CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   Solving with FLUENT in a reduced domain (https://www.cfd-online.com/Forums/fluent/147954-solving-fluent-reduced-domain.html)

manalis February 2, 2015 06:18
Solving with FLUENT in a reduced domain
 
Hello,

Does anyone know if it is possible to solve with FLUENT a part (which is changing dynamically) of your domain and not the full one? Could this be possible with a UDF? I was thinking about this option as I have melting/solidification problem that I am dealing with and it could be helpful to be able to solve only the fluid domain and a small part of the solid, thus avoiding the time-consuming solution of the full domain.

Thanks in advance.

manalis February 20, 2015 03:42
Any thought or idea is welcome!

CeesH February 20, 2015 05:10
Hi

I think you need to be a little bit more clear on what you want, but in many cases the answer is yes:

1) if the domain is periodic, you can use this and only solve one periodic section of the domain.
2) if you are interested at looking at a small part of the domain, while keeping the rest constant, just make a geometry of the specific part and impose the proper boundary conditions (if they are complex or transient, use a UDF for your boundary conditions)
3) there are options for combined RANS/LES if you mean you want to solve one part of the domain with significantly more detail than the other

But as mentioned, I'm not sure if any of these things is what you want?

manalis February 24, 2015 03:47
Quote:
Originally Posted by CeesH (Post 532624)
Hi

I think you need to be a little bit more clear on what you want, but in many cases the answer is yes:

1) if the domain is periodic, you can use this and only solve one periodic section of the domain.
2) if you are interested at looking at a small part of the domain, while keeping the rest constant, just make a geometry of the specific part and impose the proper boundary conditions (if they are complex or transient, use a UDF for your boundary conditions)
3) there are options for combined RANS/LES if you mean you want to solve one part of the domain with significantly more detail than the other

But as mentioned, I'm not sure if any of these things is what you want?
Hi,

To be more specific, I would add these: The domain isn't periodic. It is a closed circular cavity (2D), where an initially solid material is being heated and it is gradually melting. The flow is laminar, so I don't use any turbulence models.
The unusual thing in this case is that when you are using the melting/solidification model, your fluid domain changes dynamically (it has been transformed to fluid from solid) but you cannot know in advance which will be the new fluid "part". There is a criterion to figure out solid from liquid, it is a variable called "liquid fraction". So, actually my question is if it is possible through a UDF to solve every time only the domain (from previous timestep) that has the required value (in this case "liquid fraction" = 1).

I hope that I made it a bit more clear than before. Thank you very much for your help anyway.


All times are GMT -4. The time now is 06:11.