Ideal xfmr(s) for connecting 480/277Y inverters (60kW total) to 12470/7200 grid

Status
Not open for further replies.
Hello, this is my first question on this forum. I've found so many good answers here I'm a little confused!

I have a rough draft (very rough some would probably say) of a one line drawing if anyone wants to see it.

About the inverters: 60kW total (which is also 60kVA...?, because the PF of the inverters is =1. Or...1 / 0.8 leading?0.8 lagging)
Consisting of four 15000TL inverters:
Output phases / line connections 3 / 3-N-PE
Nominal AC voltage 480 / 277 V WYE

The grid is 12470Y/7200.
We have single phase, 7200V to 120/240 service now, 200 amps.
It goes through one 25kVA xfmr, which is 208 amps, 4% over the service level. Not quite enough.
So...60kW of inverter power, when the current single phase (7200V) inverter on the pole is 25kVA, isn't going to sit well, as in, if we turned that PV on, BOOM, not good at all!

The lowest amperage 480/277 service offered is 100A.
The 12.47 to 480/277 xfmrs seem to be VERY expensive.
It seems like it would be less $$ if the POCO went with three 7200 --> 480/277 transformers, but I don't know their prices yet on everything.
BUT...60kW doesn't match up easily with common sizes of xfmrs.

20kVA or 60kVA xfmrs aren't common. 25kVA, 30, and 75 are.
Going by things I've seen here, there are different ways to do this. I'm looking for simplest.
The POCO is OK with a line side connection, so....
how important to both the grid people AND the overall balance of both system and grid is it to match the xfmrs and inverters as closely as possible?

SO...which one here? If they are OK with not going for a whole 100A of xfmrs on the pole, A seems "low", C seems "high" or "hot", B seems like it may be best match overall and for 100A service, but B is another 100+ panels and a lot more paperwork.

A. Four 15000TLs - 60kW - 72A ---> three 25kVA xfmrs?
B. Four 20000TLs - 80kW - 96A ---> 3 30kVA xfmrs? (96A being that 4% under thing again)
C. Five 15000TLs - 75kW - 90A ---> 3 25kVA xfmrs?
D. Three 24000TLs - 72kW - 87A ---> 3 30kVA?
E. Totally over the top- TWO 24000TLs with the same amount of panels as A, for a 1.66 ratio of panel watts/inverter output. ---> One 50kVA transformer. 4% under again. File that under "so crazy it might work", I have to try a simulation with that one. Hmm.
F. other?

They aren't kidding when they say these inverters are "flexible"!
There was a comment in this forum about using two larger and one smaller transformers for the grid connection, not sure how one would "balance" that to the grid's benefit.
I also saw something about having a Delta between the two wyes to...trap harmonics?
Also something about "wild leg" wiring for 120/240.

The going from 480/277 Y to 120/240 part seems straightforward enough, any thoughts there would be appreciated too.
This, for instance?

http://www.temcoindustrialpower.com/products/Transformers_-_General/HT0213.html
or...this?

http://www.temcoindustrialpower.com/products/Transformers_-_General/HT0193.html

We're getting another xfmr to supply the current 120/240 wiring off of the PV output- the PV will be going in line side. If we do this right, with some batteries, there won't be a drop of power used all year.

But then...too much inverter will sit there wasting power at night, which adds up eventually.
Too little is of course also a bad idea, but...
whew, it really gets complicated!

Thanks, if anyone can help out here.
These inverters seem like you could program them to put out a few kWs while you sleep and THEN turn on the coffee maker at 7AM!


http://www.wseas.org/multimedia/journals/power/2013/035702-208.pdf
2. Step-up transformers for conventional PV plants
The cost of the step-up transformer for a PV plant can be evaluated as a composition of four contributions, namely:
? initial cost
? cost of the energy wasted due to transformer
overloads
? cost of energy wasted due to transformer
efficiency
? cost of energy wasted due grid instability

2.3 Energy losses due to transformer efficiency
The power wasted due to no load and copper losses
during the k-th interval can be computed through the
following equation, assuming constant the
amplitude of the inverter output voltage:
(equation won't paste correctly)
 
a HA!

a HA!

