Grid tie inverter - parallel AC sources 1

[aegcmac]

I need to grid-tie a bunch of about 50 small inverters that process DC input of say, up to 17 Vmp and 6.5 Imp and back feed this through a small sine wave inverter rated to +/- 120 watts into an existing 120 VAC / 60 Hz AC grid. The AC grid is not a public utility, rather the end-user's AC supply, so IEEE or U.L. standards for back feed do not apply. The key here is the functionality, not bureaucratic approval of the precise quality of the back feed.

It's easy to get DC to AC, and relatively easy to get the AC in a sine wave, but synchronizing to another AC sine wave source for a parallel connection seems complicated.

Is it that complicated? If you could live with harmonic distortion of up to 5% - 7% and efficiency of the conversion from DC to AC in the 85% to 90% range, is it really that difficult to process the DC and parallel these two AC sources?

Or, if we have the small sine wave inverter, we need to parallel it with the AC grid on site. The key here is being able parallel the two AC sources.

A company in Europe, for example,

http://www.mastervolt.com/sunmaster/soladin.asp

has a mini grid-tie inverter that processes the DC input from a solar panel, but not available in 120 VAC/ 60 Hz. They seem to be doing it almost exactly as I need, but in more/less a Radio Shack form of package, with a nominal 24 VDC input and the output is European 230 / 50Hz. I don't need the packaging and all the bureacratic approvals, just need the function in a 120 VAC / 60 Hz environment.

Can anybody help?

 

[Bung]

Don't sine wave inverters for this type of application automatically follow the "mains" frequency? Of oourse, if they are stand-alone without any ac driving source (genset or utility) then you would have a problem mainitaining the frequency, but the inverters should all lock in to each other and follow along. But some are deliberately set to drift to guarantee that they drop off under "loss of mains" conditions, so these types will all follow each other into shut-down.

Bung

Bung
Life is non-linear...

 

[aegcmac]

Bung

Yes, a grid tie inverter locks with the AC main and in the event of loss of the main or say, like a brown out where line voltage drops below a certain level the inverter instantly isolates from the main.

So whether the inverter synchronizes with a public utility AC main or a user-owned AC main this feature is still desirable.

The difficulty seems to be with getting the two AC sources to synchronize fequency etc with minimal distortion, since each AC source seems to more/less be on its own clock - so to speak, like getting two clocks to tick at precisely the same time.

aegcmac

 

[jbartos]

Question: Would you consider rotary converters?

 

[busbar]

It may help to provide the more information on the context of the application, if intended for something besides a public utility. What are the anticipated individual and total capabilities of the inverters? What are ratings of the non-inverter part of the system?

 

[aegcmac]

I need to synchronize a much smaller sine wave inveter to an AC main fed by a larger 5000 watt sine wave inverter, both are 120 VAC 60 Hz output. The small inverter output is between 70 and 140 watts, and it's supplied by a DC input source that's variable, a photovoltaic panel, so when the panel is in full sun I need to backfeed the small inverter into the larger main. I'm good on every aspect of this except the synchronization of the two AC sources. I need a PCB that can sync these two AC together automatically over say 5 to 30 seconds, and not manually. So, it seems it has to sample the main, possibly adjust the smaller inverter's voltage and frequency to match the main and then connect the smaller inverter at precisely the right time to keep the phase angle as close to zero as reasonably possible as well as disconnect if the main is lost or the PV panels stops producing acceptable input to the small inverter. Disconnection from the main or disconnection due to lost PV panel input doesn't seem to difficult, it's the synchronization of the two AC sources.

There are much larger auto-sync units available from Basler, Beckwith, etc, but these are like a tractor trailer for auto-sync/switching several large mains and large generators, etc, when all I need is a little VW bug.

aegcmac

 

[buzzp]

I am no expert on these types of issues but I would think a sync check relay should work in conjunction with your breaker. Also, you may want to consider the voltage amplitude of the two systems as well before closing in the breaker. Just a little thought.

 

[aegcmac]

Yes, thanks. I'm thinking I have to be able to sync-check voltage, phase angle and frequency, but here's where I get confused, like does the AC main then dictate to the small inverter in this process, so the circuit has to be able to force the smaller inverter into sync with the main or does it simply bring both of them together in frequency and phase angle and when so, it closes the breaker and locks them in together?

But if the AC main is at 120v and the small inverter is at 117 doesn't this drag down the AC main? I can't see the circuit tryng to lift the 117v to the 120v, so I'm not sure what happens here to sync the voltages. Although, it would seem the voltage difference would be the least important of the three; frequency, phase angle and voltage, if they were in a fairly tight voltage range.

