Some time ago, over some beers with a fellow EE who's much more experienced than me, we were discussing the unique challenges present in utility-scale power generators, like the ones in nuclear power plants. A single generator can be in excess of 1000MWe, and has all the more exotic stuff like hydrogen cooling, etc.
In passing, my friend told me
At least they don't have to spin at 3000 RPM, they have many poles on the stator to reduce the need for rotor speed...
Which I thought was a good solution to an engineering problem.
It seems, though, that they don't actually use it often.
When reading about the Chernobyl disaster, I found that the RBMK reactors, and indeed even the modern VVER varieties, use generators spun at 3000 RPM. Excerpt from Wikipedia:
The turbine and the generator rotors are mounted on the same shaft; the combined weight of the rotors is almost 200 t (220 short tons) and their nominal rotational speed is 3000 rpm.
Also, from the Chernobyl NPP page:
The Kharkiv turbine plant later developed a new version of the turbine, K-500-65/3000-2, in an attempt to reduce use of valuable metal. The Chernobyl plant was equipped with both types of turbines; block 4 had the newer ones. The newer turbines, however, turned out to be more sensitive to their operating parameters, and their bearings had frequent problems with vibrations.
Indeed, on the night of the Chernobyl disaster there actually were two tests being done about the same time: the famous turbine-rundown test we all know about, and a quite obscure vibration measurement test (it's not widely described, but search for "vibration" here if interested).
Another data point: the only time I went on an inside tour of a working NPP (a VVER design) I noticed that one of their instruments in the control room was sometimes dipping into the redline area, and when I asked, we were told it's the vibrations indicator.
So all in all it seems that vibrations could be quite the hassle, and one has to wonder why they don't utilize the half- or one-third-speed approach, which will allow for 1500 or 1000 RPM on 50Hz grids.
So the technology to use lower generator speeds is there, but why is it only used for halving the speed, why not divide further? What are the downsides of lower speeds?
The question may be not appropriate for this StackExchange, as the reason may be purely materials/mechanical issue (less RPM needs higher torque, ...). I'm interested if there are EE reasons that make the higher RPM desirable.