It is currently Tue Nov 24, 2020 7:22 pm



Post new topic Reply to topic
Author Message
ekstra   ara
 Post subject: Arc flash study with dirt ground path
PostPosted: Wed Sep 26, 2018 4:05 pm 

Joined: Thu Jul 19, 2018 2:08 pm
Posts: 10
Hello,
I am working on a study at a facility where there is no copper or steel connection between the utility ground and the MCC ground bus. Any single phase fault current will have to travel through dirt to return to the transformer. I have measured the resistance from the utility to the MCC ground bar at 0.75 ohms (this is about 30 feet away). Of course that impedance varies with temperature and moisture.

I have tried to explain to my client that the lack of an effective ground path would have a big effect on the study results, especially with respect to single phase faults.

Is there an industry-accepted means of handling a dirt path? Any suggestions?

Thanks
Andy


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Thu Sep 27, 2018 6:29 am 
Arc Level

Joined: Mon Jan 18, 2010 11:35 am
Posts: 524
Location: Wisconsin
AndrewKGentile wrote:
Is there an industry-accepted means of handling a dirt path? Any suggestions?


You could treat it as ungrounded.
Also, make note of the NEC violation for not having the equipment properly bonded. Dirt is not considered as an effective path for fault currents for 1000V and below.

BTW, how did you measure the resistance of the dirt.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Thu Sep 27, 2018 6:39 am 

Joined: Thu Jul 19, 2018 2:08 pm
Posts: 10
Thanks for your reply. I used a 6V lantern battery and an ammeter. I shorted the MCC ground bar to the utility ground and measured the current flow, which of course was through the dirt.

It's a disturbing situation. My clients wants his labels, and I'm trying to tell him I can't confirm they are correct. I did run a simulation using a resistive grounded transformer with 0.7 ohms resistance. The clearing times of everything went up significantly. But the facility has so little fault current that the clearing times were already near 2 seconds. Coincidentally the arc flash IE only changed slightly.

Do you have a code reference for the dirt path not being an effective path?


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Thu Sep 27, 2018 10:40 am 
Sparks Level
User avatar

Joined: Fri Jul 08, 2016 10:01 am
Posts: 227
Location: Indiana
Could use a bit more info.

Are you talking about from the POCO transformer to an MCC? Voltage? Is this even a grounded wye system? Does it have a grounded neutral? Corner grounded delta? Ungrounded delta? High leg delta? Single phase?

Where I work we still have several old 240V UNGROUNDED delta services that feed nothing but motor loads. The switchboards have ground fault indicator lights on them. WIth no faults present all 3 phase lights burn dimly. When we have a ground fault the light for that phase goes out and the other two lights burn brighter. We have to start opening switches until the fault goes away and we work our way downstream from there until found.

Nothing happens when one phase has a ground fault. It just creates a pseudo corner grounded system. Only if and when a second phase goes to ground do we have fireworks.

Again, more info on the install is needed.

Thanks.

EDIT: Oh, I was able to dig up a pic of a couple of our switchboards that have the fault lights on them. See below.

Attachment:
ungrounded delta lights.JPG


I bought a new switchboard for one of these type services a few years ago. Here were the particulars.

Attachment:
New MEA 240V board.JPG


You do not have the required permissions to view the files attached to this post.

_________________
SKM jockey for hire


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Thu Sep 27, 2018 1:37 pm 

Joined: Thu Jul 19, 2018 2:08 pm
Posts: 10
The utility transformer is a 3 pole mounted cans, grounded wye. The utility ground is at the pole. The MCC building is about 25' away from the utility pole. There is no copper or steel connection between the utility ground and the MCC ground. All single phase fault currents have to flow through dirt.

My question is... is there an industry accepted method for modeling this? I'm inclined to tell my client that the study results are no reliable because they are based on a solidly grounded system. I don't think dirt ground is considered solid.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Thu Sep 27, 2018 4:23 pm 
Sparks Level
User avatar

Joined: Fri Jul 08, 2016 10:01 am
Posts: 227
Location: Indiana
AndrewKGentile wrote:
The utility transformer is a 3 pole mounted cans, grounded wye. The utility ground is at the pole. The MCC building is about 25' away from the utility pole. There is no copper or steel connection between the utility ground and the MCC ground. All single phase fault currents have to flow through dirt.

