Undreground connection for 3000 Amp @ 4160 V

[toto2002]

Good people,

Does anybody know of a practical way how to bring 2500 Amps @ 4160 V underground connection to MV Switchgear ?!!?!

My problem is that after applying the ampacitis from NEC Table 310.77 Detail 3, the net result is about 7 circuits of 750 KCM.

And here starts the problem:

I can opt for 9 x 500, which would give me an ampacity of 2950 according to the table, but detail 3 only considers 6 conduits in a ductbank.

I can opt for 9 x 750 which will give me ampacity of 3555 (MV-105), according to the table for 6 conduits.

Here comes the second problem:

a/If I go with 500 KCM option - my ampacity is unknown, but for sure is less than the breaker rating of 3000 Amp.
b/If I go with 750 KCM option - my ampacity is still unknown, it will be less than the breaker rating, and the conduit should be 5".

Having said 5" conduit - here comes the third problem:

The bending radius of a 5" conduit is minimum 60". NEC allows me to go maximum 36" deep counted from the top of the first conduit layer of the ductbank.

On the up side - my SWGR rep says that they very rarely see so many conduits coming into a 3000 Amp SWGR. So apparently either I am missing something or everybody else knows something that I do not know.

Some people suggested using buss duct. However, I do not see a way of putting buss duct without building a tunel, so that the electricians can walk.

PLEASE HELP. I FEEL LIKE I AM CHASING MY TAIL HERE.

Thanks
Ted

 

[jghrist]

I think that you should seriously consider bus duct instead of cable for this much current. Why does it have to be underground? Is this for a main feed from a transformer to the switchgear? Not only will there be problems getting the cables into the switchgear and terminating them as your switchgear rep said, but how are you going to terminate them at the transformer?

If you do use cable, I would get an experienced engineer to design the installation instead of relying on tables. Test of on-site soil thermal characteristics should also be made for the calculations.

Bus duct would make a neater, more reliable connection.

 

[mc5w]

There are some design figures in annex B, figures B.310.3, B.310.4, and B.310.5 for high capacity duct banks. However, you might have to pull in some polybutylene or polyethylene tubing that is hooked up to some water chillers to handle maximum demand. DO NOT use hose clamps with polybutylene tubing.

Another problem with underground duct banks is arranging corners so that all conductors of the same phase are the same length.

Mike Cole, mc5w at earthlink dot net

 

[KWman]

I have only seen one very high ampere cable installation similar to what you are considering. It was at a Dow Chemical plant in Texas. The engineering team had sought the services of a specialized engineering firm to do the cable arrangement, to assure load sharing betweent the cables and conductors.

My recommendation is to use buss duct if at all possible. Sounds like you are committed to an underground approach, but I would explore other major change options (like a raised foundation for the switchgear?) that would permit an above ground route... as so superior to a tunnel.

 

[toto2002]

Thanks guys for the response.
Let me expand a little on my problem.

This is a dual ended configuration. Under normal conditions each section of the bus and the associated feeder will carry about 1200 amps.

Under emergency conditions, which would be loss of one of the supply sources, the total load can be covered by the remaining source. And here starts the problem with the feeder ampacity, because it can potentially be loaded with 2500 Amps for more than 3 hours.

I don't have a way to go overhead, because that would mean crossing about 80 feet of parking lot.

Does anybody know something about flexible buss duct? I have heard about it but cannot find it.

Thanks again

 

[BJC]

Have you looked at using precast trench?
Whether you can use it I think will depend on the loading in your parking lot. I have used it it for 12 Kv and I knew there was no heavy truck traffic on it. Most of the precast trench will take normal traffic. If a rubber tired crane shows up and puts an outrigger pad on it --that might be a problem.
If you can find a Utility Vault catalogue look at the 5684 trench and similar types.
There are a variety of widths and depths available in precast and you would be advised to get help from a civil or structural engineer.
What you are doing is replacing the ductbank with a concrete tunnnel system. You can install cable tray in the trench to support the cable.
If you use tray you can us a higher ampacity for the cable, maybe using smaller and/or fewer cables and there is less chance of damaging it in installation.

 

[fluxcap]

Get a copy of IEEE 835 (formerly ICEA P-46-426).

It has tables that cover your application, or you can do the calculation manually.

 

[alehman]

The precast trench approach may also work with cablebus.

If you must use conduits, you don't have to stick to the NEC tables, but you should seriuosly consider finding someone experienced with these calculations. Special software is available to run the calcs for configurations other than the NEC tables.

http://calcware.com/

The problem I have with thermal calcs (or using the tables) is finding someone to do the soil test mentioned by jghrist. This is the only test method I'm aware of:

http://www.trtester.com/

I would be interested to know if anyone has a source for testing service or has experience with this device.

