Accuracy of the heatsink material ?

[Anonymous]

Hello Forum,

i tried to compare the special heatsink material to a build heatsink (basic fin heatsink). With the same material parameters, i found out that the build heatsink is about 1,3-1,5 times hotter than the heatsinkmaterial. It was tested for a natural convection setup with 3 different geometries and with laminar and turbulent. Additionally i figured out, that in some cases the laminar setup for the heatsink material diverged.

Has anyone made similar experiences ?

And what would be the more accurate ?

[Jon.Wilde]

Hi,

How have you set up the model to capture natural convection?

Was the heat sink well meshed in both examples?

We tend to use the Mixing Length model, please see the guide here: Natural Convection

Thanks,

Jon

[Anonymous]

The natural convection setup is the same i used in this post: http://forums.autodesk.com/t5/cfd-general-discussion/cfd-ignores-heat-flow/td-p/6036878

The mixing length model gives nearly the same results as the standart k-epsilon. i chose the k-epsilon, because it was said : suitable for most applications

I used the automatic meshing.

In my further researches i will not be able to use the heatsink material because of simple variantions in the heatsink geometrie. Would u say, the automatic mesh with he mixing length model is accurate enough for standard heatsink simulations ?

Thanks  for the very fast replies!

[Jon.Wilde · Accepted solution]

I will reply quickly when I can 🙂

Check out this webinar we ran during last week on meshing - I suspect you need to think about the mesh through gaps (you can sign up to the webinar series using the sticky post on this forum too). You might also like to think about a mesh sensivity study (also in the webinar).

Otherwise, yes you should be fine. Sorry for missing that you already had one of these models running previously.

Thanks,

Jon