Recently, a client asked if there was something I could point to in the NFPA 70E Standard that specifically prevented a qualified electrician from working on a disconnect after it was shut off, but with the line side left energized?

I asked exactly what he meant? He said our electricians sometimes need to disconnect one of our 480 volt welders for repair or to be serviced. We sometimes have more than one welder on a circuit, so shutting off the breaker to disconnect a welder affects other welders or machines. So, my question is this: is there a specific section in NFPA 70E that prevents the above described activity? Please remember, the electrician has verified that the load side of the disconnect is off. However, he's left the line side energized.

In article 130.2(A) it states that the only "justifiable" reasons to perform hot work on anything are: additional hazards (created by turning off the power), infeasibility (it is impossible to perform the necessary work with the power off), or the equipment is less than 50 volts (and there is no increased exposure to electrical burns or electric arcs).

From the way you describe it, the line side of the switch would still be hot, and he would be exposing himself to an electrical hazard. Since it doesn't sound like it meets any of the criteria from 130.2(A) it is not permitted by 70E.

There is nothing, in NFPA70E, that prohibits this activity.
You need to determine if an arc flash hazard exists (70E 130.5). I think most people would agree that moving tools and conductors near uninsulated lugs could be treated different than when the lugs are shielded from accidental (e.g. brushing) contact.

When working inside of the AF boundary appropriate PPE must be worn. In your case the line side of the disconnect would be the location for calculating the incident energy.

Two different issues. First off, energized work is defined as making circuit modifications to energized equipment. This obviously also captures the idea of operating equipment as well such as operating a disconnect but this appears to be more of a definition issue...operating equipment has never been intended to be treated as energized work. So working on the de-energized side of anything is not energized work except for testing it for absence of voltage because it is treated as energized until tested dead. This is justified by the fact that it is part of the de-energization process and also no energized work permit is required for voltage testing.

Second we get into the issue of what hazards exist when performing work on the load side of an open disconnect. There are two issues to be aware of. There are shock protection boundaries to be concerned with that exist only if there are exposed, energized conductors. This means that the conductors are not insulated, isolated, or guarded. With most disconnects they are not fully insulated or isolated but they are recessed to the point where they are protected against accidental contact. In this case there is no shock hazard and thus the boundaries do not exist. Just because you can visibly see metal does NOT count. In fact since a lot of industrial equipment has metal housings and metal conduit, it should be obvious that the criteria for shock hazards doesn't trigger just because there is a metal surface present that can under some extreme circumstances become energized.

The second issue is that of an arc flash hazard. It is separate from the shock protection boundary but goes hand-in-hand with the concept of protection against inadvertent contact. The definition of an arc flash hazard specifically requires that employees are interacting with energized equipment in such a way that they can create an arc. This is not an absolute definition but rather that there is more than a small chance that it could happen. If not then the arc flash hazard boundary doesn't exist and PPE is thus not required for arc flash. There are fine print notes within 70E that mention that properly maintained and operating 600 V class MCC's for instance aren't normally an arc flash hazard and the task list in 70E-2015 has an extensive list of examples where an arc flash hazard does not exist and arc flash PPE is not required.

Again, the way that these are defined in 70E is based on probability. If there is only a remote chance of an arc flash or shock, then the hazard "does not exist" using 70E's terminology.

There are situations where neither, one, or both hazards (and boundaries) exist.

In the case of an open, locked out disconnect, generally the energized parts are guarded and/or shielded so neither set of boundaries exists.