Backfeeding transformers with grid interactive micro-inverters

Status
Not open for further replies.

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
.

Frequency problems cause inverter to throttle down...
If frequency varies outside of trip points most inverters just shut down. SMA inverters that are designed and set up to coexist with SMA Sunny Boy battery inverters will intentionally throttle back with frequency changes so that they don't overcharge the SB's batteries in off grid mode. The SB changes the frequency on the AC bus to accomplish that.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
If frequency varies outside of trip points most inverters just shut down. SMA inverters that are designed and set up to coexist with SMA Sunny Boy battery inverters will intentionally throttle back with frequency changes so that they don't overcharge the SB's batteries in off grid mode. The SB changes the frequency on the AC bus to accomplish that.
I meant Sunny Island and SI, of course. The coffee hadn't kicked in yet.
 

Smart $

Esteemed Member
Location
Ohio
Great, thanks.

Here's another question for you or anyone else, loosely based on the title of the thread.
There are two bits of info provoking the question- my POCO says that 4160V service is "no longer standard but may still be available".
They also say that for => a certain size PV system which involves service upgrade (which happens to be the precise size proposed in one case) the customer "may be charged for *some* of the cost of the xfmers.
oh,wait, 3 bits, they also ask "is the customer supplying a xfmr" on the application.

So say the grid is 12470 L-L and 7200V L-N.
I'm assuming the possible 4160V service is delta, because it doesn't say possibly 4160/2400V.
I'm also assuming that when 4160V delta was available, it was supplied by 7200V wye.

One thing which IS available is a high-efficiency 480/277V (which is the voltage of the PV inverters and a standard POCO service) to 4160 delta xfmr.

If standard 480/277 service is supplied by a 12470V delta to 480/277V xfmr setup, would a 7200V L-N/grid >>> 4160 delta to 480/277 wye xfmr setup have any advantages to it?

That way the customer would be paying for the 4160 to 480/277 xfmr, and the grid would be taking care of the 7200V wye to 4160 delta end, however they did that when it was standard.

I'm not sure what the POCO uses for standard 12470V to 480/277 service xfmrs, but if the customer is going to pay for something, might as well be something...better?
Wouldn't a high-efficiency xfmr in that spot send more of the PV output to the grid overall, and pay for itself after....not too long?
Someone would have to run the numbers, but I can't fathom a double step down scenario as being more efficient than a single transformer step down scenario.
 

wwhitney

Senior Member
Location
Berkeley, CA
Occupation
Retired
I don't know, seems like there could be some current lost on the neutral
Current doesn't get lost. Charge is conserved, so you have Kirchhoff's law.

If someone goofed and hooked up a 208V inverter to a xfmr with the 203 tap going out- that would affect the PV output.
Would it? I thought the inverters are able to efficiently adjust their voltage output to match the grid for small deviations like 203 versus 208. It would mean a little more current for the same power, so the I^2R losses would increase slightly, but that's it.

Cheers, Wayne
 
ggunn-
I meant Sunny Island and SI, of course.


But what about the Tripowers? Don't they do the same, so if there was some kind of frequency problem in the grid (or xfmrs), it would affect the inverters?

Smart$
Someone would have to run the numbers, but I can't fathom a double step down scenario as being more efficient than a single transformer step down scenario.


For instance-
If the POCO was going to provide a standard 12470V to 480/277 transformer(s) that were 96% efficiency, but instead the customer bought a 4160V to 480/277 that was 98.75% and then the POCO's was 4160 to....7200V?
If not 7200V, 12.47KV....but if that POCO one happened to be 97.5% due to....some reason, the 98.75% and 97.5% together would be better than a single 96%, wouldn't they?

Here's another question- if the PV output is much more than the load, and the PV is outputting power 54% of the time (annually), it's really only pushing out to the grid say 40% of the time, because there's a variable load (which can even exceed PV output in late afternoon at times).
BUT- the PV is also putting out (annually) 3x the annual load.

Which is more important- the fact that the xfmrs are running towards the load (step down) 60% of the time OR the fact that there is 30,000kWh going to any loads and 60,000kWh going OUT thru the xfmrs (step-up).
It has to be the latter, so you'd therefore want to design any xfmrs as step UP?

Current doesn't get lost. Charge is conserved, so you have Kirchhoff's law.

Would it? I thought the inverters are able to efficiently adjust their voltage output to match the grid for small deviations like 203 versus 208. It would mean a little more current for the same power, so the I^2R losses would increase slightly, but that's it.

That's the first I've heard of Kirchoff. I think it may depend on the inverter model/company?
All I was really thinking was that (fault) current might sneak by the meter thru neutral somehow.

But besides that....if the grid side voltage was high, the inverters would adjust and put out less current?
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer

But besides that....if the grid side voltage was high, the inverters would adjust and put out less current?
Within their operating voltage range, the GTI will try to put out the same amount of power (whatever is available from the PV or limited by the GTI's rated output.
So if the grid voltage rises the GTI can deliver that same amount of power at a lower current, yes.
 
whatever is available from the PV or limited by the GTI's rated output.

