g'day,
> HZ moving around causes far more problems -- makes peoples clocks go slow
> and fast. for that reason, HZ is very tightly regulated - if it goes over
> for a period, they'll make it go under to even it out.
Sorry for the nit-picking... Grid AC frequency is tightly controlled for
entirely different reason - imagine what happens when you have two
generators on the same grid, out of phase from each other. (In really
large grids, like ex-USSR's United Energy System wave propagation delays
make the whole synchronization dance quite interesting,
i tend to think of electricity grids in the form of "organised chaos". its really just one big phase-locked-loop setup where the producers are always second-guessing what the consumers want.
in the four power-stations i've worked on the construction of (doing automated power-station controllers), we used automatic-phase-loop balancing that removed the "manual" part out of closing a circuit-breaker of a generator coming online - it'd do it for you, and it'd only do it when it was "in-phase".
the hard part them comes dealing with a "black start" - ie. where there has been a total power failure and there is no phase to sync on to.
i've also had the somewhat dubious honor of getting the automated power-station stuff wrong too. the somewhat dubious honour i can claim is that i caused a blackout of an entire nation. (the Kingdom of Tonga).
its very difficult to physically close a circuit-breaker when its out-of-phase. if you do manage to do it (eg. you have a hydraulic or compressed-air based breaker that DOES have enough force to manage to close it out-of-phase), then obviously "bad things"(tm) happen to that generator.
legend has it (and i don't know if this is true or not) that there is a generator (or, at least the Alternator) in the Swan River in East Perth (Western Australia) that 'jumped'
particularly
considering that turbines start rotating faster if load drops, etc).
that is true of all engines. its always a supply/demand driven system, with the engine typically always operating at a fixed RPM. if supply exceeds demand, the engine spins faster (and HZ goes up). if demand exceeds supply, the engine slow down slightly (and HZ goes down).
the role of the governor is to try to keep a stable frequency while governing the amont of fuel being added to the engine.
There other, as imprtant, reasons to keep frequency stable: for example,
phase difference determines the direction of energy flow in an inter-tie
transmission line! (see, for example, analysis of coupled oscillators in
Feynman's lectures on physics). And there's a whole can of worms in
keeping right the angle between voltage and current 
the kVARs, yes.
i've was also at one installation where the kVAR balancing was awry between generators in the same power-station. the [electronic] governors were basically 'battling' with each other but never getting in sync with each other. the net-effect there was that we had the HZ so far out (and oscillating) that it was evident by street lightbulbs pulsating. (thankfully, that didn't occur on a power-grid, but at an isolated mine-site).
Actually, a lot of what grid control automatics people do could be a very
well worth to learn for the network people. Grid control requires very
fast redistribution of the load to keep parts of grid in sync; miss the
time window, and you have to live in panic mode, effectively shutting down
and patritioning grid to protect equipment against cascading effects.
power grids typically always have trip-alarms based on all sorts of things -- KVars out of sync, Frequency out of range.
things tend to 'work' providing there is a large enough pool of producers and consumers. its large load-spikes and load-dips that cause things to break.
i still find it incredible that power is as reliable as it is. (i don't live in Califormia, so i can make that statement. ).
in either case, i think this discussion has probably exceeded its usefulness.
cheers,
lincoln.