3-Phase Transformer Problem

PWDickerson

Senior Member
Location
Clinton, WA
We installed a 12.5 kW PV system at a commercial site last year, and we are having transformer problems. The system looks like this: The output of three 3.84 kW 240V single phase (no neutral) inverters are connected in a 3-phase 240V delta configuration. The AC output of each inverter has OCP consisting of a fused disconnect with 20 amp fuses. The combined output of the three inverters connects to the primary side of a 15 kVA 240D primary/277-480Y secondary transformer. There is OCP on the 277/480 side of the transformer in the form of a 3-phase fused disconnect with 20A fuses (initially), and there is another 3-phase fused disconnect with 20 amp fuses (initially) at the utility service 600' distant from the PV system. Note that we are not back-feeding a step-down transformer. The transformer was designed as a step-up unit.

The 20A fuses on the 480V side were not sized correctly initially. They were sized for the PV system output, but the transformer manuf. suggested larger fuses to accommodate the inrush current that occurs when the transformer is turned on.

I was at the site last week and noted that one of the inverters was not operating and showed "Grid NA" error. Upon investigation our electrician found that two of the 20A fuses at the 480V 3P disconnect at the transformer were blown, and the neutral wire connecting to the transformer XO conductor had overheated. The transformer manuf. thought the transformer might still be OK, but when we replaced the blown 20A 480V fuses with 30A fuses and powered up the transformer from the 480V side with the inverter disconnects open, we immediately blew a 30A fuse indicating damage to the transformer. The 240V fuses protecting the inverters never blew.

At this point it seems like an inverter disconnected for some unknown reason, which fed imbalanced current into the transformer, and that led to excessive neutral current on the 480 side of the transformer, excessive current on some of the 480V phase wires, which blew the fuses, and eventually to overheating and damage to the transformer.

These events highlight a design flaw in the system. We should have installed common trip OCPD's at the transformer. I have much more experience in designing smaller residential system, and I would very much appreciate comments on best practices for installing a system in this environment.

BTW, I would have loved to spec an inverter with an output that matched the utility voltage, but due to very generous state incentives for locally made equipment, the single-phase inverters were required.
 

GoldDigger

Moderator
Staff member
We installed a 12.5 kW PV system at a commercial site last year, and we are having transformer problems. The system looks like this: The output of three 3.84 kW 240V single phase (no neutral) inverters are connected in a 3-phase 240V delta configuration. The AC output of each inverter has OCP consisting of a fused disconnect with 20 amp fuses. The combined output of the three inverters connects to the primary side of a 15 kVA 240D primary/277-480Y secondary transformer. There is OCP on the 277/480 side of the transformer in the form of a 3-phase fused disconnect with 20A fuses (initially), and there is another 3-phase fused disconnect with 20 amp fuses (initially) at the utility service 600' distant from the PV system. Note that we are not back-feeding a step-down transformer. The transformer was designed as a step-up unit.

The 20A fuses on the 480V side were not sized correctly initially. They were sized for the PV system output, but the transformer manuf. suggested larger fuses to accommodate the inrush current that occurs when the transformer is turned on.

I was at the site last week and noted that one of the inverters was not operating and showed "Grid NA" error. Upon investigation our electrician found that two of the 20A fuses at the 480V 3P disconnect at the transformer were blown, and the neutral wire connecting to the transformer XO conductor had overheated. The transformer manuf. thought the transformer might still be OK, but when we replaced the blown 20A 480V fuses with 30A fuses and powered up the transformer from the 480V side with the inverter disconnects open, we immediately blew a 30A fuse indicating damage to the transformer. The 240V fuses protecting the inverters never blew.

At this point it seems like an inverter disconnected for some unknown reason, which fed imbalanced current into the transformer, and that led to excessive neutral current on the 480 side of the transformer, excessive current on some of the 480V phase wires, which blew the fuses, and eventually to overheating and damage to the transformer.

These events highlight a design flaw in the system. We should have installed common trip OCPD's at the transformer. I have much more experience in designing smaller residential system, and I would very much appreciate comments on best practices for installing a system in this environment.

BTW, I would have loved to spec an inverter with an output that matched the utility voltage, but due to very generous state incentives for locally made equipment, the single-phase inverters were required.
How did you wire the POCO side (480V side, and for the purpose of transformer engergization the primary side.)

Even if you have a wye service and a wye wound transformer, the X0 on the primary side should generally NOT be connected to the POCO neutral or to ground. The transformer case must be grounded though.

When you have a secondary delta winding and a wye primary with the center point connected to neutral any imbalance in either the POCO voltages or the secondary "load" can cause high neutral current on the POCO side.

You are correct that the 480V OCPD has to be sized to allow for transformer inrush and initial magnetizing current. Because the inrush is dependent on the physical design of the windings and core, using the designed secondary side of the transformer as the primary (energizing) side can result in 5 or more times the "normal" inrush.
You may need to tailor the time curve of the primary OCPD using fuses or adjustable breakers to be able to survive that.

PS: The behavior of a common core three phase transformer with one (or in the case of wye even two) phases open is not at all what you might naively expect. There is an extensive thread on that topic featuring some analysis by Phil Corso.

The upshot is that some form of sophisticated electronic phase loss detection may be needed, especially if you use fuses rather than a common trip breaker!
 
