PV Array (microinverter) with Generator

[jaggedben]

That's because we don't leave the grid breaker on for extended periods of time...it's used as needed, and we never let if reach the point where the system would have surplus power to sell back. You can't do that through the same meter you get your supply from, just by jamming it back into the circuit from your controller...it won't generally pass through the meter, and in some cases like ours, it's a digital meter they read from the office and this reverse push generates an error message on the meter. In order to sell back, you have to have a sell-back system with the local grid, usually involving a separate meter and lines for this reverse flow, or perhaps it can pick it up the reverse flow from the grid connection and sell that back at some predetermined (but usually a pittance) rate through the 2nd meter. It costs extra for this gear and it's more to go wrong and for the tiny amount you earn from sellback, it just isns't worth the grief you have with the local utility, PREPA.

The energy certainly passes back out to the grid through the meter. Whether and how your digital meter counts it is another matter. If PREPA wanted to let you sell back they could probably just reprogram your meter, or at the most they might have to give you a new one that supported the required software, like those on a million or so services in California.

 

[winnie]

The energy certainly passes back out to the grid through the meter. Whether and how your digital meter counts it is another matter. If PREPA wanted to let you sell back they could probably just reprogram your meter, or at the most they might have to give you a new one that supported the required software, like those on a million or so services in California.

And if PREPA wished, the meter could generate an error message if there were significant reverse energy flow...but I absolutely agree, the meter cannot directly _prevent_ reverse energy flow, otherwise it couldn't function with anything that doesn't have a unity displacement power factor.

-Jon

 

[caribconsult]

The difference is that you can backfeed a utility.

Not down here in Puerto Rico, you can't. We have several friends who have various types of solar, and everyone of them has a separate meter for 'sellback.' As I've mentioned several times, we don't want that type of system for numerous reasons. I don't know enough about grid operations to explain why a separate meter is used here but not in other places, this but I know we can't sellback through the existing 'use' meter.

 

[jaggedben]

Not down here in Puerto Rico, you can't. ...

Physics does not care about legality. The energy will go somewhere, whether it's counted and credited or not. The question came up in the context of why you cannot use a motor generator as a voltage reference. One reason is that you cannot backfeed a generator. You can always backfeed a utility, doesn't matter what their rules are. There's always a limit to how much you can backfeed, but the limit is almost never zero like for a generator.

Oh, and your friends with separate meters are still backfeeding the utility.

 

[caribconsult]

Yes, our friends with two meters are backfeeding the utility but it is with PREPA's consent, it is wired so that the backfeed doesn't light up at least the local transformer, perhaps not the primary overhead, so that PREPA's guys won't get electrocuted if they're working on the line and expect it to be dead, but I don't want to get into a debate over this.

Instead, I have a question and it may sound like a dumb one but I don't claim to know everything about electricity. The question is: why would you try to backfeed a generator? To what purpose? If you back feed the grid, you're selling them something and will get paid for it. What would you accomplish trying to backfeed excess PV production to a generator? Get it to spin around without the motor going? I don't get the purpose.

 

[caribconsult]

The difference being, he doesn't. The PV array and inverter are not grid-tied.

You get it. I don't think some have read the fine details of what I have posted, or I'm not explaining well...I've said it repeatedly, that our grid is present but not connected unless needed. We don't try to backfeed the grid EVER, not from the PV array, not from the PW, and not from the gen. And we have a nice big switch with one handle that transfers the whole load of the entire house directly to PREPA, all @240v/200a, if we need to say, run a welding machine, or the electric clothes dryer. The Tesla just runs alongside in idle, charging if there is sun. And similarly, we have a second switch which will transfer the main load center of the house to the gen, and disconnect it completely from PREPA and the Tesla. There's no connection gen<->grid or gen<-> Tesla. And just for in case there's not much sun on a given day, we have a breaker that will send PREPA to the gateway and charge the battery, but this only gets turned on when needed. Total flexibility. If I had it to do all over again, I'd go for the same setup.

