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.

 

[busbar]

electricpete, it doesn’t take a very big resistor connected from a transformer’s secondary [480V] wye point to ground for mitigating the overvoltages associated with low-voltage arcing ground faults. At that voltage, a rule of thumb is to allow 1 ampere/MVA of the serving transformer’s capacity. The resistor could be shoehorned into a small switchboard cubicle [or even a transformer throat] that will do the job. (Folks have even mistaken them for enclosure heaters. :/ )

The upper limit I'm familiar with is 4160V, but a couple of decades ago the IAS Transactions carried a paper about formal [and successful, in the authors’ view] experiments on high-resistance grounding done with 15kV-class distribution. Based on the cited ANSI reference below, it is a valid application for generator/GSU-transformer combinations, and they go up to what, ~23kV?

The purpose of the resistor is to offset ‘charging current’ from the three phase-to-ground capacitors that the system insulation naturally forms in cables and windings. It can be easily measured in an operational system.

A formal description states, “the power dissipated in the resistor is equal to, or greater than, the zero-sequence reactive voltampere loss in the zero-sequence capacitance of the [motor] windings, [cables and busses] and the windings of the transformer.” (ANSI C37.101 §5.1)

For the sake of other readers, if one fault develops, then you’ve got to watch out for is a second fault on another phase. At this point, overcurrent devices start operating, often in pairs. If allowed to progress to this degree, the whole concept is now useless. You need to have fixed monitoring equipment for this application, and respond to it in a timely manner, to be effective in giving high-level service continuity. Suffice it to say, the more stuff connected to a single bus, the more difficult it is to trace a fault. In plants with MV distribution, it’s normal to have multiple ~750-1500kVA secondary unit substations scattered about to serve localized LV loads. This inherently limits circuit lengths and numbers of motors for each unit sub.

Recent coverage of this issue is—High-Resistance Grounding of Low-Voltage Systems: A Standard for the Petroleum and Chemical Industry, IEEE Transactions on Industry Applications, Vol. 35, No. 4, July/August 1999.

 

[electricpete]

Busbar - thanks for the info (star-worthy). In our case the system is fed by wye/wye transformer, so no opportunity to establish neutral ground (except corner delta ground). But as I mentioned, we also haven't seen any problems with the system... do you think we're just lucky?

Here is an interesting discussion of ungrounded system with ground fault from IEEE 62.92.3-1993.
"The voltage and currents between phase conductors, ground, and neutral are affected by the equivalent capacitance of the circuit and equipment windings to ground". With no intentional conductive path to ground, this capacitance establishes a return path for ground fault current as follows: from ground, through the capacitances of the unfaulted phases, to system neutral, and out the faulted phase to the fault. If fault resistance is low, the predominant impedance is capacitive; current zero occurs at the fault at voltage crest. It becomes possible for high voltage to reionize the arc path and forthe arc to restrike. Such an intermittent fault may be established with the arc restriking every half cycle, equivalent to switching a capacitor every half cycle. A cuumulative builduip may the occur if the recovery rate of insulation strength increases after each extinction (a situation not likely to occur in open air arc, but prevalent within confined spaces within multiconductor cables, raceways and machine windings). High transient peak voltages will occur and will be limited either by the insulation recovery rate and strength at the fault, or by the system insulation strenth."

#1 - can anyone explain that in English.
#2 - They talk about a neutral... is the same thing possible in our delta system?

 

[busbar]

On your ‘vital loads’ system, is it at all possible that there is a resistor somewhere in the secondary-side switchboard for this purpose? You mention a wye-wye tranformer; do you know if there's an XO bushing on the bank?

#1 - can anyone explain that in English. I think that’s essentially being described at this reference: About two-thirds of the way through thread238-6870 I posted an incident that was published in Beeman’s 1955 Industrial Power Systems Handbook that discussed a sustained 1200-volt measurement, effectively pushing the system ‘neutral’ point way outside the 480V phasor triangle. I’m guessing that’s what eats insulation for breakfast. With resonant series-LC networks and motor/transformer iron, it’s probably a real nasty harmonic-filled waveform. Seriously, in a staged ground fault at 480V [Don’t try this at home, kids.] the arc noise is fairly quiet, but it sounds very unusual.

#2 - They talk about a neutral... is the same thing possible in our delta system?
Various IEEE papers and standards talk about using a separate, dedicated zigzag autotransformer or wye-delta bank to establish a neutral for resistance grounding. The relative kVA rating of the bank can be quite small, because you only need to handle charging current; typically less than 5 amps for a low-voltage system.

 

[electricpete]

I said wye-wye but I meant delta-delta. (Our ungrounded system is fed from a delta-delta transformer).

 

[electricpete]

Another factor to consider... the Navy has had many years of experience with ungrounded three-phase systems near 480vac. But I have never heard of the multiple-motor failure event in Navy ships.

