CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Cell/nodes needed for CHT (Urgent) (http://www.cfd-online.com/Forums/fluent/63956-cell-nodes-needed-cht-urgent.html)

 tengra April 24, 2009 11:56

Cell/nodes needed for CHT (Urgent)

Guys:
I have to conduct conjugate heat transfer problem of a system containing some cast iron parts. The thickness is not constant which forbids me to use FLUENT wall thickness feature. So I have to mesh the whole solid domain. My question is how mnay cells (at least) I need to put accross the thickness of the solid material? Thank you for your help.
MB

 Ralf Schmidt April 24, 2009 12:21

Hi!

this is very difficult to say without knowing the geometry and the BC...

But, that are definitely less than you would need for a fluid domain (If temperature gradients are not so big)

Ralf

 vishyaroon April 24, 2009 12:38

I agree with Ralf. If the temperature gradients are not big something like 3 - 4 cells should also be good. Also since you'll see just conduction in the solid the temperature variation it is going to be linear.

Meanwhile here in my company CFD is defined as "Color For Dollars"

 tengra April 24, 2009 12:40

Thanks Ralf for ur time.
So here is how the geometry would look like (in word):
-hot gas (source of energy in the problem) is flowing through a steel pipe
-the pipe is inserted inside an irregular shaped cast part (iron)
-there is another part (aluminum) which is wrapping the the cast part
-so some portion of the cast part will be exposed to atmosphere, some portion will have coupled wall with aluminum part
-the external surface of the aluminum will be exposed to atm.
-so the heat will transfer from hot gas to cast iron part and then to the aluminum part and then to the ambient
-BC: there will be convection type BC on all the exposed surfaces
I hope you will able to visualize the problem now.
Thanks.
MB

 tengra April 24, 2009 12:43

Quote:
 Originally Posted by vishyaroon (Post 214093) I agree with Ralph. If the temperature gradients are not big something like 3 - 4 cells should also be good. Also since it is conduction the temperature variation it is going to be linear Meanwhile here's in my company CFD is defined as "Color For Dollars"
What does that mean? No value from CFD? Just curious.
Thanks

 vishyaroon April 24, 2009 13:57

Quote:
 Originally Posted by tengra (Post 214095) What does that mean? No value from CFD? Just curious. Thanks
Well on the other hand you get more color (information) for the dollars you invest

 Ralf Schmidt April 24, 2009 16:22

coupled wall with aluminum part

Hi,

take care, that the heat transfer between the aluminum and the iron is calculated correctly. If both parts are just "put" together, there might be no perfect contact and you end up having a finite heat transfer between both (in contrast to an infinite heat transfer - that is conduction...) . That contact situation is difficult to describe and as far as i know, there are no reliable methods to calculate the heat transfer.

For the number of cells, I agree with vishyaroon. But, if you have enough computational power, take more cells, about 6 or 7. As i said, that depends on temperature difference and thickness of the solid shell.

and... colours for dollars... thats a good one!

Ralf

 All times are GMT -4. The time now is 17:28.