CFD Online Discussion Forums

CFD Online Discussion Forums (
-   STAR-CCM+ (
-   -   set wall temperature (

Reunion January 18, 2012 06:45

set wall temperature
Hey all,

I have created a box surrounded by air.

One side of the box I have at a set temperature (25 degrees above ambient).

When I add an emistivity to the side of the box with the set temperature or an interface to model convection from the side of the box to air at ambient, the temperature of the side of the box reduces.

How can I set the side of the box to stay at a constant temperature?

I am modelling steady state if that is any help.



willimanili January 18, 2012 10:43

I dont know why, but it isn't possible to specify the thermal specification of an interface. But the following way should work:

1. Create an interface between your hot box boundary and your surrounding air boundary.
2. Create the volume mesh
3. After your volume has been meshed delete the interface under the interfaces node.
4. Now you can set the thermal specification for both boundaries to temperature and specify the temperature value under physics value for each boundary.

I hope i could help.

Reunion January 18, 2012 11:27


Thanks for your reply. I think the only energy source option for an interface is a heat flux. But I'm not 100% sure.

Would that work? What happens when you delete the interface between the two regions?

Can't you set the two temperatures without deleting the interface?



willimanili January 18, 2012 16:58

Hi Reunion,

1. You are right, the only setting for thermal specification you can apply at an interface is the heat flux as an energy source. It would work, but do you really know which heat flux is necessary to let the temperature of the interface stay at the temperature you want to?!

By the way when you get the solution you will see that the heat flux isnt constant all over the boundary, so you will never get an realistic result if you apply an constant value for that heat flux. Maybe you can work with an field function if you know about the distribution of the real heat flux at the interface, but to determine this distribution depends on to many factors.

2. When you delete an interface after volume meshing there will be no more contact and interchange between the boudaries building the interface. Each boundary will stay for its own, as there were never been an interface.
The only reason to create an interface before is to avoid intersection or duplicate face problems at the meshing process. Especially if you created your regions with boolean operations at the parts or regions level.

3. As i know there isnt any way to specify the temperature of an interface. But maybe anyone else can tell us?!

If i have understood you right and you just want to simulate the convection between the box and the surrounding air, where the side of the box stays at an constant temperature, do the follwing:

Create at each region (box and air) an seperate boundary for the side of the box and the opposing area of air respectively. Create all the interfaces you want and especially the interface between the both boundaries created as described above. Generate your surface and volume mesh. Delete the interface where you want the temperature to stay at an constant value. Change the thermal specification to temperature and specify the value for this temperature at the physics values for each boundary (box and region). Now you can start your simulation.

Reunion January 25, 2012 12:25


Thanks for your reply.

I found out that you are unable to set the temperature of a region and have an interface. The only possible way is add a small heat flux after each iteration to try to keep the temperature constant.

I thought of another way of attempting a set temperature for the box by making one side a set temperature and putting the thermal conductivity up to a significant amount (factor of 1000). When I did this, the heat transfer from the box to the air (the interface) gave me X watts while if I get a report on the heat transfer from the air to the box (at the interface) I get -Y watts. The absolute value of Y is actually greater then X. This is strange since I would presume the heat transfer should be similar as it is the same interface but it seems like it could make a difference.

Essentially should I be taking the reading of the heat transfer at which interface? the box to the air or the air to the box?

(To note; the heat transfer from the box to the air varied an awful lot and went from 100W to 20W, while the heat transfer from the air to the box was at a near constant -35W)

madhuvc January 25, 2012 15:56

just a thought..what if you model your interface as a fully developed interface.

Reunion January 25, 2012 17:46


fully-developed interface is of use for a tube or channel fully developed flow. with the pressure jump or mass flow rate known.

I used a contact interface as it permits heat transfer between a fluid and solid region.

This is from the help file in star-ccm there is a copy of the fully-developed interface here and the contact interface description here.

Should I not be using a contact interface? Which is more appropriate?



willimanili January 29, 2012 07:25

Yes, you should use an contact interface. And that your heat transfer reports for the interface gives you different results is really strange and not explainable to me.
So a last time, do it the way i have described to have an boundary at an set temperature.
By the way if all the boundaries of your box are in contact to the air and should stay at constant temperature, then there is no need to model the solid box. Just model the air and set the temperature at the boundaries where the air should have contact to the box at the temperature you want.

All times are GMT -4. The time now is 09:21.