I agree that you cannot use a resistor if there are line to neutral loads.
I've never seen a neutral grounding reactor *on a generator* of any kind in the past thirty years, so jwerthman's comments are interesting to me. Maybe I live a sheltered life. My observation on grounding reactors was based strictly on personal experience so maybe there are more out there than I realize.
But I must take issue with jwerthman on the use of grounding resistors on paralleled generators. This is commonly done and generally causes few problems with resistor heating, although it is a concern that must be addressed. The heating effect is the square root of the sum of the squares. A 10% third harmonic current increases the heating effect by only 1%. I have seen many industrial installations with three or four generators in parallel, each grounded through a separate resistor. It can be argued that this is not an optimal design, but I have run across only one installation where resistor heating became a concern. As I mentioned in my earlier post, ANSI standards do not require any continuous rating for a short-time rated resistor, so any neutral current must be taken into account when specifying the resistors. Of course, in reality, any 10-sec resistor does have some continuous heat dissipation capability, even if not on the nameplate.
I do agree that neutral switches can be used, although these become yet another maintenance and operational concern.
I feel a bigger problem with multiple grounding resistors is the increased groud fault current and resultant damage for stator ground faults.
Seems like there is always something to be learned (or re-learned) about system grounding.