3V0 Detection/Protection 2

[HeinzM]

Gentlemen, can a ground fault on the primary side of a delta/delta main transformer (1MVA, 13200/380) be seen/detected on the secondary side of the transformer? I say, NO.
We would like to avoid MV rated PTs to detect a utility ground fault. No high side Wye-G PTs are installed.

 

[davidbeach]

No, there's nothing on one side of a delta-delta that indicates the presence of a single ground fault on the other side.

 

[JensenDrive]

Is your 380V system a source, and you are protecting against backfeed into a fault?
Is the 13kV (utility) source effectively grounded? If they have a ground fault, wouldn't they trip? If even a single fuse tripped, you would single phase the delta-delta xfmr, and on the secondary get half voltage magnitude across 2 phases, and your phase sequence would be half negative, half postive.
If the utility is an ungrounded source, this is a rather unusual utility, but I suspect they have ground sensors already.
If the utility trips, and your 380V source tries to backfeed, will the load mismatch (your site trying to feed utility loads that may be large compared to your generation) trip you off?

 

[HeinzM]

The utility want to protect against feeding a ground fault on their side. I have no idea how they protect their feeder and under what conditions they trip off line.
Besides, the generation source is solar inverters, which are setup according IEEE 1547 and stop when the line is gone. Utility mandates the 3V0 and they win!

 

[HamburgerHelper]

Not with ground current. I suppose you could see negative sequence currents.

 

[davidbeach]

Things fail. Anti-islanding schemes rely on there being a distinct imbalance between load and generation; there isn't always. The utility doesn't have visibility of the inverter.

If there's a ground fault on the 13.2kV side, the utility will trip. Between the time the utility trips and the inverter goes off line there will be overvoltages on the unfaulted phases that can only be detected by voltage as there is no current to the fault. The only current would be due to an arrestor conducting or something flashing over.

 

[JensenDrive]

Now I see the problem.
There is a trick to do it with a single VT to ground. If the monitored phase goes to ground, you see low voltage. If the monitored phase is an unfaulted phase, voltage goes high. Hence, trip on any voltage off normal. I suspect a VT on every phase will give better comfort to the utility though.

 

[rockman7892]

Wouldn't a broken delta PT on the primary side indicate the presence of zero sequence voltage 3V0 across the broken delta thus indicating a ground fault?

 

[HeinzM]

Yes, Rock.
But that's the expense they want to avoid. The PT cabinet measures about 4' (cube), has to be set on a concrete pad, protected by bollards. And the 15kV insulated PTs cost a few bucks, too.
And all that for (proven ) clearing of the generation within cycles of losing the line.

 

[marks1080]

We do just as described by JensenDrive for scenarios where we have an ungrounded primary and only a single-phase VT.

We use standard set-points for most applications of:
UV - 0.70pu
OV - 1.3 pu

It's not 'gold plated' option, but at least its cheaper than buying 3 VTs

 

[jghrist]

I think a time delay would be needed to keep from tripping the generator for a single-phase fault on the primary for which the utility feeder breaker would not trip. This could be a fault on a fused tap or on another feeder. A single-phase fault would drop the phase-ground voltage. A time delay may not be acceptable to the utility. You might be able to get by with two VTs and trip if either both voltages are high or if one is high and one is low.