 

[GoldDigger]

Just to tie up some lose ends:

1. One of the reasons for a separate production meter alongside the normal service meter is so that any solar energy returned to the grid is valued at a wholesale power producer rate and not the consumer rate that is used for a net metering scheme. A serious disadvantage of this for the consumer is that you can end up paying the differential between the two prices even when your local PV power is only meeting the needs of your local loads.
2. A separate production meter in series with main/revenue meter may be required when renewable energy carbon offset credits are available in the area. The carbon credit is the same whether the PV is feeding local loads or netting back to the grid.
3. In areas with a marginal or overstressed distribution system backfeed into the grid, especially if substantial, can increase the instability of the network as well as cause sudden large voltage changes as weather reduces the PV output momentarily. Some of these disturbances can occur even if there is no net backfeed, because it can still produce sudden changes in the load seen by the network, but it is usually worse when backfeed is involved.
4. Even without backfeed, use of PV inverters that output only at unity power factor can cut the resistive current to the customer to zero bugt leave the reactive current to be supplied by the grid. If the power factor of the local loads OR nearby loads is low enough this can lead to a very very low power factor (approaching zero) on portions of the distribution network. Again this will happen regardless of how the metering is done but can be worse if backfeed into the network is allowed. Generators are typically capable of completely supplying less than unity power factor loads. PV inverters for use in Hawaii are now required to be able to output reactive power on command from the utility, since in some areas PV contributes more than 50% of the power production during peak solar hours.

 

[Carultch]

Instead, I have a question and it may sound like a dumb one but I don't claim to know everything about electricity. The question is: why would you try to backfeed a generator? To what purpose? If you back feed the grid, you're selling them something and will get paid for it. What would you accomplish trying to backfeed excess PV production to a generator? Get it to spin around without the motor going? I don't get the purpose.

It's not that you are necessarily intending to backfeed a generator. It is more that you set up the conditions so that if circumstances don't go as expected, you might backfeed the generator.

The solar source depends on the weather and time of day/year, and the load on the generator depends on user behavior. Both of these are transient, and not entirely predictable. You can predict the solar source to some extent based on historical weather data and the typical meteorological year (TMY). You can predict the on-site load to some extent based on the expected application of the building and energy use modeling. In concept, you could strategically size your solar so it is most likely to be less than the instantaneous power consumption throughout the year, but the possibility exists that it doesn't go as planned. You could have zero load and full PV source, because of the bad luck that loads diminish to zero on a bright & sunny day. That scenario will backfeed the generator, if the solar is permitted to run in generator-mode.

This is why the simple solution is to connect on the utility side of the transfer switch, and shut off the solar by default when in generator mode. Assuming the most likely case that you'll use the utility more often than the generator. If connecting to the common terminal of the transfer switch, you'll need a control system to #1 identify that it is in generator-mode, and #2 to dynamically limit the solar production to keep the generator operating within its allowable range of operating power. A "minimum import" control, so that the generator is never sent below a given percentage of its nameplate value, as a result of solar production.

 

[caribconsult]

Thanks for the input, guys, but I still don't see an answer to my question: what would you accomplish by trying to backfeed a generator?

And to GoldDigger, you just verified what I had said about needing a separate meter for backfeeding the grid (and the additional wiring involved) and that's one of the principal reasons we don't do it that way. I love this forum...there's so much knowledge here, I learn a lot. Also, our setup is wired so that if the gen is being used, the Tesla system is disconnected and in 'idle' mode, where it can charge if there's sun, but there's no load and no interference between the Tesla and the gen or the grid. We don't have a conventional transfer switch to select either the grid or the Tesla....we're full time on the Tesla and manually switch to the grid on an 'as needed' basis.

 

[Carultch]

Thanks for the input, guys, but I still don't see an answer to my question: what would you accomplish by trying to backfeed a generator?

You wouldn't accomplish anything of value. It is detrimental to do so, and it could damage the generator if enough backfeeding happens.

As I mentioned previously, you aren't intending to backfeed the generator, but if you set up your interconnection so that if you are unlucky enough with the timing of source and load, that you will backfeed the generator.

 

[winnie]

As I understand the desire to run the PV system with a generator is to reduce fuel consumption during a grid outage.

