Shattered ACSR Conductor

[jghrist]

One of our clients had a tree fall on an overhead 12 kV line, breaking two phases. Four spans toward the substation, two conductors of the same phase broke and fell to the ground. We suspected that the initial fault (3300A) caused the phases to slap together on the other span causing a burn down. There wasn't a lot of sag and the conductors were on 8 foot arms, but the phases that burned down were on one side of the pole with 24" spacing.

A strange phenomenon was observed after the fault. There were dozens of small straight pieces of conductor strand scattered around where the second conductor break occurred. See photo at

http://www.utilitytec.com/office use.html

(click on IMG 0387.jpg). Have you seen anything like this? It's like the conductor shattered. The strand ends are not stretched like in a tension break or burnt. The conductor may have been thrashing around on the ground (part asphalt roadway, part concrete sidewalk, part grass) before the fault was cleared.

One theory is that the conductor had experienced overheating in the past and had annealed and become brittle. If this is the case, replacement of all of the conductor may be required. I tend to doubt this theory; the strand I have is not brittle, and according to a Belden glossary, the definition of anneal is:

Anneal
To soften and relieve strains in any solid material, such as metal or glass, by heating to just below its melting point and then slowly cooling it. Annealing generally lowers the tensile strength of the material, while improving its flex life and flexibility.

Weather was 40°F, windy (gusts to 30 mph), cloudy, but no rain or lightning. Conductor is #1/0 ACSR. Total fault clearing time was 32 seconds, with seven apparent instantaneous relay operations and six apparent breaker failures to operate. The relay (ABB IMPRS) fault records are confusing and we are consulting with ABB to make sense of these.

 

[Skogsgurra]

I have seen a similar thing when a multistrand copper wire snapped off and hit a concrete wall with high speed. There were lots of strands coming lose and we thought it was like a toffee bar that normally is quite soft and can be bent easily but gets brittle when danged into a hard surface.

Gunnar Englund

http://www.gke.org

 

[wareagle]

Well 32 seconds is way too long to clear a fault. Do you have ground relays to monitor a ground fault? 3200 amps is not a large fault.

 

[jghrist]

Yes, 32 seconds is too long. The breaker should have tripped on instantaneous (pickup 1800A phase, 1050A ground) in less than 0.1 sec. Line set for reclosing of 2 sec, 15 sec, and 30 sec. After first reclose, instantaneous is cutout and time overcurrent should trip in 0.8 sec for a 3300A phase fault or 0.83 sec for ground fault. We think there was a high impedance fault with the conductor on the ground, with intermittent short bursts of current as two phase conductors touched each other or the neutral.

This became a real mess, with all four 12 kV circuits from the substation having faults because of flashovers of insulators and jumpers burning down. One circuit had a hole burn in a single phase regulator and an oil fire that cooked one breaker. The bus flashed over. The substation is served at 12 kV from the power supplier's 100 kV - 12 kV substation. To top it all off, the suppliers 100 kV breakers failed to trip and the fault was finally cleared when the supplier's sudden pressure relay called on gang operated disconnects (non-loadbreak) to open under fault conditions, causing an arc to ground and tripping of a 100 kV tie line that is tapped to serve the distribution sub.

We're trying to piece things together with ABB IMPRS relay records that don't make a lot of sense. Records from three circuits show multiple events with BFO (Breaker Fail to Operate) on all circuits. We can't get records from the fourth circuit because the relay was cooked in the fire. Testing of the three circuits after the fault shows all relays and breakers operating properly.

 

[waross]

I have seen similar wire fragments on a smaller scale. It was small transformers failing under short circuit conditions. (250 watt ballasts.) The connections were set wrong and 240 volts was applied to one 120 volt winding with the other 120 volt winding short-circuited. Most of the fixtures had a quantity of wires about 3/4" to 1" long, blown out of the transformer. I assumed that it was the result of a combination of overcurrent (heat) and magnetic forces.
It may be a related effect or maybe not.
respectfully

 

[alehman]

When a fault occurs on an overheaad span, the conductors can whip about quite violently (I saw this happen a couple of years ago and remember being alarmed by the violence of the forces). If the fault took 30 seconds to clear, maybe the breaks are a result of metal fatigue. Just a thought...

