How do you prevent back feeding the grid during outage?

[donrichards]

I assume this has been answered but I can't seem to find it so feel free to just answer with a link to the correct discussion if that's the case.
I'm really new to this site.

Just wondering how an inverter (or whatever hardware it's supposed to be) prevents back-feeding power to the grid when the grid is down? If I were to get a grid tie inverter, how do I guarantee it won't back-feed electricity to the grid if the grid goes down?

 

[Texas-Mark]

If I were to get a grid tie inverter, how do I guarantee it won't back-feed electricity to the grid if the grid goes down?

If it's a true grid tie inverter, it won't run without grid input. That it is how it is deigned.

 

[sollap]

Any inverter that is UL 1741 compliant is designed for anti-islanding. That means it will not backfeed a grid that is not supplying steady power. When you power it on, you'll have to wait about 5 minutes while it evaluates the grid. It's won't let you begin to backfeed until it's completed it's evaluation. Once it allows you to backfeed, if the power goes out or becomes unsteady, it'll disconnect.

 

[Bubba1]

Put the main breakers in the off position.

 

[zanydroid]

Put the main breakers in the off position.

Technically correct. But How does this help answer the question?

 

[jasonhc73]

Any inverter that is UL 1741 compliant is designed for anti-islanding. That means it will not backfeed a grid that is not supplying steady power. When you power it on, you'll have to wait about 5 minutes while it evaluates the grid. It's won't let you begin to backfeed until it's completed it's evaluation. Once it allows you to backfeed, if the power goes out or becomes unsteady, it'll disconnect.

UL1741 is actually fairly old.
UL1741SA is basically what all modern grid-tied inverters use.

To answer the OP...
A physical disconnect with the main circuit breaker or a safety disconnect switch is the only way to guarantee it won't backfeed.

UL1741SA inverters have current sensors at the grid connection to ensure that the inverter doesn't backfeed. This is still software controlled, and susceptible to incorrect settings.

All UL1741 inverters measure the frequency of the grid connection. As long as the frequency is within operational range, they match this frequency and operate as necessary.

There are all kinds of neat settings that you can totally get wrong.



There is only one setting you can get wrong.

 

[jasonhc73]

If it's a true grid tie inverter, it won't run without grid input. That it is how it is deigned.

Newer grid-tie inverters with UL1741SA standard work without grid input, and island themselves from the grid. There is no physical disconnect, they can just not backfeed, thus isolateing the load from the line.

 

[zanydroid]

The OP was about grid down backfeed. As long as the island looks semi normal it should work. Of course it’s probably impossible to test all combinations of micros, string inverters, and battery inverters sitting on your island. Fortunately at least they are all regular low frequency and high frequency inverters (which are sort of part of the UL test standard), and not random generators

Backfeed with grid up is a lot more complicated and easy to mess up. In software, CT setup, etc. with perfect setup It is a fact of life that there will be a little bit of excess backfeed above the cap that you set. The best that can be done is compensating for excess backfeed later.

 

[zanydroid]

UL1741SA inverters have current sensors at the grid connection to ensure that the inverter doesn't backfeed. This is still software controlled, and susceptible to incorrect settings.

My UL1741SA micros absolutely do not have CT anywhere remote. a few do have the option.

 

[Ampster]

UL1741SA inverters have current sensors at the grid connection to ensure that the inverter doesn't backfeed. This is still software controlled.......

I think it is most likely in firmware and uses a number of tests to verify there is a grid. It tests grid frequency, impedance and voltage. I don't believe it uses current sensors because that would mean current would have to flow which is what the anti islanding is trying to prevent.
When the grid is up it undoubtedly uses the current sensors.

 

[zanydroid]

I think it is most likely in firmware and uses a number of tests to verify there is a grid. It tests grid frequency, impedance and voltage. I don't believe it uses current sensors because that would mean current would have to flow which is what the anti islanding is trying to prevent.
When the grid is up it undoubtedly uses the current sensors.

