This discussion comes up from time to time. The calculated incident energy is directly dependent on the arc duration which is normally defined by how long it takes an upstream protective device to operate. What would you do if the upstream device’s condition was questionable and it was not certain if it would trip/clear correctly - if at all.

You can select MORE THAN ONE answer if appropriate.

What would you do if a protective device in an arc flash study had questionable operating characteristics?

• Ignore it and use the next device upstream
• Use it anyway
• Replace it
• Exclude that circuit from the study
• Something else (please explain)

Something else - test it to verify and fix if necessary.

I voted for "Something else". IEEE 1584 recommends a maximum arc duration of 2 seconds, but caveats that with a comment about increasing that arc duration if egress is somehow limited or obstructed. Thus, your "worst case" is most likely due to the questionable protective device timing-out at 2 seconds. To be thorough, you could also look at the case of the next upstream device, but that device would most likely trip quicker than 2 seconds, resulting in a lower incident energy. This is the most conservative approach and would account for any unknown impedance internal to the protective device, i.e. arcing of the circuit breaker contacts trying to open, but ultimately not opening. Of course this worst case method also makes a stronger case for replacement/repair of the questionable device.

Since nothing is perfect, there a small (sometimes infinitesimally small) risk that a downstream device will fail to clear so if we expect the device to operate within its rated parameters, use the device ratings, but be aware that if it fails..... If the downstream device is not reasonably expected to operate within its parameters and it can be replaced - then by all means replace it as soon as possible. Until replaced, the arc flash warnings, approach boundaries, work practices, etc. should be based on the upstream device's ratings. This could be hours, days, weeks or in a recent case at my plant, almost nine months.

I am voting for something else. An analysis isn't worth the paper it is printed on if the devices used in the study aren't relieable and operate within the manufactures' range. Having a questionable device--- in my mind it should be tested verified and fixed when necessary.

Something Else - in some cases replacing or testing the device is not in one's control (ie consultants). In that case in the report it is clearly stated that the results are based on the device being maintained and many times a recommendation made that the PPE used be one level higher. I have seen that done before.

Replace it. There is no justification for leaving a known potential safety hazard in place (if replacement is prohibitively expensive, then test/repair).

Totally disagree with the "PPE 1 level higher" approach. This is ridiculous. PPE "0" to PPE "1" = 4/1.2 = 3.33 margin. PPE "3" to "4" = 40 / 25 = 1.6 safety margin. If you are on the bubble in the first place, none of this matters. Better to reanalyze with a "2 second rule" and see what this results in. In my plant they are still using the "just wear a 40 cal suit" rule. About 25% of the buses would be safe if wearing PPE as listed on the label without modification either because it's a fuse and thus low probability of failure, or the rating didn't change substantially. However 50%+ exceeded 40 cal/cm^2. Thus the "40 cal rule" or whatever you want to call it (PPE+1, etc.) completely fails with half of the buses. Granted I didn't figure in relative exposures (some buses are more important than others) but at over 1000 buses it's not really practical to figure that in.