intermittent arcing faults on ungrounded systems - 480vac 10

[electricpete]

We've heard some interesting things on thie forum about intermittent arcing faults on ungrounded systems.

Does this apply to 480vac systems or only to higher voltages? My memory was that a part of the phenomenon had to do with establishing and extinguishing an arc as the voltage to ground (accross capacitance to ground) oscillated in one of the phases. But also it seems that this type of arcing might not be possible in 480 volt systems due to very samll propensity for arcing.

Also, are there any features incorporated into the design of modern ungrounded systems that alleviate this problem? The reason I'm asking is... my facility has ungrounded 480-volt system for vital loads... but we have not experienced these problems over the course of 12 years since facility waw built.

 

[jburn]

Just found an interesting article that is related to this topic. It talks about, on high-resistance grounded systems, some engineers are selecting a 2kV cable where a nominal 600V cable could be used. They feel the extra insulation thickness handles voltage anomalies during fault conditions for longer periods of time.

Well worth reading. Again it is a reference to having a voltage anomalie when a ground occurs on such a system.

http://www.southwire.com/tech/pubs/pcu/600vtoolkit.html

 

[electricpete]

Well, I've just been searching the eng-tips archives to dig up info on ungrounded system intermittent arcing faults.

Lots of good info. Many many horror stories about 480vac ungrounded systems gone berzerk. Meanwhile, here we stand with probabaly 1,000 or more motors on 480vac ungrounded systems and I've never seen anything remotely resembling this. And by the way, it's a pretty standard design among nuke plants and I have heard anything from the other plants.

So now the question: why would some plants be susceptible and some not?

I can only come up with a limited number of ideas:
1 - you guys are lying. Probably not... there are too many of you saying it.
2 - All of those stories involved motors much inferior to ours... maybe 1950's motors... maybe open drip proof motors out-doors?
3 - Maybe there is a difference in the systems. We have discussed that capacitance can affect the behavior of resistance grounded systems... maybe it plays a role in ungrounded systems as well? Our plants have no large capacitances: transformers within a few hundred feet of motors => relatively short cables, no power factor correction caps, no surge caps at this voltage level.

Can anyone familiar with the problems occuring on ungrounded systems tell me if those systems had any of these factors?

 

[peebee]

One more possible explanation, e-pete: your environment is probably not as corrosive as some other industrial installations (like steel mills or chemical or paper plants).

 

[peebee]

And here's another thought:

What's your maintenance response on the first (not second) experience ground fault on your ungrounded system?

1. Correct the ground fault that day.
2. Correct the ground fault during the next scheduled shutdown.
3. Correct the ground fault after the next unscheduled ground fault.
4. Um, how do we know if we have a ground fault?

The closer you are to #1, the less often you'll experience problems with overvoltages, etc.

Also, you nuke plants usually build things to be rather bulletproof compared to some other installations (with good reason), which would make it seem natural that you'd experience fewer insulation failures. That said, it surprises me that you would not employ more resistance/reactance grounding methods to provide for better ground fault relaying.

Just curious, besides lights, what other measures do you typically provide on your ungrounded systems -- thumpers, remote alarming, anything else?

 

[pablo02]

Although many of you speak of having ungrounded systems, aren't they really grounded through a high impedance? (e.g. the ground detection system?) that's why we could tolerate the first ground, but the second was a nightmare...

it's when we lost the ground reference that problems really showed up...

am I off base here?

 

[alehman]

I think the main concern is for domino-effect failures which result from high transient overvoltages when a second ground fault causes substantial arcing. Apparently the current level of the first fault is normally too low to cause damaging voltage transients. I suspect in nuclear plants, good maintenance practice normally prevents a second ground fault from occuring.

 

[electricpete]

thanks for all the comments. I'm still working through the issue.

Does anyone know the new address of that southwire article mentioned by jburns. I was able to get into

http://www.mysouthwire.com

By registering you get access to many docs, but I didn't see that one.

 

[Shortstub]

See thread238-48579 & thread238-45575.

 

[busbar]

electricpete — Perhaps significant headroom can be gained with a warm, dry, indoor environment {including periodic exorcisms.}

Spent time in a wet-process plant with fourteen 750-1500kVA secondary unit substations—all but one with 480V ungrounded {wye!!} secondaries. The classic account in the 1955 Industrial Power Systems Handbook by Donald Beeman is very real. Old habits [since World War II as far as anyone could tell] die hard. 6L6.net//localuser/busbar/dat/beemaIPSH6z.doc

 

[jbartos]

Suggestion: Reference:
IEEE Std 141-1993 IEEE Recommended Practice for Electric Power Distribution for Industrial Plants, (Red Book),
Section 7.2.1 Ungrounded Systems describes the current industry standard views on ungrounded systems with References to:
IEEE Std 142-1991 (Green Book)
GET-3548 available on

http://www.ge.com

 

[electricpete]

Thx all - a lot of good comments.

