Hey, The dream is at hand if you have the money that is.
Here is a technical article on a 1 kw generator running off truck exhaust
http://www.hi-z.com/papers/ICT%201994%20-%20Performance%20of%201%20kW%20TE%20Gen.pdf
The generator uses seventy-two of Hi-Z’s 13 Watt bismuth- telluride thermoelectric modules for energy conversion.
Here is the company that makes'em
http://www.hi-z.com/index.php
I am sure we will advances in size and efficiency but probably not in price reduction.
miwyl
I spoke with the people at Hi-Z a while ago. The devices they sell are quite nice, but also quite expensive and it is difficult to get good efficiency out of them. They have a "quantum well" version that is due out sometime in 2010 (according to them) which will increase efficiency fourfold in the same size package as the HZ-14 device they sell. The hope is that the pricepoint will be better, too, but I suspect that is a function of the market and not a function of the costs of goods.
The current HZ-14 I think is $125, plus hardware, plus heatsinks - I think I figured it out to be around $200 per HZ-14 unit, which is 13 watts maximum efficiency. So, best case scenario is $15 a watt. Considering things usually aren't best case, I'm betting on something like $22 a watt in my equations - ouch. A 13hp Changfa 195 let's say is $800, plus an ST-10 at $500, plus $500 of components to complete - you're at $1800 for ~7kw, or $0.25 per watt with reasonable assurance of delivery. So TEGs are about 88 times more expensive than a Changfa-style generator per watt. Ouch. Of course, the TEGs will run on any heat source you can feed them - Changfas will choke a bit on coal (though I will be trying a wood gasifier/pyrolisys configuration in the next year or two, which the Changfa WILL run on.)
If the costs could be cut by 75% for TEGs, and the efficiency quadrupled, these would start to be appealing. But right now they're mostly interesting for the concept and not for the application - just too expensive and fiddly. Even if cheaper, one would have to build a very custom exhaust system with high-precision machined surfaces to accept the TEGs so that heat transfer would be efficient. Probably some sort of exhaust tube that was made of extruded titanium or highly resistant stainless on the inside with copper cladding to "funnel" the heat towards the TEGs. They have to be easy to clean (soot _greatly_ reduces heat transfer) thus the internal surfaces need to be easily and agressively washed somehow, and you'd be hoping for as much surface area as possible so lots of convolutions to capture heat... and soot.
Then, you'd ideally need a source of cold water to put on the other side of the TEG. Maybe a ground loop? The greater the heat difference between the two sides, the more energy you get - Carnot's equations. Then you need perhaps a series of TEGs, with larger, more efficient ones at the hottest point of the system to generate the real power, and then less efficient ones further up the line to power batteries or something. Oh, and then you need DC-to-DC converters because all these different TEGs in the exhaust heat stream will be putting out different voltages and amperages and you'll need to normalize them into something usable.
So this all starts to look pretty complex. Nice idea if it were really cheap, but it's not - telluride/bismuth ain't falling off trees. I have some hopes for the quantum well variants, but that's some time in the future.
JT