My first reaction would be - no, since the arc is downstream from the transformer. However, CAFIs weren't part of my EE curriculum
. I've wondered about the arcs produced by the motor in my Bosch hammer-drill, though.
Kiril - I think GFCI, AFCI, and CAFI can be summarized as:
GFCI - intended to protect personnel by tripping on current flow of 5-6ma out of the circuit - presumed to be through a person to ground.
AFCI - intended to protect property by tripping when a series arc is detected - presumed to be a fire-starting event. (And, indirectly, to protect personnel from the ensuing CO and fire - they were first required in bedrooms.)
CAFI - intended to protect propery by tripping as an AFCI does, but also when a parallel arc (i.e., to ground) of 5A is detected. Both events are potential firestarters, but if the parallel fault is to ground via a person, it's a fatal current.
Note that none of these devices provides surge protection for attached devices. In a perfect world, a single device would provide protection from every imaginable fault, and would be inexpensive, but this is an imperfect world. The paper you cited is an interesting way to go about attacking the technical problem; I doubt we'll ever consider it inexpensive, but if it's attached to an airplane it doesn't have to be.
Probably the best way to get all protections now is lightning protection, a whole-house surge protector, CAFIs at the panel, GFCI receptacles, and purpose-built surge protection at the device. I've got 3 out of 4 (no CAFIs yet) and haven't fried anything yet while living in the most lightning-prone area of the US. (see
http://www.lightningsafety.noaa.gov/lightning_map.htm)
Mike (EE, A&P, ex-IBM, fwiw)