Troubleshooting 480V Ungrounded Delta Systems

[sanders11]

I have a 480V ungrounded (delta) system. From voltage readings taken, it is apparent that there is an unintentional ground on C-phase. There is no ground indication system (lights or otherwise). I am curious about troubleshooting methods to determine where the ground is coming from.

I am familiar with the obvious method of turning breakers on/off to see if the ground disappears. This is not a popular option due to the interruption in production. We have tried this on equipment that we were authorized to do so with no success.

I have recently been told of a Bender grounding locator. I was curious if any of you have had any experience with this system and whether it is useful.

I am also interested in any other methods that anyone may be aware of.

Thanks!

 

[templdl]

The simplest thing that I can think of is to remove one load or branch circuit one at a time by opening a breaker. As you eliminate each one close the breaker and go to the next. At least you should be able to identifying which branch circuit the grounded 'C' phase may be on.
When you ID the offending circuit start isolating each load on that circuit.

 

[charlie b]

Re: Troubleshooting 480V Ungrounded Delta Systems

I am familiar with the obvious method of turning breakers on/off to see if the ground disappears. This is not a popular option due to the interruption in production.

Remind the "Powers that Be" of the loss of production that might be encountered if another ground occurs on a different phase. I suspect it would be on the order of weeks or months, if a line-to-line fault were to cause an explosion in the service transformer or in a major panelboard. Then remind them that such a failure carries a high risk of injury or worse. You can't sit on this one, and you can't let the bosses sit on this one either.

 

[iwire]

Re: Troubleshooting 480V Ungrounded Delta Systems

Not having any experience with 480 ungrounded systems I have a question I have been meaning to ask for a long time.

Remind the "Powers that Be" of the loss of production that might be encountered if another ground occurs on a different phase. I suspect it would be on the order of weeks or months, if a line-to-line fault were to cause an explosion in the service transformer or in a major panelboard.

Why is a double fault so feared, I have read other knowledgeable men stating the same fears as Charlie.

The reason I ask is because a line to line fault on a grounded 480 system is not unheard of but I have not seen it cause anymore damage then a string of OCPDs to open.

I do not understand. :?

 

[coulter]

Re: Troubleshooting 480V Ungrounded Delta Systems

...Why is a double fault so feared, ...
....I have not seen it cause anymore damage then a string of OCPDs to open.

I do not understand. :?

Excellent question. I've got a lot ot time with industrial systems, but little with ungrounded delta. It has been pretty rare in my travels. I too would think it would open the OCP.

carl

 

[templdl]

An OCPD will trip IF the fault current reaches a high enough magnitude to trip the breaker. Can that be guarateed? Probably not.
And, if the C ph is grounded at one place in the plant and the B ph goes to ground in another part of the plant can one determine with any certainty what path the fault current will take? The current can flow through any conductive path at random between the two points. Should the current be high enough to trip the OCPD would you mant that current path to be random?

 

[charlie b]

Re: Troubleshooting 480V Ungrounded Delta Systems

Why is a double fault so feared. . .

OK, you caught me. It is a fair question. So here is my evasive answer. :wink:

If an ungrounded 480 volt system experiences a fault on one phase, and if it is not detected and repaired before another fault takes place on a different phase, then the total fault current can be no higher than would occur from a line-to-line fault at the transformer. You are right about that. OCPD(s) should open and clear the fault. You are right about that too, or at least you would be if everything was designed, built, and maintained correctly.

But not all jurisdictions require a fault current analysis to be performed, sealed, and submitted by a PE, prior to approval of the plans for a building's distribution system. Of all the possible types of faults (e.g., single line to ground, line to line, double line to ground, three phase), the one that is most severe depends on a variety of things. These include the supply from the utility, the rating of the transformer, the sizes and lengths of feeders, and the number and ratings of motors. If these were all properly taken into account in a system analysis, and if the OCPDs were selected and set on the basis of the worst possible fault, then the most severe consequences of this "double fault" should be the tripping of one or more OCPDs.

But faults do take place, and in some cases equipment does get destroyed. How can this be? Why do the OCPDs not prevent damage to equipment? I can't say.

