NEC and NFPA70E list equipment that should be included in an arc flash analysis, i.e. electrical equipment such as switchboards, panelboards, control panels, meter socket enclosures, and MCCs. This has led to lots of discussion about what "such as" could mean.

I was wondering if this had been thought about backwards, i.e if there was any equipment that should be excluded from an arc flash analysis? Maybe something like single phase 120V lighting fixtures, or 120V household receptacles?

The new wording in IEEE 1584-2018 seems to indicate there is no such equipment ("less likely" it says, but still possible). Or at least no one is willing to say. Since equipment is not electrically isolated until it is verified isolated with a meter, this presents some problems. Technically, even if I've opened the upstream protective device it's not "safe" until it's verified. The verification process would therefore require opening the equipment, which means wearing PPE. So does every lighting fixture and small piece of equipment really need to go through the data-collection and analysis, because the equations use 3-phase, there's a greater-than zero chance that the incident energy might 1.2 cal/cm2 or higher.

Additionally, I can't really be confident the table methods are usable either due to the possibility that the current might be low enough that the clearing time of the protective device would exceed 2 cycles.

Should a safety policy really include ALL electrical equipment in an arc flash study? If not, how can it be justified?

The short answer is no, not all equipment must be included in an arc flash analysis. The long answer, in my practice is:

I include all 480V and greater distribution equipment and if the incident energy is greater than a value that myself and the client agrees on at that point we go a step farther. In some places that could be 8cal/cm2 and others it could be 2cal/cm2.

We do something similar on the 208/120V side, we go to the first panel after a transformer. Typically we do all transformers, but depending on the customer we may place a cut-off at certain transformer sizes, say 45kVA. At those low sizes they won't allow enough energy to pass to sustain an arc.

Integration into an electrical safety program

Where you stop in the arc flash analysis depends on how this analysis will be used in the electrical safety program. For example, if a client decides to only go to the distribution panel, but there will be energized work at the motor disconnect switch, I would expect that the electrical safety program would state the the upstream label shall be used for that work. So if the label says its 12cal/cm2; that is the PPE that the worker will wear downstream.

The client could, if they have engineering staff at that location, make the decision that if the upstream device is greater than 10cal/cm2 that a specific calculation would be completed for the work area, and that energy level and PPE would be indicated on the Energized Work Permit.

I hope that helps.

Cheers,
Jeff M

Our Electrical Safety Program sets the lowest threshold for equipment to labeled for arc flash hazard at any device or circuit at 480V or below protected by a OCPD with a rating of 20A or below. For that equipment employees are instructed to wear PPE as they would for a label stating 1.2cal/cm^2 or below (long sleeve, non-melting natural fiber, safety glasses, hearing protection, etc) until de-energization and LO/TO has been verified.

This came about by maintenance electricians asking for guidance on PPE for changing bulbs, using meters to test 120V outlets, etc. We looked at many common 20A MCCBs with the widest conceivable range of arcing fault current and could not find any combination that produced an incident energy rating over 1.2 cal/cm^2.

The flip side is we now are obligated to label everything (disconnects, panels, etc) protected by a 30A OCPD or larger. It's an ongoing process.

Testing performed by IEEE - 1584 committee found that sustainable arcs are possible but less likely in three-phase systems operating at 240 V nominal or less with an available short-circuit current less than 2000 A. See Section 4.3 for more information.

One technique to short cut a tedious study of many small circuits may be to develop a target number. Let us say 7 calories, draw a constant energy line for (aka energy boundary) for that 7 calories and then plot the various CBs (the ones most at risk of producing high energy) within the applicable fault current range. If the CBs are to the left/below the line then they all produce less that 7 calories, 8 calorie PPE is good enough (chose a margin that suits your policy). No need to figure out if a particular exposure is 4.3 calories and another in 6.7 if you are using 8+ calorie PPE. All you need to know is that the Ei is below 8, or whatever cutoff level you have selected.

Whatever brand software you use for the system studies should be able to do this in one way or another.

But you have to be careful with small circuits with low fault current, they may allow an arc to sustain a long time because the current is "low" and the protection reacts slowly.

It is true that the tests done by the IEEE indicate that at 120/208V arcs self extinguish often and are hard to establish... but to some degree that is a factor of the gap and enclosure. I have seen pictures of residential load centers that melted on the wall into a metal blob... that arc lasted more than a few milliseconds!

The new 2000A AIC rule is way different than the old 125KVA rule. When you consider that small lighting transformers general have low impedance the KVA size goes considerably lower. An impedance of 4% @208 3ph, would limit KVA size to 29 KVA . So that means a 30KVA xfrm now needs to be considered. If you can find a xfrm at 5% impedance, you could move up to 36KVA, but with nominal size ranges that means a 30KVA is ok, but a 45 KVA is not.
This is really a considerable jump from 125KVA. Makes you question the validity of either the first or last testing method and how likely that arc is actually going to occur in the workplace.

There was no first testing method.
The 125kVA limit was chosen based on anecdotal evidence only. I recall the original IEEE report contain a single data point for all 208V installations.

NFPA70E 130.59N(A) says a risk assessment must be performed taking into consideration:
1 the hazard (IEEE 1584 is one available method)
2 the likelihood of the event or injury (this is very subjective, industry statistics may be considered)

Now if we just had some clear guidance, from NFPA70E, on how to evaluate the Likelihood portion. For the most part I see many companies simply saying any possibility is bad therefore the calculated hazard must be protected against. However, there are some that go with the older methodology based on the industry statistics for arc flash injuries occurring, while wearing 1.2 cal/cm² PPE. This rational is similar to how most people treat single-phase 120/240V cicruits.

Opinion - Since IEEE1584 algorithms do not work at 208/240V (no testing to substantiate the results ) wouldn't it make more sense to use the NESC value of 4 cal/sq cm that is based upon substantial testing. If you follow some of the unofficial testing done on 208V, in many cases they cannot get the test completed since the arcs self extinguish??

How many on this forum have personally seen (not heard about) an arc flash burn on the torso or face in a 208V or 240V system. The same applies to single phase, how many burns at 18" have you seen. My point is, again opinion, TOO many labels make the an arc flash study less effective.

I use the NESC table for single phase. My reasoning is the NEC requires a label and generic label means and says nothing for protection. My job is to protect the public.

I believe the 2018 edition of IEEE1584 was based on an sufficient number of 208/240V testing, such that they changed their handling of small transformers.