 

[alehman]

Also in the case I witnesed the adjacent phases did slap against one another several times rather loudly, creating another fault when the breaker re-closed. I was about 100ft away or so - too close for comfort.

 

[jghrist]

I think you're right about the conductors slapping together. Either this or the conductors hitting the pavement caused the shattering.

Actually, we don't think the fault lasted for a full 32 seconds. The relay records show 7 events in a span of 32 seconds. We think there were short bursts of current when the conductors touched, followed by low currents when the conductors were laying on the ground or swinging around in air. The first five reported trip times for the circuit with the downed conductors were less than 0.02 seconds, followed by two time-overcurrent relay events with trip times of 0.641 and 0.454 seconds. Problem is, the first five events also showed BFO, indicating a failure to trip. I think the first 4 events were short bursts without the breaker tripping. The BFO indications have me baffled. This would indicate that the relay issued trip signals.

When all was cleared, the breaker for the line with the downed conductors was open and the closing spring discharged. The other three breakers were closed.

 

[alehman]

Sounds like the relay probably worked correctly but the breaker didn't open. That seems wrong. Hmm... may be time to get the breaker tested. Is your station battery in good shape?

 

[jghrist]

The breakers use capacitive trip devices. Post-fault tests indicate that the breakers will trip without AC power. Breakers were tested two months prior to the event.

All of the relay event reports that we were able to download indicated multiple Breaker Failure to Operate (BFO) events along with relay operations. The other circuits all had fault events with time-overcurrent (51) trip times very close to the relay time-current characteristics, but each event had a BFO record as well as a 51 trip record.

ABB confirms that the trip time indicated is the time between current exceeding pickup and decreasing below pickup. If the breakers failed to open, why would the current stop in the amount of time indicated by the relay T-C characteristics?

Relay event records:[tt]
Circuit Event Device Tr time Ph A Ph B Ph C Neut Time
11201 34 BFO 0.433 5367 5839 5850 4583 7:22:35 PM
11201 34 51P 0.433 5367 5839 5850 4583 7:22:35 PM
11201 35 BFO 0.462 3508 5741 5658 3793 7:22:37 PM
11201 35 51P 0.462 5620 5741 6275 4948 7:22:41 PM
11201 36 BFO 0.487 4993 5068 0 4312 7:22:46 PM
11201 36 51P 0.487 4993 5068 0 4312 7:22:47 PM
11201 37 BFO 0.454 5191 5667 0 3202 7:22:47 PM
11201 37 51P 0.454 5191 5667 5573 3202 7:22:49 PM
11201 38 BFO 0.437 6500 6370 4 6203 7:22:50 PM
11201 38 51P 0.437 6500 6370 4 6203 7:22:51 PM
11201 39 BFO 0.529 6268 4 2 6293 7:22:52 PM
11201 39 51P 0.529 6268 21 4 6293 7:22:53 PM
11201 40 BFO 0.416 5754 6321 6302 5769 7:22:54 PM
11201 40 51P 0.416 5754 6321 6302 5769 7:22:55 PM
11201 41 BFO 0.416 5994 6283 6239 6139 7:22:55 PM
11201 41 51P 0.416 6200 6283 6239 6235 7:22:58 PM
11202 28 BFO 0.012 3344 3170 161 60 7:22:03 PM
11202 28 50P 0.012 3344 3270 3147 1950 7:22:06 PM
11202 29 BFO 0.012 3253 3353 299 1820 7:22:07 PM
11202 29 50N 0.012 3253 3353 299 2112 7:22:08 PM
11202 30 BFO 0.020 253 2879 240 2492 7:22:08 PM
11202 30 50N 0.02 3202 3174 255 2492 7:22:09 PM
11202 31 BFO 0.008 2350 3346 3058 17 7:22:10 PM
11202 31 50P 0.008 4199 4740 3848 3140 7:22:19 PM
11202 32 BFO 0.008 2904 4337 4063 2476 7:22:22 PM
11202 32 50P 0.008 6502 5389 4063 4628 7:22:22 PM
11202 33 BFO 0.641 7048 5221 3757 1822 7:22:24 PM
11202 33 51P 0.641 7048 6508 5675 4882 7:22:34 PM
11202 34 51P 0.454 240 5323 5272 3260 7:22:35 PM
11204 43 BFO 1.833 2 1508 2 1484 7:22:33 PM
11204 43 51N 1.833 2 1508 2 1484 7:22:47 PM
11204 44 BFO 1.383 150 1731 142 1957 7:22:55 PM
11204 44 51N 1.383 155 1731 155 1957 7:22:56 PM

