From 1584 Introduction "The arc-flash hazard calculations included in this guide will enable quick and comprehensive
solutions for arcs in single-phase or three-phase electrical systems either of which may be in open air or in a box,
regardless of the low or medium voltage available."

From 1584 section 1.2 "Single-phase ac systems and dc systems are not included in this guide."

From 1584.1 section 3 "Calculations can be run for single-phase equipment using IEEE Std 1584â„¢, assuming it has three-phase service that will yield conservative results."

Question 1:
Can you clarify the single phase analysis method?

Question 2 (mostly in jest, but partially serious):
How many sentences may I copy and paste before violating the restrictions of the documents?

Section 3 explains it. Model the single phase system as a 3 phase system and use the result even though it's not 3 phase. Obviously section 3 is a bit at odds with section 1.2 though.

Some very early arc flash modeling suggested using 1/3 of the three phase result for single phase. However the folks over at Mersen published a couple papers looking at among other things, 3 phase arcs with and without phase barriers, and also consequently at low voltage achieved some single phase-only cases where the arcs did not propagate into 3 phase arcing. One of the take aways from this is that the single phase arcs weren't all that much lower in magnitude of heat released than the 3 phase arcs. This is a little indirect but without having actual test data for comparison, it seems to suggest that making the conservative assumption that single phase and 3 phase arcs have the same incident energy is not so far fetched.

The incident energy models published in 1584-2002, with the exception of the Ralph Lee model, are based on statistical analysis of actual arc testing. There hasn't been a large enough volume of single phase arc testing completed to publish a statistically valid incident energy model for single phase applications.

There are a number of manufacturers and even end users that have conducted handfuls of arc flash incident energy tests on single phase, DC, resistance grounded, and other system configurations. These results are useful to inform the engineering community and the IEEE 1584 working group, but there is not enough data available to make anyone confident (statistically speaking) that these tests have captured a representative sample of a worst case event.

In most situations single phase transformers are quite small so it is not too inconvenient to use the 3 phase model and overprotect the worker, knowing that the results will be conservative.

Back when the 2002 Edition of IEEE 1584 was published, there was a lot less known about single phase arc flash. It was understood that sometimes a single phase event could escalate to a three phase event so the model should always consider three phase arc flash for three phase systems. Do all single phase arc flash events escalate to three phase? No, but there is no way to know for sure. 1584.1 continues to reference this approach until more is known.

Our company has taken the approach of not studying single phase. Since there is no real data to support it. We state that we do not recognize any of the current modeling methods and cannot stand behind a best/worst guess.

Kinectrics has done testing on single phase arcs and is in fact the basis for NESC arc flash tables. I believe work was originally done by Ontario Hydro. It appears that OSHA will be recognizing this work as an accetable method in the future. The range of testing does include low voltage. I do not see anyone referencing this work or software. Is there any controversy about the adequacy of this work or software?

Yes, there is lots of information on the Kinetrics data. I spoke with one of the two authors of that program. He said that it was based on some sort of theoretical analysis and was not based on empirical data. So...not quite the gold plated standard that we'd like it to be. Best use is above 15 kV where the empirical data doesn't exist anyways so choices are Ralph Lee (known to be off by orders of magnitude), Duke Heat Flux, or Kinetrics Arc Pro.