[Anne-Sophie]

Hi all,

I am looking for a way to implement the mass diffusion coefficient for a solid in Fluent.
In my case, I am modelling a stack of potatoes as a porous medium (fluid= mixture of air and water, solid=potatoes) and I want to define the diffusivity of water in the potatoes.

Does anyone know where to define this property?

Thanks in advance for your advice!

[vinerm]

Are you modeling potatoes as real solid or as fluid? For a real solid, one option would be to use UDS. Another one is too use energy equation to represent the diffusion of water, i.e., assume temperature to represent water concentration in potato.

[Anne-Sophie]

Thank you for the tips, Vinerm! I am new to the concept of user defined scalars (actually, I am pretty new to cfd in general). Do I understand it correctly that this uds will represent the water content in the potato and will be distinct from the water in the air? I mean: in this way there is no link between the water in the potato and the water in the air?

[vinerm]

You can have scalars that are different in both or a scalar that is continuous. It depends on the user.

[Anne-Sophie]

So if I want the moisture in the air to be coupled with the moisture in the potato, I should not use species transport with a mixture of air and water vapor, but I should use only air and a uds, to represent the moisture, that is solved in all zones? Is that correct?

[vinerm]

If you want to use species transport, then you have to model potato as a fluid zone. A species cannot enter a solid by the definition of solid in Fluent. Usually, the former method is preferred, i.e., if someone wants to study diffusion of moisture or some other species in a solid, then the solid is represented as a porous zone. Specie can diffuse as well as get convected in a porous zone. If you want to model potato as a solid, then UDS is a good option and you do not need species transport. Do note that by default diffusion coefficient of UDS is assumed as laminar; without the effects of turbulence. As far as solid is concerned, this should not worry you though.

[Anne-Sophie]

Ok, I will try that out. Since I want to model the entire pile of potatoes as a porous zone, I think the UDS is my best option, since a porous zone inside a porous zone is probably not an option.



Thank you for al the information and advice, Vinerm!


Kind regards,

Anne-Sophie

[vinerm]

Actually, porous inside porous inside porous is always there. If same values are used for resistance coefficients, then it behaves like one porous zone. However, if you specify resistance coefficients as functions of space, then you can develop any kind of porous zone.

If we look at the real physics, then it is actually capillary phenomenon that makes the moisture flow in potato and not diffusion. However, the effects can be modeled using diffusion.

[elwinxavier]

Originally Posted by vinerm

Actually, porous inside porous inside porous is always there. If same values are used for resistance coefficients, then it behaves like one porous zone. However, if you specify resistance coefficients as functions of space, then you can develop any kind of porous zone.

If we look at the real physics, then it is actually capillary phenomenon that makes the moisture flow in potato and not diffusion. However, the effects can be modeled using diffusion.

Hi Vinerm, I'm also doing moisture diffusion (moisture diffuse into electronics packaging, like PCB), I've set the geometry as Fluid, and use UDF on diffusivity, but the moisture get get fully absorbed in <3 second as compared to hours (it suppose to happen gradually in hours)
I've turn on Species Transport, Use Porosity, use Viscous Resistance

What could be the reason for this? Do I need to use VOF, Eulerian Model?