[TommyM]

Hi All,

I am modeling a component with fins generating a power of 50 W.
I am not interested in how heat flows through the solid, so I consider only the boundaries of the finned component and focus the analysis on the heat transfer on surrounding air.

I tried two different approaches (see picture attached):

Case A) Heat source is applied to all the boundaries of the finned component.

Case B) I created a Shell Region from the boundaries of the finned component and apply the heat source to these shells.

As you can see on the picture, shells allow a uniform heat transfer, while case A exhibits an abrupt peak (irrealistic) of temperature between fins.

This trend is very odd, since the two approaches should be equivalent, more or less. This test makes me think the modeling of case A is incorrect.

Do you have any suggestions?

Thank you in advance,

Tommy

[LuckyTran]

I'm guessing you divided 50W by the exposed surface area of the fins into some form of W/m^2? This is a very unrealistic heat flux profile for fins with with constant cross section. But yeah, this is not a good way to model a finned heat sink even if you don't really care about the heat conduction through the solid.



It also depends what the thickness of the shells are. Very thick shells will promote way more shell conduction. But if you're going to be using shells then you might as well mesh the solid region too. Hopefully you put in some reasonable numbers for the shell thickness.



Probably the best thing you can do without meshing the solid region is to just set the wall temperature profile for the fins (but you won't be able to match exactly 50W this way either). You can get the heat transfer coefficient that you get from your simulation to help you do this calculation. It's not a trivial calculation, but the theory and formulas required are easy to handle.

[TommyM]

Originally Posted by LuckyTran

I'm guessing you divided 50W by the exposed surface area of the fins into some form of W/m^2? This is a very unrealistic heat flux profile for fins with with constant cross section. But yeah, this is not a good way to model a finned heat sink even if you don't really care about the heat conduction through the solid.



It also depends what the thickness of the shells are. Very thick shells will promote way more shell conduction. But if you're going to be using shells then you might as well mesh the solid region too. Hopefully you put in some reasonable numbers for the shell thickness.



Probably the best thing you can do without meshing the solid region is to just set the wall temperature profile for the fins (but you won't be able to match exactly 50W this way either). You can get the heat transfer coefficient that you get from your simulation to help you do this calculation. It's not a trivial calculation, but the theory and formulas required are easy to handle.

Hi Lucky Tran,

thank you for your reply. I agree with you, it is not a good way to model finned heat sink.

The trouble is that I know the part generates 50 W but I do not know how many components it contains and how it produces heat.

Regarding my first post, as you said, shells provide more heat conduction and "smooth" the peaks of temperature caused by the fact that all the surfaces are assumed to exchange the same amount of surface heat flux (W/m^2), and this assumption is unrealistic.

You suggest to set wall temperature profile, but I do not know it. And I will obtain again a constant surface heat flux if I set the same temperature on all the walls, right?

Tommy

[LuckyTran]

Go to the fin equations and get the spatial temperature profile and apply it as a BC. It'll be 1D but it'll be way better than your constant flux or constant temperature assumption.

You could also use the same fin equations to get the local heat flux and apply that as a BC instead of a constant flux.

Originally Posted by TommyM

You suggest to set wall temperature profile, but I do not know it.

Now you know it. No more excuses.

[TommyM]

Originally Posted by LuckyTran

Now you know it. No more excuses.

I did not want to appear unfriendly, but if I did it accidentally, I am sorry.

I read the notes you posted about fins heat transfer. Unfortunately, I do not know the wall base temperature T_0 and the fluid temperature T_inf (because the case I am modeling is actually a closed cavity containing several heat source components, so T_inf is strongly affected by fins heat transfer).

Anyway, I could obtain such temperatures and match the power of 50 W by following an iterative, trial-and-error procedure. The formulas you mentioned will be useful to achieve this goal.

Thank you for your help.

Tommy

[LuckyTran]

Analytical thinking is breaking down the overwhelming complex problem into smaller problems that you can handle.


You could honestly ballpark sooo many of these numbers and still do wayy better than what you already have with a constant heat flux.


T_inf is gonna be the ambient air temperature. The base temperature is gonna be the temperature load point, which is probably around 50 °C +/- 10 °C.