If anybody knows exactly how to do this I know a company that would pay a few $K for the PCB

 

[busbar]

It is possible that off-the-shelf equipment (so-called grid-tie inverters) with “bureaucratic approvals” will do what you have described, and that in-place IEEE/UL/NFPA standards could save reinventing the wheel. You might check with vendors’ applications people on lists like that at

http://www.consumerenergycenter.org/erprebate/equipment.html

and explain your needs.

 

[aegcmac]

Thanks, but that's actually why I'm here trying to see if anybody knows how to do it on a smaller scale. I can get a sine wave inverter that will grid tie, but they start on a scale of 2500 rated watts and cost $1500+ each because they also have internal battery chargers and auto-transfer switching for UPS applications. There was a company that had a small 100w grid tie unit in the U.S., but they're gone. So there's a huge gap between what's out there as far as scale and cost and what a specialty application like this needs. The only thing close is the unit I first mentioned in Europe, but it's not available in 120v/60Hz, and they won't make 50 just for me. So, that's why I'm trying to look everywhere for a more scaled solution.

 

[busbar]

Long shot — you might try

http://www.exeltech.com

 

[Bung]

Ther are lots of small inverters for connecting solar systems of the size you describe to ac systems - try BP Solar, for one (in Australia) and there would be lots of them in Europe. I don't know why; sunshine seems to be in short supply there. Even Tandy might have something?

The utility I work for has several dozen of these things in customer premises, and they just switch themselves in (and out) pretty much as they please.


Bung
Life is non-linear...

 

[aegcmac]

Thanks Guys.

Yes, Exeltech would be long shot. They know how to do it, but again, their whole product scale is for much larger applications, so like the other MFRs they don't offer it in a smaller scale and would have no desire to do it for just 50 units.

In the U.S., there was the one small 100w grid-tie inverter, so there are many of these grid-connected throughout the U.S., but they're out of the business completely. There are a few of these units floating around, but not enough for what I need and they're 100% epoxy-filled-encapsulated, require twice the input voltage I'm needing so I can't go there. All other U.S. units capable of grid-tie are scaled much, much larger.

I'll probably try to talk to a few of the European MFRs making the small units to see if they have something or maybe send out an RFQ into Korea, China etc, where these units are likely being made in the first place. Or go to Plan B.

 

[aegcmac]

Thanks BUZZP:

I've checked further into sync check relays and may have found something I can work with. It may not be a home run or exactly as I need it, but I think it will get me to third base, and I think I can get it the rest of the way with a couple small, inexpensive peripheral items.

 

[jbartos]

Suggestions to aegcmac (Electrical) May 7, 2004 marked ///\\It's easy to get DC to AC, and relatively easy to get the AC in a sine wave, but synchronizing to another AC sine wave source for a parallel connection seems complicated.
///Not quite. In old times, three bulbs were used. When all three bulbs were dark the synchronized switch was turned on. It is assumed that the inverters will have a highly accurate clock for triggering circuitry. Alternately, an electronic firm or an electronic hobbyist may be contacted for a custom made solution.\\

 

[GOTWW]

They, were selling solar panel piggy back invertors. Could not see them with a quick search. One of three idea's.
(1) Too many "guerrilla" solar folk were attaching A/C plugs and covertly back-feeding the grid.
(2) Instabile chatotic behaviour resulted from unit to unit cross-talk.
(3) Too many farmers got electrocuted, handling 110Vac wiring connected to panel-invertors. (input or output, only 5 ma is all it takes, night or day, beer can or not).

The parallel invertors I have seen require a master that provide supervision and syncornization, by way of a cable.

Generally codes require a single of grouping of disconnects for personel safety. For, aegcman If your main invertor shuts down then, its all up to the childern to make up their own rules.

 

[aegcmac]

On a privately owned system there's less of a safety issue since (at least in this case) the operator(s) know and fully control the system, where in a public grid you have all kinds of potential issues. The modern grid tie or grid interactive inverters shut down in like 30ms if the line drops out too much or shuts down - an anti-islanding feature. I think the bigger problem is folks using fueled generators during power failures without adequate circuit isolation, auto-transfer switching etc, but I'm sure there are home-brew inverter connections that don't have the anti-islanding safety provision or adequate transfer switching too.

When two or more battery based inverters can be and are paralled, yes, they do have a Master and the rest are slaves, usually via a control card, etc and of course they have to be identical units. I just discovered Honda now has a few small generators like their little EU1000, 2000, 3000 that can be paralled to another EU1000 etc. They're promoting these units as having "inverter technology". But the modern grid-tie inverter designed solely for back feed that doesn't use a battery eliminates the Master/Slave requirement and just processes all available power from solar panels and each handle their own intertie, so you can just add as many as needed, different size or brand units, etc, as long as they're made for the correct line voltage and frequency.

