I've heard it stated by AF study engineers something to the effect of: 'if a label shows an astronomic calorie value, it's clear the person doing the study has no idea what they're doing'.

That said, what approach(es) do you all take when an existing system has particular components that have obscenely high incident energy levels? This is particularly common with smaller commercial Main Service Distribution gear, where the available energy from the utility to the entire busing is orders of magnitude higher than any PPE accounts for (in my current study, I'm talking excess of 10kcal on a 480V, 400a service). The feeder busing runs through each section of the MDP so the sticker needs to reflect what's there.

The only 'parlor trick' I know of is the 2-second rule, which drops the incident energy down to 22.9 cal, but given it's in a small room with one (crash bar-equipped) door, I'm hesitant to apply it here. Seen too many safety videos of electricians getting blown up. I guess my question is, how would you approach the customer with this situation?

I wish I still had access to and permission to share some of the video evidence I was privy to when I was doing forensic/expert witness work for fires and explosions suspected of an electrical cause. Once case from a few years ago involved a fault at the line side of a 480V, I think 3000A from memory, switchboard. Might have been 2500A or even 4000A. I just don't recall now. It was fed from two utility company transformers in a vault in a spot network configuration.

The switchboard was in a room roughly 20-25' wide and maybe 40-50' long with a 15'+ high ceiling (floor slab of floor above). It was a fairly large room with a lot of other electrical and mechanical equipment in it. Corridor video surveillance captured the event. At least what it looked like from outside the mechanical/gear room. Their was a pair of locked, solid wooden, steel framed, commercial type double doors from the corridor into the gear room. These doors were roughly 20' away from the main breaker section of the switchboard. The corridor doors blew open, ripping the deadbolt out of the door and a giant fireball erupted from the room. I mean giant. Again, about 20' away from the gear. Some of the bolt-on panels of the gear were blown out and very deformed of course. If a person happened to be anywhere in that room during the event, not just within the limited approach boundary, and who knows what the calculated arc flash boundary was, I think they would have been lucky to survive. The whole room had to have been engulfed. The pressure wave blew out several of the corridor lay-in ceiling tiles.

If those engineers that claim high incident energies mean someone doesn't know what they are doing saw that video I think they would change their minds.

Circling back to this (and touching on a common theme talked about here), when a study indicates a generator breaker will never trip because the fault current is below 2kA (causing it to time out in software), what are the options? In a study I'm working on now, the LT pickup is 320A, the ST pickup is 3200A, and the INST is 7200A(fixed). How does one determine how far one could drop the ST pickup before I'm in danger of tripping under inrush conditions of downstream pumps?

Add the pump motor starting curve to the TCC and coordinate the breaker with the curve.

For example

Thank you! In this particular case, the generator breaker and ATS are the points with objectionable incident energy, but there is an MCC feeding 3 motor circuits downstream and a 480V panel>transformer>208V panel situation (feeds no motors). Only one of the three motors has a VFD (which also has an RL reactor upstream, but no maintenance bypass).

Would you, in such a case, add the inrush current of the two non-VFD-controlled motors together, along with the inrush of the transformer, and adjust the curve to the sum?

Maybe. Not enough information. I can for sure model multiple motors on one feeder thereby combining their starting curves into one. I'm not sure if that is possible to add in the transformer inrush too. I might be able to build that into the controller but I've never tried. Also sizes of each would affect how I would approach it.