There are
threads on the outback power forum regarding
AC coupling of battery backed and non-battery grid-tie inverter systems, including when the grid goes down. It is trivial to have both types of system operating together when the grid is up, but it gets interesting when the grid goes down. Battery backed inverters are designed to isolate via a transfer switch to power critical loads on a sub panel under these conditions, but what if a non-battery grid tie inverter (or other source of AC generation) is on the islanded sub panel, too? It turns out that the battery based inverters are generally good enough to spoof the non-battery grid-tie inverters into operating as though the grid is up. This allows you to access solar power from a non-battery PV inverter when the grid is down, as long as you have at least a small battery based inverter to spoof the grid. The battery based inverter doesn't even have to have any PV directly attached to it (typically connected via a DC charge controller directly to the battery), since they typically have a battery charger/maintainer built in that will operate off AC.
The challenge comes in matching loads to generation, basically the larger grid management problem recast in your own tiny microgrid. The non-battery PV inverter will try to generate as much power as it can get out the of the panels whether the power is needed or not on the microgrid. If the PV power is insufficient to meet loads, the battery inverter will make up the difference from stored energy in the batteries, but if the solar is too much, the battery inverter actually starts drawing power, unregulated, into the batteries in an attempt to keep the AC parameters in spec. This can't go on for long and isn't good for the batteries, so the solution is to implement a dump load to dissipate excess power (water heatering, etc), or the solar PV inverter can be disconnected. There are automated solutions to accomplish both of these, but this is fussy stuff to set-up and not exactly plug and play. Each non-battery grid-tie inverter has its own approach for anti-islanding (the safety feature that shuts the inverter down when the grid goes down) This feature must be fooled by the battery based inverter and there may be anomalies between inverters of various manufacturers. No guarantees that this will always work or that there won't be AC waveform quality issues, etc.
While I agree with previous posts that the extra cost and maintenance associated with battery backed systems is generally not justified for homes on a reliable grid, it is nice to have back-up power available and it is frustrating for a PV owner to be in the dark with the rest of the grid when it happens. It is particularly ironic for a PV and EV owner to be in the dark, since they have all the components (some a couple times over) to provide backup power, but because of the lack of intentional integration between their components, this functionality is missing. I used to try to defend my interest in a battery backed grid-tied PV system with the intention to use solar power 100% of the time, by essentially operating off-grid at all times except when the PV array was generating more than the house was using. However, I've pretty much come around that this purist conceit is just that, and lacks technical merit. It is dubious to install batteries into a system that will potentailly see little practical use and conjuring the need for them by intentionally disconnecting from the grid, isn't much better. The more rational approach is to use the batteries for a task they are uniquely suited, such as mobile electric supply in an EV, and piggyback household power back-up when the need arises. That model allows much better utilization of the battery over its life. Unfortunately, the existing hardware isn't well designed around this combination of uses, but hopefully some of this synergy will catch on.
For now, one can consider using battery based inverters such as the new
Outback Radian in a mobile application (not its design intent). For $4000 you get an 8kW 120/240VAC grid tie capable inverter and ~6kW battery charger. If one was building a conversion EV it is easy to spend $2000-4000 on a charger alone. The down side is that the Radian (like all household battery inverters) uses a very low battery voltage by EV standards (48VDC). This makes for underwhelming EV performance unless one is satisfied with a golfcar/NEV around town type vehicle. It is possible to do series/parallel switching of the pack to operate at higher voltages when driving as an EV and lower voltage when charging and inverting, but this is a kludge to cover for a gap in system integration. An EV like this that can can also connect to your home in the event of an outage and spoof the grid to keep you non-battery PV system operating is an interesting combination of features (even though you still need the load management layer to pull this off). An inverter that can take a battery voltage of 400VDC off the chademo port would be even more intersting, and it could be combined with a higher rate charger for the Leaf. But that's new kit for someone to build.
Back on the topic of crazy things to do with existing hardware, it occurred to me that the battery voltage of the Leaf is within the DC input spec of many non-battery based PV inverters. Besides confusing the MPPT function, I wonder what would happen if a Leaf battery was connected to the DC input on a PV inverter. I guess it would just push the rated power of the inverter onto the grid for as long as the battery lasted. Not particularly useful unless you fancy gaming the utility to sell you power at 6 cents at night and push it on the grid for a 30 cent credit during the day. Still, a grid spoofing battery based inverter for outages could be set up with a very small battery pack just to bring the microgrid alive and then the non-battery PV inverter could use solar during the day and the Leaf battery over night to provide the majority of the power, assuming you could come up with a sensible way to manage the load matching in the EV to microgrid case (dumping excess power to heat no longer makes any sense when the source is a battery).
Howdy
micro inverters - there's plenty of real-world experience going back to the 1970s. Apparently the devices have been very reliable since the mid-1980s.
