whether or not what has occured is arcing damage, cavitation, corrosion or even porisity must be determined with consideration to all the info you can aquire. If an definite decission is not possible, then use info to indicate most likely.
given my back ground is related to the mechanical aspects of assembling generators, my information provided will be biased towards such and I will provide discription of the mechanism that would cause arcing.
the frame of the generator is grounded through out its lenght.
the rotor of the generator WILL devolp an electrical potential during opearation. to prevent this potential from becoming a current, the turbine end of the rotor is grounded with a contacting brush. the collector end components with the posibilty of current arc over are provided with an insulator. this would be the oil defectors, the bearing and the seals.
If the insulation becomes damaged or contaminated with low resitance oil, the potential on the shaft will drive a loop current. given the thin film of lube oil (either at the bearing or seals) provides a very small seperation, arcing will go through this film.
some methods of insulation do not allow the quality to be verified onced assembled. this seems to be the normal now. some manufactures provided double insulation so that the section between the two insulators can be measured to verify both insulators are working.
for the single insulator (or if one of the double type) fails, the effectiveness of the insulation is verified by measuring the potential at the collector end of the rotor. this needs to be a relative reading since the normal potential varies between units. I have normally seen 50 to 250 VAC
the turbine end of the rotor should be grounded, this is verified by also measuring shaft voltage and it is expected to be zero. IF this ground fails, and the turbine end voltage increases, the collector end potential will increase, but to a greater potential and can lead to failure of the insulation, thus devolping a current loop, with arcing across the oil film
as for the ring ID dimensions, as for being equal for the air side and the gas side ring, the differance expected is only 0.002" so were they they within 0.001" of each other (example the air side diametrical clearance is 0.006", with the gas side diametrical clearaance of 0.008")
thus to consider arcing as a possible cause, the quality of the seal insulation must have been poor. given no damage to the adjancent bearing, the bearing insulation was still effective