Phantom fault currents

[muniengineer]

We have two feeders coming out of substations that we continue to have issues with but have no explanation for. The similarity between these feeders is they leave the substation via (3) 500 MCM copper tape shield conductors with a bare 4/0 neutral all pulled into a 6" PVC conduit. They then rise above ground for a few spans before dipping back underground to get under roadways. From there they both return overheard and continue on. Once feeder is at a 4kV station and the other is at a 15 kV station and they are on opposite ends of town. Anytime we get a phase to phase or a phase-phase-ground fault on either one of these feeders, we get a second operation that always shows fault current levels way above any possible fault currents that our modeling shows us. It is always around 6,000 amps on 15kV and 7,200 amps on the 4kV feeder. This is true no matter where the initial fault occurs on the feeder. If the feeders experience a phase to ground fault, everything operates as normal and we have not seen the second operation with the phantom fault currents. Has anyone ever experienced anything similar? I have tried to lean to some type of circulating current being induced but I am having a hard time making that plausible because of the independent neutral. The neutral is attached to the overhead neutral and then directly back to the substation ground grid. Any help on figuring this out would be appreciated.

 

[bacon4life]

What is the fault level at the first overhead span? Are the overhead wires or jumpers loose enough to slap together during a fault?

 

[waross]

Any grounded wye:delta transformers on the circuits?

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

 

[magoo2]

How is the overhead portion of the two circuits arranged? Are the two circuits on the same poles/structures?

If it's a horizontal crossarm type of construction, we would typically have the 15 kV above the 4 kV. With a phase to phase fault on the 4 kV, the hot arc tends to travel upwards and it would sometimes result in a fault on the upper circuit. I can't see it going the other way unless you had some cutouts that expelled their ionized (conductive) blasts to the lower circuit.

Like bacon4life suggests, any phase to phase or even phase to phase to ground faults produce conductor motion on both conductors of the circuit. It forces the two conductors away from each other and, when the circuit breaker or recloser interrupts, the conductors swing back toward each other. The force per unit length is the same, so it is possible that the longer spans will contact each other when reclosing takes place. From examining digital relay outputs, we've had cases where the initial fault is further out but the subsequent fault in closer in (and a higher current magnitude). This phenomena is discussed in a couple of IEEE papers and it depends on the fault current magnitude as well as the conductor size, spacing and the tripping characteristic.

 

[Marmite]

It sounds like a classic cross country earth fault. Is your system impedance earthed, or solidly earthed? Impedance earthed systems using an arc suppression coil/NER/NEX are more prone to cross country faults.

Regards
Marmite

 

[muniengineer]

We have considered all of the obvious scenarios and have actually gone to every pole between the UG dips to inspect crossarms and insulators. We have even hi-potted the cables to 25kV and have found no issues with the UG cable. Our spans are typically less than 100 feet because of the municipal setting and the need to serve residential loads. Conductor sag is not an issue because of the short spans. There are no signs that OH conductors ever made contact. We have a whole mix of transformer configurations but there are no wye:delta transformers close in to the subs. In the case the other day, the original fault current was around 3,700 amps on B&C but the subsequent fault current was nearly 6,000 amps on B&C. As mentioned above, we see a second fault of this magnitude after every Phase-Phase or Phase-Phase-Ground fault regardless of what the initial fault current is. We operate a 4-wire grounded wye system.

 

[MKFPE]

Hi muniengineer,

If I understand your description these two circuits are completely separate geographically.

The only common feature is perhaps the construction of the overhead spans.

I am curious regarding your faults being above what your modeling shows. These levels are similar to what we have for close-in or bus faults on our current 12kV system on the larger transformers and that we had on our retired 4kV system.

Similar to bacon4life and magoo2 we have also found phase slapping in a span just outside a 12kV substation where the line rolled from horizontal to vertical construction. Apparently this had been happening for a long time and we were unaware until we upgraded to SEL-351S relays.

We had two cases of approximately 2.5kA L-L faults about 2 miles out on the feeder followed closely by 4KA L-L faults just outside a 10MVA station (per our fault study).

Based on the 4kA faults reported by the relay we sent a crew to inspect the circuit close to the sub. Up-close inspection of the wire showed the arcing damage. That span has been rebuilt.

Apparently phase slapping is more common than I thought.

Good luck from another muni engineer.

 

[magoo2]

It would help if you sketched a one-line for this. From fault calcs, where would you expect a fault magnitude of 3700 A and one of 6000 A?

What is the phase spacing and conductor sizes?

 

[muniengineer]

Here is a single line of our most recent issue along with 3 SEL event files.

 

[waross]

Do you have re-closers on the circuits?
Although it may be stretching the bounds of coincidence for it to happen every time, If the first fault is near symmetrical levels and the second fault is asymmetrical, the 2.5 kA:4kA ratio may be reasonable.
Otherwise, look for contact close in, possibly in the sub itself.

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