Fluent natural convection

[luckyjones86]

Hi,
I'm dealing with an heat sink in an air enclosure. Attached to the heat sink there's a chip with an internal heat source.
I'm setting the Fluent model with the "laminar" behaviour for really low Reynolds number.
My target is understand how much is the influence of natural convection on the heat sink final temperature.
So, I'm conducting a steady state simulation with the Boussinesq Model and laminar behaviour.
It's correct?
I made the same simuation in Ansys steasy state thermal package indicating an average convection coefficient (20 W/m^2k alluminium - air). The results are really different from fluent simulation output. Why??
Thanks so much!!

 

[L_K]

What is the ambient air temperature? Is there also forced convection in
addition to natural convection?

You can output heat transfer coefficient in CFD post, probably also in Fluent.

By the way have you considered radiation in your models? Many times in natural convection its influence
cannot be neglected.

If you can post some pictures of the setup, it would help.

 

[IRstuff]

You say you are doing natural convection? 20 W/m^2-K seems too high for that.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[luckyjones86]

Dear L_K thanks for the reply. There isn't any forced convection. The heat sink is inside an air enclosure. Below i post the picture.
The ambient temperature is 26 °C and i want understand how much is the effect of a natural air convectioni for the Heat sink temperature distribution.
Clearly this effect is not present in the same simulation made in ansys workbench steady state thermal.

 

[luckyjones86]

Besidees changing the viscous model from standard k-epsilon to laminar the final temperature distribution change so much (15 °C)-
I shoud use the laminar model , right?
Thanks again

 

[L_K]

The dimensionless parameters for natural convection are Ra and Gr numbers, Re is for forced convection
(regarding the original post).

Those numbers should tell you whether the flow is laminar or turbulent.

Also mesh near the walls can affect the results. Meshing should be done taken
into consideration the turbulence model you are using.

 

[luckyjones86]

ok, so if my Ra and Gr numbers indicate a laminar flow the right viscous model is the laminar, right?

 

[IRstuff]

Natural convective heat transfer coefficient is typically under 6 W/m^2-K

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[luckyjones86]

Until now conducted the simulation with the Boussinesq Model for not transitory simulation. Now i would check the results comparing them with the same results of a transitory simulation. For the density variation which parameters i have to set?
Thanks again.