I had read a brief article on the old light bulb approach - thanks. It was a brief overview, so I'll have to do more digging on this. It does seem that the very best place to make the parallel between our larger inverter and the smaller ones as they would need to come on line would be at the point of rest. I'm thinking that if the voltages are real close and pretty stable and the frequencies are generally always within 1 or 2% of each other, then the real main (most important) issue is to eliminate phase angle. The place I'm still a bit confused is that assuming both AC sources are at rest, can I be sure that when we connect them with a rapid switch that they both go toward negative (or positive), for example or is it still possible one could go positive and the other negative from this rest point and I end up with serious distortion. Maybe that's what the three light bulbs help establish. I admit I'm also not 100% clear on the first few milliseconds of AC, whether AC always first flows toward the negative or the positive or is it random. So I'll try to better understand the light bulb concept and how with this concept we can use the sync check relay. The relay requires sync of phase within I believe 5%, so I'm also wondering rather than bring the two active AC sources together if you couldn't just use an ON/OFF switching device that turns on the smaller inverter's AC for a second and if the sync relay likes it then we simultaneouly bypass the switch (or keep it closed) and the sync relay closes to lock them together, and if the sync relay doesn't like it after a second then the switch turns OFF/ON to try again. I'm wondering if with this type of almost random OFF/ON if sooner or later you can't get the phases close enough to trip the sync relay. Problem is of course, this could take a few seconds, minutes or hours etc. It would seem you would have a say, 1 in about 120 chance of getting it right each time at 60 Hz. Probably a completely crazy thought, but I thought, well, 30 tries a minute means that after 10 minutes you have 300 attempts, which is nothing for a high quality switching circuit, and a 10 or 15 minute lag would not be an issue in this project, but it would definitely be like a dice roll if possible at all. I guess it could potentially take days to get a phase sync. So a more controlled, faster way to sync them is best. I'm just brain storming all possible ways to do it relatively simply and automatically without needing complex logic circuits, numerous ICs and such. I'll get more info on the old light bulb method, and I am trying to find someone who can help complete the design on this like a hobbyist or someone who can consult privately (discreetly if need be). Our company will compensate, but the EE firm's executives will try to make ten times what the person makes who actually designs it, so they'd rather compensate a hobbyist or the actual designer than the executives.

 

[busbar]

aegcmac — You are probably very familiar with this stuff..

http://www.basler.com/downloads/intro_synch.pdf

http://www.basler.com/downloads/ula1bull.pdf

http://www.basler.com/downloads/Ump5bull.pdf

Many responders may be ingrained in the utility frame of mind dealing with rote high-value equipment damage and personnel safety.

 

[busbar]

[continuing] A problem may be the idea of multiple IEEE-929/UL-1741 components {a/k/a “consumer electronics”} for paralleling duty being a contemptuous and wholesale bastardization of some deep-rooted reliability and preservation issues, and not widely understood or accepted.

 

[aegcmac]

busbar:

Yes, thanks, I had gotten the Basler info after buzzp mentioned the sync check relay which was very helpful too. If it weren't for the over-scaled +/- List $1000 each for the auto-sync units or if the mini inverter I first mentioned was available in 120v/60Hz I'd be done. But now it's becoming more of a personal challenge to solve this for this scale of a specific, unique project. It's a private system, so we can do whatever we want if it works, and fortunately I have the time to pursue this a little while.

For the public grid, I do think the inverter MFRs are basically forced to kiss the A of the public utilities and in the process they are jumping through all kinds of hoops for IEEE, UL, etc, for the grid tie, so I can understand why they don't want to offer a very small, inexpensive unit that really can't be mainstreamed right now. The reality is that the leading inverters are extremely reliable, but they have to grid-tie to a grid that isn't. I hear some home owners are getting a constant 126 volts, some 112, etc. I imagine this happens to industry as well. There does need to be an acceptable uniform standard for back feed of course or else a sleezy MFR could just dump a bunch of garbage into the public line, but as usual, once the buracracy gets set in motion it just grows larger and larger, so MFRs are more/less scared to get too creative past what they already have IEEE or UL approval for. The costs to design and build is nothing compared to the cost for these approvals, which adds cost to the end product.

The fact is that if you had thousands more of these modern inverters grid tied with like solar, wind and hydro for example, you'd actually improve the public grid, but instead of clearing the way and encouraging more of these most states and public utilities throw out all the road blocks at the MFRs and also force the end user to obtain permitting, line tie approval, install a new meter, etc.

Since each small inverter has to be able to parallel to the main on its own, I haven't checked yet to see what the per unit cost would be for 50 of the BE3-25A Basler auto-sync units, with I think a List of like $1000, but I do have some of the sync check relays on their way to me to experiment with once I figure the best way to try to get the two inverters in sufficient sync to close the relay without a manual adjustment requirement...if I can. I'm willing to kill a few inverters or whatever in the learning process, but also hoping to find someone who knows exactly what to do before I create too much smoke