Is it possible to compose some kind of database for particular species (H2O, CO2, etc) and use a fixed composition mixture of this species in Saturne? (With automatic calculation of properties based on P and T.) Or it anyway requires something like CoolProp to calculate mixture properties?

Sorry for offtopic, it's just connected with the EOS theme...

BTW, Viva, Saturne I composed a test case with:
=> Aerodynamic resistance (head loss model)
=> Heat exchange with the wall
=> Distributed heat source
=> Irregular inlet velocity profile
=> SST Turbulence model
=> Near-wall mesh layers (for Y+ less than 1.0)
I chose Steady with SIMPLE scheme. And the only option I should "tweak" was gradient calculation (Least squares method over partial extended cell neighborhood instead of default). Calculation was successful, I checked results and found no problems comparing with k-epsilon variant (without prism layers), although results was not the same because of mesh refinement in near-wall region.

[Yvan Fournier]

Hello,

What I would recommend is to use CoolProp it to compute points in the database you need outside Code_Saturne, build a tabulation (or fit polynomials if the properties are regular enough),
and use that through user subroutines (or even mathematical expressions defined through the GUI in cas of polynomials). As CoolProp calls might be slightly expensive (each call must parse character strings, for a start), the planned integration is to use it to build arrays of values in a given range, then use tabulations (we could make that transparent by extending the tabulation range "on demand"). Things might evolve when the implementation is really done, as we will look more into details then.

Of course, we have no prior experience with CoolProp, so we are very interested in your feedback, but can offer no guarantees (though the project seems quite active, so I am optimistic).

Regards,

Yvan

[Antech]

Hello, Thanks for your advise. For now I'll use polynoms (built-in expressions in Saturne) based on thermophysical properties calculated with other programs (it's not a problem because we use these properties extensively in other calculations). Hope that Saturne will incorporate CoolProp in future.