arrester failures 2

[alehman]

We experienced failure of the arresters I mentioned in thread238-213060. The utility feeding this switchgear is 4800V ungrounded with about 21kA sym available. The circuits are from a common substation bus, ungrounded, and do not have recloser operation. I selected an intermediate class polymer 6kV gapless MOV arrester, rated 40KA pressure relief. Their 10kA, 8/20 crest voltage is 16.7kV. They are installed in separate compartments in metal-enclosed interrupter switchgear, rated 15kV.

This is a main-tie-main setup with the arresters in separate back-to-back compartments. The side was blown out of the C-phase arrester on main #2. About 15 minutes later there was a fault in the arrester compartment protecting main#1 and its C-phase arrester was destroyed. I can't tell if the origin of the second fault was internal to the arrester or external. There was considerable arcing in both arrester compartments, but I don't know exactly when it occurred. Fuses on both mains were found blown, but again I don't know exactly when. The utility reported a ground fault for roughly the time interval between our two events.

The utility engineers and local electricians are suggesting a higher voltage rating. The arresters protect oil-insulated transformers rated 60kV BIL that are about 50 feet away.

Any thoughts on changing the arresters to 9 or 10kV ? Any other thoughts on the failures or replacement arresters type selection?

 

[Kiribanda]

alehman,

If I have understood your post correctly,

1)Is 6 kV=MOV of the said arresters?
2)How did you install arresters with MOV=6 kV for a 15 kV ungrounded system?

 

[alehman]

1. Yes. 6kV arresters.
2. The system is 4.8kV ungrounded. The switchgear is rated 15kV.

Thanks,

 

[magoo2]

I'm assuming the 6 kV is probably a duty cycle rating. The corresponding MCOV would be 5.1 kV.

With voltage regulation, the voltage could be higher than nominal. 5.1/4.8 gives you 1.0625 pu, so you could tolerate up to 6.25% higher voltage.

The polymer arrester failure description with the side blown out sounds like an overvoltage failure. The polymer housing usually splits and you get a power arc external to the blocks.

I'd be inclined to replace it with a 9 kV duty cycle rating (MCOV is 7.65 kV).

The protection characteristics would be roughly 1.5 times the 16.7 kV or 25.1 kV. I estimated this based on the ratio of the arrester ratings (9/6). Even if the surge doubles, you have almost the required 20% margin.

 

[alehman]

Yes, 5.1kV MCOV. This was as initially recommended by the manufacturer, the utility engineer, and as recommended by IEEE 62.22.

The utility states they maintain +-5% voltage. They are trying to obtain their substation voltage recorder data for the time leading up to the event.

 

[opmgr1]

I would not be inclined to change the arrester to 9kV. This is for higher system voltage and would continue to be guided by IEEE 62.11 table 1. I need to ask some questions though and the answer should point to the problem.

What is the BIL of the entire system? is this more, less than or equal to the arrester BIL?

What is the ground resistance at the point where the arrester is grounded?

 

[magoo2]

opmgr1
It's good to have noble preferences but you're not offering alehman any alternative. Leaving a 6 kV rating in will probably destroy more arresters.

Sure the 9 kV rating is for higher voltage systems, but it will work. You're sacrificing some protective margin, but you should be OK.

System BIL - line should be higher than the transformer and other equipment (if any) should have their own arresters.

I'm not sure the ground resistance would change anything if we're dealing with a sustained overvoltage issue.

 

[alehman]

opmgr1,
Switchgear is 95kV BIL (except for the 6kV arresters), cables are 15kV 133% EPR, transformers are 60kV BIL. There is an extensive grounding system, but I do not know the resistance.

My thought process was that the normal line-ground voltage should be significantly <4.8kV since the utility does not allow ground faults to persist.

A 9kV arrester in the same product line would have 25.0kV 10kA/8x20 crest.

 

[opmgr1]

Magoo,
I dont agree with your post. If the problem is one of sustained over-voltage The approach is to find out what is causing the overvoltage (temporary or sustained) and fix that, not putting bandage on the situation by increasing the arrester duty cycle rating.

Alehman, looking at your post re L-G voltage are you saying that this system is an un-grounded wye system and that the C-phase to ground voltage is close to 4.8kV?

 

[alehman]

The system is 4.8kV nominal, ungrounded. I don't know the actual voltages, except at the secondary of my DY transf's.

 

[rbulsara]

I would look for the source/cause of the overvoltages first and correct that. Switching transients/ Lightning/resonance??