I wonder what other factors make some ungrounded systems susceptible to this phenomenon and some apparently not? Perhaps it has to do with the capacitance etc of the system? Or maybe a very specific type of intermittent arcing fault is required to initiate the event.. which only occurs in certain types of equipment?

 

[jburn]

I have been responsible for the electrical maintenance in industrials plants for over 30 years. Plus I spent 3 years as an electrician on a Navy ship. Over that time I have worked in 4 different plants and all of them had ungrounded 480 volt distribution systems (delta/delta 480 Volt transformers) The plant where I am presently working (21 years) has 38 different 480 Volt unit substations. The plant is old, started in 1939. We find grounds on our system several times a month. I can't say that I have ever seen a multiple-motor failure event as described in Beeman's book. However, I do feel that in all of these plant, and I would also have to add the ship I served on, we had, in my opinion, a high rate of motor failures. I have seen some very sever damage to the electrical system, which I concluded was cause by having a ground on the system, and then having a second ground. I posted a question several weeks ago about alarms for ungrounded systems, and I now have a project to install an alarm system. For the past 30 years, I have always insisted that my folks check each substation daily, looking for grounds. In the old days, each substation had a bank of lights, and all the shift electrician had to do to check for grounds was drive past the each substation. Today, the light banks are gone, so they have to walk into each substation. Consequently, they are only getting checked once a week, if that often.

If you have an ungrounded system, and are not checking for grounds, and have not had a problem, I would say you are just lucky. What is your rate of unexplained motor failures?

 

[busbar]

During the 70s I worked in a tank-farm winery that had 14 unit substations, all with 480Y secondaries, and only one had its XO bushing tied to ground for an electrode boiler. One unit sub that regularly wasted 500-Hp motors, another did in 200-Hp motors all too often.

From an insulation-damage/transient-overvoltage standpoint, it was pure, unadulterated Beeman. [They'd done it that way since the 40s, and they were not about to change.] You wouldn't believe the manhours spent pulling motors, baking them overnight to get megger readings >1M and reinstalling. I got good with varnished-cambric, rubber and vinyl tape, though.

It got comical at times. Of course, sub-50-IQ hose jockeys almost always kept everything wet, and no one seemed to know about NEMA 4/4X enclosures.



[ref Chapter 6, DONALD BEEMAN Editor, Industrial Power Systems Handbook, McGRAW-HILL BOOK COMPANY, 1955]

 

[dwraith]

High resistance grounded systems are quite common in Europe, particularly in France. Hence Merlin Gerin, part of the Schneider Group, have a large range of solutions for ground fault monitoring. Try this link for more information.

http://195.25.64.11/HomePage.htm

 

[electricpete]

busbar - I got your e-mail but was unable to reply... kept getting error messages from your server. I would definitely be interested in getting those references you mentioned.
Thanks!
electricpete

 

[jbartos]

Suggestions:
1. There some good posting covering this and related topics in this Forum
2. To </images/new.gif> electricpete (Electrical) Jul 18, 2001 marked by ///\\\.
We've heard some interesting things on thie forum about intermittent arcing faults on ungrounded systems.
///Probably overvoltages caused by arcing, damaging electrical equipment and causing EMI emissions.\\Does this apply to 480vac systems or only to higher voltages?
///The phenomena caused by ungrounded systems, frequently used in more remote past, are common to all voltage levels, in general.\\My memory was that a part of the phenomenon had to do with establishing and extinguishing an arc as the voltage to ground (across 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.
///It may be realizable, but it is difficult since one has difficulties to discern small arcing from various noises.\\\ Also, are there any features incorporated into the design of modern ungrounded systems that alleviate this problem?
///There is some effort developed to use the digital signal processing to detect arcing and trip the protective devices.\\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.
///The problems still may came; especially, when the insulation becomes to deteriorate, and current creepages become higher. Then, you may contemplate the high resistance grounding or whatever may be next better and feasible.\\\

 

[redtrumpet]

To the best of my knowledge, there is little that can be done for ungrounded systems to alleviate the effect of overvoltage during a ground fault (other than oversizing insulation to at least 1.73 times the phase voltage - still doesn't help on arcing overvoltages like the Beeman reference). That is why high-resistance grounded systems were developed.

I have seen high-resistance grounded systems used up to 15 kV, 400 A (although this is arguably a low-resistance grounded system). The rule of thumb of 1 A/MVA of transformer capacity arose because the capacitance charging current of cables and equipment in systems is usually within this range, and the neutral grounding resistor must be sized greater than the charging current. I have also seen this rule of thumb modified to read 0.5 A/1000 kVA for low-voltage systems and 1 A/1000 kVA for medium-voltage systems.

