Switchgear Failure Vacuum Bottle & Arresters 3

[Chad12]

Hello all, I have another interesting event to pass along..

Recently I was testing a 15KV Cable (EPR 133% insulation). This cable was a combination of old and new cables spliced together and terminated at several switches, with a total length of about 5,000 feet. The customer requested a test voltage of 13.2kv (I know...it's low). Voltage was first raised to 6.6kv, held for 1 minute, then raised to 13.2kv and held for five minutes. The average leakage current for each phase was about 30 microamps. Upon re-energization of this circuit after testing, the Square D VR Vacuum circuit breaker had a bottle failure on the A Phase. Upon inspection of the rear of the cubicle it was noticed that the arresters were burned. The cable was retested and the C Phase arrester was found to be shorted. The assumption (not mine) is that the arresters were damaged by the overpotential test. These arresters are 10KV rated.

Chad Snyder
Director of Operations
American Power Technology

http://www.ampowertech.com

 

[jghrist]

Arresters don't have a high withstand voltage - that's why they are used to protect equipment. Exposing a metal-oxide varistor to a voltage above it's TOV (Temporary Overvoltage) rating will make it fail.

13.2 kV is 1.57 times the MCOV (Maximum Continuous Overvoltage) rating of a 10 kV arrester (8.4 kV). An Ohio Brass Type PVR arrester will withstand a 1.57 overvoltage for less than 0.1 seconds according to the catalog.

At least it didn't fail violently and injure someone.

 

[dpmac]

Is it possible that 30 microamps of DC current could damage what is probably an intermediate class arrestor?

 

[rbulsara]

Agree with jghristm, who beat me to the post!

But there is something else which may not be in your favor.

Regardless of the actual cause, it appears that basic precautions/procedures were not followed for the hi pot testing. Cables must be disconnected from all equipment including the surge arresters, switches and breakers prior to hi-pot testing. This itself could put you in an awkward situation.

Also why hipot test both new and old cables together? I would have only hi-pot tested new cables, disconnected from everything else and only meggar tested the old (that too when isolated).

Was the 13.2 kV applied between ground and the conductor or between phases? AC or DC?

Surge arresters have something called maximum continuous operating voltage rating, which is about 84% of the SA rating. You may want to check with the SA mfr. So if you exceed 8.4kV for a certain period (this is what I am not sure of, how long), you could damage the SA.

10kV SA are selected because the L-G voltage would be around 7620V for a 13.2 kV L-L system.

Different type of SA, silicon vs. MOV may also have different behavior to overvoltages.

Rafiq Bulsara

http://www.srengineersct.com

 

[rhatcher]

I am very skeptical that you are to blame for this.

Your test results of 13.2kV (DC?) with a leakage current of 30 micro-amps indicate an insulation resistance phase-phase and phase-ground of 440 mega-ohms. This is a very good reading by most standards.

If a surge arrestor or anything else in the circuit failed during your test, I would expect the high potential tester to trip. Period.

If your test gave the expected results with no trips of the test set and steady leakage current (or stable IR reading) as the voltage was increased, then there was no failure of any (ANY) downstream equipment during the test or as a result of the test.

With respect to the Square D VR Switchgear, check page 31 of the Square D VR switchgear instruction manual that is linked below. The high potential voltages for field tests of 15kV gear are 27kVAC or 38kVDC. You didn't come close to that. In fact, you didn't exceed the test voltage for 4.16kV gear.

http://ecatalog.squared.com/pubs/El...r, Metal Clad/Masterclad Series 5/6055-30.pdf

I'll finish with the final bit of skepticism...Why did they have you perform this test in the first place? If this was a routine test as part of a regularly scheduled maintenance cycle then that is one thing. But...If someone called out of the blue for a 'random' test and then afterwards reported an equipment failure and blamed it on you then, I would be very skeptical.

That is just my opinion based on experience.

 

[LionelHutz]

You didn't short the surge arrestor. 30uA @ 13.2kV = 0.4W which is not enough to overheat and damage the arrestor.

A higher energy transient will cause enough heating to damage some of the bulk material the arrestor is made from. Exposed to enough such events they do fail.

If 30 micro-amps damaged the surge arrestor then that arrestor would be useless for protecting against a transient with some energy behind it.

As for the voltage rating of them - They are sized for line-neutral voltage. Check that the Y or neutral point of the transformer feeding this power system is grounded. The system should have 15kV rated arrestors if it is not solidly grounded.

 

[jghrist]

The arresters were found shorted after the test. They may have been damaged by the test but didn't fail until reenergization of the circuit.

The other explanation is that they failed immediately after the test completely by coincidence. I don't believe it.

Maybe the arrester failed sometime before the test and the Square D VR Vacuum circuit breaker was open during the test.

The arresters should not have been exposed to the overvoltage. As Rafiq said, they should have been disconnected before the test.

 

[dpmac]

How do we explain the failed vacuum bottle?

 

[LionelHutz]

It still doesn't make any logical sense that an arrestor of that size would be destroyed by the application of a short duration 30uA current. A GE arrestor of that type is rated at 28.5kV for a 10kA, 0.5uS surge.

The above is 142.5J - to reach this same energy with what you applied (assuming all 30uA was through the arrestor) would take 356 seconds or 5.9 minutes. Even then, that would not cause the final destruction of the arrestor unless it was near the end of it's service life. And if it did, your tester would have tripped.

If 30uA for a short period of time could damage a large surge arrestor then they'd be failing every time a power system was turned on.

Besides all this, it doesn't even sound like you reached the turn-on voltage of the arrestor. I expect the 10kV RMS rated arrestor would begin to conduct at closer to 14kV. At the turn-on knee point, the arrestor begins to really conduct so the test set voltage almost stops rising completely and the current begins to rise very quickly. Been there, done that and didn't short the arrestor.

 

[LionelHutz]

How do we explain the failed vacuum bottle?

That's a good question - the phase C arrestor is shorted and the phase A bottle fails. I don't know how to explain that. Could have been that a 3-phase phase-phase or phase-ground fault event occurred and phase A just happened to be weak and let go.

 

[jghrist]

There are some unknowns:

1. When the circuit was reenergized after the initial test and the vacuum bottle failed, what tripped the circuit?

2. Was the vacuum bottle failed with contacts welded closed?

3. Was this an ac or dc test?

4. Was the vacuum bottle replaced before the retest?

5. When the cable was retested and the arrester found shorted, did this cause the test set to trip immediately?

6. Were the vacuum bottle and arrester OK before the first test?

7. The ØA vacuum bottle failed after the first test and arresters were burned. Were all three arresters burned?

8. Were the other phases of the VR breaker open after the first test? If so, what tripped the breaker?

9. The ØC arrester was shorted after the second test. Was the breaker tripped? If so, what tripped the breaker?