What Bobbi says. Plus:
You could feed the panel into the 12V directly via a simple regulator that only supplied energy when the PSU power-good signal was asserted, showing that the supply was active, and the regulator could be arranged to never cause voltage to rise above 12.1V or some other limit.
You'd need to be sure that when PV input was applied that the PSU did not backfeed in unexpected manners. A diode in the PV feed stops backfeed to the PV but stopping psu backfeed may or may not be as simple as adding a diode in the PSU's 12V output to the GPU.
If you are doing this "from scratch" you want to determine if it is economic compared to grid power. If you receive local subsidies and feed in tariffs (FITS) it may be. Otherwise maybe not. And it may depend what you cost your time at.
At an average of 3 full sunshine hours per day over a year at a typical location you get about 3 units x 365 = 1095 hrs/year or say 1000 units of electricity per year from 1 kW of panels. So if electricity costs N cents per unit effective cost including any FITs subsidies etc then this will save you N x 1000 cents a year or 10 x N dollars.
So if your electricity costs 20 cents /unit you save 20 x 10 = $200/year in power costs. Say $300 outside.
A very optimistic cost for installed panels* would be $2/Watt so 1 kW of panels would cost $2000 installed so would cost $2000/$300 =~7 years to break even if your money is worth 0% if invested. That's at 30c/unit and $2/Watt installed and 0% on your money. Adjust as required. * I can land panels here ex China for about $1.30/Watt at my door. Then add installation costs, mounting hardware, cabling, any control equipment, any regulatory costs, and my time at $1/hour. Add ongoing cleaning and possibly costs for hail or other damage long term.
If you feed direct DC then your panels will provide 12v/18V ~= 66% of rated output at full power unless you use an MPPT or other downconverter at extra cost. A decent grid tie inverter should beat that even allowing for the 18VDC - 240 VAC and 240VAC-12V1DC conversion BUT adds extra cost.
All up, unless you are subsidised hand over fist or are off-grid, chances are that using your own PV is not worthwhile compared to using grid power.
If you can use night-rate power for part of your grid energy this would be even more true. If night rate power is much cheaper than peak load power then bitcoin mining at night may prove more economic while taking longer.
You are describing a Solar Hybrid SOMETHING that may or may not exist in todays vast Marketplace? Step back and revisit a Block Diagram of a Complete Solar Panel Whole House System. You describe several Blocks or Components of the complete system and seem to want all of them in one Device which may be possible today but not practical for the DIY person that dwells within you. It is my impression your main question is of a "Charge Controller" who's function is to charge your Battery using your Solar Panels as the Charging Power Source. Todays Charge Controllers decide for you when the Mains need to take over, when the Battery is fully charged by the Solar Panels, and also when to divert the Solar Panels output to a Dump Load if and when your Batteries are completely charged. The Inverter is the expensive item who's sole purpose in life is to Convert the DC Battery Voltage to a usable AC Voltage ( not necessary for LIGHTING Circuits ONLY ) and Current your household appliances will agree with exactly like your Mains Power from your local Utility Company 230 VAC at 50 Hz I suspect. In an effort to thoroughly answer your question in the form of a recommendation.... Do not look for a single Component that "does it all", if it fails you will loose everything. Rather purchase a quality Charge Controller, a quality Inverter of size and capability to handle your entire anticipated Load, which will indeed be expensive! Most of all, read up / study the topic and fully understand what each individual Block or Component of the Basic,Entire System is responsible for and mostly how they interact to form a Complete System before making any investment. Read, understand, and then make an informed decision / plan to purchase equipment.
Best Answer
I think the odds of that working are very slim.
The UPS is much smaller than the inverters, so it will struggle to maintain a stable 208V 60Hz. That's especially true if it's not intended to be back-fed.
It will be very difficult to get the load to exactly match the generation. Too much load, and the UPS will shut down, followed immediately after by the inverters. To little load, and the inverters will attempt to push out current into no load.