[sdiy] WAY OT: US power failure - Was: Giga-Ohm Resistors--Where?

[J. Larry Hendry]

----- Original Message -----
From: Magnus Danielson <cfmd at swipnet.se>
> But where do you get the 1000000V power supply to run the circuit?
Ehum... we have a power line expert here... (I still expects him to give us
a insiders geek-tour of the big powerfailure earlier this year...)

Larry Hendry writes:
OK.   I am guessing that might be me.  If not, I'll have a go at it anyway.
Those who consider it too OT, you were warned in the subject line. :)  And,
I warn you again, this is long.

First, some basics, and analogies that make the whole thing somewhat easier
to understand.  And, of course, this is from information I have been able to
gather from the "inside".  The "official" report will no doubt come from
some politician.

The delivery of power between generation and load is a constant state of
balance.  And, it seems that generation and load are never centered in the
same place.  Therefore, transmission lines are utilized to transport power
between generation and load centers.  Since we are talking AC in most cases,
the two components of supply and demand that must be balanced are both true
power and apparent power.  These represent the components of current that
are in phase with voltage and out of phase with voltage respectively.  These
are referred to as WATTS and VARS.   WATTS is a pretty common term most
people understand.  However, VARs (volt-amps-reactive) is not something
thought about too often by people that work mostly in DC and low voltage AC.
The phase angle of the apparent power current will lead or lag the voltage
by exacty 90 degrees, depending on whether the load appears capacitive or
inductive.

Both of the components of power WATTS and VARS must be kept in balance
between supply and demand AND have sufficiently strong connection conponents
between them.  Let's examine the effects of out of balance conditions
between these two components (on a basic level):

In an inter-connected system, generated WATTS must closely equal load
(consumption).  The result of unbalance is a change in the frequency of the
interconnected system.  When load exceeds generation, frequency begins to
decline.  When generation exceeds load, frequency begins to rise.  So,
frequency is not a fixed quantity.  It is normal in the USA / Canadian
eastern interconnected system for frequency to vary from 59.95 Hz to 60.05
Hz.  Variances of a few hundredths of a cycle occur when sudden changes in
load or generation occurs.  Then, automatic generation control systems of
all utilities respond to these changes in frequency by bias of their
interconnected interchange schedule (wholesale transactions between
utilities).  The result is that every utility adds some to generation when
frequency is low, and every utility subtracts some when frequency is high.
These small bais amounts are in the background and work in addition to the
scheduled power transactions.

To understand frequency regulation, think of the cruise control on an
automobile.  The car engine is the generator, the drive shaft is the power
transmission system, and the wheels represent the load (which is of course,
friction and many other components).  As the car travels along a flat road
(constant load) the fuel supply to the engine is constant.  But, when the
car comes to a slight incline, the car begins to slow down (load is
exceeding generation).  The control adds fuel to then engine (increasing
generation) and the cruise speed in maintained (frequency is moved back to
the set point).

Now, lets look at the VAR balance.  That is more diffuclt to regulate and
understand.  Because, most all of the losses on a power system are created
by "L" and not "R."  The basic L and C components of a power system are:
The customers loads (all motors consume VARs)
The utility generators (can appear as "L" or "C" to the system)
The utility transmission system (lines and transformers)
The utility installed capacitors used for correction of too much "L"
The utility installed shunt inductors uset to compensate for too much "C"

The balance of the supply and demand of VARs is considerably more
complicated.  I could write and entire book on this subject alone.  Most all
of the "losses" on a utility system are related to out of phase currents.
But, let's simplify at this point and say when the system gets too
capacitive, voltage rises.  When the system gets too inductive, voltage
declines.  So, in addition to balancing the supply of true power supplied
with load, apparent power must also be balanced.

To help those having trouble with the concept of how apparent power is
represented by actual loads, think of a couple of things.  One, think of a
motor.  In a motor, the WATTS (in phase current) do the work - i.e. supply
torque to the shaft to transfer power to the mechanical load.  The VARS are
currents that are drawn by the motor, 90 degrees lagging the phase angle of
the voltage, that create the electromagnetic field without which the motor
cannot operate.  Now, for many of you, this may be an over-simplified and
incomplete analysis.  But, I am trying to reach the point of this note,
without teaching too much "AC power 101."

The biggest variable in this balancing act is the utilities own transmission
system.  A transmission line, unloaded by WATTS, is very capacitive in it's
character.  The "R" component is low and we will ignore it for this part of
the discussion.  Those wires strung out over miles parallel to each other
make one heck of a capacitor.  When a transmission line is energized from
one end, current begins to flow into this capacitive load.  That current
leads voltage by 90 degrees.  The line is "apparently" producing VARs.
Whether you consider that load or a source depends on your perspective.
When the oppostie end of the transmission line is conected to loads, WATTs
begin to flow on the line.  That current flow creates electromagnetic field
around the wires.  As the power transferred across the line increases, the
intensity of this field increases (Mr." L" is showing his contribution to
this puzzle).  Eventually, the currents consumed by creation of the magtetic
field exceed that of the currents of the capacitive characteristics of the
line and it begines to consume VARs. The L current and C current are of
course 180 degrees out of phase and cancel each other.  So, if I put in 500
MW (megawatts) and 100 MV (megavars) at the source end, I may get 500 MW out
at the receiving end, but the 100 MVARs and possibly more, will be consumed
by the line (creating the field).

