What I have read elsewhere is that the grounding resistor should be able supply a minimum of 2 times the capacitive that would flow through a ground fault but no less than what it takes to actuate ground fault alarms and ground fault protection for equipment.

However, ground fault current also needs to be limited to a level that will not cause ground faults to burn clear. If you are using a shielded extension cord so that cord cuts a nail punctures will be shut off by ground fault protection, it makes no sense if the fault burns clear. Actually, one failure mechanism for a cord is that a wire strand breaks and then works its way to the surface or to the other conductor in which case the conductor shielding will cause that strand to ground fault and the ground fault MUST NOT burn clear.

One thing that electrical engineers sometimes forget is that the grounding resistance also needs to withstand a power cross with a control voltage such as 120 volts. This would mean a 397 volt resistor for 277Y480 impedance ground and 267 read 480 volts for 347Y600 volts. The ground fault current may or may not be enough to blow control circuit fuses. Control circuit fuses also need to be able to interrupt the higher voltage but not necessarily at short circuit levels.

Mike Cole, mc5w at earthlink dot net

I've read many IEEE-published and other articles on high resistance and low resistance grounding. Only one that I've come across (so far) stated the following, which could be miscontrued to mean that the grounding resistor should supply two times the capacitive current that would flow in a ground fault: "the current I must be greater (or equal) to twice the network capacitive current in the event of an earth fault in order to limit overvoltages" (from Schneider Electric Cahier Technique No. 62 - Neutral Earthing in an Industrial HV Network, page 8).

However, this document does not say that the resistor current should be twice the capacitive current, but rather the total ground fault current should be twice the resistor current. The total ground fault current is equal to the square root of the sum of the squares of the resistor current and the capacitive current (or "system charging current"), since they are each 90 degrees out of phase with the other.

There is consensus over the past 50 years in the literature of high resistance grounding: The resistor current must be at least equal to or greater than the system charging current (defined as the capacitive current which flows into a bolted ground fault if the system were ungrounded). In low voltage systems the resistor current typically ranges from 1 A to 5 A. The system charging current typically ranges from 0.1 A to 2 A. High resistance grounding may be used in medium voltage systems up to 5 kV. In such systems, the resistor may be sized up to 10 A.

As long as the resistor current is equal to or greater than the system charging current, transient voltages will be limited during intermittent arcing ground faults.

As regards "a power cross (of a 480V 3-wire high resistance grounded system) with a control voltage such as 120V, causing the neutral grounding resistor voltage to rise from 277V to 397V":

It is known that the accidental connection between a medium voltage system and an ungrounded low voltage system, such as in a 4160V:480V transformer where a phase conductor from the primary touches a phase conductor on the secondary, may cause the secondary neutral to rise to a maximum of 2400 + 277 = 2677V above ground.

If the 480V system were high resistance grounded, however, there would be a path (through the 277V neutral grounding resistor) for the ground fault current to return back to the 4160V source; the ground fault protection on the 4160V system would trip.

If a 120V control voltage phase conductor were to touch a 480V high resistance grounded system phase conductor, the neutral grounding resistor voltage would not rise to 277 + 120 = 397 V; rather the 480:120V control transformer core would likely saturate from overvoltage and blow a control fuse. Neutral grounding resistor do not have to be rated for more than line-to-neutral voltage according to IEEE Standard 32, "IEEE Standard Requirements, Terminology, and Test Procedure for Neutral Grounding Devices".

Only if the 480V system were ungrounded could I imagine some voltage shift on the 480V system neutral during accidental contact with a 120Vac system, but that has nothing to do with the rating of a neutral grounding resistor.