can UDF change the boundary type or turbulence model?
 

Hi,

I am trying to change the turbulence model I am using depending on the flow speed. For example, I want to use a laminar model when flow speed is 0, and a realizable k-e model when the flow speed is 4 m/s. If possible, how can we do that through UDF?

It is not possible by UDF. It might be possible by an interaction scheme-UDF, but it seems quite complex.

I can't help you, I don't know a lot about scheme.

Anyway, what you are trying, if it is possible and you can implement it, I think that I will not be numerically stable, besides it does not make sense to me.

If I where you, I would check in your ANSYS Fluent manual the chapter called "How do I choose my turbulent model", for further information.

Do you think it would be numerically unstable if I keep changing the models? But if the real condition is like a process have both larminar and turbulent flows at different time, would it be OK if I just use the turbulence model?

Doesn't the turbulent model give nearly similar results as the laminar model when you apply it to a laminar situation?

In that case I would just use the turbulent model everywhere. You will use some unnecessary computation time on calculating the (small) turbulence in the laminar zones, but I would not worry about that.

The results are different, I don't understand why. I thought in the turbulence model, when velocity is low, there are some mechanism that can depress the turbulence model. But it looks it's not like that.

The attached figure show the difference. It plots the temperature, the upper one is using a laminar model, the lower one is using a turbulence model. The velocity is 0. All the other boundary conditions are the same.

Attachment 36848

Did you try the rng k-epsilon?

I see, the correct initial value of k and e seems very important, I used 1 as the initial value before. Looks like that is not correct.:p