Here's something that may be worth having some discussion on:

We all know that IEEE 1584 contains a statement (paraphrasing here) that systems fed by one transformer rated at 125kVA or less and the secondary voltage is below 240V need not be analyzed. Now, what about a generator that meets the same criteria? Should that be exempt also based on the criteria for a transformer?

Thoughts/comments/opinions?

Hmm, really good question wdb!
A generator has a sub transient, transient and a steady state current. This should result in a higher current that decreases overtime back to the steady state current.

I would think that in some conditions the arc flash will sustain until the current drops and their isn’t enough heat development to ionize the air to conduct the arc flash. Since the electric field (Don’t know if this is the right translation) isn’t strong to begin with the arc will extinguish.

If this is the case then you might want to only calculate the generator for the first X cycles. How long is something that will vary between generators.

I used SKM to do a quick arc flash analyses on a diesel generator of 47kW 208V for 0,5 seconds and got an incident energy of 1,1 cal/cm2. If for some reason the protection device does not trip within 2 seconds then the energy would be 4,2 cal/cm2. If I would assume the arc will self extinguish within let’s say 5 cycles the energy would be 0,21 cal/cm2.

I do not have the specs for a generator with a higher rating then 47kVA and less then 125kVA.

Since a transient and subtransient disappear after a few (about 5-10) cycles, these would boost the initial energy released by the system. So although it's usually very close (<5% differences) because SKM takes this into account, you can't just treat time as linear based on results from this software.

As to self-extinguishment, the complication here with a generator is that you simply won't have much in the way of available fault current which severely limits things. As reported recently by I think a report at Battcon a DC arc at 130 V with a 20 kA fault current could not be sustained over 0.8 seconds. There was a single case of a sustained 208 VAC arc in the test data for IEEE 1584. More recent testing using "barriers" and different sized boxes are in a published report on Mersen's web site that show sustained arcing at 208 V down to about 4.5 kA with narrow (1/4") arc gaps, some of which definitely reach the 1.2 cal/cm^2 treshold.

I would suggest reading up on the various laboratory tests done at <250 V. Right now there is some evidence that the "125 kVA" threshold is invalid so trying to extrapolate beyond the stated rule is going way out on a limb at best. If you can find relatively valid test data though that performed at test at or beyond your conditions (voltage, currents), then you can probably safely use that reference for the "self-extinguishment" threshold. There are no equations for this right now for AC. There is an equation for DC based on the known measured transition point between the low and high current arcing regions for DC. However so far the DC models also tend to grossly overpredict the actual available fault current.

I tend to agree, generators don’t seem to sustain a fault as compared with a transformer connected to a infinite bus….