Grounding in conduit

[nsbelectrical]

Say you have 4000amp feeder breaker feeding 4 conduit sets of three 500mcm cables at 34.5KV on a 800 ft run (lets assume pvc buried 6 ft in the ground). So, we have 4 conduit runs with three 500mcm cables per conduit.

My question is what size grounding cable does the NEC require to run in each conduit?

Be Safe.

 

[DanDel]

Per NEC T250.122, it is based on the O/C device, which you haven't listed. (This system would require relaying.)Assuming it's 4000A, it should be 1-500kcmil per conduit.

But a 4000A C/B at 34.5kV? That's almost 250MVA.

And four 500kcmil/phase? Per NEC T310.78, three circuits of 500kcmil at 35kV is good for 315A.

 

[nsbelectrical]

The overcurrent device would be a micro-51. Are you saying it depends on the settings of the ground fault relay to determine the size of the grounding cable?

It is actually twelve 500kcmil/phase. Need current carrying capacity for over 3000amps.

Be Safe.

 

[DanDel]

Look at T250.122. It's based on the overcurrent device (phase protection).

Will all due respect, this type of system is usually designed by a professional engineer who wouldn't be asking these questions. This is basic NEC which an electrician should be familiar with, also.

 

[nsbelectrical]

No offense taken. I have looked at 250.122 prior to this post, and I am familiar with it. You are driving home a point which leads to my next question. I am having trouble with the interpretation of T250.122. Hypothetically, lets say we have a single cable/per phase (thus 1 conduit) to pull the 4000amp. In this case we would need one 500 kcmil for grounding.

Now, lets say we are splitting that through 4 conduits (because of the cables current carrying capacity). Do we still need 500 kcmil/conduit? and if so why?

Be Safe.

 

[davidbeach]

Yep. If the fault is in that conduit you need that conductor to carry all of the fault current.

 

[DanDel]

nsbelectrical, the actual code reference to this issue is 250.122(F) 'Conductors in Parallel', either (1) or (2), based on your application.

 

[nsbelectrical]

Thanks,

I have concluded that I would be in violation if I did not have a minimum of one 500 kcmil ground conductor/conduit in this application.

However, I dont think this is the common practice. I have come across engineering firms that often do not even run ground wire in conduit. They put one ground wire in a seperate conduit & a ground wire outside the duct bank or concrete. Then tie them together with the ground shields at the terminations.

Be Safe.

 

[alehman]

If you use listed metal conduit installed in an electrically continuous system, an equipment ground conductor is not required by the NEC. If separate ground is also not required by the NEC for listed MC cable or other approved grounding conductors per 250-118 are provided.

 

[nsbelectrical]

According to 250.122(F)(2) does the cable have to be a multi conductor cable? Or would the case I have mentioned, 3 500 kcmil/conduit, be able to be sized in accordance with Table 250.122 based on the trip rating of the GF protection of the equipment (of course meeting the 3 conditions specified).

Be Safe.

 

[alehman]

I believe it did. I'm not sure which version of the NEC you have. The rule for sizing equipment grounding conductors in multi-conductor cable based on ground fault protection no longer exists in the 2008 NEC.

 

[Electic]

It is surprising that for such current capacity you are not using cable tray or cable bus duct (?)

Whether allowed or not (there used to be some limit on using larger conduits for grounding conductor) I would NOT choose to utilize conduit as ground return path for such circuits. One missing conduit fitting could result in a tragic transfered potential.

 

[alehman]

I agree with the above comments on using conduits for the equipment grounding conductor. I'm thinking for a circuit of that length, it's probably not a realistic option.

I'm also wondering about running 800ft with 4000 amps. If it is underground, thermal derating may be a significant consideration as well.

 

[nsbelectrical]

We are using pvc so conduit as a ground is not an option. We will be using 500 kcmil/ conduit. It is underground and thermal derating played a large role in the layout of the duct bank.

