About thermoelectric generation

[luben]

Hi all, i was trying to use a thermoelectric board to generate electricity from a presistently hot surface. I wonder if there are any effective ways to boost the power generated? I tried to use a heat sink on the cool side of the board to maintain the temperature difference, but the effect is still minimal. Any idea is welcome !

i believe the board is using some Bismuth telluride semi-conducting material... i am thinking about applying pressure between the two faces to see if this help...

 

[mgb_phys]

You simply need to the maximum temperature difference across the junction.
So:
Hottest possible hot side
Coldest possible cold side
Best possible thermal connection to hot side (heat sink grease + pressure)
Best possible heat sinking on colde side. eg heat sink vanes/radiator in air or running cold water.

 

[luben]

thanks for the ideas. i am trying to use thermal grease to enhance the contacts of both sides. there is some improvements, but not by much. The maintaining of this temperature gradient is not very economic in itself. I found the heat quickly passes across the thin board (which conductivity rather like glass) to the other side and neutralize the gradient. Perhaps... i need something that can quickly take away the heat from the cold side.

About to try convection, but seems not very promising when the working fluid is air. The worst is, forced convection with air often require energy supplies... evaporative cooling may be an option here

 

[mgb_phys]

Maintining a cold sink is always a problem - cold running water is best, that's why pwer plants are near rivers/oceans.

I was using peltier stacks to cool detectors and had a similair problem. You have to keep the cold side cool but any fans would create disturbance in what I was trying to measure.
It's also very difficult to remove a small amount of heat ( a few watts ) with a heat sink since they rely on getting hot to generate convection currents to help the cooling.

We had to use recirculating water and a chiller in the end - not really practical for you.

 

[luben]

come to same conclusion here. Just tried some natural convection settings, but the results are... terrible... force convection by air (a simple CPU fan) greatly increase the efficiency of cooling (about 40% power boost), but this is not an option since we don't want to use electricity in the system. how unfortunate ~ But at least, we somehow know where the bottleneck is now.

water will be a nice coolant choice, we are just afraid of leakage problem if anything unexpected comes up during operation... perhaps we may try some heat pipes next time. It's been very helpful talking with you. Generated many ideas, many thanks.