Single phase inverters connected to a three phase system

[electro7]

Hi,

I have been trouble shooting a solar system that is quite convoluted. It has single phase inverters connected to a three phase system (delta with a high leg). It is residential. The homeowner has not seen a change in his electric bill since putting in the second solar system. The first solar system was installed in 2006 with three single phase inverters, two of which were connected to his three phase system. The other to his single phase system. (He has two meters, one single phase system and the other which is three phase.) He did see about a 50% decrease in his bill after installing the first solar system.

The second solar system was installed in 2013. It has three single phase inverters, two of which are connected to the three phase system. (The biggest concern is the 3-phase system inverters and the true up bill he is getting from that meter.)

I have checked all the read outs on the inverter for "energy total" and everything seems to be working fine. The utility meter tech was there today and found that the meter is working correctly. We even turned off all other loads and checked what the meter was reading vs what was reading on the inverter screens for instantaneous power and it added up pretty close between the four inverters that are connected to that meter. As a matter of fact I think the meter was reading more that what the solar was putting out.

I checked voltage in a lot of configurations- two inverters off, the other two on; the other two invertes off, and the other two on; all inverters on and all inverters off. I am thinking that the inverters are throwing the voltage off so that there is a higher imbalance between phases. Therefore causing the motors to run hotter, harder and more inefficient. I am wondering if somebody could confirm this for me and help me out with a solution? Here are the readings I got:

No Solar On: Readings at the garage sub-panel ("Garage solar")
A-B 246; B-C 243; C-A 243 (Avg 244V, .8% Vmu)
Readings at the solar in field ("Field Solar") (240V to 208V step down transformer, 2-208V single phase
inverters connected)
A-B 212; B-C 213; C-A 215 (Avg 213.33, .78% Vmu)

"Field Solar" On, "Garage Solar" Off: Readings at the garage sub-panel
A-B 247; B-C 243; C-A 244 (Avg 244.67, .95% Vmu)
Readings at solar in field
A-B 213; B-C 215; C-A 217 (Avg 215, .93% Vmu)

"Garage Solar" On, "Field Solar" Off: Readings at the garage sub-panel
A-B 252; B-C 245; C-A 246 (Avg 247.67, 1.75% Vmu)
Readings at solar in field
A-B 211; B-C 215; A-C 216 (Avg 214, 1.4% Vmu)

All Solar on: Readings at the garage sub-panel
A-B 251; B-C 245; C-A 246 (Avg 247.33, 1.48% Vmu)
Readings at field solar
A-B 214; B-C 220; C-A 222 (Avg 218.67, 2.14% Vmu)

I did not take any readings at the service. I tried switching on of the two inverters in the garage sub-panel from phase C-A to B-C phase at it tripped both solar breakers in that sub-panel. Not completely sure why that happened. I tried it to hopefully correct the voltage imbalance.

The two "field solar" inverters are connected to phases C-A and B-C. The two "garage solar" inverters are connected to A-B and B-C.

Can anybody help me??

 

[GoldDigger]

The very first thing to confirm is that both meters are bidirectional and have been configured in the billing software for net metering.
The voltage readings are not going to tell you much. Amp-clamp readings (including at the inverters) will tell you more but still will not tell you for sure the direction of the power flow at the meters.
Are the inverters telling you that they are producing energy?

Voltage imbalance will not be a problem as long as all voltages stay within the inverters' tolerance.

 

[electro7]

The meter tech was there and confirmed the meters are working correctly. It did show the energy from the solar being back fed into the grid when we turned off all loads on the three phase system and left on only the solar.

Yes the inverters are showing power output at each inverter lcd screen.

I thought that connecting single phase inverters to three phase systems can cause voltage imbalance which can ruin motors. I am thinking that the motors are running hotter and less effecient.
http://ecmweb.com/code-basics/interconnected-electric-power-sources

Let me know. Thanks

 

[pv_n00b]

The voltage measurements are not going to tell you much unless the inverters are shutting down due to voltage being out of tolerance. I would put a data recorder on the electrical system to record PV production, load, and energy to/from the utility for a week or so and compare that to what the utility says it is seeing. You can rent one if you don't have one around.

 

[GoldDigger]

A voltage imbalance in the range you are seeing should not. IMHO, be seriously affecting your motor, but you could measure the phase currents to check on that possibility.
The easiest way to reduce the effect of the unbalanced production is, unfortunately, larger wires.
If you could get the inverters rearranged so that you are putting balanced production on the three phase service using both PV systems and then pile up all of the remaining inverters on the single phase system your motors might be happier too.
Can you tell us more about the model and output power of each of your inverters it might help.
Do you have your arrays split into sub arrays or strings for each inverter? Are they similarly oriented?

 

[jaggedben]

Have you eliminated the possibility that the customer is using more electricity than prior to the installation(s)? Were rates raised? Has someone done a careful analysis of the bills?