I do no like to entrust the safety of personnel and equipment entirely to the action of a automatic protective function. It is far better to take action to prevent, then to rely on protection. If you find one fault on an ungrounded system, and if you do not aggressively seek to isolate and repair that fault, you are essentially relying on protective action on the part of the OCPD to prevent injury or damage from the next fault. I see that as being irresponsible. Any company that places a higher priority on productivity then on safety is being irresponsible. I am not above telling the "powers that be" that they are risking life and major equipment repairs, even if I strongly suspected that the most likely impact of the next fault would be a couple of tripped OCPDs. I would be telling the truth, but nobody has sworn me to tell all of the truth. :wink:

 

[coulter]

Re: Troubleshooting 480V Ungrounded Delta Systems

...If you find one fault on an ungrounded system, and if you do not aggressively seek to isolate and repair that fault, you are essentially relying on protective action on the part of the OCPD to prevent injury or damage from the next fault. I see that as being irresponsible. ...

Charlie - I absolutely agree with you. One should agressively persue the first fault. However, to get back to the physics of the question about why is a double fault on an ungrounded system considered more dangereous than a single fault on a grounded system:

...And, if the C ph is grounded at one place in the plant and the B ph goes to ground in another part of the plant can one determine with any certainty what path the fault current will take? ...

Also true. But why is this different than say a corner grounded Delta. My experience with impedance grounded 480 is that the few times I have seen two non-related faults, half the time the second fault doesn't trip a CB. Yes we were aggressively persuing the first fault - some times a light fixture full of water (ice) is hard to find - especially when its -40F and 40 knots.

... You are right about that too, or at least you would be if everything was designed, built, and maintained correctly.

But not all jurisdictions require a fault current analysis to be performed, sealed, and submitted by a PE, prior to approval of the plans for a building's distribution system. ...

Also true, but also true about grounded systems

...I do no like to entrust the safety of personnel and equipment entirely to the action of a automatic protective function. ...

My response here is not aimed at charlie, but all of us: I think we all do this - trust safety to the action of a automatic protective function - every day, every time we turn on a light switch, - every time we get in our cars and don't check the brake lights are working, - every time we drive to work trusting the automatic protective nuts behind the wheels of the other vehicle are functioning :lol:

carl

 

[charlie b]

Re: Troubleshooting 480V Ungrounded Delta Systems

. . . why is a double fault on an ungrounded system considered more dangereous than a single fault on a grounded system

On a center-grounded wye, there is a 73% higher voltage for the L-L fault. For a corner-grounded delta, there is probably no difference in the level of fault current. However, instead of the fault current travelling via an EGC to the source, it will travel along any and all metal paths between the two fault points. That makes it more likely that a piece of equipment (or a human) located between the two points will have a bad day.

 

[coulter]

Re: Troubleshooting 480V Ungrounded Delta Systems

...I am familiar with the obvious method of turning breakers on/off to see if the ground disappears. This is not a popular option due to the interruption in production. ...

Yeah you are right. I can imagine the attitude if I were to walk in to the head of production and tell them I needed to shut off selected pumps - causing say a 10% to 20% reduction in production, over the next week to trouble-shoot a ground fault - especially if I couldn't say this was a life safety issue - with a straight face, telling all the truth :shock:

...I am curious about troubleshooting methods to determine where the ground is coming from.....

The Bender equipment I have seen (very few) works on the principle of impressing a DC current on the AC. I've now told you more than I know about Bender So, first I would call Bender and ask what they have and how it could be used to trouble-shoot the location of a fault. If there is a piece of test equiipment out there, I would go after that first.

Failing all attempts at finding a piece of test equipment, here is something we are planning for testing the ground fault equipment on a new impedance grounded 480 installation that you might draw on:

Our plan is to solidly ground an MCC bucket - one phase, load side of CB, and turn on the CB. And then check the response and functioning of the ground fault detection equipment. For this system, there is no risk of a malfunctioning or mis-designed OCP causing a problem, but if a bonding connection were missing, I could be elevating non-electric parts above ground potential. In my case, I can clear the area to limit the risk.

In your case, you would have to ground through an impedance, say 100 ohms to limit the current, and find some way to know you weren't energizing metal frames and even plan to clear the affected areas during the tests. And you would need a contactor and pulser, and a clamp-on ground detector and you are building a piece of test equipment.

I would not recommend this unless you really understood the process, physics, risks, risk mitigation ... ...

Would I do it? Maybe, after I figured out how to do it safely, with control, and cost effectively and there wasn't a piece of production test equipment available.

carl

 

[coulter]

Re: Troubleshooting 480V Ungrounded Delta Systems

...For a corner-grounded delta, there is probably no difference in the level of fault current. However, instead of the fault current travelling via an EGC to the source, it will travel along any and all metal paths between the two fault points. That makes it more likely that a piece of equipment (or a human) located between the two points will have a bad day.

charlie -

I don't know. For this discussion, I not seeing much difference between a corner grounded Delta and an un-grounded delta - other than the corner grounded delta has an intentional first fault.

