Do you perform an arc flash study on a facility that has a single phase service with no transformers? What if the facility has a 3-phase 208 volt service and no transformers? Would you still perform a study?

Thanks in advance.

Yes.

Actually I would still perform an Arc Flash Risk Analysis.
Although I might not end up using the IEEE1584 calculations for a service fed by utility transformers <125kVA.

You might not be using software to do calculations but that doesn't mean that the study/survey can't be done. It's more a matter of determining whether or not it meets the "125 kVA exception" and thus what areas might need fairly minimum (4 cal/cm2) PPE vs. those areas where no PPE is needed.

Thank you for the replies.

Hi Paul

I could see performing a calculation on a 120/208 3 phase system, however on a single phase, lets say 120/240 single phase, how does the arc sustain on single phase that could possible result in a burn that could cause serious injury?

Thanks Paul

Best regards

See the articles section on low voltage arcs. It just barely sustains somewhere between 200 and 300 V if conditions are right. There has been a fatality recorded by OSHA in 2009 that I referenced in that article which was a 240/120 situation so we can't discount it out of hand at least.

However I agree that this is undoubtedly a rare condition but so far defining when it is sustainable (at any voltage) and when it is not has confounded almost everyone. I can suggest that there is a research group that has published an ESW paper once a year where they have been doing actual testing on MCC's and have come to some startling conclusions on how and when sustainable arcs form in 480 V gear. So my contention right now is that we can't really define sustainability, and further we can't even define incident energy because of that but what we do have is enough evidence out of the equipment specific testing crowd that they have given guidance for this in the tables in IEEE C2 (NESC) which essentially gives a rating of 4 cal/cm2 for everything under 300 V based on testing, and the existing guidance in IEEE 1584 which has a current limitation (and no PPE required) whereas IEEE C2 does not.Thus combining the two we have a 2 level system...if the equipment meets the IEEE 1584 "125 kVA" requirement, no PPE required. Above that, 4 cal/cm2.

Thanks Paul.

Actually I recall now, last year you had sent to me a white paper regarding 120V single phase arc flash and I believe the fatality was in a residential situation at the panel.

Len

Prior to that there is research that went through the OSHA statistics and put together a fairly comprehensive study going back about 10-15 years. It was published publicly through the ESFI as well as via an ESW conference paper. Given the huge amount of 120, 120/208, and 120/240 V equipment deployed I think it is safe to say that the conditions are possible to where we can exceed the Stoll curve at the face/chest area but very rare. Since we can't predict arc stability and since IEEE 1584 empirical modelling is questionable at best at those voltages, that pretty much leaves the Lee method as the only way to calculate any kind of result as an upper bound. But Lee grossly overpredicts the incident energy by a factor of 2 to 5 if we make the "stable arc" assumption (extrapolate from when it extinguishes all the way to 2 seconds) and by orders of magnitude more in many cases when we don't extrapolate.

So where does this leave us? My contention is that we need to rely on tests done on actual equipment, not using theoretical results that have nice pretty equations that are not representative of reality. If we go down this rabbit hole all the way it would seem to suggest then that we have to actually test every piece of equipment for valid results. However we don't need to go that far. There are already established standards given in IEEE 1584 and IEEE C2. C2 gives a blanket requirement (4 ATPV PPE, no face shield) whereas IEEE 1584 gives a much more restricted "no PPE" guidance. These were established via laboratory testing and are consensus standards. So we don't need to do the test work in most cases because someone has already done that for us.