CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Heat transfer coefficient (http://www.cfd-online.com/Forums/main/4289-heat-transfer-coefficient.html)

Eugene January 14, 2002 05:18

Heat transfer coefficient
 
Hello! I'm student and have some problem... Can somebody help me: were I can find some information how calculate heat transfer coefficient from cfd data: (1) when I use wall function and boundary gradients (2) may be another methods? My problem is heat sink for cooling semiconductor elements with force convection.

Paul Ramshaw January 14, 2002 06:07

Re: Heat transfer coefficient
 
Dear Eugene,

Flotherm is the market leading CFD thermal modelling software used in the electronics industry worldwide. Please refer to http://www.flomerics.com for more information.

Michael Malin January 14, 2002 06:58

Re: Heat transfer coefficient
 
The wall-function formulae employ a local Stanton number St from which the local heat transfer coefficient (htc) can be determined from St*density*relative velocity*specific heat. The formula for St depends on the actual empirical correlation used. For an example see http://www.cham.co.uk/phoenics/d_pol...od/enc_tu8.htm ). Commericial codes usually make provision for the output of the htc, but if you are using a research code you will have to calculate it yourself.

Eugene January 14, 2002 10:01

Re: Heat transfer coefficient
 
Hello! Can you advice me how to calculate Heat transfer coefficient from Phoenics. I have STAN and SKIN for wall function 'log-law' and KE-EP turbulence model. My case is air flow in channel. For boundary cell I have: (1) Alfa1=Ro(air)*U1*Cp(air)*STAN were U1-fist point in air. (2) Nu(real)=STAN*(Tw-T1)/(Tw-T(overal))*U1/U(overall)*Re*Pr were U(overall)=Integral(Udx)/Integral(dx) T(overall)=Integral(U*T*dx)/Integral(U*dx) (3) Alfa(real)/Alfa(1)=(Tw-T1)/(Tw-T(overall))

But this algoritm is not working in conner area. And STAN have have strange profile ->1/Alfa1... Why?


Michael Malin January 14, 2002 12:48

Re: Heat transfer coefficient
 
STORE(HTCO) should elicit storage and printout of the heat transfer coefficient.

I suspect that the printed corner Stanton number refers to one of the corner faces (whichever was calculated last) and is not a weighted average of the two corner faces.

I don't understand what is meant by: "And STAN have have strange profile ->1/Alfa1...". Provide a more detailed and precise explanation.

I suggest that you submit PHOENICS questions to the PHOENICS Discussion Group rather than the Main one.



All times are GMT -4. The time now is 03:36.