[jpina]

Originally Posted by vinerm

This is how convection coefficient condition is implemented in CFD tools. There is nothing wrong with this. However, they are using a different coupling described in next section and that coupling can be implemented in Fluent using C UDF.

Which section do you mean? Of the full thesis or the last paper?

At the full thesis they take into account the change of the wall temperature. For me it would be enough to consider it a fixed temperature like the first paper I linked:

http://www.google.es/url?sa=t&rct=j&...HbYVnchyWdRUFQ

How could I define the coupling where I set the h=5000 W/m^2/K of the fluid-solid and a fixed solid temperature?

Thanks a lot vinderm, I really appreciate your help.

[vinerm]

In section 3.2.2, they have mentioned about coupling with mold wall temperature. So, you need to look at, may be, immediate next section after 3.2.2

[jpina]

Originally Posted by vinerm

In section 3.2.2, they have mentioned about coupling with mold wall temperature. So, you need to look at, may be, immediate next section after 3.2.2

You are right, but as a first and easier approach, I would like to consider the wall temperature as a constant. Like the first paper of the author:

http://www.google.es/url?sa=t&rct=j&...HbYVnchyWdRUFQ

[jpina]

By the way, if I don’t want to model the solid temperature, I can make the fluid surface in Design Modeler and using “named selection” name the two edges as Wall.

Then, once in Fluent, I can set the two edges as Wall BC and define a Heat Flux.

The mould enters at 523K and the walls are at 353K (this is a temperature difference of 170K. I can set the heat flux as:

Q=-5000W/m^2/K*170K=-850000W/m^2

In a next phase, I should make an UDF for:

Q=-5000W/m^2/K*(Tfluid-Twall)

There is no need to define a solid zone in ANSYS Design and have a coupled boundary. Am I right?

[jpina]

Maybe I should use DEFINE_HEAT_TRANSFER for setting an h value?

http://www.afs.enea.it/project/neptu...udf/node39.htm

[vinerm]

As far as I know, there's no such macro.

[vinerm]

DEFINE_HEAT_FLUX is not used to define heat flux or convection coefficient, rather how heat flux and temperature field interact

[vinerm]

I was under impression that you're doing a fluid flow only simulation and not conjugate. Considering that there's only solid mold on outside, certainly there cannot be an h. It can only be temperature or heat flux. You do not need to use h. Rather do a fluid flow simulation without solid part and use a constant temperature condition on mold wall. Do not worry about h at all.

[jpina]

Dear vinerm,

how could I proceed?

Maybe should I write a UDF using define_profile for defining the q=h(Tfluid-Twall)? But if I set this in the wall heat flux BC, the Twall won’t be a constant…

[vinerm]

Just apply constant, uniform T at wall.

[jpina]

But with uniform wall temperature BC I cannot set the heat flux to my desired value because it is calculated, am I wrong?

[vinerm]

No, with uniform temperature you'll get some flux. Just compare it with expected value. Most likely it should match because you're using numbers from paper where they already did this exercise.

[jpina]

Dear vinerm,

I tried to set a fixed temperature and I looked at the heat transfer coefficient results. The h turned out to be around 80, far from the 5000 W/m^2/K.

Should I try to set the wall BC to heat transfer and develop a UDF with DEFINE_PROFILE and:

q=5000*(Tfluid-Twall)

Where Twall reads the wall temperature at each time step?

Thanks a lot!

[vinerm]

h being reported is as per velocity of polymer and its properties. By changing heat flux, you will affect only temperature. h will still remain same. Objective should be to observe heat transfer. If this is as expected, then you'd observe freezing else, either reduce T_wall or increase velocity to increase h

[jpina]

I'm sorry vinerm but I don't think that your proposal is what I want to model.

At the fluid-wall side we have:

q=h(Tfluid-Twall)

If I fix the Twall at a temperature (wonder 200K) where h=5000W/m^2/K I will be faking the results, because the calculation will then be:

q=5000*(Tfluid-200)

Instead of:

q=5000*(Tfluid-353)

Where 353K is the temperature I want to fix at the wall!

[vinerm]

If q, h, and T_wall are fixed, there's no way you can adjust T_fluid, since, that gets fixed, too. There has to be at least one degree of freedom to solve otherwise problem has enough constraints to have only one solution

[jpina]

Originally Posted by vinerm

If q, h, and T_wall are fixed, there's no way you can adjust T_fluid, since, that gets fixed, too. There has to be at least one degree of freedom to solve otherwise problem has enough constraints to have only one solution

No, no; I only want to fix h and T_wall. But I don't know how!

[jpina]

Hello,

I've found that the fluid side heat transfer coefficient for laminar flows is calculated with:



Maybe a good approach to setting the h to a desired value would be to set a fixed temperature and modify the "k" value in order to obtain the desired "h".

In fact, the paper says somehow "because of the micro-scale, we increase artificially the h value". Since seems that ANSYS Fluent doesn't allow to artificially increase the "h value" like ANSYS-CFX, I can artificially increase k value. Do you think that this is a good approach?

Thanks a lot!

[vinerm]

But that would mean changing material properties as I mentioned earlier. Even CFX does not allow that. No. CFD tool will allow fixing h on fluid-solid interface where fluid flow is being simulated. That renders the whole idea of simulation useless. Heat transfer coefficient for laminar flows is simply given by conservation of thermal energy, conduction on one side and convection on other. Fluxes must be equal for conservation.

[jpina]

So it looks like that the only way for artificially increasing the h for the fluid-solid side is increasing k.

Just one thing, what do you mean with:

"CFD tool will allow fixing h on fluid-solid interface where fluid flow is being simulated. That renders the whole idea of simulation useless."

Haven't we said that CFD tools DOESN'T ALLOW fixing h on fluid-solid interface?