Which OC device for IEEE-1584 Arc-Flash

[sparkyguy545]

Using IEEE-1584, if you have a 2000 A MCB switchboard, do you utilize the instantaneous trip time for the MCB in the switchboard to calculate the PPE label, or are you required to use the upstream device (feeder) trip time? At a recent seminar, I've heard that you must use the upstream device ONLY due to the fact that the line-side of the MCB could be flashed. OK, this is the safest condition since the line-side of the MCB could flash. However, it makes construction difficult because SWBD's with large MCB's would be required to keep the Main CB in a separate section in order to have a lower-rated (safer PPE level) distribution section. I can also see the argument this is overly-conservative and you should utilize the MCB instantaneous trip to calculate the PPE for the eitire SWBD. I don't have the entire IEEE-1584 to refer to, so references appreciated.

 

[jim dungar]

There is no definitive statement addressing this condition in IEEE1584 or NFPA70E. IEEE1584 is nothing more than calculations to determine cal/cm?, it has almost no interpretations in it

My company uses the upstream device.

 

[zog]

Using IEEE-1584, if you have a 2000 A MCB switchboard, do you utilize the instantaneous trip time for the MCB in the switchboard to calculate the PPE label, or are you required to use the upstream device (feeder) trip time? At a recent seminar, I've heard that you must use the upstream device ONLY due to the fact that the line-side of the MCB could be flashed. OK, this is the safest condition since the line-side of the MCB could flash. However, it makes construction difficult because SWBD's with large MCB's would be required to keep the Main CB in a separate section in order to have a lower-rated (safer PPE level) distribution section. I can also see the argument this is overly-conservative and you should utilize the MCB instantaneous trip to calculate the PPE for the eitire SWBD. I don't have the entire IEEE-1584 to refer to, so references appreciated.

Well I would think you would use the upstream OCPD for doing the calc on the Main breaker itself and the Main breaker settings for the rest of the switchgear.

Just my thoughts, I dont think there is a solid answer on that.

 

[jim dungar]

Switchgear is usually compartmentalized, so a fault is not likely to propagate from the breaker cubicles to the bus cubicles and then to the main breaker line side lug compartment.

However, switchboards have no compartmentalization, often you can see the line and load side of all the devices. The UL SUSE line side lugs barrier are to prevent contact, not arc propagation.

 

[sparkyguy545]

I agree entirely. As you stated, true IEEE Switchgear has each Power CB in a dedicated, individual compartment. Some say that's why it is so expensive, part of what you are buying is safety. However, Switchboards and Panelboards are another thing. Once a Panelboard cover is removed, both the Line and Load side of the Main CB are naked in the tub. An AF on the Line side of a Panelboard MCB cerrtainly isn't reduced a bit by a fast instantaneous trip on the downstream MCB. However, in a larger Switchboard (that's why I chose it for my example), you often have the Main CB in a separate section from the distribution CB's. Some isolation is certainly better, but the cross-bus typically is still an open space between sections. An AF in one section would be expected to instantly propagate into any other section open to it. The only device clearing an AF on the Line-Lugs of this Swbd would be the upstream transformer primary fuses with clearing times well into many seconds. I just ran some comps for my 1500 kVA padmount & 2000 A Swbd and get over 62 cal/cm2. My software notes this is the approximate equivalent of 1.9kg (4.2 pounds) of TNT. Clearly, very frightening stuff. Apparently, there just isn't a clear answer for Switchboards and Panelboards.

 

[zog]

Ah ha, he said switchboard, as usual I was thinkiing switchgear, whole different story. I guess Iwire is right, my head is always in a switchgear cloud.