My question is... is there an industry accepted method for modeling this? I'm inclined to tell my client that the study results are no reliable because they are based on a solidly grounded system. I don't think dirt ground is considered solid.


So, no neutral was brought in with the service conductors?

Do you have a pic of the transformers that we can see the wiring configuration?

_________________
SKM jockey for hire


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Fri Sep 28, 2018 4:29 am 
Plasma Level
User avatar

Joined: Wed May 07, 2008 5:00 pm
Posts: 826
Location: Rutland, VT
Most likely the utility would have run quadraplex over.

_________________
Barry Donovan, P.E.
www.workplacesafetysolutions.com


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Fri Sep 28, 2018 7:03 am 
Arc Level

Joined: Mon Jan 18, 2010 11:35 am
Posts: 524
Location: Wisconsin
AndrewKGentile wrote:
The utility transformer is a 3 pole mounted cans, grounded wye. The utility ground is at the pole. The MCC building is about 25' away from the utility pole. There is no copper or steel connection between the utility ground and the MCC ground. All single phase fault currents have to flow through dirt.

My question is... is there an industry accepted method for modeling this? I'm inclined to tell my client that the study results are no reliable because they are based on a solidly grounded system. I don't think dirt ground is considered solid.


Arc flash calculations, based on IEE1584, do not use single phase L-G faults, so that fault path is not important. The IEEE1584 formulas do care whether the enclosure is solidly grounded or not. Model the MCC as ungrounded and apply labels.

For the NEC violation look at the last sentence in 250.4(A)(5). For at least several decades the language has been unchanged, "The earth shall not be considered as an effective ground-fault current path."


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Fri Sep 28, 2018 5:10 pm 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
Two things. With regards to how to model, IEEE 1584-2002 supports either three phase faults on grounded or ungrounded systems. As reported there was something odd going on with the ungrounded tests and it resulted in a slightly higher incident energy. Further testing has shown that this was simply a numerical anomaly and in the proposed new draft, this has been deleted but for now if you are going to follow the standard, ungrounded systems have a slightly higher incident energy. In the revision to IEEE 1584, this difference is going away. The reality is that arc faults rapidly progress from single phase line-line or line-ground faults to three phase faults within 1 to 2 cycles at most in the testing done on the subject. The air rapidly heats up to the point where it becomes conductive enough around the other two phases that it becomes a full on 3 phase arcing fault very quickly except under the most extreme conditions. Mike Lang's results also show that the incident energy from a purely single phase arc aren't that much different from a 3 phase arcing fault either. It is certainly not 1/3 of the 3 phase result which was suggested early on. See for instance the very readable papers from Mike Lang on Mersen's web site. So this is just to give you a heads up that from an arc flash point of view, this doesn't matter. The reality is that observation of arcs shows that during the part of the cycle where phase B is at zero, the arc jumps to the enclosure from phases A and C so you get a line-ground-line arcing fault for part of 120 degrees, then back to the typical line-line arcing for the other 240 degrees. The important part of this result is that whether or not the system is grounded really doesn't even enter into the picture. Given the relatively high impedances typically encountered in grounding compared to the very low impedances of an arc (a few ohms), it's no surprise then that the only impact that grounding of any kind has is that if the impedance in a line-ground fault is high enough, it can significantly reduce the likelihood of forming an arc flash. But once arcing gets going, it doesn't matter anymore. Hence the reason IEEE 1584-2002 treats everything as a function of the three phase bolted fault current.

