CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   heat transfer at walls (

ioana May 11, 2004 10:29

heat transfer at walls
Hello everybody!

I have some difficulties in modeling appropriate wall boundary condition for a room with a fire.

The walls are composed of heat conducting bricks of 0.1m thickness which have the following material properties: specific heat 840 J/kg K, thermal conductivity 0.69 W/mK and density 1600 kg/m3.

I tried to use the Heat transfer coefficient but I don't know how should I calculate the individual convection heat transfer coefficient...Could you help me?

Have a nice day,


Glenn Horrocks May 11, 2004 18:35

Re: heat transfer at walls
Hi Ioana,

The heat transfer coefficient is a function of the flowfield more than the solid properties. Let CFX work it out. It would probably be better to express the wall boundary condition as either a heat flux or a constant temperature. If these options are a bit crude, then you should consider doing a conjugate heat transfer simulation (CHT) where you model the temperature in the wall as part of the simulation.

Regards, Glenn

ioana May 13, 2004 04:23

Re: heat transfer at walls
Hi Glenn!

Yesterday I have tried to set the walls at fixed tempaerature as you said, but I don't like the results. The temperature is too low in the room ... I don't have any experimental data regarding the walls so I cannot aproximate the heat flux... Can you tell me a little bit more about this CHT simulation that you suggested. I am a beginner in cfx and I didn't understand, but I really want to give it a try.

Have a nice day, Ioana

Bart Prast May 13, 2004 09:58

Re: heat transfer at walls
What Glenn is saying (correct me when wrong), is that in CFX you can also model the heat flow inside the solid material in CFX. You do not have to assume anything then on the interface, just apply the proper material properties. i.e. you can model the brick walls then also. You can then assume the temperature on the other side of the walls to be constant

Glenn Horrocks May 13, 2004 18:25

Re: heat transfer at walls
Hi Bart, ioana,

Yes, Bart is correct. When you do any simulation with CFX you have to put a "box" around your area of interest (in your case, the air in the room), and describe the boundary conditions at the box boundary. Sometimes it is difficult to know what boundary condition to impose, as there is no simple model - for instance the wall you mention. In that case, as long as you have the computing capacity to do it the best way forward is to include the unknown boundary inside your "box" (as Bart suggests), so now your external boundary conditions are either simple enough to be described easily in CFX, or far enough away from the region of interest that it does not affect the region of interest significantly.

That is why I suggest a CHT simulation. For help on CHT simulations have a look at the tutorials, for instance the "Conjugate Heat Transfer in a Heating Coil" tutorial should help you. There is also some stuff on CHT simulations on the CFX-community website.

Regards, Glenn

Robin May 13, 2004 20:38

Re: heat transfer at walls
Hi Iona,

Although you could do a CHT calculation, the heat transfer coefficient boundary condition should suffice for what you are modelling.

Glenn: when you supply a heat transfer coefficient at a boundary, it is in reference to the heat transfer on the outside of the wall, not the fluid side.

A heat transfer coefficient is defined as:

h = q/(Tw-Te)


q is the heat flux per unit area, Tw is the wall temperature and Te is the exterior temperature.

At a minimum, h should account for the heat transfer through your wall. If we assume the wall exterior temperature to be Twe, then linear heat conduction through the wall can be calculated as:

q = k * (Tw-Te) / (dx)

Insert this into our equation for h and you get:

h = k / dx

For you wall properties, this equates to:

h = 0.69 [W/(m K)] / 0.1 [m] h = 6.9 [W/(m^2 K)]

Now your boundary condition will at least account for the conduction through the wall. You could even write the boundary condition in terms of the variables you have by first declaring them as expressions:

LIBRARY: EXPRESSIONS: k = 0.69 [W/(m K)] wall thickness = 0.1 [m] h = k / wall thickness END END

Note that if you have a variable wall thickness, you could just make "wall thickness" a function of (X,Y,Z).

If you also want to include the heat transfer coefficient from the exterior of the wall to it's surroundings, you would have to add it in series to your heat transfer equation. If external radiation is important, add that too and so on. In the end, however, you should be able to come up with an appropriate value of h.

Best regards, Robin

ioana May 17, 2004 09:13

Re: heat transfer at walls
Glenn and Bart I do thank you, you both for the help. I understood what i have to do.Thankssssss!

Have a nice day, Ioana

ioana May 17, 2004 09:26

Re: heat transfer at walls
Hi Robin,

Thanks a lot for your response. I wil try also this path. Have a nice day, Ioana

All times are GMT -4. The time now is 11:54.