Do you mean during peak ok hours you’ll operate off battery power?
I’ve read something about PGE charging customers for using solar - arguing that solar customer aren’t paying into the system for keeping up the grid. Are you referring to that scenario and do the batteries help with that in any way?
In terms of maintenance this is a really simple setup since there are no batteries to maintain. The inverters do not require periodic maintenance and the panels only need to be washed off periodically to keep them operating at their peak. In the time we've been operating like this we lost one inverter and one panel which was damaged from a falling rock. So easy to maintain, and trouble free.
The question then was "how much does the power company pay for power that you produce?" The terms and conditions of what you pay, and what the power company pays, is nominally the "rate plan."
When we started, this was new to PG&E and we were on a plan where we stopped getting monthly bills, instead the mechanical meter would run forward when we were drawing power and backward when we were generating power. Each month we'd have a 'net power' which could be positive (used more than produced) or negative (produced more than used)and every 12 months that was summed up. If the number was negative they would just zero it out and roll over to the next year (free power for them), if it was positive they would charge a stepped rate based on total power used for the year. Once they got "smart" meters installed they got creative with the plans, we ended up on a plan where they pay us a wholesale rate, bill us at a retail rate, and total $ up instead of actual power used. This works out better for them and extended the time it took for the system to pay for itself.
Powerwalls can (and in our case will) completely disconnect you from the electric grid. They don't buy any of your power and you don't buy any of their power, hence no rate plan. If you size the system you can be pretty sure you won't ever be without power (even with a series of cloudy days) and you can add a natural gas fueled electric generator[1] (we would still have gas service) which could charge the powerwalls in a pinch.
I've got all the feeds instrumented so I can tell exactly how much power the house is using and the panels have generated (fed into an influxDB time series database) and using that data have been planning for the retrofit based on our usage over the last 15 years.
The Powerwalls double the initial installation cost but since I'm not paying margin (selling wholesale and buying retail)to PG&E the actual value delivered is higher and so it has a better rate of return. Of course I can only speculate on the lifetime ownership costs of Powerwalls (much like I had to do with the inverters which I had in my spreadsheet being replaced every 10 years since that was the warranty on them).
[1] What I really wanted was some Bloom Energy fuel cells for that but they don't really have a 15 - 20kW rated one, it is too small.
On the plus side, you can put them in a subterranean vault if you have space on the property for the set back limits.
PG&E still plays rate-plan shenanigans. Besides the retail/wholesale stuff, they charge $10/month just for the grid connection. They're currently lobbying CPUC to raise that to $60/month.
Also the "time-of-use optimized" setting in the Tesla app isn't actually that optimized. It's unaware of the retail/wholesale issue, and hence simply tries to maximize the amount you ship back to the grid in peak hours when it'd be better off minimizing total consumption. It also sometimes doesn't discharge the PowerWall as much as it could (leaving solar energy on the table), and it charges it with grid power when it could easily use solar energy. I've found it's better to just use the "Self-powered" setting, where it charges the PowerWall as soon as you have excess energy over the home consumption, starts discharging as soon as there's a shortfall, and continues until the PowerWall reaches the reserve level you set.
Note that solar generation is incredibly seasonal. I'm currently generating about 25 kWh/day. In January, this is more like 4 kWh/day. So depending on your shade levels, you might have to put on 6x as many panels to be entirely grid-independent vs. grid-connected with battery backup. We're sized so that we can power a full normal workload from about Apr - Oct, which at least covers fire season, but would have to conserve significantly (i.e. forego loads of laundry and electric appliances) if we had an extended outage in winter.
In the summer, the sun is higher in the sky, so it clears the hill earlier (by 8 AM or so) and never hits the trees. Both arrays produce the full day, and the days are longer, and the sun is incident at a steeper angle anyway.
My situation is a little weird, but it also applies to many urban or rural settings where you'd have a tall building or series of tall trees that block sunlight when the sun is low on the horizon, but that the sun would clear when directly overhead.