 

[cuky2000]

The duct bank configuration can be modeled relatively easy using commercial software such as AmpCalc based in the same equations (Neher and McGrath) used in the NEC. For details, see the enclose site or search in the net for similar product.

http://www.calcware.com/index.html

 

[wareagle]

toto2002
There is a solution to your problem if you are using 310.77.
In detail 2, the ampacity of 3 conductors per conduit, 3 ckts, 500 kcm is rated at 375 amps. By separating the duct banks by 24" you can install 9 circuits and rate the 500 kcm at 375 amps. The total ampacity for 9 x 375 = 3375 amps.

Table 310.77 as well as the other tables in this section use
a 100% LF. In my opinion this is not necessary. Using a 75%
LF raises 500 kcm to 419 amps which would require 7 circuits.
You would have to make the determination if 75% LF would be
applicable. According to Okonite Cable Co. Load Factor is
the ratio of the average hourly load to the maximum hourly
load for a given period, usually 24 hours.

 

[dpc]

I have done this at 3000A, 4160 V, using Trenwa pre-cast trenches, Okonite C-L-X cable, cable spacers in the trench to maintain constant separation and special thermal sand for backfill. Okonite calculated the resulting ampacity, using their Neher-McGrath software.

As BJC mentioned, you have to be careful about traffic on the trench covers. Basic Trenwa section can't handle any traffic, but there are surely equivalent systems that can. But you may have some Code issues with depth below grade that we didn't have to worry about because of our location.

It wasn't my favorite design, but it worked, and no problems after about 8-10 years, AFAIK. The tricky part is transition from the trench back into the real world.

For cable bus, do a search for "Calvert".

 

[wareagle]

toto2002
Please respond to the proposals to your post.

 

[jraef]

Just so you know what you are getting into, check out this old thread.
thread238-93479


"Venditori de oleum-vipera non vigere excordis populi"

 

[toto2002]

Good People,

Thanks for all of your valuable input, which I will consider when I design similar installation with less spacial constraints. I hope this will not be soon.

Let me tell you how I solved my problem, as of yesterday:

1/ As mentioned before I will need the full ampacity of the feeder of about 2400 Amps only if one of my two sources fails, and I have to run full capacity, which is highly unlikely, since one of my six pumps is supposed to be standby.

2/ Under normal conditions I will be running only about 1250 Amps per feeder.

3/ That being the case I called Okonite, and they ran McGrath calc for me. The result was that 9 sets of 500 KCM,in 4" conduits at 7.5" on center can take 2300 Amps at LF 100%. This is sufficient for my situation.

4/ I will ask Okonite to perform the calculation with LF of 75%, in order to determine the setting of the breakers.

So basicaly that solved the problem in my very particular application.

While researcing the case I found out the following information, that you may find useful:

a/ In the 2005 edition of the code - detail 4 is missing.

b/ CalWare - is the recomended software for the purpose. Unfortunately it costs about $1900, and my boss told me to basically take a hike.

c/ IEEE 835 is the other solution. From my understanding it contains about 3000 tables on 3200 pages, that can help calculate the ampacity on MV to HV. It costs $430 ( non-member) $380 member. I am still fighting for the money to get it.

d/ If such high ampacities at MV are incidental to your work, probably the best bet is to call Okonite, ask for applications engineer and ask him to run several scenarios. They were very co-operative. Calling Southerwire does not give the same results - they go by the code.

Thank you once again for your co-operation and sound engineering advise. Next time I will state upfront that I am going in trench, and I need the room - else no power.

Regards
Ted

 

[cuky2000]

PROBELM 1: Does anybody know of a practical way how to bring 2500 Amps @ 4160 V underground connection to MV Switchgear ?

SUGGESTION 1: Consider using cable compartment in the SWGR.

PROBELM 2:
a- 500 KCM option - my ampacity is unknown, but for sure is less than the breaker rating of 3000 Amp.

b- 750 KCM option - my ampacity is still unknown, it will be less than the breaker rating, and the conduit should be 5".

SUGGESTION 2:The estimated ampacity calculated is as follow :

Cable Insulation Rating> MV-90oC MV-105oC
500 kcmil 1,980A 2,140A
750 kcmil 2,450 A 2,530 A
1000 kcmil 2,580 A 2,800 A
[sub] See the enclose link for additional details.

http://cuky2000.250free.com/UG.jpg

http://www.okonite.com/engineering/3-conductor-underground-ducts.html

[/sub]
PROBELM 3: The bending radius of a 5" conduit is minimum 60". NEC allows me to go maximum 36" deep counted from the top of the first conduit layer of the duct bank.

SUGGESTION 3: Double check the allowable bending radii. The enclose guide may be helpful

http://www.okonite.com/engineering/bending-ratios.html

If still can not be solved, consider design a transition vault to be located under the switchgear cable compartment.

 

[cuky2000]

For load factor 75%, here is an estimated values for 500 kcmil cable in a duct bank. I hope this could help.