Boundary Conditions for Internal Fan in a Sealed Box

[Anonymous]

Do I need to set inlet and outlet boundary conditions for and internal fan running inside a sealed box? If so, where would I place the boundary conditions? I am following the tutorials for Electronics Cooling and the Application Best Practices - Internal Force Cooling Sealed. However the best practice does not include any information on how to set up the internal fan.

[srhusain · Accepted solution]

You should not place any flow boundary conditions on the surfaces of an internal fan model. The fan properties such as flow rates and rotation speed drive the solution.

[Anonymous]

That's what I thought. The only boundary conditions I have are for total heat generation on the silicon chips inside the enclosure, and a film coefficient on the exterior surface of the enclosure. The fan is set at a constant flow rate, but 0 rpm is set, and a slip factor of 1.

[srhusain]

If you know the fan rotation speed, you can specify that along with a slip factor in the range of 0 to 1. A typical value would be 0.6 if you do not have manufacturer data available.

Hope this helps.

[Anonymous]

Basically I'm trying to simulate placing an ATI HD 5570 video card into a sealed enclosure. I want to run one scenario with the fan running inside the sealed enclosure, and the second with the fan turned off (still in sealed enclosure). I'm not looking to evaluate the fan, but just get an idea of what the effect the fan might have. I assume running the fan will increased the overall temperature inside, but hopefully lower the highest temperatures on the gpu to give an even spread of heat.

For the forced convection, I assigned air to fixed, turned on auto forced convection, and turbulance set to epsilon-k. I've ran the model with the fan set at a constant flow rate of 29 CFM (RPM 0, Slip Factor = 1). The model seemed to look pretty good as far as the air flow pattern. My only boundary conditions are total heat generarion on the chipset and a film coefficient on the exterior surfaces of the enclosure. The temperature rise was substantially more than I would think. The insided of the enclosure was at 90C.

I am running the passive convection now, I assigned air to variable, did not enable auto force convection, and switched to laminar flow. I also built a large air volume surrounding the enclosure using the geometry tool. I suppressed the fan in the model and ran a few interations, but I didn't like the way the model looked with a void where the fan sits. So I stopped and unsupressed the fan and changed the material type to air (variable). I starting running new interations and CFD froze up. Not sure why. I think I will just make another model without the fan at all, that way the void fill will be consistant thoughout the internal volume.

Thanks for the input.