Never give up! It isn't the 4% I mentioned before, it's 10%.

One question semi-answered. 60kW of this particular inverter must go thru a 66kVA transformer minimum.
The maximum apparent power of all inverter (SPV = ?SPVi) connected to the low voltage side is:
? Less than or equal to 90% of the rated power of the transformer (SXMFR) SPV ? 0,9 ? SXMFR.

http://files.sma.de/dl/7418/STP24-US_MV_Trafo-TI-en-10.pdf

But I'm still not sure about whether "less than" is enough of an effect to worry about.
60kW thru a 75kVA xfrmr would be 80%.
Hmm.

Hey! Wrong voltage here, but they do make them.
Would this be a "special order" type thing? I don't want to want 2 months for it...
http://www.lcmagnetics.com/transfor...three-phase-auto-transformer-66-kva-pn-19063/
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
I'm really not qualified enough to talk about transformers but I have a couple comments and questions.

1) Seems like you are not adding the NEC required 125% factor on your required amperages. So, for example, your four 20000W inverters at 96A would really require a 120A or larger service.

2) If the POCO is going to have to upgrade the service or install an entirely new one, then of course you want to plan on the PV being connected line side. That detail should really should not require their permission, although I understand some jurisdictions have trouble with the idea of just following the NEC.

3) Why not just tell the POCO that you want a 480/277 service, and you will backfeed X amps, and see what they tell you is necessary? The good part of this is that they are totally responsible for the transformer and connections being specified correctly, right? For that matter, why not let them do the work? (Is there even an option?) I guess they may gouge you, and I don't know how it works in your area with respect to who contracts for installing services. But personally, I would rather just give specs to the POCO and let them engineer it, than be responsible for figuring it out.

4) Related to number 3... Where would the service point be? And how will it be metered?

5)
If we do this right, with some batteries, there won't be a drop of power used all year.

But then...too much inverter will sit there wasting power at night, which adds up eventually.
...
...
These inverters seem like you could program them to put out a few kWs while you sleep and THEN turn on the coffee maker at 7AM!

Whose inverters are you talking about here?
 
1) Seems like you are not adding the NEC required 125% factor on your required amperages. So, for example, your four 20000W inverters at 96A would really require a 120A or larger service.

2) If the POCO is going to have to upgrade the service or install an entirely new one, then of course you want to plan on the PV being connected line side. That detail should really should not require their permission, although I understand some jurisdictions have trouble with the idea of just following the NEC.

3) Why not just tell the POCO that you want a 480/277 service, and you will backfeed X amps, and see what they tell you is necessary? The good part of this is that they are totally responsible for the transformer and connections being specified correctly, right? For that matter, why not let them do the work? (Is there even an option?) I guess they may gouge you, and I don't know how it works in your area with respect to who contracts for installing services. But personally, I would rather just give specs to the POCO and let them engineer it, than be responsible for figuring it out.

4) Related to number 3... Where would the service point be? And how will it be metered?

5) Whose inverters are you talking about here?

Howdy, thanks in advance.
1. Could you elaborate on that 125% factor? I tried looking that up, and it seemed to be about fuses. Sorry if I was confusing, I'm not getting into the fuses yet.
Just kind of wondering...having a transformer twice the size of your output would be wasting money, but also less efficient electrically.
But you can't go 1:1, 75kW into 75kVA of xfrmrs, that's too hot. I guess that would even depend on the type of xfrmr. I'd like to get rid of the old oil-filled can and put something more efficient, perhaps on the ground. More on that.

