That's pretty cold. Here in Florida we don't see those temps ever so I don't have experience running equipment under those conditions. However, I do know my two systems generate their own heat during normal operations so would suspect that a system you put together in your shed would "self heat" from losses via normal operations. My Growatt 3k inverter document says the "no load power consumption" is <60W. That combined with the charge controller efficiency loss (1-95%=5%) and the inverter efficiency loss (1-90%=10%) the unit (in theory) would generate ~75W as heat powered on and idle. if the MPPT is pumping 200W of solar panel output into the battery, that 5% loss turns into another 10W of heat. If the inverter is powering a 200W load from battery (or solar), another 20W of heat gets added to the air surrounding the inverter. At 3.412BTUs/Watt that's 255 BTU's heating the shed/cabinet/box your system is enclosed in.
My 3k inverter is in at cargo compartment in my RV (with the battery). Here's a graph of the inverter temp (measured inside the inverter) from the last 24 hours:
Here's the power graph for the same period:
As you can see: Just charging the battery and powering the near constant 40W load in the RV (nothing's running) can produce useful heat.
I personally would have heater pads for the LiFePO4 batteries with a thermostat. I'm not sure how effective the heater pads are. As i recall they were in the 5-10W range so powering them from the inverter output should be fine though I don't know if a 5W heating pad is enough to keep things above freezing when its -20F. you can easily test it with a bucket of water sitting on you heater pad...
I think the main risk is the batteries draining, the system shutting down and then you have no heat to keep the batteries from freezing. If there's anyway to have a backup heating source (propane/kerosene heater or smudge pot around) around, I'd consider it.
Hope this helps.