Intermittent earth fault 1

[Yanten]

hi guys, im new here, asking for technical help.

We have a newly repaired 18 year old 33KV cable (repaired by jointing), it was repaired due to struck by hard machine, the length is approximately 2000M. After successful repair and testing; hipot (VLF) testing (24KV, leakage is 14mA) by our cable contractor, we are facing a trip on MCGG52 relay (51N). We had a tripped after 2 days and after 20 days from the date of testing. I just want to know what testing can be more efficient to capture the fault. Thank you in advance for your advice.

 

[Marmite]

An intermittent fault on a 33kV cable is an extremely unlikely scenario. Cable insulation doesn't self heal like air insulation, and usually it faults permanently. On a 33kV system, generally the energy released into a fault is significant and permanent damage results.
Do you have any overhead line connected to the cable?
Initially I wondered whether you had a pilot wire protection scheme and perhaps the phases had been inadvertently rolled in the repair, but looking at the MCGG52 manual, it is just a simple 2 pole overcurrent plus earth fault relay. Were the settings altered to speed up tripping as a precaution when the faulty cable was returned to service?

Regards
Marmite

 

[waross]

We did repairs to a damaged direct burial cable on a 13,200/7620 Volt circuit. After a time one phase started blowing fuses intermittently. It may go several days or a week or more between failures. This was on a small island in the gulf of Mexico and not a lot of test equipment was available.
We were unable to locate the fault and eventually replaced one phase conductor.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[Yanten]

Thank you sir/s for your reply.
@Marmite, yes we had our pilot wire but our cable contractor didn't touch anything on it and we have only a direct burial cable. Unfortunately I don't want also to increase the setting relay since the feeder was paralleled to 2 more feeders with the same setting. I attached herewith our PMS readings, wherein we can find here that the current carrying of the repaired R-phase cable was dropping consequently my neutral current increased tripping my 51N.

My plan is to ask the cable contractor to re-joint the cable, replacing the whole cable was too expensive and will involve lot of works so we don't want to do it.

 

[stevenal]

Parallel feeders with individual protection? Ohms law can be pretty harsh in this arrangement. Tiny amounts of impedance difference can result in large amounts of current difference. Even a well done splice will introduce some impedance.

 

[Yanten]

Yep, 3 paralleled feeders, 2 bus tie close with individual protection, one thing I have observed was one the reactive power of one phase is different from other 2 phases not like the other two feeders which have a symmetrical value. Got some meeting with the cable contractor yesterday but they just recommended to test the cable sheath with 30KV dc.

 

[Marmite]

Maybe worth checking the end to end continuity and impedance using a ductor. Compare the impedances between phases to see if you have an open circuit fault developing.
Regards
Marmite

 

[stevenal]

Not sure what voltage testing will accomplish, since this is a problem of impedance and current division, not insulation. Check impedance as Marmite suggested to see if the splice was made up properly. You may find, however, that current division between two un-spliced cables and one spliced may always be unbalanced enough to cause your sensitive relay to trip when no fault is present. If so, you can either insert splices in the other two cables, or adjust your 51N pickup values.

 

[jghrist]

Are you measuring ground current or neutral current? If you are measuring ground current with a core balance CT including the neutral cable if any and not shields, then unbalanced phase currents shouldn't cause a trip.

 

[stevenal]

jghrist,

I assumed he was using a simple residual CT connection or calculation, but I don't see the core balance CT making a difference. Consider two parallel cables feeding a balanced three phase load. Introduce an impedance in A phase of cable 1, and you will see a drop in A phase current on that cable. If you sum the three phase currents you will see a non-zero value. Meanwhile A phase current in cable 2 has picked up the slack, creating a summation that is of equal magnitude and opposite polarity to that measured in circuit 1. If the impedance is large enough, both cables will trip due to this circulating 3I0. OP had two cables to make up the slack, so only one cable tripped.

 

[jghrist]

Steve,

The core balance CT, if it includes three phases and a neutral, will not measure unbalanced load 3I0 neutral current, only ground fault current. Increased impedance in one phase would be similar to a decreased load in that phase.

 

[davidbeach]

Feeder A and feeder B; both should be carrying 100A, all positive sequence. Instead feeder A has 150A on phase A and 100A on B and C due to high impedance connection on feeder B that now has 50A on phase A and 100A on B and C. Both feeders still have 120 degree phase angle separation. How is there not 3I0 current on each of the two feeders? Both have 50A of 3I0 and the two have 180 degrees between them. Hard to tell which would trip first, totally eliminating the 3I0 from the other. Make it three feeders and the one with the high impedance will always be the first to trip and then there will be no anomalies on the other two.

 

[stevenal]

Davidbeach is correct. This 3I0 exists only in the phases, so including neutral current in the calculation makes no difference. My assumption above was balanced loading, so IN=0. This assumption was only used to simplify the concept. If unbalanced loads are connected phase to neutral, then IN should split between all the cable shields according to ohms law, but it will not necessarily be equal to the sum of the phases in each cable run.