Back feeding the generator is _not_ the goal, but an unfortunate failure mode of trying to do this with a PV system designed to back feed the grid.

Jon

 

[jaggedben]

Thanks for the input, guys, but I still don't see an answer to my question: what would you accomplish by trying to backfeed a generator?

What you'd accomplish is getting out of doing additional work you didn't include in your bid. e.g. if there's already a PV installed in the panel you're backing up.

And to GoldDigger, you just verified what I had said about needing a separate meter for backfeeding the grid ....

Nothing he said substantiates that. 100 bucks says your friends' twin meters are connected to the same transformer.

 

[caribconsult]

You wouldn't accomplish anything of value. It is detrimental to do so, and it could damage the generator if enough backfeeding happens.

That's what I thought and I couldn't understand why any one was trying to figure out how to do this. I got it now.

 

[winnie]

Re: separate meter requirements.

Standard meters are bi-directional; power can flow in either direction through the meter hardware. AFAIK there is no good way to prevent this, meaning that if one had a standard meter and connected PV on your premises, you might very well end up exporting power to the grid.

The fact that the meters are physically bi-directional does not change tariff requirements.

The utility could easily set up their standard meters so that they only register power flow in one direction, so that when you are exporting power you get no credit. Electricity flows to the grid through the meter but the meter doesn't count backward.

The utility can require multiple meters as part of their connection requirements, or can have meters that separately total consumption vs production. Two meters in series, one designed to only rotate in one direction, one in the other would be a perfect example of a system that the utility could use to charge one consumption rate and pay a different production rate.

-Jon

 

[caribconsult]

Re: separate meter requirements.

Standard meters are bi-directional; power can flow in either direction through the meter hardware. AFAIK there is no good way to prevent this, meaning that if one had a standard meter and connected PV on your premises, you might very well end up exporting power to the grid.

The fact that the meters are physically bi-directional does not change tariff requirements.

The utility could easily set up their standard meters so that they only register power flow in one direction, so that when you are exporting power you get no credit. Electricity flows to the grid through the meter but the meter doesn't count backward.

The utility can require multiple meters as part of their connection requirements, or can have meters that separately total consumption vs production. Two meters in series, one designed to only rotate in one direction, one in the other would be a perfect example of a system that the utility could use to charge one consumption rate and pay a different production rate.

-Jon

I think you understand what we face here in Puerto Rico. Our utility (PREPA) uses a digital meter. There are no wheels spinning around, or rotating number dials....it's all digital and can be read from the office, so our utility bill is accurate, which is a huge advance for us. They used to just 'guestimate' bills and everyone was getting overbilled. I don't know if the meter passes current in both directions, but if the system is in a sellback mode, (battery full, lots of sunlight, grid connected to gateway) which we've experimented with several times just to see what happens, an "ERROR" message appears on the meter LED screen. I don't know what they see at the head office but it can't be good. And there seems to be no way to tell if it 'spins backwards,' since you can't read the numbers when the error message is on the screen. And what I don't need at all is PREPA coming my house and asking us do we have this permit, or that permit, (they're permit-crazy here) or any number of nonsensical questions. We're happy to pay the average $13/month bill for the convenience of having PREPA when needed. As I said previously, we have friends here will full solar setups (some Tesla, some not) and ALL of them have two meters - one for consumption, the other for sellback. They had to do it like that since they got their systems as part of a gov't program that financed their systems to a large percentage, and those were the rules. How these two meters connect to the main circuits I have no idea, and I don't really need to know this. We decided 'no sellback' when we purchased the system and I'm very satisfied with that decision. We're here over 18 years, and we had a mountain of gear (TV's, microwaves, computers, phones) fried from spikes, surges, outages, brownouts, over-voltages and the like from PREPA and all that has ended by our not being connected 100% of the time to PREPA. That's worth a lot more than the couple of cents/Kwh they give you for sellback. Not to mention the danger of frying the gateway due to some electrical a-bomb coming up the lines from PREPA.

