Your concept appears to be good. Since the additional micros would only be used in grid-down situations, you would want them to do proportional curtailment as the frequency increases, or you should measure the frequency and use your variable current control to do the curtailment yourself.
I am curious how much you have thought through the controls and the details of the circuits to do the variable current control on the DC side of the micros.
The system can perform in two modes: "Grid Power On " - power arbitrage and "Grid Power Off" PW2 extension.
Arbitrage simply charges during high solar time and discharges in the evening when the new tariffs at PG&E will be after solar time. So the system will move the abundant non-peak solar power to the evening peak time to arbitrage the rate difference. Its not much but its a side effect of the design.
The tricky part in both modes is to properly coordinate with the PW2. My thinking is to try and create the illusion to the PW2 that the system is "erasing" the home vampire power and low end power ( < 500W per leg - lights etc..). by measuring the home power on each leg, the current limit can be used to track the home load in near real time. It quietly charges while there is good solar and the PW2 has not shut off the solar inverters with high 60Hz because its fully charged. The UL1741 micro inverters would be compliant so its no different than my 18 M215s on my solar. When the PW2 raises the 60Hz, the micro inverters ( MI) would shut down. If I buy MI's that meet the new curtailment spec, they would also be curtailed if the PW2 slowly raised the 60Hz - just like the solar inverters.
Charging would "look like" a home load on the 240V with a BMS function which stops on battery fully charged or high temperature alarm ( or any other alarm that makes sense). I'm hoping the PW2's only raise the 60Hz when they are fully charged and see excess solar power to shut down the solar until they discharge to 95%. When the PW2's are cycling from 100% down to 95% and back, these MI's would only produce power when 60Hz was normal. It should be ok as long as the PW2's only raise the 60Hz when they are fully charged.
I have a design engineering background so the variable current limit should not be a problem. I would either use an Aurdino or a Raspberry Pi or maybe both. They would send an 8 bit code to an eight bit DAC which would produce a voltage between 0-5V which would control the pass transistor. Non of the circuit controls should be a problem but my largest concern is whether there are any "gotchas" where coordinating with the PW2 becomes a problem like the 60Hz issue I've mentioned. I'm also a bit concerned with the delay in the current/power monitors and the radio delay to the main controller. If someone turns on an oven in the house, both 500W MIs would be turned on in less than a second and when the oven turns off, that one second delay could cause a spike in the MIs output voltage if the PW2 doesn't soak it up.
I've located a low cost power supply and constant current which should be less than $150
Sign In | AliExpress Mobile
Sign In | AliExpress Mobile
The power supply would be DC 18V, AC 185-265VAC and 1500W. The combined efficiency should >90% even with the Tandem power supply and Boost constant current converter.
The Tesla battery is a 74p6s configuration with a fully charged ( 4.1v per cell) of 24.6V
Another issue would be adding a real BMS should balancing be needed. If the batteries in the Tesla Module track well once balanced once, than it might be reasonable to leave balancing off and just manage over-charging, low battery, and temperatures. The controller could have an 6:1 analog selector and ADC so it can monitor the balance and report any problem.
I have a Raspberry Pi and Aurdino and a LoRa transceiver. Its mostly software with a few circuits and many off the shelf from the Aurdino world.
Once my PW2's are installed, I can monitor when they raise the 60Hz to see if thats ok. I could also build a simple version using a much smaller battery before buying $1000 Tesla pack.
I just don't know if the PW2 software can detect odd things going on. Only way to know for sure is try it.