[EvanOscarSmith]

Hello all,

I'm having issues with a conjugate heat transfer problem I'm trying to run in CFX. I've modeled natural convection heat transfer through a small cascade made up of fins on a flat plate, and taken the time history at a certain point in the flat plate over a period of time (5 minutes).

I've also measured the pressure loss across the cascade model (the one with the fins) as zero. Given that the flow is so slow, it seems reasonable that there may not be a pressure loss.

I've then tried to replicate this result using a porous domain instead of modelling the fins (thus saving on mesh complexity). RMS Courant number of 1, y+<1. I've calculated the following parameters:

interfacial area density = total surface area of the fins / total volume of the fluid zone

porosity = total volume of blades / total volume of fluid zone

heat transfer coefficient = 4.36 W/m2.K (taken from the average HTC from the aforementioned fully meshed fin model)

loss model = none (since there's no pressure loss, I think I'm right that there is no requirement for a loss model).

Here's my question / problem:
Taking the time history of the plate at the same point, I get very different results for the porous and the fin model. The temperature of the point in the porous model shows a much faster heat transfer rate than the fin model (3-5 times faster).

Am I doing anything that's obviously wrong? Sometimes when you look at a problem for so long you can miss obvious things...

Is there a particular setting for the porous-solid interface model that I should be using for this type of problem? I have left it as "conservative interface flux" with heat transfer left as the default, which is "none". Do I need to use the thermal contact resistance model, for example?

Any help would be greatly appreciated!

Thanks,
Evan

[ghorrocks]

Given that the flow is so slow, it seems reasonable that there may not be a pressure loss.

No pressure loss means no flow. This sounds like there is something wrong with your simulation. So I would definitely not use your fin model as the basis of tuning your porous model as I suspect your fin model has a major error in it.

Have you read this FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F