2. I was wondering about the line side thing originally, but if it is better...they should of course allow it! Seems like they are OK with it. Maybe the guy already knew how snazzy these inverters are (SMA Tripower US models)

3. "see what they tell you is necessary?"
That's sorta what I'm worried about- necessary doesn't...necessarily always mean "best".:?
Plus, I'm sort of trying to impress them, in order to speed things up- it could take them 2 or 3 weeks to design it, but if I show them something, we might be able to skip that whole part.

4. The service point is...weird. It's 250+ feet from the road, so we have 250+ feet of wires straight from 1 of the overheads, two poles, to the 25kVA xfmr, then overhead wires from there to a barn with the meter outside a shed and main panel inside the shed. We might be able to sneak it underground from the bottom of the pole instead. The PV power is coming from the other side of the barn and going into the grid, and will feed the barn through a 480/227 to 120/240 single phase or maybe 240V transformer. The 60kW of PV is sized to cover the usage in December/January, which ends up, because of the seasonality of the load, being 4x the usage in June. Which = $$. Which is nice for us...BUT we end up putting out a ton in summer when the grid is strained, so good for them too.

I suppose most residential xfmrs are on public property. I did impress the guy a little by figuring out the feed wires are 7200V. They sent me a pre-application info sheet that said "12.47kV radial feeder".
Side note of historical trivia- rural feeders used to be single phase 7200V, fed by 3 phase Delta 7200V. But as demand grew, they changed them all (or most) to 3 phase 4 four wire 12.47kV by adding another wire for neutral/ground.
Kind of a neat trick!
The early rural systems
were 7200-volt delta, three-phase in most cases, with
7200-volt single-phase branch lines. These systems
have now given way to 7200/12,470Y-volt three-phase
four wire systems. This change became necessary due
to the fact that existing 7200-volt delta, three-phase
lines became more heavily loaded and also due to the
fact that rural lines have been extended to much greater
length than formerly.

http://www.cooperindustries.com/con...urces/library/201_1phTransformers/R201902.PDF


5. These puppies. Germany does have a pretty solid rep in the engineering things department.
http://www.sma-america.com/products...s-20000tl-us-24000tl-us.html#Downloads-137610

Thanks again!
 

jtopham26

Member
Location
US
usually only systems large enough to justify the added cost are interconnected at medium voltage. you should consider some single phase inverters with a single transformer. I would recommend Schneider electric inverters which are sized to even output current increments 50/75/100a.
 
usually only systems large enough to justify the added cost are interconnected at medium voltage. you should consider some single phase inverters with a single transformer. I would recommend Schneider electric inverters which are sized to even output current increments 50/75/100a.

Howdy- I'm not quite awake yet, but I don't get what you mean by the first sentence there. The grid voltage is 12470Y/7200, so any type of inverter has to be stepped up to either/or 12470 / 7200V (3 phase / 1 phase).
With the transformer we have now, I guess we could step the 480/277 3ph output down to 2 legs of 120V and put it through like that, but that doesn't make sense, and limits us to 25kW.

We could get 120/240 output inverters, but that's sort of a lot less efficient than 480/277 3 phase output, and I don't think you can go over 600V DC with the Schneiders?
The SMAs, you can run them at ~789V DC or half that, and it doesn't make too much difference. But compare the SMA output to other companies output at below 600V DC/120 or 240AC, and there is a difference.
That's using a simulation like SAM or PVSyst, or SMA has a good one on their site.

I think I see what you're saying-
If we got one 45kVA 7200V - 120V xfmr, it would be 624A @ 120V.
We could put 6 100 amp Schneiders into it.
But why not just get 3 25 or 30kVA xfmrs and go with 480V 3ph inverters?

It's the 12.47kV to either 480 or 240 or 120V xfmrs that are big $$.
Along the lines of $20G for a single 12.47 >> any of those (I think even $22G for the 12.47 >> 120 model.)
Whereas 7200V >> the same are around $5G each.
I could be off there, but I guess that makes sense due to amount or size of wire?
Of course the grid people will have real wholesale prices, but still...