Last edited:

PWDickerson

Senior Member
Location
Clinton, WA
Thank you for your comments GoldDigger. We have identified the need for a common trip breaker on the utility side of the transformer to prevent single phasing in the event 1 or 2 legs of power are missing. We have the equipment on order and will be installing it at the site shortly.

It is not clear to me if a common trip breaker on the PV side of the transformer would also be a good idea. In the event that one of the three inverters shuts down, is it possible that the neutral current in the XO could be excessive without tripping the utility-side breaker?
 

GoldDigger

Moderator
Staff member
Thank you for your comments GoldDigger. We have identified the need for a common trip breaker on the utility side of the transformer to prevent single phasing in the event 1 or 2 legs of power are missing. We have the equipment on order and will be installing it at the site shortly.

It is not clear to me if a common trip breaker on the PV side of the transformer would also be a good idea. In the event that one of the three inverters shuts down, is it possible that the neutral current in the XO could be excessive without tripping the utility-side breaker?
Yes and no. As long as the primary side is properly protected and the phase voltages are balanced, the neutral current should never be any greater than one of the phase currents.
But, once again, for your delta winding on the 240V side, the wye point on the 480V side should not be connected at all. Otherwise a small voltage imbalance on the POCO side can cause a combination of neutral and phase currents which could trip the primary protection even with no load connected. The neutral current would still not be higher than any single phase current, but if you wired the system with a reduced size neutral you could have a problem there.

You should also check with the transformer manufacturer to see whether they approve a reverse connection of the transformer and what precautions they recommend or require if it is supported.
 

jaggedben

Senior Member
I'm not going to offer any help, but I'm trying to put this all together and understand if I've gathered correctly...

......I was at the site last week and noted that one of the inverters was not operating and showed "Grid NA" error. Upon investigation our electrician found that two of the 20A fuses at the 480V 3P disconnect at the transformer were blown, and the neutral wire connecting to the transformer XO conductor had overheated. ...

At this point it seems like an inverter disconnected for some unknown reason, which fed imbalanced current into the transformer, and that led to excessive neutral current on the 480 side of the transformer, excessive current on some of the 480V phase wires, which blew the fuses, and eventually to overheating and damage to the transformer..
...
When you have a secondary delta winding and a wye primary with the center point connected to neutral any imbalance in either the POCO voltages or the secondary "load" can cause high neutral current on the POCO side.
...
PS: The behavior of a common core three phase transformer with one (or in the case of wye even two) phases open is not at all what you might naively expect. There is an extensive thread on that topic featuring some analysis by Phil Corso.
If this were a delta with no transformer then blowing two fuses would have caused all three inverters to lose phase voltage and stop operating. (Blowing one fuse would cause two of them to stop operating).

So is it correct that if the neutral were not hooked up on the utility side of the transformer, there would have been no voltage on the PV side, and all inverters would stop operating?

Is it indeed plausible that the two inverters that stayed on did not stray out of the proscribed grid-tie voltage windows because the neutral was connected? Would the voltages not be that different?

The upshot is that some form of sophisticated electronic phase loss detection may be needed, especially if you use fuses rather than a common trip breaker!
hmmm. 705.100
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
I'm not going to offer any help, but I'm trying to put this all together and understand if I've gathered correctly...





If this were a delta with no transformer then blowing two fuses would have caused all three inverters to lose phase voltage and stop operating. (Blowing one fuse would cause two of them to stop operating).

So is it correct that if the neutral were not hooked up on the utility side of the transformer, there would have been no voltage on the PV side, and all inverters would stop operating?

Is it indeed plausible that the two inverters that stayed on did not stray out of the proscribed grid-tie voltage windows because the neutral was connected? Would the voltages not be that different?



hmmm. 705.100
If one phase went out, why wouldn't both inverters that were connected to that phase quit working? Why just one of them? They were phase to phase, not phase to neutral, right? The way I read it there was no neutral on the inverter side of the transformer.
 

PWDickerson

Senior Member
Location
Clinton, WA
I may be recalling incorrectly as to how may fuses were blown. I am more concerned with making sure that we implement the correct design changes to the system to prevent this from happening again. My plan is to disconnect the HO conductor on the 480 (utility) side of the transformer, and install a 3-pole common trip circuit breaker in place of the existing 3-pole fused disconnect. If I do this, it should solve the problem.

Next question, what is the correct size circuit breaker on the utility side. If I install a 25 amp breaker (15 kVA @ 480V 3ph = 18 A x 1.25 = 22.5 A), am I likely to experience nuisance tripping from the utility in-rush current? The transformer primary OCPD can be sized at 250% if the secondary OCPD is at 125%, but with a PV system the power can flow in both directions and I believe over-current devices must be calculated in both directions. That being the case, is it ever possible to size a PV transformer greater than 125%?

Will the in-rush current really be significantly higher because the utility is connected to the secondary side of the transformer? I read an article recently that when a transformer is reverse-fed, it is significantly less efficient than one that is fed in the intended direction. Given the necessity to use single phase 240V inverters to feed a 277/480 service, How would other people design such a system?
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
Oops. I meant to ask if it is ever OK to size a PV transformer Breaker at greater than 125%
I am pretty sure that you can put whatever OCPD you want on a set of conductors as long as the ampacity of the conductors is such that they are protected by the OCPD.
 
Top