[Ric_]

Good afternoon,
I have set up a simple simulation for a chemical reactor producing methanol (packed bed reactor). My current simulation works fine, but, in view of more reliable results, I am trying to introduce non-constant kinematic diffusivities in the "Fluid Models" window (through the option "Transport Equation" in "Component Models"). My expressions for diffusivity depend upon temperature, pressure and composition (like most literature correlations).
When I introduce kinematic diffusivity as an expression, I get the following error message:

"Parameter 'Kinematic Diffusivity' in object '/FLOW:Flow Analysis 1/DOMAINefault Domain/FLUID MODELS/COMPONENT:Ar' has been assigned an expression that references the following unavailable variables: Molar Fraction" .

It either seems that I have recalled my molar fractions in the wrong way (but I have already done it several times for reaction kinetics and it works) or that functions depending on composition are not allowed in the context of Kinematic Diffusivity. Could anyone kindly tell me whether I am making a mistake or there is a trick to solve the issue?

I have attached the file of interest (I have introduced only one non-constant diffusivity, i.e. the one of Argon, "DiffeffAr" in the attached file).

Thanks again,
Riccardo

[ghorrocks]

I am no expert on this, but isn't normal practice to specify diffusivities and viscosities by defining the diffusivity and viscosity of the components, and then CFX uses the ideal mixture law to get get the diff/vis of the mixture? You appear to be defining the diff/vis of the mixture.

If you want CFX to use a different function than the ideal mixture law you need to define that in the mixture material.

[Opaque]

Are you modeling laminar, or turbulent flow ?

For either flow, if you have a mixture with only 2 material components, you should not have a problem or anything to be concerned about. However (a big one), if you have more than 2 material components, I do not see what influence the "molecular diffusivity" has on the solution for turbulent flow since it will be swamped by the turbulent contribution.

D_effective = D_molecular + Turbulent Viscosity / Schmidt Number

Schmidt Number ~ 1 for gases (usually @0.9)

Now, if you have a laminar flow only, then Ansys CFX does not provide an interface to input the D_ij coefficients needed to properly account for material diffusion. That is, Ansys CFX assumes a diluted mixture, so be careful.

The error you got is probably due to a limitation on what variables are allowed to be used for an expression on a specific widget.

Search in the forum on how to override the Dependency List for a give widget/parameter in Ansys CFX