Smell of ozone from MV switchgear 7

[Thedroid]

We recently had our GE Magne-blast 1200A breakers serviced during our annual outage. It's been about 3-4 weeks, and we're starting to notice the smell of ozone coming from the back of one of the switchgear cabinets.This breaker provides power to a transformer and also a MV motor through a parallel feed. The original set of cables is shielded, and the other set was added on at a later date and is unshielded. They run through seperate conduits, and are pretty close to the same length. I'll say within 10ft. They both terminate on the bus of the load break switch, and another run of shielded cables to feed the mill soft-starter is also connected to the same bus.

Today we racked out the breaker and removed the back cover to look for any sign of corona effect or tracking. We could not find any obvious damage to any of the insulation, but did notice that some of the cables were bent sharply, and also contacting the switchgear cabinet.

The unshielded cables were the ones contacting the enclosure, so we shortened the cables by about a foot, and made a support for them out of a fiberglass insulator to keep them away from any metal. The original terminations were real sloppy. The ends looked like the had been wittled instead of stripped, and the fabric material under the insulation had been left hanging at the termination. We redid these terminations, and wrapped the cables with scotch 70 about 8 inches back. We inspected everything that was visible, and could not find any signs of the insulation deteriorating.

We then inspeceted the breaker, for any signs of tracking or powder, or anything out of the ordinary, and could not find anything. This breaker was just hi potted and ductor tested a month ago, and was found to be ok.

After our inspection, we put the breaker back in service and noticed the smell of ozone coming from the back vents with about 5 minutes. The smell is there even with no load on the circuit.

Are there any other areas I can check, and how big of deal is this. I feel that it's a definate sign of a failure coming, and also the smell of ozone is choking.

 

[electricpete]

If it really is ozone, I agree with you it is a sign of possible failure - not something to be taken lightly.

As a small comment - OSHA has limits on ozone - somewhere around 20ppb which is far below level where can easily be smelled.

You can get a very cheap hand-pumped sniffer by "Draeger" (the kind with disposable tubes) to test for ozone gas and determine the concentration. Perhaps there could be smell coming from some other source like offgassing of some coil varnsih or other coatings which was mistaken for ozone.


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[alehman]

If this is coming from the gear and it accumulates that quickly, it sounds like a serious problem. I'd contact an electrical testing firm that has partial discharge testing or ultrasonic measuring equipment to help locate the source. And do it right away.

 

[Thedroid]

No new varnish or coatings were applied. I've smelled ozone generators for odor control, and would say that this is the same smell coming from the back of the switchgear. Definately not a burning smell, more of a bite you in the back of the throat smell. It goes away after opening and removing the breaker, and comes back right away after racking the breaker back in. We are going to try another breaker tomorrow, and see if that solves the problem. If not, I suspect the line or load side disconnects of the switchgear, as these have not been inspected. Taking down the utility is probably not an option, especially after completing the shutdown. I'm going to look into some forms of PD testing to possibly find the cause of the problem

 

[Thedroid]

alehman,

My thoughts exactly. I'm predicting a failure in the equipment soon if we cannot locate and repair the problem. The smell of ozone is strong enough to be choking, and is definately coming from this cubicle. I will recomment PD testing tomorrow, and hope that it does not fall on deaf ears. We just went to a reduced work week this week thats expected to last 2 months, so any additional expenses will not be appreciated, but a flashover in the gear would be far more catastrophic.

 

[pwrtran]

Sounds like quite a lot ionization there caused by field stress at some point. I would keep eyes more on cable and the terminations. Is your shielded cable grounded at both ends? Can you also do a PD test on the cable?

 

[ScottyUK]

Although it's likely too late to help you in this instance you might want to consider one of the TEV instruments manufactured by EA Technology in England. I'm not sure who else makes this type of device. I've used them and they give a good indication of incipient PD problems. Look in the lower half of this page TEV Instruments.

Based on what you're describing entry to the switch house should be restricted to the absolute minimum, and then only for people wearing suitable PPE.

----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[Zogzog]

Sounds like you have a potential major problem, like memntioned by others PD detection is your best bet here, these guys are great and use similar equipment to what ScottyUK posted.

http://www.nooutageelectricaltesting.com/

They will be able to pinpoint your poblem in no time. You want PD detection in this case, not PD testing.

I would also call the testing company that serviced your gear, I bet they come out for free to look at it, I know I would for sure if I were them.

 

[Thedroid]

I sure hope not. I will pass on the name of the company recommended today, and maybe we will find something else after further inspection. How critical are the lenghts of the the parallel lines coming off of the breaker? One of our guys seems to think it's a problem.

 

[dpmac]

Have you taken the breaker out of the cell and removed the barrier boxes to do a visual inspection of the breaker its self?

Are the rear load side bus bars where the cable connections are made supported or clamped by GPO-3 (red or orange fiberglass)? Look closely where the bus bars are clamped or supported for corona damage.