To me, it seems "not enough", if we're getting upgraded service for the purpose of cranking out PV power, to stay with single phase.
They're going to be running 250+ feet of wire to our pole, and it will be all three mains from the grid, so the power is going from inverter level to grid level on our property, coming in 3 phase.
Thanks for the suggestion, I will try a simulation using your setup and let you know the results.
 

jtopham26

Member
Location
US
I really don't understand your scenario I guess... if you have an existing service just backfeed that with your inverters. you shouldn't need a transformer. can you post a single line of your existing installation?
 
I really don't understand your scenario I guess... if you have an existing service just backfeed that with your inverters. can you post a single line of your existing installation?

Yep, here you go- the service now is just 200 amp 120/240 single phase, meaning two lines of 120 to get the 240. It comes from a single 7200V to 120V can on a pole type xfmr which is 25kVA.
I guess that puts out 2 lines of 120V from the single overhead line from the grid (grid has 3 lines of 7200V and a neutral, so grid setup is 12470Y/7200). We get one of those three and the neutral at our pole now.
Meaning, I think, the 7200V to 120/240V xfmr simply puts out 2 lines of 120 and you combine them at the panel for 240V.

That's too small to backfeed with 60kw (which is also 60kVA when the inverter power factor is =1.)
So we're getting 480Y277V service, 100 amps, which would be ~90kVA worth of xfmrs.

I suppose we could get a 75kVA single phase 7200V to 12/240V, but the inverters are 3 phase and 480/277, while the load is one phase, and the PV output is = to or 2x the load, so going 3 phase into the grid and then getting a smaller xfmr for the load side seems more efficient than putting the 3 phase PV Power into a transformer to make it one phase and then back into the grid through the correctly sized xfmr.

90 kVA xfmr is kind of too much for the 60kVA inverter output- I'm wondering how much of a problem that is.
I'm assuming that we'd get 3 xfmrs (7200V to 480/277) instead of one (12470 to 480/277) when we go to 480/277 3ph, just because one is seemingly more $$.

Hopefully, three 25kVA xfmrs to make 75kVA total will be..."not too much" size wise. 80% PV output to xfmr size seems like a good number if 90% is the max.

Thanks for your help.
 
Hey man- this is a typo- i meant too much TRANSFORMER will sit there wasting power at night. The inverters don't waste any.
Whoopsie.

What I meant was:
You've got a xfmr sized to the PV output.
So at night, when the output=zero, and the load is zero or close to it, the transformers are wasting watts, like say .3 kWh or something.
So that's an issue for the grid, because it's on their side of the meter- they don't want to oversize the xfmr too much.
Say the load at night is a few refrigerators and clocks and just small stuff- when the refridges aren't running, the load is almost nothing.
If you have a battery bank it might be =zero.
So when load =zero and output is also zero the xfmr is sitting there on the pole like a lump, wasting more and more the more oversized it is.

5) me- "But then...too much inverter will sit there wasting power at night, which adds up eventually."

you- Whose inverters are you talking about here?
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
Howdy, thanks in advance.
1. Could you elaborate on that 125% factor? I tried looking that up, and it seemed to be about fuses. Sorry if I was confusing, I'm not getting into the fuses yet.

Your fuses or circuit breaker for the inverter is required to be sized at 125% of the inverters rated maximum output. (See 690.8(A(3) and 690.9(B). or see 705.60.) The service is required to be rated for the total value of the OCPDs for the inverters. (See 705.12(A).)

Thus if you have a 20000W inverter that connects at 480V 3-phase, it presumably outputs 20000/480/1.73 = 24A. (Technically it is whatever the inverter spec sheet actually says, but it's usually close to the calculation.) 24*1.25=30A for the breaker, so if you have four of them your service is required to be rated 120A not 96A.
 
thanks! i get it now.

thanks! i get it now.