Second, I don't believe you have an ungrounded system based on what you're reporting. I do ground testing all the time. Mines are required to test every ground rod and the equipment grounding conductors from the MCC to the various motors annually so I have quite a bit of experience recognizing "good" and "bad" grounding. There is no regulatory requirement for grounding except that under NEC if the first ground rod is over 25 ohms, drive another one. There is no restriction once the second ground rod is driven but obviously the conceptual idea here is to get under 25 ohms. Typically grounds with good conditions such as near buildings with undisturbed dirt tend to run around 5-10 ohms for common medium to heavy clay soils. Once in a while I'll run into a ground that measures up to around 100-200 ohms but generally speaking if it's over that, the grounding is broken somewhere and needs to be investigated. On the low end by way of example except for 3 electrodes that had grounds that were broken (and it was obvious they were broken), with a very extensive ground grid for obvious reasons, a grounding survey at a power plant that was built about 10-20 years ago measured between 0.3 and 1 ohm everywhere. IEEE Standard 80 for substation grounds recommends less than 1 ohm. Also getting back to those equipment grounds generally they measure between around 0.2 and maybe as high as 2 ohms but rarely any higher than that.

Also, low resistance grounds usually use a 400 A resistor which on a 480 V system works out to around 0.7 ohms. This is with an intentional resistor mind you. A high resistance ground usually has a resistance of at least 10-11 ohms at 480 V and up to around 100 ohms.

Finally when looking at peg grounding which is what you describe where someone simply drives a peg (ground rod) everywhere that a ground or neutral is needed, it's not as bad as what you think. The resistance of a wire is linearly proportional to the length but Earth works differently. It's efffectively a 2-dimensional system so following IEEE Standard 80 as the distance increases resistance on a given path through the Earth increases linearly as well. However the number of paths is increasing with the square of the distance so actually we get the somewhat counterintuitive result that resistance through a peg ground is proportional to the inverse of the distance. Typically at around 1 mile the resistance of a wire is about the same as the resistance between two ground rods. Beyond 1 mile peg grounding actually gives better impedance than a neutral or ground wire. However over very short distances (30 feet) obviously this effect does not apply.

So now we come full circle. You reported 0.75 ohms. I'm not sure how you measured it but I'll just say this. In my experience there is physically no possible way that you got a 0.75 ohm resistance through 30 feet of Earth no matter how good soil resistivity is. It simply can't happen. That's the same resistance as what you'd typically see on an equipment bonding jumper to a motor or on a very large and extensive ground grid for a substation or generating station, or a low resistance grounding resistor. But the only way you're going to get 0.75 ohms through Earth is if you are measuring it over several miles. Or put another way according to Wikipedia drinking water runs between 20 and 2000 ohm-meters. Dividing by 10 meters (about 30 feet) gives us a resistance of around 2 ohms if you measured resistance between two points in a relatively clean lake, which is much higher than 0.75 ohms and a much more severe case than pretty much any soil measurement over that distance. That's why I'm saying it simply can't happen. So we either have to question the result or the conclusion.

Questioning the result, maybe it's that your testing wasn't done correctly. The standard way to do this kind of testing is to first take a roll of #14 THHN and measure the resistance of the roll using a low ohm bridge for accuracy. All multimeters no matter how bright the yellow color and no matter how big the price tag need not apply here...multimeters don't work accurately below 1 ohm in the realm of milliohm and microohm meters. Second use the roll of #14 THHN as a "meter extension" to connect to one ground rod and measure the resistance to the other ground rod, subtracting out the #14 THHN spool impedance. Low ohm bridges typically put out up to around 10 A of current though so even though your methods aren't exactly text book, I'm going to take it on faith that you probably did it right.

The second explanation is simply that the ground grid exists but since it's buried you can't see it. And keeping with common practice probably someone went ahead and tied the pole ground into the ground grid since there was such a nice source of low impedance ground grid nearby, far better than wrapping several wraps of #6 around the butt end of a pole and burying it. This is pretty common practice when it is available and convenient for use but once done it all gets buried and nobody knows it's there. Utilities are more practically oriented in many cases and don't follow NEC as rigidly since usually NESC applies which is more performance instead of prescriptive oriented. So the second explanation is simply that you DO have a ground connection, it's just not visible without excavating the ground grid.

