(19 percent of Carnot efficiency, which is the standard unit of measurement for the efficiency of a heat engine).

Great way to dodge saying that the actual efficiency of the device is piss-poor.

Carnot efficiency between 100 C and 25 C is just 20%, and 19% of that is 3.8%

100 C isn't really waste heat either - it's hot enough to heat houses in a district heating system. it would be a total waste to cool it down to 25 C to get a tiny bit of electricity out.

@Eikka: In a car engine 2/3 of total energy are wasted as heat, so you can get 7.6% more power by using the refrigerator's heat. The exausts temperature is much more than 100 C, the Carnot efficiency would be higher if you can use that. In a big electric plant gigantic quantity of heat below 100 C are generated, and you can't use that heat for houses in summertime, every single point of efficiency could save millions

@Eikka: In a car engine 2/3 of total energy are wasted as heat, so you can get 7.6% more power by using the refrigerator's heat.

That sentence appears to make no sense.

In a big electric plant gigantic quantity of heat below 100 C are generated, and you can't use that heat for houses in summertime, every single point of efficiency could save millions

In a big powerplant, the waste heat generated comes out significantly below 100 C, because the lower the temperature at the condenser, the higher the efficiency of the generator turbine. The coolant water temperature is typically under 30 C as it exits the powerplant. Environmental considerations also limit cooling water temperatures to below 32 C to avoid hurting plant and animal life.

With such a small difference in temperature to ambient conditions, the size of the device needed to extract even small amounts of electricity would raise the cost to millions, and you essentially save nothing.

Suppose for example you got river water at 15 C flowing into the plant, and the outflow is at 32 C - the maximum the law allows.

The Carnot efficiency between these two heat reservoirs is already at 5.6% maximum, and 19% of that would be 1%

So 1% of the waste energy could be recovered by extracting all the heat from the cooling water stream.

Meanwhile, if we assume the steam turbine to operate between 400 C steam and 32 C at the condenser, the Carnot efficiency is 54.7%. If we drop the condenser and the cooling water temperature to 20 C by pumping more water through, then the efficiency improves to 56.5% - an improvement of 1.8% points.

The remaining 5 C temperature difference isn't large enough to run any practical thermoelectric generator.

This is why powerplant waste heat is not a great source of energy. At best it can be pumped around irrigation channels to thaw fields in the early spring and extend the growing season a bit.

Likewise, in a CHP plant, the heat for the district heating loop isn't taken out of the power turbine's condenser. Instead, the heat is taken from the steam in the primary loop, so the heat output and the electricity output can be independently adjusted and the power turbine can always run at maximum efficiency.

In a car's engine two parts out of three are wasted as heat. With a 100 KW engine you waste 200 KW of heat, with a 3.8% pyroelectric generator you can recover 7.6 KW assuming the radiator's coolant temperature of 100 C, more if you use the higher temperature exausts gases. A 7.6% efficiency improvement may not sound great, but you can drive the AC compressor and all the electrical services with it, and get rid of the alternator too

with a 3.8% pyroelectric generator you can recover 7.6 KW assuming the radiator's coolant temperature of 100 C

1) a 100 kW engine isn't running at full power all the time. The actual average output is closer to 10 - 20 kW.

2) Only a very small portion of the waste heat goes out through the radiator. Modern engines actually struggle to put out enough waste heat to run the interior heaters. The vast majority goes out the tailpipe.

BMW already tried to harness that heat using a rankine cycle turbine that extracts waste heat out of the exhaust, which is more efficient than the pyroelectric generators. The necessary heat exchangers to extract the heat out of the flow of gas were too big and too expensive.

1) If you need only 10 KW from the engine you'll get 760 W from this generator. What's the problem?
2) The more heat you can use from the exaust the better, you'll have more efficiency if the temperature is higher than 100 C. And the heat exchanger is not a problem, it's just a bigger tailpipe shaped like an intercooler