[/tt]
Circuit 11202 was the one with the downed conductors.

 

[alehman]

I'm not sure I totally understand this event log. Have you ever seen a trip event that does not include 'BFO'? I see BFO's on the other circuits as well. Maybe the monitoring contact is mis-wired?

 

[jghrist]

"Event" is a serial number for each circuit. "Tr Time" is time from start of current to end of current. "Ph A", "Ph B", "Ph C", and "Neut" are currents in amperes.

When relays were tested post-fault, we got BFOs on one relay that had a breaker operating time setting erroneously set to zero. When we set it to 125 ms, we got no BFOs. I thought that this relay was a spare that was replacing the one destroyed relay, but I'll have to check to ensure that it was not one out of a breaker involved in the faults.

The other relays tested were set properly and testing produced no BFOs. Records from past incidents have not produced BFO events except in one case where we know that the breaker did fail to trip.

Note that on Circuit 11202, there are 7 events in 32 seconds. The breaker could not have physically tripped and reclosed this many times in 32 seconds; not enough time to recharge the closing spring that many times.

 

[stevenal]

Agreed, the breaker did not trip and reclose for each 50 or 51 event. 0.008 trip time? That's either an extremely fast breaker or the current dropped below pickup on its own. Is it possible the relay only closes for the duration of the programmed breaker operate time (0.0s)? Other relays I'm familiar with open their contacts once the breaker fail condition has been determined.

 

[jghrist]

I confirmed with the Electric Superintendent that was helping with the testing that the relay that had the zero breaker operating time setting was a spare, not one in service at the time of the fault. That means all relays in service had a time of 125 ms; meaning none of the BFO events make sense.

The Electric Superintendent was home, on Cct 11203 (with destroyed relay) at the time of the fault. He observed three trips and reclosures on the circuit before everything went dead. A customer on Cct 11204 observed two trips and two reclosures, along with two voltage dips. This is consistent with the relay records, ignoring the BFOs.

The breakers are not 1/2 cycle breakers (we couldn't find any of these ;-)). In fact, the instantaneous is supposed to have time between 0.022 and 0.032 seconds. I think we had very short bursts of current when conductors slapped together, and the breaker did not trip until the fault on the Cct 11201/11202 tie switches occurred.

One customer reported seeing a "fireball" moving down the line near the original fault point on Cct 11201. One theory was that the arc moved from the original point down the line to the 11201/11202 tie switches and then to the station. I initially discounted this report because I don't think an arc of this length (at least 24") on a 12 kV line could sustain itself long enough to travel a mile, especially on a windy day. I'm beginning to doubt my doubts! Ever hear of anything like this? Unless the arc had a high resistance and limited the current to a low value, it would be picked up by the relay as a long duration current. According to a formula by A.R. van C. Warrington, the arc resistance with a 24" arc length and 800A current would be 1.5 ohms. This would be consistent with an 800A current on this circuit. 800A is below the instantaneous pickup and would take 20-30 seconds to trip on time-overcurrent. This wouldn't explain the BFOs, but it would explain how the other circuits could get involved in the fault.

