Question about Bus differential protection of 12kV switchgear and coordination with the Utility

[bdn2004]

We’ve got a 12 kV feeder into the Plant from the Utility that feeds a switchgear lineup. The switchgear lineup has bus differential protection that is set to trip at 5 cycles and that is the time determining the arc flash current for the feeder sections of the bus. We are to provide an arc flash label.

But what about the incoming section of our switchgear that will lie outside of the switchgear bus differential zone? Will the Utility also have bus differential on their lineup that is set to trip their main?

I’m just now looking into this and was thinking somebody may have some insights on how this is typically done.

 

[crshears]

Depending on the age and/or sophistication of the utility's relaying, could be [a] as rudimentary as just low-set instantaneous overcurrent protection for the first trip, with low-set blocked for a few to several seconds upon auto reclose, if so equipped [often avoided on feeders with a high proportion of cables and very little overhead line]; could be impedance based feeder protection; could be [c] the latter plus load encroachment blinding, especially on feeders with a significant amount of distributed generation [for the most part I still encounter the more developed protection schemes used only on HV circuits that operate at voltages > 100 kV.

Feeder differential schemes are relatively uncommon, at least in my experience; that being said, my utility does still have numerous pilot wire schemes [a quasi-differential protection] in service in the heavy industry area on the waterfront of a medium-sized city.

Others may have seen vastly different ways of doing things...

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[dpc]

There's no standard. If the utility has a local breaker, your bus diff relay can and should trip their breaker if they will allow it. Then the main breaker can be covered if the CTs are located correctly. Your bus diff zone should include the main breaker. The problem is clearing the fault.

If that isn't possible, you will have to determine the utility fault protection and come up with a fault clearing time based on the expected fault current. You would then have two arc flash levels. One for the switchgear and a another level for the main breaker.

It is not common to set an intentional 5 cycle delay on a bus differential trip. Remember that the total arc time includes the relay operating time plus the breaker opening time, typically 3 or 5 cycles. And possibly plus additional time if tripping through a lockout relay.

 

[bdn2004]

I’m looking at the existing model and settings. It’s not shown schematically just a time on that incoming bus. So that 5 cycle arcing time I’m seeing is most likely the Utility’s breaker closing time + relay time.

If the bus differential relay is around the main and all the feeder breakers. And the way I understand it trips on a differential between input vs output...What keeps it from tripping the Utility breaker on any ground fault? Doesn’t it have to be coordinated with the feeder breakers, and therefore have a time delay?

 

[jghrist]

The utility protection will have to be slow enough to coordinate with your feeder protection. You probably also have time overcurrent relays on your main breaker, coordinated with the feeder protection. This would provide backup protection for the feeders. If the utility serves more than one customer beyond their protection device, their interrupting time would be slower than your main overcurrent relay.

Also keep in mind that the bus differential will not cover faults in the outgoing feeder sections that are past the bus differential CTs. You will have to consider the feeder protection interrupting time in determining arc flash incident energy at the feeder cable terminations. If you are only providing labels for the feeder sections, all you need is the feeder interrupting time.

 

[bdn2004]

I'm obviously not understanding how the bus differential relay works. I'm looking a drawing of a similar sub, there is a 1200/5 CT around each outgoing load feeder and a 3000/5 CT around the incoming feeder before the main breaker. Their outputs going to a summing bus, and the 87 relay. The output from the relay to the Utility's feeder breaker. And btw the Utility's feeder serves no other Customer.

Wouldn't a ground fault on any of the feeders cause an imbalance and therefore a trip ?

 

[davidbeach]

I hope the 3000/5 isn't being summed with the 1200/5 CTs, that will cause a trip for feeder faults.

In a properly implemented bus differential that relies on summing CT outputs (as opposed to a low impedance bus diff like the SEL-487B where every CT is individually connected to the relay) all of the CTs need to be sized the same. A fault in the bus diff zone results in measured current at the relay and a trip. A fault on the feeder should have the same current entering at the main and leaving at the feeder and those sum to zero, no trip.

 

[dpc]

If the fault is within the differential zone and the differential relay is tripping the utility breaker, there is no need for time-based coordination on a differential trip. If the utility protection is simply an overcurrent relay, then it will have to be slower than everything downstream. You need to figure out the utility breaker protection scheme and then work out settings and then lastly, calculate incident energy.

 

[cranky108]

I believe most utilities do have people you can talk to. So ask your question to them.

They won't give you any advice, but they can answer questions on the protection into your plant on their side of the meter.

But to be specific, you need to talk to the right person. The customer service person is not that person. Most utilities have an account manager that can get you in contact to the right person.

In our utility, we specify that type of information on all arc-flash requests.

 

[bdn2004]

I have another question however about this....

The differential relay test report shows the relay operates in .083 seconds. 5 cycles.
This opens a 3-cycle circuit breaker. Does that 5 cycle clearing time that is shown in the test report include the 3-cycle breaker opening time ?

 

[dpc]

It depends on the test setup. I would refer to the relay specifications and get the maximum operating time for the relay, then add in the breaker opening time (on the nameplate), plus a safety factor.

 

[cranky108]

When asking that question, also look at what other devices maybe between the relay and the breaker.

I would expect the other devices would not take much operating time, but it's better to check then be blind.

Maybe a lock-out relay in between, which should take a 0.25 cycle to operate, but some types if not maintained can take much longer, or not operate.
This might also be a maintenance note.

We don't test the relay and breaker as a unit for a timing test. So that to me would be unexpected, but not impossible.
I.E. I would add the operating times as an assumption.