model convection

[burri]

hi

i'm modeling in a FEA program the combustion chamber of a engine( piston and air and some cooling fluid inside the piston)). And for using a convection coeficient between the piston and the fluids i have to introduce a sink temperature, so i cannot use it. But there is another funcion of the program called gap conductance.
In the manual of the program the said: The conductive heat transfer between the contact surfaces is assumed to be defined by q=k (Tf - Ts) .
Do you think that i can use k as convection coefficient ?

thanks in advance

 

[Maxham1]

Hi Burri,

Newton's law of cooling is q = hA(Tf-Ts), where q is the heat rate, h is the convection coefficient, A is area, Tf is the fluid temperature and Ts is the surface temp. You probably know that already, but what I wanted to point out is that in your equation above there's no term for area. So the k term contains the area. That is, k = hA.

I don't know if that helps, but I thought I would mention it.

Dave

 

[ko99]

Not sure what you mean by "use k as convection coefficient"

However, if the gap is small enough, natural convection is effectivly prevented and you can assume heat travels through the fluid via simple conduction:

q = K A dT / gap

q = watts
K = fluid conductivity, W/m-K
A = contact area, m^2
dT = temperature drop across fluid, K
gap = meters




ko (

http://www.ecooling.biz)

 

[Maxham1]

ko99,

Excellent point you make that if the gap is small enough the heat travels by way of conduction and not natural convection. I found this out by accident and have been intrigued by it ever since. We have a design where we have about a 1/4" (6mm) gap between a heated surface and a cover plate. We initially had foam insulation between the two surfaces, thinking that would eliminate the natural convection. But one day, while testing, we removed the insulation and the temperature of the heated surface went up. The only thing I could figure is that the natural convection wasn't as important as I originally thought. I ran a Flotherm model and saw how small the natural convection really was.

Burri: My posting above applies for convection. If you can consider the problem as conduction, which I think is what you need to do due to the program's requirements, then I think the your k is related to k099's K as follows:

k = K*A/gap

(By the way, you could relate the convection equation to the conduction equation with h = k*A/gap).

Warning: Check the units of your program variables. For example, if the q you describe is heat flux (heat rate per unit area) then your k is related to K099's K as:

k = K/gap.

 

[ko99]

If I recall correctly, free convection can be neglected when the Nu < 1.

Nu = h * gap / K
h = heat transfer coefficient, W/m2-K
K = fluid conductivity, W/m-K
gap = meters

Of course Flotherm is an even better way to determine the convection contribution.

ko (

http://www.ecooling.biz)

 

[corus]

The k used in the equation refers to the gap conductivity which has typical units of W/mC. For convection therefore q = k/x(Tf-Ts) where x is the gap distance. The equivalent heat transfer coefficient, h, (W/m^2 C) for gap conductance is therefore k/x. Those who are suggesting using the area in the evaluation of h are wrong.

corus