Also, if the utility is PG&E (and possibly with others), you should be able to look up 1 hour intervals (at least) of energy consumption or export on their website. While limited, it could tell you a lot more about potential intermittent problems than a one time check.

Perhaps a monitoring system would also be worth the customer's investing in.

That's all the advice I got.

 

[electrofelon]

Wow what a confusing mess the whole set up is. Strange it has two services in those configurations, I dont quite get it. Probably a long shot and not a fix to your problem, but any chance in just changing it to a single service? I'd almost volunteer to help for the satisfaction of it!

 

[electro7]

Shoot, I thought I was on to something with the inverters causing a voltage imbalance between phases. And therefore causing motors to run hotter and more ineffeceint causing more usage. Did anybody check out that article by Mike Holt? He even said in that example, that a 1.64% Vmu vs the 1.96% Vmu, when connecting those two inverters to those two phases, is an increase of 20% of the maximum unbalanced voltage. And that you do not want to do that. And so in the example he moved one of the inverters to two different phases and it lowered the voltage imbalance.

So to check this though, you (GoldDigger) were saying to check current on each phase of the motor? Should they read the same? Is there a healthy range if they are off a bit? (I wish I was up on motors more)

I do have the bills and will take a closer look at them. The homeowner said he has not changed anything except took out a wine cooler and somebody changed some timers on some stuff, so he was going to look into that. It is PGE and the meter tech saw in his 15 minute interval chart there was a spike between 11am to 2am of about 1800 watts. So we know something is cycling there. The homeowner is going to check into that.

So on the "Garage Solar" there are 72 Solarworld 250w mono with 3- Power One 6000w inverters. Two of these are connected to the three phase system. 2 Strings of 9 per inverter. Installed in 2013.
The "Field Solar" has 9 Strings of 12 Photowatt Ontario 220w panels going to 3 PV Powered 4600w inverters. Two of which are connected to the 3 phase system.

Any thoughts on why the 2 breakers operating the two Power One inverters in the garage, which are on phases A-B and C-A, tripped when I moved the breaker from phase C-A to B-C? If it had anything to do with C being the high leg I would have thought they would have tripped in their original position. I will say the last company who installed these two inverters used two-pole 20A breakers when they should be two-pole 35A, which I will change.

Thanks again for the help!

 

[GoldDigger]

A small voltage imbalance can cause a larger current imbalance and the current imbalance is the real problem contributing to vibration and overheating.

 

[electro7]

Okay, gotcha. So are you thinking it possibly could be the voltage imbalance that is causing the motors to run hotter and less efficient, that could be a cause to the problem (as I am)? When measuring current on each phase of a three phase motor they should be close to even, eh?

If this could be part of the problem I have no idea why the two breakers tripped when I changed the one to phase B-C. I did this to try and fix the voltage imbalance between phases.

Maybe I am chasing something that is irrelevant. Please let me know if you think so.

 

[ggunn]

Why would inverters cause a voltage imbalance? Inverters match the voltage they see on the AC line.

Here's a wild guess... Sometimes on a high leg service the capacity of the high leg is significantly lower than the capacity of the other two phases. Could it be that a smaller service conductor on the high leg is causing a voltage rise on that phase when the inverters are running?

Most of the time when I have interconnected with a high leg service, I did so with breakers feeding the other two legs and the neutral as if it were a single phase 240/120V service.

 

[electro7]

Well, I am not sure why inverters cause a voltage imbalance. But in the article I listed the link to earlier in the thread, Mike Holt talks about it. If you look at the readings I took, the voltage between phases does change with inverters connected vs when they are not. Not sure the reason behind it but must be a component in the inverters that causes it?

The high leg conductor is the same size as the other two.

 

[jaggedben]

Any thoughts on why the 2 breakers operating the two Power One inverters in the garage, which are on phases A-B and C-A, tripped when I moved the breaker from phase C-A to B-C? If it had anything to do with C being the high leg I would have thought they would have tripped in their original position. I will say the last company who installed these two inverters used two-pole 20A breakers when they should be two-pole 35A, which I will change.

Odd stuff here, I have questions and comments.

1) Did the breakers trip immediately when turned on (i.e. short) or did they trip when the inverters turned on (e.g. 5 mins later)?

2) I find it odd that you have Power One inverters on a high-leg delta, since I was told by a Power One engineer that they cannot operate without a neutral. So they'd work on the C-A phase only, is what I was told (assuming B is the high-leg following the usual convention). So maybe I (and he) are wrong, but take it for what it's worth.

3) If those 20A breakers never tripped then that suggest some serious misdesign of the system. If they did trip at some point, and the homeowner is not telling you or somehow didn't know, then that could be your explanation of lost production.