However, this is only part of the story. An adequate tripping ratio is required to ensure operation of the ground-fault relay on arcing faults and on impedance-limited or part-winding faults. I have seen figures of anywhere from 5 to 10 times the charging current recommended to allow an adequate tripping ratio. The minimum I would use is a 5 A resistor on an LV system.

A high-resistance grounded system can easily be added to any ungrounded system by either grounding an accessible neutral point, or the neutral point of a zigzag transformer connected to a system without accessible neutral, through the resistor.

If you use two resistors and a timed contactor to short one resistor out, you can &quot;pulse&quot; the ground fault current and use a clamp-on ammeter to trace cables to find the fault location. This can save the expense of purchasing separate ground-fault relays for each feeder, and works well if your electrical code allows you to run your system without tripping. In Canada, we can alarm on ground fault only on systems 5 kV and less, and 5 A and less.

You can monitor ground fault current at the neutral only, or at feeders by using a core-balance (zero-sequence) CT. If your system is such that you have to trip on fault, selective coordination is easy to do, using time-delay only ie. the pickup current on all relays can be identical, just progressively increase the delay on upstream relays.

There is a Canadian company, Federal Pioneer (now a branch of Groupe Schneider/Schneider Electric), that makes a relay system that will monitor several feeders. Upon occurrence of the first ground fault, the system alarms only. Upon occurrence of a second ground fault, which of course is now a phase-to-ground-to-phase fault with no resistor limiting of the fault, it will trip the ground fault on the feeder with the lowest-priority instantaneously. This relay probably works with ungrounded systems as well, and could help limit damage when the ground fault escalates (as it inevitably will, given enough time, due to the overvoltage stress on unfaulted equipment).

 

[redtrumpet]

Whoops - I'll correct myself before somebody else does it for me. The Federal Pioneer selective tripping relay I mentioned in my last post won't work with an ungrounded system (obviously), as it is unable to sense a ground fault on an ungrounded system. Little bit of a brain fart there.

 

[jbartos]

Suggestion: The latest development in integrated relaying is bringing the output of the single phase distribution transformer located in the transformer or generator neutral directly to the multifunction integrated microprocessor protective devices, e.g. Siemens SIPROTEC 4 7UM611/612 &quot;Multifunction Generator Protection Relay&quot;

 

[electricpete]

to respond to jburn

We do have ground detection in the form of 3 lights which have unequal intensity when a ground appears. I'm not that familiar with the history of how many grounds we have found and cleared (I can't recall of any), but I have only heard of one motor failure. This is among probably 100 480 volt motors which operate periodically for short periods of time over a period of 10 years.

I believe this is standard design for safety buses in nuclear plants. At the time of design this was felt to be more reliable in terms of ensuring that motors would be available when needed. We also have features like motors that alarm (vs trip) on overload, but that's another story.

I am still curious how it is that our design seems to work so well in spite of all the horror stories from other ungrounded systems.

 

[busbar]

A couple of ideas -- a wet environment will increase the likelihood of insulation failure. If the ground lights are big enough, they will provide some damping of transient vervoltage.

 

[davidm]

At my power plant, a nuke, we have 600VAC three phase delta ungrounded supplying lots of equipment, and no arcing problems as you are asking about. No, you aren't lucky, you just don't have hard grounds. The arcing over the capacitance of the insulation just isn't going to happen at these low voltages unless you have a serious problem with insulation degredation due to damage, age and heat related failure, etc.

 

[electricpete]

davidm - that makes sense to me.

 

[jburn]

From above thread...

.....&quot; unless you have a serious problem with insulation degredation due to damage, age and heat related failure, etc.&quot;

True, On a 480 volt system, you should not have arcing, until you start having insulation problems. But where you do have an insulation problem, you can have an arcing contidion, when it is not a solid ground.

In an older plant you do have a problem with insulation degredation due to damage, age and heat related failure, etc. We have had to replace 480 volt cable runs, due to ground faults. The insulation had deteriorated, due to age, and moisture, and steam leaks (heat)in the area over a year ago. Most grounds are found in either a motor or in the pot head of a motor. We may average two or three grounds a month. Thus it is important to detect the grounds and clear them ASAP.

 

[busbar]

For those with access to IEEE papers, fairly recent discussion of this subject is in the paper: John P. Nelson and Pankaj K. Sen, High-Resistance Grounding of Low-Voltage Systems: A Standard for the Petroleum and Chemical Industry, IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 35, NO. 4, JULY/AUGUST 1999

An important corollary to LV hi-res grounding is the single-pole interrupting rating of overcurrent devices, covered in: George D. Gregory, Single-Pole Short-Circuit Interruption of Molded-Case Circuit Breakers, IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 35, NO. 6, NOVEMBER/DECEMBER 1999