Power Grid Experiment

[Mike O'Dell]

stabilizing a large AC grid is "complex" in the 
mathematical sense of the word. the last big
blackout in the northeast US and canada was 
caused by an insufficient amount of "imaginary power"
in the system. the "voltage" on the grid is not
a simple scalar number - it's really a vector
with a real component and a reactive component,
the imaginary part. keeping the real part of the
voltage where it needs to be requires the grid
to generate enough imaginary power  with the
right sign to cancel the reactive power generated
by all the loads with crappy power factors.
if they can't keep the net reactive power
nulled-out, the voltage vector starts to
swing away from lying on the real axis. 
while the magnitude of the vector is still
the same, the real part starts decreasing
and it is the nature of most loads that they
tend to become more reactive when the real
voltage starts dropping away from the design
target. this, of course, exacerbates the situation
and if they don't get a handle on it, it's off
to the races and the phase angle protection relays
at the power stations start disconnecting the plant
from the grid, and the viscious circle just 
spins faster into darkness.

this is referred to as "voltage collapse"

given that this is all about big LC networks,
the reactive power reflected from the loads
or created by the generators is sensitive to
frequency. it's much harder to keep the reactive
component under control if you have to keep the
frequency within a few hundreths of a Hertz while
you're doing it. by allowing the frequency to shift
you can take advantage of what amounts to the
Q of the grid, moving further on the skirts
around the resonance.

it's really about giving them more freedom to
keep things balanced.

with the possible exception of the Texas Power Pool
(whatever it's called now), the other three big pools
*really* need to be subdivided and interconnected with
DC interties. it's easy to make the power flow the direction
you want with DC and bloody hard to do it with AC.
by breaking up the big grids into smallerr ones,
each one will be more stable and the macro management
can be far more effective.

for frequency stabilization purposes, the power contracts
get rebid every 15 seconds, and they'd like to take that
down to 5 and then 1. this is one possible use for grid-scale
storage systems (think biiiiiig batteries). in theory,
they should respond much faster than throttling a steam 
plant or even a gas-turbine peaker. there is also a 5 minute
power market, a 15 minute market, and a one-hour market.
these contracts are traded in real-time, all the time.

for example, since you never really know when people are 
going to get up and all go hit the ON button on the damn
1KW coffee maker, the state of California has everything
they have availble up and spinning by 5am. until the
coffee makers come on, though, they have a HUGE excess
of power, and the wholesale cost per kilowatt actually
goes NEGATIVE - the California system operator will
pay other systems to take power with the proviso that
they may need to pull back quite quickly. later in
the warm summer day, CSO will be buying peaking power
on the spot market for at least as much as they were
paying to get people to take it off their hands.
if you own some biiiiiiiiiig batteries, there is an
immense arbitrage opportunity here. take the KW from
them in the morning and get paid for it, and then
later in the day sell the KW back to them and get
paid for them again. this is a great deal! it does,
however, an immense capital expense to build the
battery systems. at the moment, the only grid-scale
batteries that are deployed even on a test basis
in the US are liquid-sodium-liquid-sulphur batteries
(known as "NaS" for the chemical symbols). they
are impressive - they run at about 300 degrees C.
a battery system that can supply 1 megawatt for
a 7 hour window (7MWH), with the inverter-chargers
and managment stuff, is about 75 feet long, 20
feet tall, and about 15 feet thick. it tips the
scales at a mere 400 tons. 

as you can imagine, when the general populace 
finds out about "liquid sodium" and "liquid sulphur",
what is eupehmistically refered to as "permitting challenges"
ensue.

oh yeah - these batteries are built by NHK and were
developed for TEC - Tokyo Electric Company.
Nuclear plans essentially have no usable fast-acting
throttle because of the immense thermal mass
so a lot of their regulation is done with the batteries.

   -mo