Heat transfer question: should I consider convection/pressure? 2

[SACAN]

I have been for some time trying to muddle my way through some heat transfer calculations. I would like to know if I should be considering convection in my thinking and if pressure would play a role.

My application uses an element on one side of an aluminum platen .625in thick with a surface area of 4856sq in, it is fastened to the platen with a .125in thick alumin sheet and insulated on the sheet side with thermal conductivity of .07 W/mK. I know the aluminum back sheet has a conductivity of 237W/mK, the platen has a conductivity of roughly 205W/mK, and the material between the platens has a conductivity of roughly .14W/mK

The platen/element packages sandwhich the material being heated/cured, pressurized to 100psi and starting at roughly 70F and heating to 300F.

I have done all the calculations for these individually, but am wondering should I be taking into consideration the heat loss due to convection from the side of these element/platen assemblies. Should a pressure of 100lbs/sq in have any factor in my calculations? AND shape... I often use a rhombic shaped platen, and I know it will lose heat differently then a rectangular shape; but I have not been able to find and relatively simple calculations to determine how much more.

Until now I have been using the "standard" numbers and watt densities that have been used for the past 40 years here, they work to a great extent but I would like to understand this a little more then I do now. Any suggestions?
Thanks!

 

[EdStainless]

The only role of pressure that I can think of would be to change the thermal resistance of the interfaces. I presume that some type of heat transfer compound is used, pressure will improve the heat transfer of the interface.

If your standard values more or less work then maybe you just need to track the deviations from 'standard estimate' and come up with an empirical correction factor.

= = = = = = = = = = = = = = = = = = = =
Corrosion never sleeps, but it can be managed.

http://www.trenttube.com/Trent/tech_form.htm

 

[SACAN]

Thanks EdStainless,

Currently this is the way I engineer heating elements... by referring to my test charts and "tweaking" the different areas in the elements to try to limit the temperature variation as much as possible. I typically use a percentage factor, but this empirical % factor can not be constant for all sizes and shapes of elements. This is why I would like to take a different approach to design, namely to calaculation of required heat and watt densities required to heat and maintain temperature in the diiferent heat areas... and limit/maintain my +/- temperature variance.

 

[IRstuff]

If the amount of power you use disagrees with the amount of power transferred with the cold side equal to 70ºF, then convection must be involved.

Conversely, if the outside surfaces of the structure are signficantly hotter than ambient temperature, you have convection.

TTFN

 

[sailoday28]

How is the pressure applied?

 

[IRstuff]

Shape may or may not matter, depending on the actual area used. You will lose heat along the edges, since there is a transition back to ambient temperature at the edges. This is generally a non-trivial calculation. A finite element analysis program would handle both the edge loss and convection.

The degree of non-curing as you get closer to the edge will give you an indication of the amount of edge heat loss.


TTFN

 

[SACAN]

>Conversely, if the outside surfaces of the structure are signficantly hotter than ambient temperature, you have convection

Yes the machinery gets quite warm, I have never actually measured it but I would estimate the machinery surface after about 1 hour of operation to be 100F or better, depending on the type of insulation package installed, and or model of machine.
Thanks for your advice on this matter.. for my purposes for now I will ignore convection from the sides to try to simplify this process.

>How is the pressure applied?

Pressure is applied with an air bladder on one side of the element/platen package. There is usually an "edge iron" that runs down either side of the splice to ensure its straightness and to help maintain a uniform thickness. I know from experieince these edge irons get very hot and conduct the heat well out of the press. I currently estimate the ammount of heat lost to these and try to compensate by rasing the watt densities in the outer perimeters of the platens. I can not really incorporate the edge irons to my calculations as some customers use different material sizes,thickness' to this end.

>Shape may or may not matter, depending on the actual area used.

Typically in rhombic designs I increase significantly the watt density in these outside "long corners" as I have seen from experience that these areas tend to lose heat much faster. Finding a balance in these areas can often be tricky... thats why I want to see about taking a different approach to the calculations of how much heat is needed.
The center areas of my platens are designed at a lower watt density and can pose a problem of heating up too slowly or they continue to build heat to an extent that can sometimes be detrimental to the curing of the splice.

And... Thanks guys for your input... appreciated!