If motor was already running and inverter picked up the load quickly enough that motor didn't stop, then it ought to keep running.
Relay transfer time is typically 30 milliseconds, so don't expect motor to stop.
Of course, a refrigerator is expected to cycle on and off to regulate temperature, so this isn't going to be a usable system unless inverter can start the motor.
When inverter turns on unsynchronized to the grid which just failed, motor would hit it hard for an instant. I would expect that high current to go away much more quickly than from a stop because motor only has to adjust timing slightly. But this could be the issue, if inverter is protecting itself.
(There are stories of larger grid-synchronous rotating generators that were accidentally connected to grid out of phase rather than in phase. They ripped their mounting bolts out of the concrete.)
The "VA" (volts x amps) of a motor is more than watts, because waveforms are misshapen and current is out of phase with voltage.
The ratings do say,
"
Rated Load Amps (RLA 60 Hz): | 12.5 |
Max. Continuous Current (MCC in Amps): | 25 |
"
I'm not clear when it is expected to draw maximum.
AC current is normally specified as RMS, root-mean-square average over a cycle. Perfect sine wave in phase produces PF = 1.0 "Power Factor", same apparent power as actual power that a resistive load would. Any other current waveform results in PF < 1.0, excess current for a given power. It is the higher current which heats things like transistors. The inverter has to protect itself from overheating, and may directly measure current to shut down quickly and prevent damage. Some cheap inverters don't protect themselves and fail.
It seems like your compressor, rated less than half the power of the inverter, ought to be able to keep operating if already running. But it may be the inverter just can't handle that large a motor load.
There is something called "reactive power", which is part of the power factor issue. Waveforms other than sine wave draw excess current during part of the cycle. Reactive power draws current from the supply, then later shoves current back in. The inverter may not be good at handling reactive loads. Motors were designed back when AC power sources were rotating generators. Transistorized inverters were designed to power various loads, but often for consumer applications without large motors.
You're going to need an inverter which can start that compressor. A "Soft start" kit might let a smaller inverter start it, but wouldn't help with this transition from grid to inverter if surge due to out of phase is the problem.