Either way, 0.75 ohms is probably about the best you can expect for a 30 foot neutral including mechanical splices (and corrosion over time) at either end. I would maybe point out that it doesn't look right, reference where in NEC it lists how to handle this, and your findings, and move on. I'd recommend "fixing it" (putting in the neutral) but since it obviously exists even if it's not done to the letter Code-wise (and you're at the line of demarcation between NESC and NEC rules), I wouldn't make a huge deal out of it. In terms of arc flash modelling as I mentioned the standard currently gives some different values for grounded vs. ungrounded systems but will soon only give the same (same as grounded) values. Whether it looks correct or not, your system is obviously effectively grounded so I see no reason to treat it any differently and I would simply model it as grounded assuming your investigation is correct, especially knowing the ungrounded fudge factor is going away. I know that this disagrees with your conclusions but I highly doubt the results here. It just seems highly improbable.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Sat Sep 29, 2018 9:00 am 

Joined: Thu Jul 19, 2018 2:08 pm
Posts: 10
It sounds like people are saying I shouldn't worry about single phase faults with respect to arc flash calculations because the Lee equations assume all single phase faults will flash over to three phase, at which point they don't need a ground path. I thought about that, but what I hadn't considered was the idea that a high impedance ground will lower the single phase fault current, thereby reducing the chances of an arc. This is an interesting idea. I agree with that logic, however that doesn't match ETAP results.

If I model the transformer as having a solid ground I get an IE of 10. If I model resistance grounded with impedance of 1 ohm, instead of solidly grounded, the IE goes from 10 to 13. However the fault clearing times and the arcing currents are identical for both cases. In fact, I printed the report, selecting all study results, and the only differences are in the IE and the AF boundary (which is related to IE). This doesn't make sense to me. I may have to contact ETAP on Monday to see how the resistor affects their calculations. Unless someone here can explain this.

So the only real safety issue with this situation is that single phase faults might not draw enough current to trip, in which case they will result in a non-clearing shock hazard.

Thanks again for all the suggestions.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Sun Sep 30, 2018 5:43 pm 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
AndrewKGentile wrote:
It sounds like people are saying I shouldn't worry about single phase faults with respect to arc flash calculations because the Lee equations assume all single phase faults will flash over to three phase, at which point they don't need a ground path. I thought about that, but what I hadn't considered was the idea that a high impedance ground will lower the single phase fault current, thereby reducing the chances of an arc. This is an interesting idea. I agree with that logic, however that doesn't match ETAP results.


Lee is useful as an absolute worst case theoretical upper limit on arc power and energy only, and should not be used unless such as with DC systems there really isn't any other option. It's also useful conceptually when training non-engineers on how arc flash works because most tradesmen learn that energy = V*I*t*power factor. One of the biggest problems with it is that with Lee arc power is directly proportional to system voltage, which is simply not true. Arc power is only slightly affected by arc voltage. Mostly arc power is related to available fault current. IEEE 1584-2002 isn't much more complicated and IEEE supplies an Excel spreadsheet to do the calculation if you don't want to grind it out by hand. The draft new model has considerably more terms in the calculations but again it's very straightforward to do it and easy to do in a spreadsheet.

Quote:
If I model the transformer as having a solid ground I get an IE of 10. If I model resistance grounded with impedance of 1 ohm, instead of solidly grounded, the IE goes from 10 to 13. However the fault clearing times and the arcing currents are identical for both cases. In fact, I printed the report, selecting all study results, and the only differences are in the IE and the AF boundary (which is related to IE). This doesn't make sense to me. I may have to contact ETAP on Monday to see how the resistor affects their calculations. Unless someone here can explain this.


IEEE 1584-2002 uses resistance/ungrounded vs. solidly grounded as a binary decision. It's either solidly grounded or it's not. And if it's not, you get a bump in the incident energy which is precisely what ETAP is reporting. The amount of resistance is immaterial, only the presence matters. Even at 1 ohm I'd argue that this is clearly solidly grounded but since we define solidly grounded as "no INTENTIONAL resistance", even if I happen to know that anything under about 5 ohms is a typical solidly grounded resistance, IEEE 1584-2002 and thus ETAP doesn't care. Resistance is resistance. This is where we reach a point where we need to understand how the software comes up with the calculated results that it does and know when to question those results, as well as how and when to fudge things to get the correct result. Try plugging in 0.1 or 10 ohms. You will get the same result, even with 10 times the resistance.

