Hi Creo, and welcome to the group.
Your suspicions and concerns are correct in most everything. Thank you for posting pix of the diagram, as it saves me the research 'on a guess' of what you really have... (it's not unusual for someone to ask, and upon a search, I find 30+ flavors of the same model, right?)
It's a transformerless inverter, so no need to do any 'conversion' antics to make it wake up... just pick
Ripple would be a concern, but note that L001 is an iron-core series inductor immediately downstream of the 3-phase bridge rectifier... that will not only 'smooth off' any 3 phase ripple to the primary link capacitors, for single phase input, it will ALSO have the effect on preventing high ripple loading from occurring at the two segments of your input rectifier array.
The only concern I see,is the same as any OTHER inverter system designed for 3-ph in... you have only 4 diodes of the bridge carrying full load current, rather than 6.
Now, see L002? That's a sensory device- it's looking at the lower leg of LINK current... but since it's an INDUCTIVE device, it's not measuring DC current- it can't... it's effectively measuring CHANGE in current. It could be set up specifically to read LOW FREQUENCY incidence of change (immense current spikes during welding op), OR it could be wide-banded, basically looking for ripple-yielding fault conditions.
See ZG101?? That looks like some sort of plug-in assembly, and it appears to be a combination noise and surge-supression component. The 2.5" wire going form GV107 to the far side of L001 is probably 'freewheeling' suppression (a highly inductive component does NOT like to have it's current supply interrupted, when it does, it adopts the attitude of an ignition coil).
Use L2/L3 and give it a whirl. I would not be surprised to see VG101,102,104 and 105 working fine at all loads, but if one DOES go, just replace it with a larger flavor. keep an eye on C001, C002, C003, and respect their terminals- I see no identified bleeder resistors across them... they MAY have them integral to the cap bank assembly, or they MAY have opted to leave them out, and rely on the output thyristors' firing on shutdown to 'crowbar' the link caps to 0V. If they did, don't TRUST it- electrolytics may go to 0V on a shutdown thyristor firing pattern, but long afterthe microprocessor is down, the capacitor electrolyte is STILL chemically equalizing, and can rebuild a charge. Make a 'chicken stick' and short them down manually before you get body parts in the circuit.