[Nikopar10]

Hello everyone! I'm doing a research about metal foams usage for heatsink applications. I have the geometry which is some kind o cylinder with holes (photos)
vista desde la circunferencia.PNG

Isometrico espuma.jpg

vista desde el rectangulo.PNG

My idea is to set a heat source/heat flux on one of the circular faces and a air flux.
My question is: How can I set a heat source/heat flux through one of the circular faces? I have tried to put some kind of a plate as a source of heat, but the DesignModeler show me this error:"Operation would result in non manifold bodies"

Greetings and stay safe!

Nicolás

[vinerm]

It is not very clear whether your query is related to source setting in Fluent or model creation in DesignModeler.

[Nikopar10]

Thanks for the reply! it's my first post.
I think that the real question is Which is the best way to place a heat source. I commented about the non-manifold error to show a way i tried in case someone might suggest that same way or how to improve it in case it is right approach.

Regards,

Nicolás

[vinerm]

There are multiple ways, however, whether it is to be a heat source or heat flux depends on whether you have a cell zone, i.e., a body or only a surface. If the balls supply the heat and you generate the bodies representing balls and mesh it, then you can apply heat source within that. However, if you only have fluid domain around the balls, then you have to apply heat flux. If you face issue with creating or meshing such a foam, I'd recommend using overset mesh. In that, you can mesh the cylinder assuming that there are no balls in there. Then you mesh the balls and overlap whole of the domain with the cylinder. You can also use multiphase flow with packed-bed option or porous zone. For porous zone, you have to determine resistance coefficients, i.e., the resistance offered by the balls to the flow. For both, multiphase packed bed as well as porous zone, you will not observe any flow direction change because all the cells are open to the flow. However, the momentum will reduce as if the balls were there.

[Nikopar10]

Thanks for the meshing advice, thankfully i have already manage to produce a acceptable mesh for such a complex geometry.

I think i have not explained the situation properly. The fluid is not fluid but solid, the balls are voids within the metal cylinder. In the simulation I generate a fluid domain that will surround the metal foam, this fluid domain will have a velocity, the goal of this simulation is to disclosure if metal foam (particulary this metal foam) is efficient enough to be considered as a candidate for HPC heatsinks. I need advice to set a heaflux through the metal foam in a similar way a computer processor would do. I hope i've explained myself in a better way

[vinerm]

If this is a solid, then you can simply apply a source term in it. Go to cell zone conditions, select the cell zone representing the foam, enable source, and then apply value for energy source. Do note that the value is volumetric.

[Nikopar10]

I have tried that, but i kinda think it does not show the proper conduction that takes place in the procces.

Maybe it's not that necessary to show that procces. I will consult my guide teacher.

Thanks!

[vinerm]

Thermal conduction will take place as per its properties, i.e., density, thermal conductivity, and specific heat. Furthermore, the void also needs to be filled with fluid because fluid will pass through the void. That's the essential feature of metal foam because it provides a very high area density for heat transfer.

[Nikopar10]

Exactly, it is the main advantage of metal foams. Ansys Designmodeler do fill the voids when i create the fluid domain, later, a boolean-substract operation is needed to avoid the volume duplication.

[vinerm]

Yes, a boolean would be required to ensure that only one region occupies a space and not both.