Quote:
So the only real safety issue with this situation is that single phase faults might not draw enough current to trip, in which case they will result in a non-clearing shock hazard.


That's generally a true statement with most systems. The usual assumption with solidly grounded systems is that the phase overcurrent protection can sucessfully trip on single phase faults. Generally speaking this is simply not true and the point at which this occurs is alarmingly low. Take for instance a fused disconnect size correctly for a 100 HP motor. The maximum FLA of the motor is 124 A using the NEC charts. Using NEC a typical RK5 (dual element) fuse would be rated for 175% of that or 217 A, and sizing up to the next standard size is 225 A. A line to ground fault will see 277 V, not 480 V, so the maximum resistance through the bonding/grounding in the event of a ground fault is 277 / 225 A or about 1.23 ohms. It's not hard to imagine that by the time you add in impedance for the transformer that in your situation it's very unlikely that fuse protection on a motor will trip in the event of a ground fault. We can easily extend this example to show that starting around 300 A or larger, circuit breakers without ground fault detection will simply never trip either in the event of a ground fault. NEC only requires us to include ground fault protection once we get up to somewhere around 800 A or larger (the rule has undergone numerous variations over the years) but realistically once you realize what typical ground circuit impedances are, anything over about 100 A seems more like a realistic maximum for expecting phase overcurrent to provide ground fault protection. However it's also a very small leap when you recognize that with arcing faults that it's going to rapidly progress into a full three phase arcing fault anyway why in most faults the lack of ground fault protection quickly becomes inconsequential except in motor circuits where ground faults are very common.

This is fundamentally why except for very small systems such as residential, it is completely unrealistic to expect phase overcurrent protection to provide any protection at all in the event of a ground fault. And once you realize that you really need ground fault protection on almost all three phase systems, it's a very small leap in terms of both cost and design to go to a high resistance ground which improves system reliability and reduces both initial and operating costs simultaneously.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Mon Oct 01, 2018 3:39 pm 
Arc Level

Joined: Tue Jan 13, 2009 5:00 pm
Posts: 540
So 0.75 was the measured value, which is way too low for a dirt only path as PaulEngr pointed out. You have an NEC/NESC compliant system with the two grounds bonded together through the neutral of the service drop. Treat it as a grounded system.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Tue Oct 02, 2018 4:58 am 

Joined: Tue Dec 31, 2013 12:19 pm
Posts: 32
bbaumer was onto the correct line of questions. You don't give enough details to say if there's a code violation or not.

1) Where is the service disconnect, at the pole or at the MCC or somewhere in between?
2) Which wires did they take from the transformer to the MCC, i.e. is there a Neutral conductor?

Normally you don't take a ground wire from the service transformer to the service disconnect in a grounded wye system. You take ABCN on a grounded wye system between the service transformer to the service disconnect. You make a Main N-G bonding jumper at the service disconnect and a that is how a ground fault gets back to the source.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Arc flash study with dirt ground path
PostPosted: Wed Oct 03, 2018 10:00 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
Jeff S wrote:
bbaumer was onto the correct line of questions. You don't give enough details to say if there's a code violation or not.

1) Where is the service disconnect, at the pole or at the MCC or somewhere in between?
2) Which wires did they take from the transformer to the MCC, i.e. is there a Neutral conductor?

Normally you don't take a ground wire from the service transformer to the service disconnect in a grounded wye system. You take ABCN on a grounded wye system between the service transformer to the service disconnect. You make a Main N-G bonding jumper at the service disconnect and a that is how a ground fault gets back to the source.


NEC allows a bonding jumper (and ground) at either the transformer or the panel, and there is usually both so we end up with the system as you described with two separate grounding systems. It is supposed to minimize lightning damage by providing a higher impedance path between the distribution grounding system and the load side of things, or at least that's why we do it in open pit and underground mines.


Top
 Profile Send private message  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 14 posts ] 

All times are UTC - 7 hours


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
cron
© 2019 Arcflash Forum / Brainfiller, Inc. | P.O. Box 12024 | Scottsdale, AZ 85267 USA | 800-874-8883