I'm trying to understand this, so please bear with me.

"(N)ever showed any evidence of being '"iced up"'. Why would they?"

In GSHP heating mode, the outdoor coil is an evaporator, while the indoor is a condenser. (Wikipedia)

The temperature of the refrigerant at the beginning of the evaporator HX will be colder than 32°F. Because the HX is so cold, at 'hard' shutdown - no electricity to run the evaporator circuit, the lines at the top of the evaporator will freeze if no antifreeze and will start to gel if antifreeze. As ice builds, it restricts the flow through the HX. After susequent starting, because of the restriction, refrigerant can't absorb as much heat from the the HX. This causes the refrigerant to boil later in the evaporator, which causes ice or gel to form further. This situation continues to progress, until the evaporator is (restricted). Once that happens, the refrigerant will start to boil in the suction line. This causes the temperature of the suction line to drop, and just like in the evaporator, freezing.

Eventually the freezing works its way all the way back to the compressor, which is where the trouble can really start. If allowed to operate in this condition for too long, liquid refrigerant can make its way back to the compressor. If this happens, the compressor can be damaged.(adapted to fit gshp water flow, from BMitch and Philip Ngai, Stackexchange, https://diy.stackexchange.com/questions/48450/why-do-the-evaporator-and-suction-lines-freeze-when-there-is-low-refrigerant)

Maybe someone would say that the water flow is adequate, the evaporator field did not freeze, the condenser circuit continued to flow and pressure was adequate in the system.

All we know for sure is that the compressors are failing and maybe a low pressure situation developed with no low pressure switch cutoff for some reason and restricted flow caused by shutting off the GSHP with no 'cooldown' existed.

Just guessing.

Respectfully submitted, B

By definition, the coldest part of any GSHP system is the heat exchanger inside the heat pump (the water coil), since that is where the heatpump is trying to extract a heat from the fluid in the pipes. The heat exchanger is relatively small and requires a larger delta T, which makes its walls very cold.

So the loop field can never freeze, because by that time where the loop gets cold enough, the water coil is already frozen shut (and the heat pump is in lockdown, because the safeties kicked in). So the scenario you describe can not really happen, as soon as the coil freezes the whole heat pump goes into lockdown, and the loop temps cannot drop further, since the coldness cannot transported from the water coil into the loop field.

No idea why the compressors were failing, we installed probably 20 NEWs units in the last couple years, they are very reliable thus far.

But the design here put a lot of stress on the compressor, since the system was set to operate on the high limit all the time. Also the loop field was doing fine, dropping down to 28F min and stabilizing, well within the secs of the heat pump.
So the loop never froze, the circulation flow was always there, not sure about the pressure (loop field or refrigerant?) but pressure is not needed in a loop, only flow, the loop cannot be frozen(!) and likely the operation of the compressor on the limit caused the premature failure.

Thought I'd report back since I've had the time to do a quick analysis of the various iterations of my system in AK. In summery, system started out as two 4 ton WF units with 10,000 feet of horizontal slinky tube. In May after 2 compressors failures, installer replaced both units with a 5.5 ton compressor along with an outside temperature sensor for control. Read the thread for more detail as to the whys and wherefores this change was made. I've graphed the electrical usage which has its own meter base. Thanks for all the help. Mike

Looks good. The biggest change seems to be the outside temp controller, making the heat pump run significantly more efficient. Looks like 20% savings so far. Not running the units at the 150F load max temp likely will have that effect.

BTW my EWT is 31.4*. Did a two day temperature profile logging a temp every 2 minutes.