(See 705.12(A).)
That was the missing link, thanks! :happyyes:

Thus if you have a 20000W inverter that connects at 480V 3-phase, it presumably outputs 20000/480/1.73 = 24A.
24*1.25=30A for the breaker, so if you have four of them your service is required to be rated 120A not 96A.

So 1 inverter at 18.1A. (The 15000TL option)
18.1 * 125% is 22.6A.
22.6 * 4 = 90.5A.
Service required = 90.5A.

Three 30kVA xfmrs - 108 amps? I think.
Three 25kVA xfmrs = 90.3A.

So!! Four 15000TLs would put out 83.8% of the capacity of 3 30kVA xfmrs!
Which seems to work, if 90% is the max.
Still wondering- if 90% is "the max"...is it also the "best"?
That would...depend on the efficiency curve of the particular transformer maybe? :huh:

Thanks again.
 
whoops i goofed again.

whoops i goofed again.

me- "So!! Four 15000TLs would put out 83.8% of the capacity of 3 30kVA xfmrs!"

No. That's not it.
Four 15000TLs will put out 72.4 amps and need 90.5 amps of fuses and therefore *maybe* 90.3 is OK?
If so, and like I said, "Three 25kVA xfmrs = 90.3A."

That means "Four 1500TLs would put out 80.2% of the capacity of three 25kVA xfmrs."
Which is still good.
I think?
 

BillK-AZ

Senior Member
Location
Mesa Arizona
This thread really needs a 1-Line diagram that shows:
  • Service
  • metering
  • 12.47KV/277/480V transformer (if used at this point)
  • Service disconnect (fused)
  • Utility disconnect that may be required, often unfused
  • AC Combiner panel with buss ratings
  • Transformers if service is not 277/480V or one large transformer
  • Inverters
  • PV array

How does one use three transformers for four inverters? Single phase transformers? How is overcurrent protection provided for the transformers?

If you work out the costs of the various NEC compliant designs, you may find that having the POCO provide 277/480V service will be lowest cost considering all the unmentioned other components.

Systems like this really need a PE.

The economics of the situation are questionable.

The inverters you mention are not compatible with batteries.
 
This thread really needs a 1-Line diagram

How does one use three transformers for four inverters? Single phase transformers? How is overcurrent protection provided for the transformers?

If you work out the costs of the various NEC compliant designs, you may find that having the POCO provide 277/480V service will be lowest cost considering all the unmentioned other components.

Systems like this really need a PE.

The economics of the situation are questionable.

The inverters you mention are not compatible with batteries.

Howdy, thanks for the questions.

Diagram attached- I copied the "sample" from the POCO's site. I guess the 100 amp fuses should be 90A?

4 inverters/3 xfmrs- Doesn't the AC combiner panel turn the 4 into one 3 phase output @ 72 amps?

Overcurrent protection- Do you mean AC backflow into the xfmrs? I think it's built in. Or if you mean DC they make a connection box.
From the manual-
To reduce the risk of fire, connect only to a circuit provided with 50 A maximum branch-circuit
overcurrent protection in accordance with the National Electrical Code? (NE, ANSI/NFPA
70).


you may find that having the POCO provide 277/480V service will be lowest cost

That's kind of what I'm hoping. Getting a bigger single phase 12/240 xfmr and stepping the 480/277 inverters down might be cheapest overall, but you lose too much of the output by going 480 to 120, especially over 25 years.
One other slightly doofy question- when people refer to 277/480 and 480/277...am I missing something? Just...input vs output? I don't think that's the answer. So it must be...480/277 is wye/star and 277/480 is Delta?

Systems like this really need a PE.

Yes, very true. The POCO guy is one, and then on my end I'm trying to do his work, then show one that the electrician knows, so he can say "yep, that's how it should go" and stamp it.
To save $$ basically. I could get a dozen different "ideas" from the same number of "installers"...but it could be that none of those are BOTH least $$ and best efficiency.

