Very high temperature heat pumps

[Cyril R]

How about aluminium chloride for the working fluid?

[E Ireland]

Problem is that anhydrous aluminium chloride reacts non reversibly with water. How about elemental iodine?

[KitemanSA]

Folks, the working fluid does not have to condense. Think Brayton vs Rankine cycle. Indeed, an acoustic heat pump might work famously.

[E Ireland]

Having stumbled across a discussion of transcritical carbon dioxide heat pumps on the internet in the context of domestic heat pump equipment I decided to see what would happen if a transcritical steam cycle was used.

We assume 285°C steam at ~70 bar. It has enthalpy of ~2900kJ/kg.
Compression of the steam to ~580 bar yields a supercritical water fluid of ~904K. This process consumes ~530kJ/kg, so ~590kJ/kg at ~90% compressor efficiency.
580 bar, 630°C steam has an enthalpy about ~3400kJ/kg.

This is fed to the gas cooler which I assume will be some for of printed circuit heat exchanger given the pressures and temperatures involved.
Constant pressure heat capacity of water is about 2kJ/kg.K
Normal design practice is to get the biggest possible delta-T on the high temperature leg, so we would seperate out heat for several temperature ranges.

The top range will be 500°C to 630°C, which would be suitable for various petrochemical and similar processes, enthalpy would drop to 2600kJ/kg at 500°C - which results in the extraction of 800kJ/kg of heat to the heat exchanger.

The middle range will be 400°C to 500°C, I envisage this being used for heating of TOSCOAL type retorting processes and similar pieces of equipment that can require active temperatures of ~480°C from steam circulating in the reactor. Enthalpy will drop to roughly 1850kJ/kg. Releasing 750kJ/kg in this temperature range.

Finally the 'bottom' range is ~270°C to 400°C, enthalpy drops to ~1200kJ/kg, releasing 650kJ/kg. I am not entirely sure what this temperature range could be used for, but I am sure our industrial colleages could find something useful.
The fluid is then an effective liquid and can thus be throttled down to 70atm through a simple pressure regulating valve with minimal loss of temperature and enthalpy, whereupon it can be used as feedwater for the reactor, indeed it is higher temperature than most feedwater and can assume a significant portion of the reactor feedwater duty.

So the net effect is:
1700kJ of reactor heat and 590kJ of compressor power will produce:
800kJ from 500°C to 630°C.
750kJ from 400°C to 500°C
650kJ from 270°C to 400°C

Net effect is 1700kJ reactor heat and 590kJ of compressor power to ~2200kJ of heat at higher temperature.
1700kJ of reactor heat is equivalent to 595kJ of electrical power, so effectively 1185kJ of power is converted to ~2200kJ of high temperature heat.

Effective COP is 1.86.
I would have to do some proper calculations rather than this semi-idealised model, but it seems there is some potential here after all, although a 580 bar steam compression cycle seems rather difficult to engineer.
I will also develop a recuperated cycle

[alexterrell]

Effective COP is 1.86.

By "effective" you mean after the inefficiencies?

I think that's a good result if we want to electrify a process - obviously beats resistance heating.

At the moment though it has to compete with co-fired gas, so gas -> heat -> electricity -> heat, gives an efficiency max of 60% x 1.86, which makes it probably not worthwhile yet.

[E Ireland]

This is not like a conventional heat pump cycle - as the heat source used could be used to generate electricity.
It is the effective COP after compressor inefficiencies if we assume the alternative is all the input steam is fed to a turbine and used to generate electricity for resistive heat.