 

[gadfly56]

I think you understand what we face here in Puerto Rico. Our utility (PREPA) uses a digital meter. There are no wheels spinning around, or rotating number dials....it's all digital and can be read from the office, so our utility bill is accurate, which is a huge advance for us. They used to just 'guestimate' bills and everyone was getting overbilled. I don't know if the meter passes current in both directions, but if the system is in a sellback mode, (battery full, lots of sunlight, grid connected to gateway) which we've experimented with several times just to see what happens, an "ERROR" message appears on the meter LED screen. I don't know what they see at the head office but it can't be good. And there seems to be no way to tell if it 'spins backwards,' since you can't read the numbers when the error message is on the screen. And what I don't need at all is PREPA coming my house and asking us do we have this permit, or that permit, (they're permit-crazy here) or any number of nonsensical questions. We're happy to pay the average $13/month bill for the convenience of having PREPA when needed. As I said previously, we have friends here will full solar setups (some Tesla, some not) and ALL of them have two meters - one for consumption, the other for sellback. They had to do it like that since they got their systems as part of a gov't program that financed their systems to a large percentage, and those were the rules. How these two meters connect to the main circuits I have no idea, and I don't really need to know this. We decided 'no sellback' when we purchased the system and I'm very satisfied with that decision. We're here over 18 years, and we had a mountain of gear (TV's, microwaves, computers, phones) fried from spikes, surges, outages, brownouts, over-voltages and the like from PREPA and all that has ended by our not being connected 100% of the time to PREPA. That's worth a lot more than the couple of cents/Kwh they give you for sellback. Not to mention the danger of frying the gateway due to some electrical a-bomb coming up the lines from PREPA.

Thanks for the additional information. However, it seems that some folks here just do not understand that when the PV array charges the Teslas that it is not connected to the grid, nor in fact is it ever connected to the grid. I can't see any reason that you couldn't charge the Teslas via a generator if you had either a manual or automatic transfer switch. I just depends on if the Tesla charger can accommodate generator.

 

[caribconsult]

Thanks for the additional information. However, it seems that some folks here just do not understand that when the PV array charges the Teslas that it is not connected to the grid, nor in fact is it ever connected to the grid. I can't see any reason that you couldn't charge the Teslas via a generator if you had either a manual or automatic transfer switch. I just depends on if the Tesla charger can accommodate generator.

That's what I'm wondering about. How is the output of the gen so different from the grid that it won't charge the PW? It seems that the PW has its own built-in limiting as to how much juice it will accept from the grid, around 4.1 - 4.4Kw or so for our single PW system when the PW is low, and then the gateway apparently shuts this down once the battery charges up. Or maybe all this lives in the PW...what difference does it make where it's located? Since the grid can supply up to 200a, or about 48Kw on demand, something in the gateway (or the PW) has to be limiting how much the PW will accept to charge itself and the run the house simultaneously. Mine seems to be limited to about 4.4Kw max feed from the grid to the PW. So how could my gen, which puts out a comparatively small 7500w, or about 31a do any damage? Shouldn't the same limiting apply and reduce the demand to about 4Kw and let the gateway divide it as it does with power from the grid? The gen puts out a clean enough sine wave to run computers, routers, EOP networks (I have 4 nodes) and so forth, so why can't you charge the PW from the gen, if you ran the gen into the gateway? It would seem the PW should limit the feed from the gen the same way it limits the feed from the grid. It would just require another DPDT center off switch at the gateway to select grid, gen or nothing to feed the gateway, and I think I'd have to not use the switch that puts the whole house on the gen...just let the juice come through the gateway. Now I have no proof of this nor any supporting technical data, but I wonder, as do you, why this is so?

I don't think I'm going to test this out because I'm sure this will totally invalidate my warranty, but I think this is a legitimate question. Suppose the grid was out, and there was little sun. That happens here in PR. The battery would be almost dead by sundown. My gen puts out enough to light up the house and still has plenty leftover to charge the battery but Tesla will tell you "No, no, you can't do that." If we had a day like I just described, we could be running the house and charging the PW from the gen, and then shut the gen off at night and use the battery and not have to listen to the gen noise. Is there something here we're not getting? How does the gateway know it's on the gen, not the grid? I'm sure some of our local experts here will chime in and explain this conundrum.

 

[jaggedben]

... I don't know if the meter passes current in both directions, ...