The concept I am trying to show here is that when one transmission line is
tripped out of service (due to some fault condition), the power
redistribuion on remaining transmission facilities will have an impact on
the supply and demand of VARs.

Finally, lets look back at our car engine, driveshaft and wheel analogy.
Remember, the driveshaft is the transmission system.  But instead of one
solid link, it is many more flexible links (the transmission grid).  There
is anglular displacement on the shaft.  As the load is increased at one end,
and the torque is increased at the opposite end, this displacement (phase
angle) changes from one end to the other.  So, it is normal on a
transmission system for phase angle to shift along a loaded transmission
line.

Now, if we start removing some of the individual strands of the
transmission, the drive shaft becomes more flexible.  The more flex in the
system, the more difficult it becomes not only to balance supply and demand,
but also to transfer the power from one end to the other and to keep load
and generation in sync.  As components of an electrical transmission path
are removed from service, the same thing happens.  Now, take it to the
extreme, and substitute a rubber band in the place of your driveshaft.  Of
course, it will not work with anything more than the smallest of loads.
Imagine that a continuously variable continuium.

Now, onto the outage:

Let's set up the basic layout.  In Ohio, a large percentage of the load is
in the north (Cleveland and surrounding area).  A large percentage of the
generation is in the south and south west (near the Ohio river).
Boiler-centered generation needs much water for steam condensation.  Indiana
is about the same with generation south and west near the rivers.  In
Michigan, much of the generation is west (lake Michigan).  Much of the load
is east (Detroit).  In New York, much of the load is east (NY City, etc.),
Much of the generation is west (Niagra).

So, if you look at normal steady-state power flows in that part of the
country, power flows basically from south to north through Indiana and Ohio.
Power flows from west to east through Michigan.  Power flows west to east
across New York.  Power flows from Michigan toward Niagra.

Now imagine one more thing.  Large transmission line loading is limited by
conductor operating temperature (that damn R thing).  As conductor
temperature rises, the conductor in the middle of the span begins to sag
(remember that metal expands when heated :) ).  This variance in operating
temperature can change the sag of that line in mid span by several feet in
high voltage lines with 1000' spans.  So without getting into all the bad
things that can happen to a transmission line, let's blame fault # 1 on a
tree (highly probable).  The line loaded toward it's upper capacity by the
transmission of power from point A to B sags into the tree and the line is
faulted and trips out of service.  It could have been another type of
failure.  But, the end result is that protective relays designed to
disconnect faulty equipment from the system have tripped the circuit
breakers and the line is no longer part of the transmission path.  This
action occurs in about 3 cycles.

Instantly, the power transferred from generation to load (neither of which
have changed yet) is redistributed along the remaining transmission lines.
Perhaps that overloads or sags another and it trips out of service.  Our
driveshafts are getting more "rubbery."  something like that happened and
much of the generation in central and southern Ohio became disconnected from
Cleveland area loads.  That power took the path of least resistance up
765,000 volt lines from central Ohio to northern Indiana and into Michigan.
It added to transmission loading in Michigan going toward Detroit trying to
get to Cleveland through northern tie lines.  Can you imagine what has
happened to the VAR balance at this time. These high loadings have upset the
balance and are voltage has declined by 5 to 10%.  Some Michigan
transmission is overloaded or oversagged by this re-distribution of power
flow and it too trips out of service.   When lines trip out of service, much
load is lost.  That is now WAY out of balance.  Frequency skyrockets to 60.2
Hz.  Now, in Michigan and Ohio, the path from generation to load is so
flexible, that the rubber band is flexing too much and they become
disconnected (called pole slipping).  Generators start tripping off line
from protective relays designed to protect the turbines from overspeed.
Generator stability trying to match fuel supply to electrical load is in
complete chaos in many plants and the balance cannot be maintained.  Plants
trip off line.  Sources of transmission and generation in Ohio and Michigan
to Detroit and Cleveland loads is gone.  The only remaining connections to
serve loads in Detroit and Cleveland are from Niagara (which normally feeds
toward Niagara).  The Niagara NY state balance is completely upset and
overloaded trying to carry Michigan and Ohio loads and massive instabilities
eventually lead to a complete collapse.

Now, much is omitted from this basic overview,and I have taken some
liberties with generalization.  And, not all facts are yet uncovered.  But,
that is the basic concept and ends toady's lesson in AC power.  I'm sure I
have lost a few and insulted the intelligence of others.  But, I tried.  So,
go easy on me. :)

Larry Hendry
28 year power system veteran.