POWER: San Mateo/San Francisco power outage report

Chris A. Icide chris at skycache.com
Wed Dec 9 16:03:23 UTC 1998

Network Engineering != Power Grid/Generation Engineering.

What you know about building a network does not apply to building a power
generation and distribution system.  Imagine what you might design if a
large BGP flap could physically destroy a POP.

A small variance in voltage or frequency can cause significant damage not
only to the transmission and generation facilities, but also to end user
equipment.  When you lose or gain significant load on a distribution
system, it causes the frequency to rise or fall.  When the frequency varies
significantly from 60Hz, you begin to catastrphically destroy generator
sets and sensitive transmission gear.  So, trip circuits are set to
identify a frequency change that is significant and to trip the gen sets
off-line.  At this point if you don't isolate the problem, it becomes a
domino efffect.  The load generating capacity keeps dropping as the demand
remains the same.  The only option is to isolate the failed grid area, and
then slowly bring the grid back on-line bit by bit.


At 09:04 AM 12/9/98 -0500, Jon Zeeff wrote:
>> breakers on the substation.  It also caused a cascade effect, tripping
>> the Hunters Point and Potrero power plants.  
>> grid primary power sources.  Aside from the human error, the system worked
>> "as designed to protect the rest of the grid."
>I suggest that causing a cascade effect and increasing the area of an
>outage isn't a good way to design a system.
Chris A. Icide / chris at skycache.com
VP Engineering/Operations
SkyCache / www.skycache.com
(v) 301-598-0500 x2235

More information about the NANOG mailing list