 

[electro7]

It is odd stuff.

The high leg is C phase. They are connected to A-B and C-A phases.

1) Breakers tripped immediately like a short. (When I moved the one from C-A to B-C phase)

2) I was wondering about the high leg and the inverters connected to it. They are using a neutral. I'm going to check again with Power One about that.

3) I did not quite understand why the other company put the 20A breakers in there. I am wondering if they did trip in the past and were not turned back on until some time after.

The E-totals since the system was installed in 2013 are:
Inverter 1: 26,032
Inverter 2: 22,267
(According to the inverter LCD screens)
We are north of Sacramento

 

[jaggedben]

It is odd stuff.

The high leg is C phase. They are connected to A-B and C-A phases.

1) Breakers tripped immediately like a short. (When I moved the one from C-A to B-C phase)

Here's my guess on this. The two A-phase conductors are faulted to each other somewhere, so when you move one of them to a different phase you have a short that affects both breakers.

I'm not seeing how this would relate to the production problems. It could be a completely independent issue. Or maybe it's related and just needs more thought to figure out.

2) I was wondering about the high leg and the inverters connected to it. They are using a neutral. I'm going to check again with Power One about that.

I'll be curious to know what they tell you.

3) I did not quite understand why the other company put the 20A breakers in there. I am wondering if they did trip in the past and were not turned back on until some time after.

The E-totals since the system was installed in 2013 are:
Inverter 1: 26,032
Inverter 2: 22,267
(According to the inverter LCD screens)
We are north of Sacramento

Okay, first, is there any difference between the Inverter 1 and 2 systems that would account for the large difference between them? (different number of modules? shading? One finished 6 months before the other?)

Next, you said 72 250W panels between 3 inverters, so that's 6000W per inverter. (You also said 2 strings of 9 per inverter, but I'm guessing you meant 2 strings of 12? or is this not confirmed?) Going by a real back-of-the-envelope number (1.5kWh per watt per year), production should be around 22,500 per inverter if turned on in the middle of 2013, or 27000 if near the beginning of 2013. Seems pretty much in the ballpark and not really an indication of a problem from the info given.

 

[ggunn]

It is odd stuff.

The high leg is C phase. They are connected to A-B and C-A phases.

1) Breakers tripped immediately like a short. (When I moved the one from C-A to B-C phase)

2) I was wondering about the high leg and the inverters connected to it. They are using a neutral. I'm going to check again with Power One about that.

Yeah, do that. If C is the high leg, and you have an inverter connected C-A, and you are using the neutral between A and B as the inverter neutral, then I can see where there might be a problem. One side of the inverter is seeing 120V phase to neutral and the other is seeing 208V phase to neutral. I am surprised that the inverter would operate at all under those conditions.

 

[electro7]

Just talked to one of the Power One (or ABB now) techs. He said some of the old inverters would work on a high leg system but they are not sure why. He recommended taking it off the high leg. He said that none of there three phase inverters will work on a high leg system. The single phase will but they recommend balancing it with three inverters having 2 inverters go to phases A-B and the other configuring to a 277v inverter going to the C leg (C being high leg). They modify settings in the inverter to make it a 277V inverter with one hot and one neutral and adjust the voltage range of the inverter.

I am going to change this today and check for the short on the B conductors. Ill be back in touch.

Thanks for all the help guys.

 

[electro7]

Okay, first, is there any difference between the Inverter 1 and 2 systems that would account for the large difference between them? (different number of modules? shading? One finished 6 months before the other?)

I think there is a bit of shading that is causing the difference. Same number of modules. They were finished at the same time.

[QUOTENext, you said 72 250W panels between 3 inverters, so that's 6000W per inverter. (You also said 2 strings of 9 per inverter, but I'm guessing you meant 2 strings of 12? or is this not confirmed?][/QUOTE]

So there are 3 inverters, I need to count the panels to make sure there are 72. That is what was on the permit pack along with 2- strings of 9 per inverter. So the math does not add up. Thanks for catching that.
The DC voltage readings on the inverters were around 362V. The STC Voc of the SW250w is 37.8v. I think there are 12 modules per string. The read out on the screen is the Vmpp which is usually a bit lower than the open circuit voltage. 31.1V for SW250's which would be 11.62 modules per string.
I think it was a mistake on the permit when it said 9 modules per string. I think its 12. Ill double check Voc.

 

[ggunn]

Just talked to one of the Power One (or ABB now) techs. He said some of the old inverters would work on a high leg system but they are not sure why.

Inverters with grounded DC, i.e., older inverters, some of them, anyway, could operate without a neutral, so phase to phase they didn't care if the service was high leg.

 

[GoldDigger]

Even for inverters where the neutral is not part of the power circuit now check for balance between the two as part of the qualification of the service. (UL requirement?)
That could be part of the difference between old and new models.