I have a customer , for whom I am preparing an arc flash hazard analysis, who has a main service switchboard that is made up of fused bolted pressure contact switches for the mains and feeders. None of the feeder switches have the same type of fuses installed, even within the same switch! The amp rating of the fuses do match in a single switch. Since the equations are three phase, I was wondering how to handle this in the modeling. Can anyone recommend an approach? I really don't want to model three cases; i.e., one case with the fuse in Phase "A", one case with the fuse in Phase "B", and one case with the fuse in Phase "C".

Because of this, and other reasons, my company created a library entry of generic non-current limiting worst case fuses. Prior to doing this we would usually compare the curves for the three mismatched fuses and use the 'slowest' one of the bunch.

In my opinion, this probability of 'mix and match' possibility, is one reason I do not like using fuses when the choice is critical (i.e. selective coordination and arc fault mitigation) in non-controlled environments.

I agree with JDB to use the "slowest" fuse, but this should be applied within the range of arcing fault currents present.

Two fuse types, even if they have different current ratings, may have some overlap. One fuse may be faster in the higher current ranges, and slower in the overload region.

Hopefully you will be performing another analysis which reflects replacement of all their fuses!

The mismatched fuses were probably installed after the originals blew. Fuses should be replaced in sets. The fault that blew the originals may have fatigued the elements of the remaining fuses enough to negatively affect their performance.