[alancalverd]

I have converted an old barn into a music room. Some of the instruments being almost as old and fragile as myself, I have installed an air-source heat pump to maintain a reasonably constant indoor temperature of 18°C.

The heat pump uses electricity to run a refrigerator system that extracts heat from the outside air and uses it to circulate warm water under the floor. So most of the time there is a temperature difference between indoors and outdoors - unless the outside temperature happens to be exactly 18°C, in which case it is unnecessary.

Suppose I have a huge thermocouple, with one junction inside the barn and the other outside. Let us suppose that Tin>Tout, which it will be as soon as I walk into the room (Tout is just above freezing tonight). So a current will flow through the thermocouple pair. OK, the voltage will be small but I could make a lot of junctions in series (a thermopile) to get more volts, and/or use a very thick wire and fat junctions to get a lot of current. Enough, in principle, to start the heat pump. Which will increase the temperature differential, and thus the power available to drive the heat pump.....

Never mind the engineering practicalities: I'm in physicist mode and assume that my alter ego can build the necessary inverters etc. so I can heat my barn to any temperature I like without expending any additional energy - it all comes from the fact that it starts off slightly hotter than  the air outside.

And why not?

[Bored chemist]

"Can I get infinite "green" heat from my barn?"
No.

[alancalverd]

I know that, but it's difficult to explain why.

[Bored chemist]

I know that, but it's difficult to explain why.

Not very.
Do you imagine a thermocouple (or thermopile) would be good thermal insulation?

Also, do you realise that one form of heat pump which you could have used is a Peltier cooler.
At that point you have essentially connected a bunch of Peltier elements together and expect one of them to spontaneously generate a temperature difference

[alancalverd]

But the heat pump moves more energy than it consumes   .

[Bored chemist]

But the heat pump moves more energy than it consumes   .

And the thermopile moves more heat than it produces.

[wolfekeeper]

Theoretically, if the heat pump and thermopile were perfectly reversible processes you might well be able to get back all the energy you used to pump the heat that performed the round trip. But they're not perfectly reversible devices. And further, a lot of the heat you pumped in leaks out through the walls, ceiling and floor, and so cannot be used to generate electricity anyway. So there's inevitably a gap.

[alancalverd]

Interestingly, a semiconductor thermocouple pair can have up to 30% efficiency, and a heat pump COP can exceed 4.That looks like a net gain!

What is interesting about this scenario is that it treats the external environment as both an infinite sink (which it effectively is) and an infinite source (ditto)!

Never mind insulation and stuff - I'm just looking at the physics (i.e. the world of weightless strings and spherical cows) not engineering (the world of wobbly girders and hungry humans). If you want to go in the direction of sordid practicality, remember that a good passive house or even a "passive school" in UK latitudes can be heated entirely by the people inside it, and if you keep the door closed, an igloo is adequately habitable until it melts.

[Bored chemist]

remember that a good passive house or even a "passive school" in UK latitudes can be heated entirely by the people inside it, a

Yes.
Remember that maintaining a temperature differential requires an energy input (which may be in the form of food).

[alancalverd]

Admittedly I have taken the essentials of life for granted, but the object of a heat pump is to add comfort - the barn already has a kettle and a toilet! When I stop eating food I probably won't care too much about keeping the piano warm, and eventually the other guys will realise that the bass player hasn't turned the page because he is actually dead.

Speaking of which, would you like to read my favorite viola joke?

[Bored chemist]

Speaking of which, would you like to read my favorite viola joke?

Only if it's a lot more polite than the cello joke.

[alancalverd]

The orchestra was on tour. On the second day the conductor fell ill. They held an emergency meeting.
Second viola said "I played this book a couple of years ago under John, and I have his notes, so I'm happy to conduct  until he recovers."
All agreed, and the tour continued.
The conductor recovered in time for the final concert.
Second viola took his usual seat and first viola said "Where the hell have you been for the last two weeks?"

[wolfekeeper]

Interestingly, a semiconductor thermocouple pair can have up to 30% efficiency, and a heat pump COP can exceed 4.That looks like a net gain!

Nope. Show your working.

[paul cotter]

Any thermoelectric generators I've seen are very low on efficiency and a lot of the heat passes directly from the hot side to the low side so they are only used on waste heat. I'm sure I've heard your joke before, Alancalverd.

[alancalverd]

Many thermoelectric systems are fairly small, with a short distance between hot and cold junctions.
But there is no theoretical limit to the dimensions of the wires joining them -  I'm looking right now at the temperature probe for a multimeter which uses a tiny sensing junction at the end of 2m of 16SWG wire to a thermistor-compensated reference junction.
So the feasibility of my absurd suggestion  depends only on the ratio of the thermal and electrical conductivities of the connecting wires.

The viola joke is a classic and I may have even repeated it in these forums. It was told to me by a conductor, who happily admitted that his profession is for wannabe percussionists who can't be trusted with two sticks.
A true bass story: the MD was introducing our brass band to a local journalist. "John plays the tune on lead cornet, Sally plays the counterpoint on the tenor horn, Ruth takes the trombone solos....and those blokes who look like bass players, well....they play bass."

[paul cotter]

I'm very rusty on this stuff so I pulled out my battered copy of "Rogers&Mayhew". They state that the max ideal efficiency is T1-T2/T1, same as the Carnot heat engine. So you would need a decent thermal differential to achieve any reasonable output. They also state that in the case of thermocouples one only gets a few μV/°K, way too low to be useful.

[alancalverd]

But as much current as you like because the current output depends on the junction area.

[Bored chemist]

So the feasibility of my absurd suggestion  depends only on the ratio of the thermal and electrical conductivities of the connecting wires.

Yes.
Specifically, your plan requires a thermal conductivity of zero and an electrical resistivity of zero.

You can stop now.

[Bored chemist]

They also state that in the case of thermocouples one only gets a few μV/°K, way too low to be useful.

It's a small voltage but that doesn't stop us using it.
https://support.celestialfireglass.com/faqs/thermocouples-everything-you-need-to-know-and-then-some/

[paul cotter]

Yes indeed, BC, they are extremely useful for temperature measurement. What I should have said is that thermocouple output will not be of use for Alancalverd's eco-barn.