Another problem with using current for anti islanding is that there could still be a voltage on that side if your goal was only to avoid pushing current back. Remember it’s not just about backfeeding an island, it’s also about de energizing it for lineworker safety.

 

[dougbert]

Technically correct. But How does this help answer the question?

he works for microsoft

 

[timselectric]

If it's a true grid-tied inverter , it won't backfeed a dead source.

 

[Ampster]

Newer grid-tie inverters with UL1741SA standard work without grid input, and island themselves from the grid. There is no physical disconnect, they can just not backfeed, thus isolateing the load from the line.

Can you give an example of some of those?
If you are referring to IQ8 micros, they cannot do that without additional hardware which presumably makes the physical disconnect. Natively, IQ8 micros are GT inverters and cannot operate without the grid.

 

[Texas-Mark]

Newer grid-tie inverters with UL1741SA standard work without grid input, and island themselves from the grid.

I don't consider that a true grid-tie. IMO a "true" grid-tie absolutely needs the grid to be up to function, and everything else is some sort of hybrid. Of course I am old school and was a beta tester for the original Xantrex grid-tie inverter.

If it's a true grid-tied inverter , it won't backfeed a dead source.

Yes, this is the way I see it.

 

[OzSolar]

IEE1547 also factors into this somehow.

My memory is prone to be faulty but at an industry conference some 20+ years ago I recall that a very qualified presenter from NREL or perhaps SMA explained at a workshop I attended that since pure grid tie inverters are only current sources they simply can't work without the grid. I could be all wet so hopefully someone qualified will jump in. If it's not true, it's certainly engaging fiction. LOL

They were politely trying to point out the irony of the roadblocks that utilities were putting through the industry through that weren't even necessary. Similar to a synchronous motor that is externally excited used on MW scale wind turbines. Again, no idea if that was 100% true back then or how true it is today.

 

[zanydroid]

pure grid tie inverters are only current sources they simply can't work without the grid

Pure grid tie inverters need a ~60Hz AC to PLL onto (ofc change 60 to 50 if you don't live in america), and that thing they lock onto does the needful to anti-anti-island.

I don't know how IQ8 emergency backup works, I think they get kicked into a mode where they self synchronize with each other, and that mode is only kicked into when the transfer switch (in that Enphase gateway or whatever box) is active.

Semantically, I think 1741 inverters that have the potential to feed onto grid all have to be grid tied, but they can have alternate modes to go into island mode. Some will require additional hardware to do so, others have it all onboard (EG the transfer switch and all necessary grid presence sensors are baked into the hybrid. VS needing extra stuff in a second box near the service).

I don't know that they're current sources. They use the PLL as the starting point of control for generating 240V 60 Hz AC, with some amplitude modulation/phase shift to achieve various goals. As far as I'm concerned they're AC voltage sources just like spinning generators.

Hybrid = Inverter with a state machine where one state is grid-tied (if it can backfeed), and one is island mode? Maybe with additional states between those to safely transition.

 

[OzSolar]

Pure grid tie inverters need a ~60Hz AC to PLL onto (ofc change 60 to 50 if you don't live in america), and that thing they lock onto does the needful to anti-anti-island.

PLL?

 

[donrichards]

Man, you guys are awesome!! Thanks for that. I wondered how they worked, if it was like a massive diode or something else. This has been super helpful. I worked on a university research team on new tech for PV efficiency improvements at UCF about 8 years ago. We found ways of increasing cell/panel performance by over 10% with some additive modifications, but we never looked at any of the other equipment so this is all new to me.

 

[Mattb4]

Unless you disconnect it a conductor will conduct. For a grid tie this means you must have a physical disconnection happen or any potential power supply must be shut off. No amount of games with software, CT or hand waving changes this basic fact. Now during parallel operation rules of parallel are in effect. If one power supply picks up more load the other power supply sees a decrease. Setups that incorporate a pseudo grid to keep alive a grid tie inverter must physically isolate from the grid while supplying house loads.