CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Cooling Simulation (

Subha March 5, 2008 05:25

Cooling Simulation
Hai all,

My model consits of a a motor housing with fins. It is cooled by a fan. I have placed the motor in a big far field domain. I have given a heat source on the inner walls and would certainly use a radiation at the far field walls. I want to do a thermal analysis of this system and don't want to give a inlet and outlet for the domain.

Could anyone suggest, what kind of Boundary Conditions should I use by considering only the rotaion of the fan?

I hope a steady state simulation would be enough.

Thank you.



andy20 March 5, 2008 07:17

Re: Cooling Simulation
Here's one idea:

Consider representing the fan as a subdomain containing a momentum source. That is a common approach for this sort of thing. Look at "general momentum source" in the CFX help.

You need to get data about the flow rate generated by the fan. Note that many fans generate not only an axial momentum, but also azimuthal momentum (known as swirl) - and that should probably be in your model. CFX allows you to define a cylindrical coordinate system centred on your fan, so that's not to bad. Oh, and don't forget the turbulence source at the fan too!

If you can get a decent amount of data about the fan - rotation speed, blade pitch, discharge coefficient etc, then that's OK. There are ways of working it all out (but they're too involved to explain on this forum).

However (as you will see from a recent question I asked here myself), if data is not available then an eductated guess about the turbulence source and swirl momentum source is often required.

Good luck, Andy

Subha March 6, 2008 05:15

Re: Cooling Simulation
Hai Andy,

Thank you the info.

In my model, the velocity profile generated by the fan is important for me. I am not interested on the rotaion of the fan. Instead of having a fan, what if I replace it with a inlet with respective cylindrical components of velocity of the fan(as you suggested may be general momentum source)?

Could you tell me how good is this approximation?



andy2O March 8, 2008 05:46

Re: Cooling Simulation
I was not suggesting you model the shape of the fan blades, or the rotation, or anything complex like that! I was suggesting you have a simple (cylindrical) subdomain in the region 'swept' by the fan, and place a stationary momentum source in that region.

I see no reason why your idea of an inlet could not also work. If you have measurements of the velocity profile generated by the fan at different radii, then your method may be better, as you have more control and you can impose this exact velocity profile at your inlet. Where will you place the outlet?

It is impossible to tell how good the two different approximations will be without knowing all the details of the fan, the geometry, the aim of the project, etc. In other words, sorry, but it is a judgement you have to make as the engineer on this task! Perhaps try and discuss it with experienced colleagues?

However, I think you could try both my momentum source idea and your inlet idea quite easily. The changes required to the CFX model should be quite small.

Good luck. Andy

All times are GMT -4. The time now is 02:32.