Is there a PT drawer located above the cell that is energized when that breaker is closed? If so have you inspected it?

As Zog mentioned, a call to the testing firm expressing your concerns should bring a couple of techs out running in no time.

Surprised you don't hear anything with that kind of odor.

Please keep us posted.

 

[BSEE1980]

I agree...you are headed for a replay of the Big Bang Theory!

You say MV: what level? 5KV is much different than 15KV class. I have seen 5KV in unshielded cable: you would never do that with 15KV.

When the PM's the breakers, did they use any cleaner on the insulators? I used to do a lot of PM's and there are certain things you should not use on the insulators. Depending on the age of the CB, they used a compressed epoxy or wrapped paper/exopy or ceramic insulators on GE MB breakers. If they used the wrong cleaner, they could have damaged the insulators.

You can detect corona (which, of course, leads to ozone) with a older style CB walkie-talkie. It needs a manual squelch. As you get closer to the source, it will breaker higher and higher levels of squelch. Start outside, away from the swgr. Just set it to recieve, turn up the squelch until it quiets down. Walk into the room: it will probably break squelch immediately. Turn up the squelch until it stops, then walk slowly until it breaks squelch again. It is not the volume, but the squelch level that you are looking for. Crude, but cheap and effective. I suggest CB frequencies (about 27Mhz): business band radios can be 800Mhz or higher and less effected. Ham would be even better as they have even lower Hz. The corona is, of course, 60Hz and maybe up to the 5th harmonic (300Hz).

Good luck and be careful!! You are wise to take action now: this is a disaster waiting to happen. Please use your NFPA 70E arc flash gear! Take this seriously! I saw the result of a LV CB failure and if someone would have been standing in front with the door open, they would have been dead.

Steve

 

[waross]

Not critical.
Example: Two cables, one 100 feet long and the other 101 feet long, load 200 Amps.
One cable will carry about 99.5 Amps and the other will carry 100.498 Amps.
Random placing of the cables in regards to each other, other phases and to ground will cause much more difference than this due to reactance issues.
Look for poor workman ship such as one strand of a cable sticking out or a dirty hand print on an insulator.
Another possibility is that the insulation on a conductor may have become brittle with age and been damaged by a bending movement that would be allowable on a new cable.

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

 

[Compositepro]

Check carefully for sharp edges on both energized parts and grounded surfaces near energized parts. Your problem could be caused by a metal shaving on a drilled hole or a loose wire strand as waross said. Corona discharge is caused by electric field concentration due to small gap or sharp edge. Look for dust accumulation that tends to collect around discharges.

 

[itsmoked]

Is there any way to observe the inside of the unit energized with the back off? In the dark,(no area lights), with a telescope any discharge big enough to put out that much ozone would be easily visible - if it's in the line of sight.

I do this with automotive ignition problems. Drive to a dark spot open the hood. Let my eyes dark adapt and if the spark plug wires have deteriorated you can see the corona glow in many places. Once it looked like a rural town from an airliner.

Anyway it would be relatively easy to set up quickly and it will show you where the problem is,(if in sight), and where the problem isn't,(everywhere you can see).

Keith Cress
kcress -

http://www.flaminsystems.com

 

[dpc]

I would not open any cable compartments without a lot of arc-resistant PPE on. Even just to look.



"An 'expert' is someone who has made every possible mistake in a very narrow field of study." -- Edward Teller

 

[Thedroid]

I've been very careful with all inspections and racking of the CB and have been wearing full PPE. I did inspect the back of the compartment live in the dark, and could not see any visible discharge or corona. The smell was less intense this morning, after reterminating the unshielded cables. I swapped the circuit breaker for our spare and the smell disappeared within a few hours. The power level is 4.16KV. The load side bus supports are orange fiberglass and the back side looks clean. These bars are wrapped with some kind of black cloth tape where they contact the insulators. The front side of these insulators would be hard to inspect, but not impossible. I'm not aware of any PT compartment in this cubicle and the CT's must be in the line side compartment. There is one large CT in the back compartment for GF protection. The testing company said they hi potted the breaker at 11KV and found no problems. We are going to pick up a few of the disposable ozone testers to make sure it is ozone we're smelling if it comes back. I'm keeping a close eye, or nose, on the gear to make sure it doesn't come back,and if it does I'm hoping we can get someone out here to do a PD detection for us.

 

[Zogzog]

Thedroid, what state are you in?

 

[Thedroid]

New Mexico

 

[Zogzog]

Sooo, humidity isnt the problem

 

[Thedroid]

No definately not. Dry as a bone out here. Ran the mill last night, and the smell is back, just not as strong as before. We are going to check closer on Monday. I don't think it could be the line side bus, because it goes away when we rack out the breaker. Is there any way to inspect the breaker spouts while the gear is energized? Also going to check the load side bus on the other side of the insulator.