So from what I gather, I can use IEEE 1584 calculation methods assuming that there actually is an exposed source (broken isulation in the conductors). That should give me at least a ballpark value.

I understand that in the manholes there is no escaping through the 2 second rule, and also the effects of the arc flash will be enhanced. So any amount of incident energy will still cause more injuries than in an open space at the same energy level.

Would you say its fair to assume that the risk is minimal if the activities being performed in the manhole do not disturb or touch the 480V cables? The activities will be only for evidence gathering (pictures) and accounting for the condition and amount of cables in it (visual inspection). Hopefully these activities can be performed without disturbing the actual installation.
Thanks again!

Arcing faults can occur regardless of whether there is exposed insulation or not. In editions of 70E prior to 2009 this area was rather ill defined in the fact that shock hazards and arc hazards are separate and distinct. That is no longer the case in the 2012 version. However the initiator is also different. Generally if you don't do anything to cause equipment to change state, the risk of an arcing fault is fairly low.



So far so good. However if the incident energy is <1.2 cal/cm^2 then obviously the arc flash boundary is less than the working distance and so chances are that the "enhancement" problem is probably minimal. I just did a poor job of describing this. However as soon as incident energy gets close to being a significant concern, the problem may be worse than IEEE 1584 calculates.



Hmm...are you going to ask me again in a court room? Because I may not want to answer.

Seriously...I'll give you an indirect answer. Look at the definition of arc flash hazard in 70E. The key words are that under conditions where the electrical gear is going to readily change state, the risk of an arcing fault is pretty good. Many times the 70E Committee has rejected various proposals because the likelihood while "just walking by" is remote. However my answer would change drastically if there is already insulation cracked and fallen off and there is a possibility of stumbling into it or knocking something loose just by crawling down into the manhole.

Even if the equipment was in good condition if for instance you have non-round manholes, my answer would change. In this case you would have somewhere between a 4% and 16% chance (based on HEART data) of an electrician accidentally dropping the manhole, and some chance of it landing the wrong way. I don't have good statistics on this last part because it would depend on your vaults so I'd have to give you a worst case analysis of around a 10% chance of an arc flash from removing the manhole in a non-round scenario. This used to be a big problem with vaults before round became the standard.

Another consideration...what's the chance that even if you try to impress on your workers that they could be killed by poking/prodding at the wiring but then you ask them to go do a thorough inspection that they absolutely can't resist the temptation to poke at the wiring?

I really recommend that for this kind of answer where there is no textbook answer, you really need to do a risk assessment using a formal risk assessment methodology. There are several that are an ANSI standard. IEC has a couple. There is also the RIA standard (my favorite for manufacturing plants), and ISA S88 (IEC 61511, my favorite for process plants). Any one of them will work as long as you are consistent in how you do it, include all the players, and document it.