[Alkain]

Hey ,
I have a problem with the simulation radiators in the room. I have to do a comparison of a panel radiator and a radiator that looks like Zehnder Charleston.

In general I have to investigate which one of them transmits more heat through radiation.

So I just have to make a water-air heat exchanger.

The panel radiator should exchange more heat through convection, while the Zehnder less.
This will cause that the panel will be cooler and consequently should transmit less heat through radiation. Emissivity of radiators will be constant.

But I have some problems.
1. How to simulate convection problem in the room ?
2. Which tubulent model to use? K-e will be good?
3. What radiation model? S2S? I have to investigate just how much radiation has taken the wall opposite the heater.
4. Any other tips?


Sorry for my English.
Regards

[Mr.Tambourine]

I am far to be an expert, but i will try to help with the few i can... Take my remarks as suggestions and hints more than anything else !

1. For convection problem, you have to setup the operating conditions so that you enable gravity (check the box and enter -9.81 in whichever axis is the vertical direction for you). You should also set the air to be an incompressible-ideal-gas (in the material panel, density section). Should do the trick.

2. I am not very good in turbulence, but i would say that it is not a key point at all in your problem (i might be speaking too lightly though). Therefore, i would take a common model like k-e and not ask myself too many questions about it

3. S2S seems to be enough for your problem. As far as :
- the air won't be considered as a participative media,
- your surfaces will be considered grey,
- your reflection is diffuse
I think you don't have to go for more sophisticated models like DO. S2S is simple, efficient, and should do the job in your case

4. I have never investigated the heat exchanger model of fluent yet. I don't know if it is fit to your problem. It may very well be, but as to me i would have probably chosen another way to do the model (depending of the data you have avilable). Again, that might be very wrong from me !