Maybe we had a combination of conductors on other spans swinging together without burning down, and an arc moving from the contact points down the line. I'll have to do some research to see if I can find a published report of something like this before I can buy this scenario. It all sounds too fanciful to me. I'd rather take the old standby and blame it on harmonics or ferroresonance first.

 

[alehman]

11204 looks reasonable, ignoring the BFO's. I'm not surprised to see the very short events considering how messy this seems to have been. I wonder why, in most cases, the BFO's are recorded at different times than the trip events. Some relays have two failure timers. Make sure you don't have any extraneous timers that aren't disabled.

The traveling arc seems very unlikely to me. I vote for harmonics.

 

[subtech]

I've seen "shattered" wire before like you stated in your first post.
In our situation, the conductor was #2 acsr and the steel core had deteriorated to the point that a very small strain would fracture the core and leave only the aluminum outer strands. On more than one occasion, these conductors had been severely overloaded and the tensile strength left in the aluminum was minimal. One trait that seemed common in our case was that all of the aluminum strands had taken on a deep grey to near black color and even seemed "etched". It was very easy to just snap off pieces in about any length with very little effort. How heavily has your 1/0 ckt. been loaded?
If your conductors were in similar condtion to ours when the fault occurred, I can certainly understand the "shattered" appearance of the remnants after this violent event.
Regards

 

[jghrist]

The #1/0 section of line is fairly heavily loaded, peak load is estimated to be 60% of rated. The strands are not brittle, they bend without breaking. As you can see from the photo at

http://www.utilitytec.com/office use.html,

the surface is quite beat up, probably from forcefully hitting either other conductors or the pavement. I like skogsgurra's explanation of brittleness when hit with force, but bendable like toffee otherwise. The parts of the steel core that I could see was not corroded.

 

[RalphChristie]

Jim

Few comments, sorry for the late reply:

Regarding conductor, I am not familiar with your conductor sizes/names, if I am not mistaken it would be called rabbit-conductor here. I've seen similar results when a dozer pulled a overhead pole (used in our mining environment to lift the trailing cables into the air for the trucks to pass underneath) accidentally into a 66kV overhead line. Strand were also shattered around the area. The line were not heavily loaded, but it was subjected to a lot of faults in the past.
On a few occasions I've seen conductors slapping together during heavy faults, and every time I was amazed by the force generated.

Regarding capacitor back-up supplies, I do not like it very much, we use a lot of it in our remote 11kV stand-alone switch-cubicles. With the back-up new, it behaves correctly during fault conditions, but after a few years in the field we experience problems, especially with faults involving more than one phase. (2ph, 2ph-E and 3ph faults) We do also do periodically testing of the units, and during tests it behaved correctly. I’ve found the capacitors to deteriorate over a period, and during tests the breaker is normally just tripped once to ensure positive operation. In my opinion this is the first place to check if you experience breaker-failures. This can also indicate the lot of BFO-logs.


[red]Failure seldom stops us, it is the fear for failure that stops us - Jack Lemmon[/red]

Make the best use of Eng-Tips.com
Read the Site Policies at FAQ731-376

 

[jghrist]

Ralph,
Thanks for your observations. I suppose that the problem with capacitive trip units being more prevelant in multi-phase faults is that the single-phase ac supply is more likely to be affected. I agree that the capacitive trip devices are likely to be at least part of the problem.

I wonder if the conductor shattering could be caused by rapid vaporization of moisture in the conductor during arcing. Problem with this theory is, it wasn't raining on the day of the faults. The violence of the conductors slapping together combined with the high arc energy probably is enough to shatter the conductor strands.