9KV MOV may help if the source and mangitude of OV is known such as switccing transient below the ratig of the MOV. On the other hand it may cause greater damage (when operates) or mask something you do not want to mask, if the source of OV is per say restriking ground fault/resonance.

 

[opmgr1]

Is this a wye or delta connected system?

 

[MxGxK]

Wouldn't referencing one phase of an ungrounded system to ground cause all phase to neutral connected equipment to see line to line voltage?

 

[MxGxK]

Correction: The phase to neutral equipment connected on the phases not referenced to ground would experience line to line voltage. The next ground fault (overvoltage of the arrester) would result in a phase to phase fault. Arcing could result in a 3phase fault and explain your blown fuses. Just a thought.

 

[alehman]

opmgr,
The utility refers to their service as 4.8kV delta.

mxgxk,
I've gone through several scenarios as to the sequence in my mind. Personnel at the facility reported two audible explosions. After the first explosion loads fed from main#2 were de-energized and there was considerable smoke coming out of the switchgear (I don't know the status of the fuses at that point). The utility says the ground fault current is about 2 to 3 amps. I wouldn't think that would be enough to explain the audible explosion. One possible scenario is there was arcing phase-to-phase after the first failure and one or two fuses blew, leaving a ground fault.

Then after about 15 minutes, there was a 2nd explosion and the remaining loads were off. This could have been caused by the arrester on the 2nd service failing or an external flash-over on the arrester due to residue from the first fault.

 

[MxGxK]

For a system sourced by a delta, arresters should be applied with an MCOV greater than the line to line voltage. Otherwise, a ground on any phase (no fault current will flow) will subject the arresters (connected phase to neutral) to line to line voltage. The arresters will fail rather quickly if misapplied.

 

[magoo2]

MxGxK
The present arresters are applied with an MCOV greater than the line-to-line voltage, but it could be close.

alehman
How much cable are you dealing with between the utility sub and the switchgear? I could envision the utility regulating around 4800 V +/-5% at the sub, but with enough cable, you'll get voltage rise at light load that will raise the voltage unless the regulators/LTC have line drop compensation. So you could be at the ragged edge as far as the 5.1 kV MCOV.

With a ground fault, 2 of the phases will see this overvoltage, as MxGxK points out. From the utility perspective, they might not detect this condition unless they have special relaying since it really doesn't draw much fault current, only some cable charging current. With it undetected, it will subject the 2 unfaulted phases and their arresters to a damaging level of overvoltage.

Maybe alehman can see if I'm off base with these assumptions. Also do you have sensitive ground fault detection in the switchgear or it is just set to alarm? This should help better define the problem.

 

[burnt2x]

magoo2 nailed it!

"The utility says the ground fault current is about 2 to 3 amps. I wouldn't think that would be enough to explain the audible explosion. One possible scenario is there was arcing phase-to-phase after the first failure and one or two fuses blew, leaving a ground fault.

No matter how many times you replace your arresters, you will get more arresters damaged until you solve the problem of arcing L-G faults in your system. 2 to 3 amps is a typical charging current of ungrounded 4.8kV systems. With the system capacitance highly charged, the threshold voltage could have been reached; re-strikes occurred 'till the arresters gave way!
Try reviewing your line protection and see if you can use the option to trip if L-G fault occurs and not just alarm. Broken-delta ground fault sensing can help.
My 2 cents!

 

[alehman]

Our switchgear is manually operated fused switches with no other protection options available. The ground fault sensing I mentioned is the utility's and located at their substation. I don't know any specifics on their relay.

I see your point about voltage regulation. Our secondary voltage runs a few percent high, but I need to verify the transformer tap settings.

I'm still waiting on voltage logs from the utility, but they have said there were no other recent ground fault events.

 

[opmgr1]

These ungrounded systems are succeptable to ferroresonant over voltage, overvoltage (as much as 1.73pu rise) as a result of earth fault transient overvoltage, and overvoltage as a result of switching switching (whether deliberate or as a result of protection from faults). If there is upstream switching as much as 10pu overvoltage burst can occur which stress the insulation of the arrester resulting in failure. Of course it is much cheaper to replace the arrester than the equipment that it is protecting.
Note if there is a transient fault there can be a build up of high voltages (up to 6 times nominal). This high transient voltage can initiate a second fault at the weakest insulation point resulting in failure. For this scenario you definately do not want to increase the duty cycle rating of the arrester. As one poster hinted at you may want to invest in SEF protection if this is a financially viable option or it may be cheaper to replace arresters.