The economics of the situation are questionable.

Ay, there's the rub. The basic idea is turn the output of $$ for the electric bill into an equal amount of input of $$! :)

The inverters you mention are not compatible with batteries.

The thought is the battery bank and inverter/s for that will be on the load side of the L in the diagram, or main service box for the account.
The transformer for the 120/240 main service box isn't in the diagram yet- it should be between I and L in the diagram, I think.
So the PV inverters won't even know the batteries are there.
Or maybe they will, not sure of the details of the digital connections between the SMA units.
But if the inverters are supplying the load, and the entire battery bank is 100% part of that...it should work?
I'm going to get SMA's help on that- hopefully "I want to buy your stuff, will this work?" will get a quick answer.
 

Attachments

  • FORUMONELINE.jpg
    FORUMONELINE.jpg
    22.4 KB · Views: 0
Last edited:
Last edited:

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
That didn't quite work, you can't read it.
Here's a link. This should do it.

http://1drv.ms/1K2odyo

Also, just found this- there is a way to hook them up to a battery bank.

Hi
Yes, it is possible to configure a Sunny Tripower with three Sunny Islands and batteries to run as a backup system. Thanks so much for reaching out!
Best,


http://www.smainverted.com/2013/07/...ilable-in-north-america/#sthash.8TbEOprN.dpuf

You understand that the Sunny Islands are also inverters, right? The Tripowers wouldn't be connected to the batteries but to the AC bus powered by the Sunny Island inverters.

There is also a limit to how much PV inverter power you can push through a Sunny Island and they are 120VAC single phase (Lx and N). You would need to have a transformer for each Tripower to convert the output from 277V to 120V line to neutral and three Sunny Islands per Tripower inverter, and then there's the cost of the batteries. Yes, you can do it but it won't be cheap. Nowhere remotely close to cheap.
 
Last edited:

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
You understand that the Sunny Islands are also inverters, right? The Tripowers wouldn't be connected to the batteries but to the AC bus powered by the Sunny Island inverters.

There is also a limit to how much PV inverter power you can push through a Sunny Island and they are 120VAC single phase (Lx and N). You would need to have a transformer for each Tripower to convert the output from 277V to 120V line to neutral and three Sunny Islands per Tripower inverter, and then there's the cost of the batteries. Yes, you can do it but it won't be cheap. Nowhere remotely close to cheap.

To follow up: Of course, this leaves you with a separate 208V three phase output for each triad of Sunny Islands, so you'd have to combine their output in an AC combiner before going into your 208VAC to medium voltage transformer.

For four 15kW Tripowers you need:

(4) Tripower inverters
(4) 15 kVA 480/277V to 208/120V transformers
(12) Sunny Island inverters
(4) large battery banks (you cannot "fool" the Tripowers into running offgrid with tiny battery banks)
(1) AC combiner panel
(1) 208/120V to medium voltage transformer of at least 60kVA capacity

And then there is the cost of the interconnection itself.

Nowhere remotely close to cheap.
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
I do not see the reason (in principle) that you need a separate trio of SIs for each tripower if you are going to connect the tripower outputs anyway.
Each tripower would independently qualify the "grid" when connected to it.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
I do not see the reason (in principle) that you need a separate trio of SIs for each tripower if you are going to connect the tripower outputs anyway.
Each tripower would independently qualify the "grid" when connected to it.
The way a Sunny Island interconnects has the PV inverter on the microgrid (protected) AC bus and all power that is not used in the protected loads panel from the PV inverter flows through the SI when the grid is up. The SI is a 5kW inverter and you cannot push more through it than that. I'm pretty sure that each Tripower will need three of them, one for each phase for a total of 15kW. I don't think you can connect all the Tripowers on the AC side of the SI's.
 
Status
Not open for further replies.
Top