Then again I think I could agree - if I am translating correctly: A grounded system with properly installed bonding is safer than an un-grounded system with poorly installed bonding. Yeah I could agree with that.

carl

 

[charlie b]

Re: Troubleshooting 480V Ungrounded Delta Systems

For this discussion, I not seeing much difference between a corner grounded Delta and an un-grounded delta - other than the corner grounded delta has an intentional first fault.

What I mean is that on a corner-grounded delta, which by the way I am presuming to be a separately derived system, there is an intentional connection from neutral to ground, right at the source. If there is a fault from hot to case on a machine in the plant, the fault current will be carried to the source via the EGC.

But in an ungrounded system, with the first (unintentional) fault taking place at one machine, and the second fault taking place at another machine, the fault current will flow from the breaker serving one machine, to the fault point in that machine, to the fault point in the other machine (via any metal path that exists between them), and back to the source via the breaker serving the second machine. The source does not see the fault current returning via the EGC. Since we cannot predict the resistance of the metal paths between the two machines, we cannot predict the fault current or the time necessary for the breaker(s) to trip. Until one or the other breaker trips, the fault current will continue, and any person standing along that fault path is at risk.

I guess the difference is this: for the corner-grounded system, the "intentional first fault" is located essentially at the same point at which all the EGCs are connected.

 

[sanders11]

Thanks for all of the discussion...

Thanks for all of the discussion...

Thanks for all of the discussion guys. I have rented a ground fault location device that is manufactured by Bender. We will begin trying to trace out the ground tomorrow. I'll let you know how things work out with this piece of equipment.

 

[ramsy]

In another thread, Coulter described the use of large CT's around conduit to find Delta faults, since they generate an imbalance. This slick idea got me pricing LEM flex probes.

A signal generator may also ring out each phase at the service, while monitoring the grounding bus, before main bondings, to isolate the EGC path. A particular phase can be rung through Xfmr, subpanels, to OCPD, while monitoring the EGC path each time.

If ringing branch equipment / outlet EGC's arn't found downstream, a directional transmitter & pickup like the Amprobe wire tracer, at-2005 may help. The at-4005 is not directional, but Iduno how the proximity sensor works, they claim its better?

BTW, for Z=grnd Delta's this Peterson fault detection relay.pdf seems to use CT's to trip faults by detecting the direction of the imbalance. Nice idea, some cl/amp meters have directional arrows that give different readings if not used the same way, but never tried using it that way to trace an imbalance.

My understanding of impedance grounded systems is they reduce ASCC=E/Z below breaker trip levels by design. The danger with the 2:1 odds of a 2nd fault occuring on a different phase, unlike same phase faults, is the entire building EGC system remains energized.

As I understand it, continuous arcing across loose equipment & raceway fittings is what causes much of the damage, but anyone in contact with raceways, plumbing, drains, drinking fauctes, & building steal also short a path to ground.

A few threads posted this week have described energized building steal, with Delta systems. It would be nice if they followed up with us on which methods worked this out.

 

[sanders11]

Update on troubleshooting...

Update on troubleshooting...

I just wanted to post an update to this thread. Over the past couple of weeks, I have been using the EDS3065 / EDS3365
Portable Ground Fault Location Equipment for Ungrounded AC/DC Systems to 690V AC/500V DC manufactured by Bender. This device has been very successful in helping to locate grounds on 480V ungrounded delta systems.

http://www.bender.org/prod.html

 

[realolman]

Re: Update on troubleshooting...

Re: Update on troubleshooting...

This device has been very successful in helping to locate grounds on 480V ungrounded delta systems.

grounds as in more than one? in your place... or found by the Bender used by other people?

 

[ramsy]

Re: Update on troubleshooting...

Re: Update on troubleshooting...

I just wanted to post an update to this thread. ..this device has been very successful in helping to locate grounds on 480V ungrounded delta systems. http://www.bender.org/prod.html

Many Thanks for the follow up. (EDS3065,EDS3365) is now on my eBay list of favorite searches.

 

[sanders11]

Multiple Grounds

Multiple Grounds

Yes in fact, there actually happened to be two separate grounds on the system. We were able to isolate and trace out each of the grounds using the system ourselves.

This ground detector was used on a system that has no ground indication system and has not been checked reguarly for such problems, so there were actually two problem areas that we found and corrected.