It does. It has to. If it didn't then it could only be because it disconnects you, in which case you'd no longer be in sellback mode anymore.

The point is that the grid can pretty much always absorb that backfeed. Of course if there's too much backfeed then grid conductors or transformers could get damaged, but it's extremely unlikely for a residential sized PV system to accomplish that in a short time span.

but if the system is in a sellback mode, (battery full, lots of sunlight, grid connected to gateway) which we've experimented with several times just to see what happens, an "ERROR" message appears on the meter LED screen. I don't know what they see at the head office but it can't be good. And there seems to be no way to tell if it 'spins backwards,' since you can't read the numbers when the error message is on the screen.And what I don't need at all is PREPA coming my house and asking us do we have this permit, or that permit, (they're permit-crazy here) or any number of nonsensical questions.

For those who want to compare notes, in PG&E territory the meters show a negative power reading or they have a little part of the display that shows 'Delivered' or 'Recieved' (by the utility that is). And if you backfeed without permission they don't care in my experience, at least if it's just a few weeks. They will just take the free energy you're giving them and charge your neighbors for it while you wait for your permission to operate. Meanwhile they still charge you for what you import. I've heard that other utilities charge you for your exports too. (This may be because they program the meters to count up in either direction, to keep people from cheating by turning their meters upside down). It all just depends how they've programmed the meters.

As far as the two meters, it may be that PREPA's compensation scheme is based on solar production and not net energy consumption. But that has nothing to do with physical considerations.

 

[caribconsult]

It does. It has to. If it didn't then it could only be because it disconnects you, in which case you'd no longer be in sellback mode anymore.

The point is that the grid can pretty much always absorb that backfeed. Of course if there's too much backfeed then grid conductors or transformers could get damaged, but it's extremely unlikely for a residential sized PV system to accomplish that in a short time span.




For those who want to compare notes, in PG&E territory the meters show a negative power reading or they have a little part of the display that shows 'Delivered' or 'Recieved' (by the utility that is). And if you backfeed without permission they don't care in my experience, at least if it's just a few weeks. They will just take the free energy you're giving them and charge your neighbors for it while you wait for your permission to operate. Meanwhile they still charge you for what you import. I've heard that other utilities charge you for your exports too. (This may be because they program the meters to count up in either direction, to keep people from cheating by turning their meters upside down). It all just depends how they've programmed the meters.

As far as the two meters, it may be that PREPA's compensation scheme is based on solar production and not net energy consumption. But that has nothing to do with physical considerations.

We're not in California, and PREPA is world apart from PG&E. Our meters don't show 'delivered' or 'received' or negative numbers in the display, they just show meter#, some other code I don't recognize, and consumption, and "ERROR" if you try to sellback. Whether the current actually passes through the meter and goes to the local transformer is also unknown to me. As far as our other friends with solar, their systems came as part of a federal program subsidy that covered a huge part of the capital expense and it had to conform to both PREPA and Gov't regulations and I don't know what those details because we weren't considering a solar system back then, but I do know they use two separate meters to measure consumption and sellback. I have no idea how these individual meters connect to anything other than the consumption meter goes to a transformer. Where the sellback meter connects is unknown to me. Also, there's really no cash involved...if you sell back, you get a credit at the end of the year against your last bill, I believe, and the sellback rate is deplorable compared to what they charge for consumption.

 

[jaggedben]

We're not in California, ...

And the OP in this thread is not in Puerto Rico. And the details of how utilities meter backfeed are mostly irrelevant to the topic of why you can't backfeed a generator. My point in comparing notes was merely to demonstrate that the utility meter is a lowly waypoint for PV current feeding out to the grid. And the grid pretty much always has some amount of ability to absorb that current since it contains other loads nearby. Whereas a generator to my knowledge has zero ability to absorb current, as it would mean spinning the motor backwards and or some other phenomenon which it's engineering and physics wouldn't allow.

People reading this thread should understand that PV inverters are typically configured by default as current sources, which means that there always needs to be somewhere for the available current to go. A grid or a battery can serve the role of absorbing that current. It's difficult to see how a generator can do that.