NFPA Arc Flash & FPB

[whycliffrussell]

When calculating the the flash protection boundary (FPB) it is based upon the panels of the equipment removed (ie: someone has the covers removed working on it). Is the FPB different if the covers on the the equipment? ie: do they terminate at the surface of the enclosure if the covers are on?

 

[dpc]

It's based on testing done with the covers removed. But unless it is arc-resistant switchgear, there's not much practical difference since there is no assurance that the door is going to stay on. If it is a situation where you are concerned about arc-flash boundary, you probably need to assume the door is open or cover removed.

 

[whycliffrussell]

Does NFPA say anthing about this? (Covers off vs. on).

The calculations that I have done are with the covers off. However, with a properly designed system (CB and panelboards with suitably sized interrupting capacities for the short circuit currents avaialble) there should be no danger wit the covers on (ie: the equipment is designed to interrupt a particular fault current, meaning as a minimum that it shouldn't blow apart while interrupting that fault current, hence specifying a suitable interrupting capacity.)

 

[WDeanN]

whycliff-
The calculations assume that the cover is off. They also do not take into account the effects of relays, breakers, or other foreign matter in the way that may direct or deflect the arc. The 1584 AFB may not even be a worst case; there are just too many variables.

As far as I understand (not that I understand much) the IAC rating is just related to the ability of the equipment to interrupt the fault, i.e. it can withstand the magnetic forces, etc, associated with the fault. An arc flash calculation assumes that something else may have gone wrong, and an explosive event has occurred. The AIC rating does not address the ability to contain the arc flash / blast. For that there is the arc resistant rating, although I think that's also a little misleading.

So, to get back to your original post, you cannot assume that the flash will stop with the doors closed. The door may be blown open, but the resulting fireball and pressure increase in the enclosed panel will find an exit. Whether it is out a loose screw hole right at the workers face, or a conduit opening in the top is anyone's guess.

 

[whycliffrussell]

What about Fault Clearing Time (FCT). I'm doing a study w/ 120/208VAC panelboards and Thermal Magnetic Circuit Breakers. The fault currents are really low (some only 1.2kA) and hence the breakers are tripping on their thermal elements (long time trip) are up to 50 seconds, hence the incident energy gets really high (+level 4 arc flash hazard rating). This sets the flash protection boundary (FPB) at distances like 150 FT!!!! This would indicate that you can't even enter the room with the equipment (let alone walk in there with a T-shirt and laptop and set up the toehr MCCB LSIG breakers). This seems over the hill to me. Is there a reasonable limit that can be put on FCT? ie: how likely is it tha the arc would sustain itself for that long? In this particular instance the 120/208VAC panelboard is fed by a 45kVA transformer!! Any thoughts or suggestions on this would be greatly appreciated.

 

[davidbeach]

I don't have time to look it up, but I believe you are allowed to set a 2 second limit on the assumption that someone would move out of the way.

 

[dpc]

there should be no danger wit the covers on

If the fault is inside the panel in question, the door or cover can certainly be blown off. Equipment testing is done with bolted faults, not arcing faults. With a bolted fault, there is no arc, so the door would stay on. With an arcing fault, the door or cover offers some protection but there is no assurance that it will stay on. There are plenty of cases where equipment doors have been blown off (taking the protective relays with them).

 

[rcw retired EE]

IEEE Standard 1584-2002 "IEEE Guide for Performing Arc-Flash Calculations" on page 6 says "Equipment below 240V need not be considered unless it involves at least one 125 kVA or larger low-impedance transformer in its immediate supply."

 

[Eleceng01]

davidbeach is right, you can limit the FCT to 2 seconds according to the IEEE Standard 1584-2002.

 

[whycliffrussell]

what are other people's opinions on limiting the fct to 2 seconds (or some other higher arbitrary number) depending on the situation? Is this a common practice when performing an arc flash hazard study?

Additionally, what are people's comments/opinions on this excerpt form IEEE 1584:

Equipment below 240 V need not be considered unless it involves at least one 125 kVA or larger lowimpedance
transformer in its immediate power supply.

I have a lot of 120/208VAC panel boards feed by 45 and 75kVA transformers (does this mean that I am to IGNORE) arc flash with respect to these panelboards? Is this also common practice?

What are other people doing? commments, Suggestions?

 

[dpc]

We typically use 2 seconds as a maximum fault duration, since it is suggested in IEEE 1584.

We follow IEEE 1584 guidelines you reference for arc-flash calculations. It says that no calculations are required for these smaller systems below 240 V. But IEEE 1584 is just a guide for calculation of arc-flash energy levels.

NFPA-70E does NOT say that these systems can be ignored. We typically label all panelboards for arc-flash hazard, using the guidelines given in the NFPA-70E tables.

Maybe someday these two standards will be better aligned, but for now, this seems to be working for us.

 

[whycliffrussell]

so if you have a 208vac panelboard fed by a 30kVA transformer you still include it?

 

[dpc]

We label the panelboard as HRC #1 based on the NFPA 70E Table 130.7(C)(9)(a). We use a Flash Boundary of 4 feet.

We don't try to calculate it using IEEE 1584.

 

[Eleceng01]

One caveat on the 2 second FCT, you need to run the study both with the 2sec FCT and without. If you look at the TCC curves, your arcing current might hit a fuse curve at 2.12 seconds so the level would actually be higher than jsut limiting it to 2 seconds. We only use the 2sec FCT once we understand why a location is so unreasonalby high.

 

[dpc]

Eleceng01

That's really the point of using 2 second maximum - to limit the maximum arc-flash energy calculated. If it wasn't higher above 2 seconds, there no real reason to limit the calculation to 2 seconds.

The idea is that longer duration arcing faults are unlikely and also any worker will not be sticking around if they can get out of the way by any means possible. Is it realistic ?- that can be argued, but IEEE 1584 says it's OK (In the Annex) so that's what is normally done.

(Assuming I'm interpreting what you said correctly)

 

[whycliffrussell]

ok. i see. so people do limit the operating times. what about the 125kVA rule (ignoring equipment 240V and less that's feed by 125kVA transformers). what are other people doing with that?

 

[dpc]

As I said, we don't really "ignore" the equipment, we just don't do the IEEE 1584 calculations. We use the NFPA table.

 

[whycliffrussell]

the nfpa short form table (the one that makes all the assumptions about fault levels and clearing times?)

 

[dpc]

NFPA 70E Table 130.7(C)(9)(a)

We use this only for the 208 V panels or 240 V single-phase panels that are outside the scope of IEEE 1584.

 

[whycliffrussell]

Does IEEE say that 208VAC panels fed by transformers smaller than 125kVA are outside its scope or to just that you don't necessarily NEED to consider such situations.

So you use the first section of the NFPA table for equipment 50 volts (not necessarily AC)up to(but not including) 208VAC, and from 208VAC to 240VAC where the system is not directly feed by at least one (1) 125kVA transformer?

I'm am doing a study in a CO for a phone company (does this imply that the 52VDC phone exchange equipment should also have a flash hazard analysis done?) The source pulls over 1000A.

Additionally, as per our earlier discussion, if there is not difference between 'covers on' and 'covers off' for panels why does NFPA distinguish between this in the first section of its short form table?