N+1 Supply: Bus Rating

[publicgood]

I am designing switchgear being used as a common output for paralleling (3) 1200KVA, 480V, 3-Phase, unity power factor UPSs. The configuration is for N+1 UPSs, so the downstream load is only anticipated to be max 2400kW; thus, 3000A bussing meets the max intended load. UPS output breakers in this switchgear are sized for the individual UPSs.

Even though downstream load isn't anticipated to grow and there is not enough mechanical cooling to support load growth either, I cannot say for sure what the future might hold. With the (3) UPS output breakers, it could be that the load increases and the parallel UPSs provide beyond the 3000A rated bus. By this, I believe I have no choice and have to rate the switchgear bus for 5000A.

This is similar to switchgear being used for paralleling multiple generators - the bus needs to account for maximum power available. I have heard others argue the UPSs and generator controllers may be programmed to avoid going over maximum design load, but I don't believe the code or UL would find these to be acceptable means of protection. I find this condition similar to gear containing main breakers with trip units - even though you can dial down the trip unit, the bus rating still needs to meet or exceed the frame rating of the main breaker.

My Question: There are instances in the code such as 220.60 for noncoincident loads. While I understand this reference is downstream of protection, I am wondering if anyone knows of similar code exceptions for my design condition.

 

[jim dungar]

I find this condition similar to gear containing main breakers with trip units - even though you can dial down the trip unit, the bus rating still needs to meet or exceed the frame rating of the main breaker.

Could you quote a reference, either NEC or UL, that supports your statement?

Likewise, where is a requirement to design for future loading?

 

[petersonra]

there are no code requirements I am aware of that would force one to accommodate possible future load increases.

 

[publicgood]

Could you quote a reference, either NEC or UL, that supports your statement?

408.36

I agree this reference is specific to panelboards and can get hairy with 240.6(C); however, I have never seen switchgear, switchboards, or panelboards with main breaker frame rating greater than the internal bussing. I imagine UL may speak to this in more detail. I would have to review further. Maybe I am wrong and it is just accepted this way by manufacturers.

 

[publicgood]

I don't disagree about future loading - the NEC is not a design manual. The downstream load is modular in nature. One may argue the multiple load scenarios is part of the Day 1 design and not future.

Maybe my reasoning to use 5000A is only prudent and not required. Or maybe the code does not directly address this subject.

Additional fodder for conversation:

1) 408.36 Exception No. 2.
2) Table 450.3(B) Note 2.

This sort of "multiple OCPDs" is evident through-out the code.

3) Definition of Feeder in Article 100. Some may interpret this as including the bussing with distribution equipment. In-fact, UL has notes in listing information referring to articles in the NEC when bus ratings are decreased within the equipment.

 

[publicgood]

On the future loading topic: this switchgear feeds multiple manual transfer switches (greater than 2400kW worth of load). These transfer switches have normal power as their other source. It is up to the user to manage which loads are from the UPS - it is very possible that the user, either on purpose or accident, could decide to disregard the N+1 redundancy and use the full 3,600kVA worth of UPS.

 

[jim dungar]

I don't disagree about future loading - the NEC is not a design manual.

The NEC is a standard that does contain many design criteria.

However, per 90.1 it is not intended "for untrained persons." The modifying phrase is important, but it often seems to be ignored, creating an 'alternative fact'.

 

[ron]

No code required reason to size the input SWGR past the design load. Protect it at its rating.

In addition, if your client is at all cost conscience, they will be sad if you went higher than needed to carry the load. As well they will likely want you to go to SWBD construction rather than SWGR.

I would suggest that 3000A input SWGR maybe too small for 2400kW of critical load. Once you consider the downstream load of 2400kW, UPS losses and battery recharge (while supporting 2400kW of load), you will likely be at the next size up in SWGR anyway (4000A).

 

[publicgood]

No code required reason to size the input SWGR past the design load. Protect it at its rating.

Thanks

As well they will likely want you to go to SWBD construction rather than SWGR.

I agree - most typically I find customers don't want to pay for the added benefits pf SWGR construction. We did walk down this path. All considered, SWGR was the determined direction.

I would suggest that 3000A input SWGR maybe too small for 2400kW of critical load. Once you consider the downstream load of 2400kW, UPS losses and battery recharge (while supporting 2400kW of load), you will likely be at the next size up in SWGR anyway (4000A).

Agree - upstream serving the UPSs and supporting mechanical is larger than 3000A.