How would it do performance and efficiency wise? The purpose is to utilize energy effectively.
Say you're building a small walk-in within a warehouse.
http://www.arthurhewett.com/id26.htm suggests you can get about 9,000 BTU of cooling at -10F coil from a 5 ton heat pump and also 17,000BTU of heat into heated space at normal indoor comfort climate.
Build a cold box inside a warehouse.
Take a plain 'ol 5 ton 3ph heat pump unit and set it up inside the cold box near the ceiling. Remove compressor and relocate them outside the box. The reason is to keep the heat rejection from compressor and CCH from bleeding into the room.
Acquire two identical "indoor coils".
Coil 1 just uses fan coil to reject heat inside the building during heating season. Coil 2 is built into a clear acrylic plastic(which serves as a sight glass to monitor the facility cooling water conditions as well) box and piggy bag into the line used for cool water going into chiller that is fed from the facility cooling tower. The lineset goes outside in a manner similar to typical HVAC setup, so I don't have to worry about oil return system.
The two coils are setup to be identical internal size, so I don't have to worry about adjusting charge.
For now, they're switched by using manual ball valves. To do winter, you pump down into air coil. To get into summer mode, you pump down to water cooled coil. If the setup works well, I can replace these with hermetically sealed magnetically coupled normally closed solenoids and use some cheapo off the shelf controller.
The reversing valve input at thermostat will get permanently wired as "heating" and I'll use an ordinary refrigeration temperature controller across R and Y to "call for cool" and the heat pump's internal controller will deal with the defrost cycling.
There's a HUGE emphasis on efficiency for residential HVAC, more so than commercial refrigeration condensing units. To be an efficient heat pump, it needs a good COP, which is perfect for this and if I can get 2.0 COP, I'm set.
The only R410A lines are in the proximity of the box and most of the complex connections are pre-made and helium leak tested at factory and if I use brazing, the entire loop is hermetic.
So, in theory, I'd have a cold box in a warehouse with practically zero refrigerant loss(like a properly installed home system) that gets a COP of 2.0 at freezer temp. Am I just thinking too optimistically?
Anyways, operating conditions from the compressor's POV isn't much different from running it year around in 0F season. So, I'm hoping it would be reasonably durable...