On Tue, 08 Mar 2011 21:26:01 +0000 (GMT), "Dave Liquorice"
wrote:

On Tue, 8 Mar 2011 20:34:49 +0000, Ian Jackson wrote:

I was surprised to see how low the voltage sometimes went (right down to
the minimum limit). When it was really cold over Christmas and New Year,
I expect the guys at the power stations were shovelling coal for all
they were worth, and that the frequency (which is of less interest than
the voltage) was well down on normal.

The frequency is very important, it's maintained at 50Hz +/- 1% ie
49.5 to 50.5 Hz nut National Grid work to an operational limit of +/-
0.4% (+/- 0.2Hz).

Obviously, any type of clock which uses the mains frequency will run
fast or slow if the frequency varies. My understanding is that the
generating people are supposed make sure that, come hell or high water,
during a 24 hour period the correct number of cycles are delivered, so
that the clocks are correct at 8am. I'm not sure that they always
achieve this!

Oh they do! Questions get asked in the house if they don't. IIRC the
last time they had to announce that mains driven clocks had the wrong
time and would need to be corrected was during the winter of 1947...

Maybe 20-30 years ago it dropped to minus 4 secs or so on one occasion
when there was a split system. It *might* have been October 1987.
Not sure if they got the clocks back to zero error at +24hours.

During winter on another occasion the frequency dropped to somewhere
in the mid to high 47's before a system split and load shedding gave a
high frequency in the north and midlands and a low one in the south.

It took a hell of a lot of work to get the two systems back in sync,
and as a consequence over the next couple of years they changed the
synchronising relays associated with a very large number of circuit
breakers across the 400kV and 275kV grid such that a second method of
automatic closure was possible dependent on slip angle and rate of
change of slip angle rather than just the slip angle. These could be
primed and left for 15mins so that if the system either side of the
breaker met the targets it would automatically close, previously the
only method was a 2 minute timer for 'perfect sync' or a manual
closure that ran the risk of exploding the breaker into copper and
porcelain shrapnel.

The long period of split running during this incident gave a clock
error such that various bits of the country were out of sync by a
second or more. None of this made the papers at the time


--