MAIN BREAKER TRIPS INSTEAD OF THE FEEDERS DURING GROUND FAULTS

[burnt2x]

Hello!

I have a ground fault protection problem that involves two different directional ground relays and two different PT's (potential trafos wired broken-delta as reference voltage sources). The feeder PT voltage rating is 190 volts and the main breaker PT 110 volts.
We have experienced ground faults during which the main breaker directional ground relay always trips instead of the feeder directional ground relay on faults of the feeder! We have tested the trip times of both relays and found the trip times are graded 200 ms apart for the same ground fault current (i.e. feeder ground fault relay set to trip earlier than the main breaker ground fault relay for the same groudn fault level)

My Q1 is:
Since we have ruled out the relay trip time settings/ grading as the culprit (by testing), is this a problem with reference voltage settings? (one relay with a 110 volt reference source and the other with 190 volt reference) Did I miss something here?
Q2:
The 110-V broken-delta connects with the main braker's 67N ground protection relay with 10% (11 volts) setting while the 190-V broken-delta connects to the feeder directional ground relay with 15 volts setting (around 8% of 190 V), are the voltage reference settings above (10% vs. 8%) correct?
The system is a substation, 4160 volts, ungrounded system.
Any recommendation is accepted.

thank you.

 

[slavag]

Hell Burnt2x.
My first Q, I'm always start with this Q in those situations:
Are you test this system with live ground fault?
that means 1ph short on the feeder side.
2. What is a current settings?
3. What is principle used for the 67N (ungrounded systems)?
Regards.
Slava

 

[slavag]

Additional Q.
What is CT you used for the ground fault detection?
I'm hope toroid CT ( around all 3 phases).
Regards.
Slava

 

[burnt2x]

Slava,
Tested each relays using DOBLE F6150A.
Relay Current settings(AS FOUND):
Main Breaker 67N function - 1.0 A; 10% V (of 110 V, or 11 volts); trip time 0.303 seconds.
Feeder breaker directional ground relay - 0.7 A; 15V (8% of 190V); trip time 0.11 seconds.
System is 4160V- 3-wire; ZCT on 3-core cables. directionality of ground relay thru voltage to ground detection using broken-delta PT; wired to trip on ground faults (not just indication)

 

[slavag]

Grate.
What is ZCT : 50/1A , 100/1A or..
I think you need put voltage setting on same voltage 10-11V.
I strongly recommend provide live test, maybe problem only in polyrity of CT/VT on the feeder.
Seeting seems good. Are you have long cables?
Regards.
Slava

 

[davidbeach]

Why two polarizing voltages? Could it be that one is wired in backwards so the feeder would only trip for reverse faults?

Also, how do you get ground fault current if the system is ungrounded?

 

[slavag]

Hello David.
For this, we always request provide live test, really 1ph short circuit.
Second. no problem with ungrounded system too, you need only enouph capacitive current. From my excpirience, 6.6kV
cable 300m ( about 100feet I think)ground fault was abot 0.5A primary.
Regards.
Slava

 

[burnt2x]

Slava,
Cable is 1000 meters! I see, we are at synch. I will verify if they can allow "live" SLG test to know line to ground amps. Also, production dept holds every decision, you know?

David,
I have recommended to the owners to take out the older 190V PT and use only one supervising voltage (Polarizing voltage)! Site electrical reply was he's not sure on the 110V broken delta PT's can accomodate additional burden. Need to verify truth of the matter.

 

[slavag]

Hi.
If main feeder tripped,thats say: enouph current.
Possible sevral problems:
1. Old PT.
2. Polarity, in this case need only cross wiring on relay or CT or VT input on 180deg.
3. ZCT on feeder not so good, that means have big phase angle error between primary and secondary current.
point 3 is big prblem. What people think, O.K. ground fault is very small, about 1 amp, I will put 30/1A ZCT and will detect this small current. Is bigest mistake.
What we do, of course, if you have newer relay with DR and fault reports. Newer relay have, what I know, inputs 1 or 0.2A for the sensetive directional earth/ground fault.
We put 75/1A or 100/1A ZCT and connect it to 0.2A input.
At commssioning time we check ZCT performence.
Avarage results for 1A primary injection:
30/5, 50/5, 30/1A ZCT phase angle error about 14-20deg.
75/1A-80/1A about 5-7deg.
100/1A about 2-4deg.
Why I asked you about 67N principle.
possible work with basic angle or I0sinf principle.
In both of cases error 14-20deg is very critical.
Actually angle between U0 and I0 in ungounded system about 90deg, with this error possible, that your meas in reverse direction. We had not once unwanted trips becouse it.
1. live SLG test show you right direction and right connection of CT/VT.
2. performence test show you phase angle error and of course amplitude error.
For information only, waht type of relay you use.
Regards.
Slava

 

[slavag]

Additional.
1. I also recommend use dumping resistor on broken-delta connection, is for protect VT against ferrorsonance in ungrounded systems. But you must again provide live test, I heard about some problems with protection operation, not so believe. We prepere now some test with utilities for check it.
2. See attached files with ZCT performance test.
Regards.
Slava.

 

[burnt2x]

Thanks slava. will get to you after digesting all you have said.

 

[davidbeach]

If you don't know where the fault was that caused the trip, it might have been on the bus so that only the main breaker needed to trip to interrupt the fault.

 

[slavag]

Hi David.
Possible more intresting case, fault was out of this substation and main feeder tripped wrong ( I had last month two cases). But I'm sure,Burnt2x check, that it was really ground fault on the feeder.
Regards.
Slava

 

[slavag]

Burnt2x.
One important recommendation. as I understand, isn't your practic SLG "live" test. You need one good relay with 59N functionality only at your lab. You check it and at time test put it in parallel with tested feeder relay with time within main feeder and tested feeder.
Good Luck.
Slava