Isn't there a limit on the DC input of the inverter also?

Anyway, say you have a 10kW inverter with the max recommended 12.5kW DC of panels.
It puts out 41.7A max at 240V.
But the idea with the 12.5kW of panels is that while you lose some due to clipping midday, the inverter is running at max output for more minutes per day.

So if the inverter is putting out its max 41.7A at 240V and then the grid voltage drops below 240, the inverter will match the lower voltage, but continue with the 41.7A, it can't go above 41.7A.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
So if the inverter is putting out its max 41.7A at 240V and then the grid voltage drops below 240, the inverter will match the lower voltage, but continue with the 41.7A, it can't go above 41.7A.

And conversely if the voltage goes higher than 240V then the inverter can put out slightly more power and still not exceed 41.7A. The DC input might have a power limit that puts a cap on this behavior, but that depends on the inverter.
 

Smart $

Esteemed Member
Location
Ohio
Isn't there a limit on the DC input of the inverter also?

Anyway, say you have a 10kW inverter with the max recommended 12.5kW DC of panels.
It puts out 41.7A max at 240V.
But the idea with the 12.5kW of panels is that while you lose some due to clipping midday, the inverter is running at max output for more minutes per day.

So if the inverter is putting out its max 41.7A at 240V and then the grid voltage drops below 240, the inverter will match the lower voltage, but continue with the 41.7A, it can't go above 41.7A.
The VA output profile changes when A equals IMAX.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
ggunn-
I meant Sunny Island and SI, of course.


But what about the Tripowers? Don't they do the same, so if there was some kind of frequency problem in the grid (or xfmrs), it would affect the inverters?

Sunny Boys have the capability (and you have to get into the firmware to enable it) to ramp down output as frequency increases or decreases from 60Hz to (I believe) 58Hz on the low side and 62Hz on the high side. At those points they shut down entirely. This is so they can operate off grid with Sunny Island inverters. In off grid mode, the SI's control the frequency of the AC bus and they monitor the state of charge of their batteries as well as the load they are supplying; as the batteries approach 100% SOC the SI's shift the frequency up or down (they alternate so that clocks on the AC bus have a fighting chance of staying nearly accurate) so that the batteries don't get overcharged. The SB's respond by ramping down their output. It's pretty slick.

Otherwise (if you don't have the off grid capability enabled) they just run full out and shut down when the frequency drifts out of the window they can operate in. No ramp.

Tripowers definitely have a frequency operating window but I do not know if they have an off grid capability whereby they would ramp down instead of just shutting down when the frequency drifts.

BTW, a transformer cannot affect frequency.
 

Smart $

Esteemed Member
Location
Ohio
...Smart$
Someone would have to run the numbers, but I can't fathom a double step down scenario as being more efficient than a single transformer step down scenario.


For instance-
If the POCO was going to provide a standard 12470V to 480/277 transformer(s) that were 96% efficiency, but instead the customer bought a 4160V to 480/277 that was 98.75% and then the POCO's was 4160 to....7200V?
If not 7200V, 12.47KV....but if that POCO one happened to be 97.5% due to....some reason, the 98.75% and 97.5% together would be better than a single 96%, wouldn't they?..
You have to consider that the efficiency rating is given at kVA throughput... not at zero or points in between. Efficiency % goes in the dump at zero throughput.

I doubt your setup will have full-rated throughput 24/7.
 
You have to consider that the efficiency rating is given at kVA throughput... not at zero or points in between. Efficiency % goes in the dump at zero throughput.

I doubt your setup will have full-rated throughput 24/7.

The thing that bugs me is that efficiency curves aren't really published by the xfmr makers.
I stumbled on a chart accidentally somewhere, it looked like under 20% of nameplate, the curve just takes a nosedive, so there's nothing you can really do about that aside from making sure you don't oversize the xfmr(s).

This one here says 50%....and then some lower voltage ones (same maker) are super-efficient and are rated so between 35-65%.
http://www.temcoindustrialpower.com/products/Transformers_-_General/HT5515.html

So...say this 75kVA model xfmr- if you had 60kW of inverters that were also 60kVA (3 phase), the 60kVA is 80% of 75kVA- you'd be above the "efficiency zone" when the inverters are at max, but then you'd be right where you want to be when the inverters are putting out 37.5kVA.

But then again, you might want 90kVA of xfmrs, not really sure there, that's the POCO's call anyway.

If the load you were running off the 60kVA of inverters was 20kVA max, you'd want 30kVA of those 35-65% models for stepping down to run the load, I guess 66% is close enough to 65%?
Product features


  • U.S. DOE 10 CFR Part 431 (and/or TP1) efficiency standards and the Canadian Energy Efficiency Regulations SOR/94-651 (and/or C802.2) efficiencies at 50% of rated load
  • UL Listed and CSA Certified
  • Efficiencies are calculated under a linear load profile
 

Smart $

Esteemed Member
Location
Ohio
I don't get into design much nowadays, but you get the gist of what I was trying to point out. Good luck with your endeavors...
 
Status
Not open for further replies.
Top