Pablo raises a good question of the effect of lights. In our case lights are driven of secondary of grounded wye primary pt's. This certainly might have some effect. Do other people that have seen the problem have similar setup? Pablo - why would we ever lose this ground reference? (unless there were overvoltages blowing pt fuses to begin with... which seems to contradict the assumption that the pt's help prevent the overvoltage).

Yes, we clear our grounds usually at most within a day or so. Do the plants that have the problems ignore their ground indications?

Based on busbar's previous comments I got a copy of Beeman. There are at least 3 scenario's mentioned. The first is associated with arcing. Does everyone here think that arcing does not apply to 480vac systems?

jb - I have reviewed the color books and they clearly express a strong opinion against the reliability of ungrounded systems. Can you by any chance provide the direct link to the ge document... I cannot find it.

This is an issue which will remain of keen interest to me. We have the systems, they operate well, and I am 100% sure we have no intention of adding a ground to these existing systems. But it remains a concern/question.

Back to the issue of pt's. Is it possible that ferroresonance is involved in some of these cases? (perhaps the initial ground provides the mild overvoltage which pushes the pt into it's ferroresonant range)
I notice we don't have any ferroresonance suppression resistors in our 480vac ungrounded systems but we have them in our 25kv isophase bus which at times is ungrounded. Any thoughts on why that is?

 

[busbar]

Although it has nothing to do with ferroresonance, sometimes 480V PTs connected ø-g are improperly specified with a nominal 277V-primary rating instead of 480V. When a ground fault occurs, two of three lights get REALLY bright for a little while, with 480V on the primary of a 277V PT until the fault is cleared.


[tt](Note thread inception of 18 Jul 2001)[/tt]

 

[electricpete]

Thx busbar. That is a good point. Our drawing shows only 480/120... guess that means I have to look at nameplate to figure it out.

btw - I think that is very relevant to the issue of ferroresonance, which depends on saturation characteristics.

 

[doctorbou]

Hi,

I am now comissioning a plant in Germany and the low voltage system in use is the new European 690V voltage.

The system here is defined as "IT system" meaning that the neutral is not grounded.

The system has the required Insulation monitors that will alarm on the occurence of first ground fault.

The plant has now been in service for six months and we are experiencing a high number of Variable Speed Drives failures which are in my opinion caused by the so called "arcing ground faults".

I have not much experience with these systems as they are not nowadays so much in use in Canada where I come from.

Condulting the Merlin Gerin Technical Publications I have noticed that French codes require the use of surge suppressors (with optional resistor) between the transformer neutral and earth, has anybody tried this on their systems? Did anybody experience problems with VFD's on such IT systems?

Regards,

Denis R. Boudriau

 

[busbar]

 
Denis — Older [and current] material describes the near-absolute need for a neutral-to-ground continuous-rated resistance whose value is somewhat below zero-sequence capacitive reactance of the associated low-voltage system.

A decent {online! free!!} paper on the high-resistance grounding is

http://www.neiengineering.com/papers/paper1JN.pdf,

although written in ANSI parlance, it briefly discusses application with drives and power-system high-frequency characteristics. Simple tuning of grounding resistor with L-C components is discussed.

To me, high-resistance-grounded LV systems serving power semiconductors need 'kid-glove' treatment, especially with rapid effort [via human intervention] at first-fault clearing.

There are also related Schneider Electric “Cahier Techniques” [173(+177)+178+204] pdfs

http://www.schneider-electric.com/cahier_technique

 

[doctorbou]

Busbar,

Thanks for the interesting paper proposed, I did not have the time to read it yet as I have to go and work on the VFD problem today. As for the Cahiers Techniques I already consulted them and they are quite interesting and practical.

From what I could see from Merlin Gerin "Cahiers Techniques" the French specifications require the surge suppressors and and optionnal resistor of around 1k ohms therefore too high to be below the zero-sequence capacitance reactance of the system. It must probably still have some "damping" effect on arcing faults overvoltages.

This LV system design with ungrounded system was actually the result of the customer's request, not our design.
We normally use the high resistance scheme with LV systems, normally the Federal Pioneer DSP MKII or equivalent. The customer apparently has another plant in Germany in which ungrounded LV system are used but they are 380V not 690V, I think that makes a big difference. (690V is a relatively new voltage in Europe)

In any case I have yet to see if IEC and VDE (German)specs require anything similar as the French specs.

Happy New Year to all!

Denis