Utility fault causing hardware failures

[BandGap1.1eV]

I need someone better versed in electrical theory to chime in.

I've got an 840 kVA behind the meter PV system that is experiencing over-voltage failures at the AC contactors in the inverters. Five of the seven inverters on site have experienced the same failure mode, in additional to many of the lighting drivers in other parts of the building. We have SEL waveforms of the utility voltage showing multiple Phase to Ground faults on the utility circuit since mid-February. I'm a bit of a dullard when it comes to the theory, so I'm wondering if anyone here could diagram/explain how a ground fault on the 13.8 kV side of a tranformer can induce an overvoltage on the secondary.

Thoughts?

 

[Elect117]

Would you mind elaborating a touch on the secondary voltage?

For overvoltage:

As load drops, voltage goes up. The overvoltage might be a sign of a large load coming on a off within the service. Or the transformer has taps and they have the voltage set too high. Or the voltage at the utility substation is high because someone used to be on the circuit that pulled more load and they are no longer pulling that load. Or the circuit configuration changed and your are closer to the substation than you used to be. I have seen all of these result in higher voltages at a customer's service.

Faults creating overvoltage is more tricky. I would diagnose the above first.

Can you elaborate a little more on the SEL values? Or what the fault duration looks like (cycles or miliseconds)?

Inrush on the transformer can trick OC protective schemes in thinking it is a fault but in actuality it is a large load coming on. You will see a dip in voltage but it will be short and I believe non-zero.

 

[BandGap1.1eV]

Would you mind elaborating a touch on the secondary voltage?

For overvoltage:

As load drops, voltage goes up. The overvoltage might be a sign of a large load coming on a off within the service. Or the transformer has taps and they have the voltage set too high. Or the voltage at the utility substation is high because someone used to be on the circuit that pulled more load and they are no longer pulling that load. Or the circuit configuration changed and your are closer to the substation than you used to be. I have seen all of these result in higher voltages at a customer's service.

Faults creating overvoltage is more tricky. I would diagnose the above first.

Can you elaborate a little more on the SEL values? Or what the fault duration looks like (cycles or miliseconds)?

Inrush on the transformer can trick OC protective schemes in thinking it is a fault but in actuality it is a large load coming on. You will see a dip in voltage but it will be short and I believe non-zero.

277/480 on the secondary side. Snips of the waveform attached.

We can tie a recent SEL event to an explosion in a utility manhole at the street. The event was captured on CCTV and the timestamps lineup with the SEL opening. The utility has not been disclosing any information on their side yet.

I'll add this feeder is allegedly a non-effectively grounded feeder, hence why the utility asked us to include 59N protection. We also put a Power Quality Analyzer in the 480V disconnect recently. If there's another event on the utility circuit we should get some good data on the secondary side.

 

[jim dungar]

Depending on the utility configuration, you might be experiencing ferroresonance or the result of an arcing fault continually charging and discharging system capacitance.

 

[Elect117]

I am not seeing overvoltage on these events.

13800/SQRT(3) = 7967V But this is RMS so Vpeak is SQRT(2) * 7967 = 11,268V which is pretty much what you are measuring.

I am seeing something happening to C phase but it is hard to tell what the event is. I am having a hard time with the current waveform graph.

Maybe you were saying that it is tripping on overvoltage on ground or neutral? That could make sense if the phase to earth fault happened after you and so that fault was looking to get back to it's source and that was both you and the substation. So a portion of the fault current could have appeared on your system as you and the sub feed into the fault, until it cleared/opened.

Do you know your location on the feeder and the fault location? Was it upstream (after you and the sub) or down stream (closer to the sub than you)?

 

[BandGap1.1eV]

I am not seeing overvoltage on these events.

13800/SQRT(3) = 7967V But this is RMS so Vpeak is SQRT(2) * 7967 = 11,268V which is pretty much what you are measuring.

I am seeing something happening to C phase but it is hard to tell what the event is. I am having a hard time with the current waveform graph.

Maybe you were saying that it is tripping on overvoltage on ground or neutral? That could make sense if the phase to earth fault happened after you and so that fault was looking to get back to it's source and that was both you and the substation. So a portion of the fault current could have appeared on your system as you and the sub feed into the fault, until it cleared/opened.

Do you know your location on the feeder and the fault location? Was it upstream (after you and the sub) or down stream (closer to the sub than you)?

Thanks for the comments. I'm not certain on my sites location with respect to the substation, but I do know we're on the same circuit that experienced a blowout in the manhole in the street. This particular utility does not have a published circuit map available to the public.

Overall I'm trying to conceptualize how to Ph-N or Ph-G fault on the primary could manifest as over voltage on the secondary. Maybe the phase loss triggered something else, like a neutral point shift, that led to over voltage on the other two phases?

 

[ggunn]

A company I worked for had a small commercial installation of seven single phase SunnyBoys, and two or three of them would occasionally trip off for high voltage. It was different inverters every time and it was only on clear days in the Spring and Fall. The issue was that the service conductors from the pole mounted transformer were too small; on cool sunny days when the HVAC wasn't running, the PV system was exporting most of its output, so the voltage rise in the service conductors would cause some of the inverters to shut down until the voltage dropped enough to where the rest would stay running.

 

[mtnelect]

Does your transformer have a standard connection between the primary and secondary ?