One issue is how do you tie together 12 500 kcmil grounds and what does it do for if the equipment ground is a 4/0 anyways?

hmm...

Be Safe.

 

[Electic]

If you are using single conductor cable, I would expect something like EIGHT parallel runs (!) probably in 6" PVC conduit and with a single 500kCMIL bare copper ground in EACH conduit. (The number of phase conductors doesn't account for thermal derating, which would increase the count).

I would also expect the equipment this terminates at to have a ground bus or ground plate capable of accepting that many terminations. Don't confuse the ground electrode conductor with the equipment ground conductor.

As others have said, and with all due respect, this should be calculated by dedicated professional study. This is more dangerous than playing with dynamite.

 

[nsbelectrical]

I have been vague to get general information b/c of the disagreemet on how to ground this system among qualified engineers doing studies - part of the issue was not so much a code issue due to the fact that in practice for this size cable you hardly ever see a 500 kcmil gnd ran in each conduit, but more about whether it is truly needed? If you get a fault will a 4/0 carry enough for a couple cycles to trip(according to NEC no)?

Electic:
Actually there is not a ground bus or ground plate for the terminations, but shortly it will be installed

If interested here is some general info about the duct:
18 - 6" pvc conduits - arranged 3 high and 6 wide. The 2 columns in the middle are ran due to spacing required between the cables to minimize heating also could be used for future expansion.

In each conduit will be 4 - 500 kcmil (1/phase and 1 for the gnd)

The spacing will require 10" off center between each column to minimize the heating and cable will fill the 2 end columns. Rows offset 15". The hottest cables will be the bottom inside cables.

The current capacity will be 300amps allowing for about 3,600 amps through the feed, but future expansion expects to pull 3 more 500 kcmil dropping it to 272 amps which gets you just over 4000 amps total.


Thanks everyone.

Be Safe.

 

[alehman]

Thanks for the additional info.

The 2008 NEC requires equipment grounding conductors to be "contained within the same raceway, cable, or otherwise run with the circuit conductors". It has been common practice to install a single ground with the duct bank either loose or in a separate duct. The fact that this has been accepted is supported by some of the duct bank figures such as Figure B.310.3 (annex B) which shows the separate ground where each phase conductor is in a separate duct.

The way I read the 2008 NEC you need 500 kCMIL either way.

 

[nsbelectrical]

Wait a second... this is exactly why I asked my initial question. I was made fun of because professionals who study this should not be asking this but now we are saying that NEC does not require a ground in each conduit ("The fact that this has been accepted is supported by some of the duct bank figures such as Figure B.310.3 (annex B) which shows the separate ground where each phase conductor is in a separate duct.") However how I read B.310.3 I am not sure it applies because we are at 34,500 Volts.

Comments appreciated.

Be Safe.

 

[resqcapt19]

While I do see the details in Annex B that show the grounding and grounded conductors in a seperate raceway for isolated phase installations, I don't find any support for that concept in the actual code rules. The Annex is not a code rule.

 

[Gianoli]

I live and work in Canada and am not familiar with the NEC. The Canadian Electrical Code (CEC) however would require the following for the installation you have described:

1. A bonding conductor in each nonmetallic raceway. – Rule 10-404 (2).

2. According to CEC rule 10-814(3) “The bonding conductor shall be permitted to be paralleled but shall not be smaller than that determined by dividing the rating or setting of the overcurrent device by the number of bonding conductors and selecting from table 16 (in the CEC) a conductor size to satisfy this result.”

3. So, 4000 Amperes / 12 conduits = 333.3 Amperes. Table 16 indicates a #3 AWG copper wire would be required in each conduit.


I have to qualify the above by stating that this is a very unusual conduit/cable configuration. Transferring a large amount of electrical energy in this manner is could be problematic in a number of respects. Ensuring equal current flow through all these parallel conductors may prove challenging. Terminating and bracing this number of conductors in a switchgear enclosure of some type may also be interesting/impossible. I would review this design very carefully prior to proceeding.