|
|
|
Illustration: Charlie Ward
°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°
If the Sun spits, the Earth fries. Humankind is ill-prepared for the furious climax of the next solar cycle.
David Appell reports
BATTEN down
the hatches, there's a storm coming.Some time
in
the next 18 months,
the Sun will turn from a relatively placid
ball of hot
ionised gases into a raging tempest of
plasma,
spitting
fireballs out into the Solar System
like an angry
god. Woe betide any
planet that gets in its way.
Should one of
those plasma storms hit Earth, the impact
could
be
devastating. Each fireball--known
as a coronal mass
ejection--is a giant
maelstrom of ionised
gases at
temperatures
of well over a
million degrees. But the
temperature is
the least of Earth's worries. The plasma
will
tear through
the Earth's magnetic field like wind
through
grass. These
wildly fluctuating fields can knock out
power
supplies, and
charged particles from the plasma can fry
the
electronic
components inside telecommunications
satellites,
bringing down
communications networks over vast areas. A
few
scientists and
engineers are preparing for the worst
while
others, strangely,
have chosen to ignore the problem. The wary
few are racing to
put in place measures to protect power grids
and
telecommunications networks, and have launched
sentinels
that sit in
space between the Earth and the Sun watching
for
storm signs. In
addition, they are developing complex computer
models to predict
which parts of the Earth might be
affected.
Others, fearing the
worst, are waiting to see what happens
to
the giant
communications networks that have grown up since
the
last big solar storm
10 years ago.
Scientists
have been watching the changing nature of the
Sun
for over 200
years and have witnessed these solar rages
every
eleven years or so.
This will be the 23rd cycle on record
and
researchers believe
it will be every bit as bad as the
last.
Six million
people in the Canadian province of
Quebec can
testify to its
effects.
The storm
struck in the early hours of 13 March 1989. It
was
not a good night to
lose power. The temperature had dropped to
-15 °C
and furnaces went quiet as six million Canadians
lost
heat and
light. After the winter sunrise, subways sat
still
for lack of
power, traffic lights hung dark and petrol
pumps
refused to
deliver.
Later
in the day, when public
officials called for an
explanation,
engineers at Hydro-Quebec, the region's
power
generating
company, had begun to suspect an unusual
culprit.
Four days earlier, a
giant bubble of plasma had burst from the
surface of
Sun. That morning it had hit the Earth,
wreaking
havoc.
Rapidly
changing magnetic fields generate
currents in any
conductors
within reach. This is how a dynamo
works--except
that the
magnetic field remains still in these devices
while
the conducting
wires move through it. When a magnetic
storm
hits the
Earth, any networks of conductors that stretch
over
the same scale
as the magnetic fluctuations act like
giant
dynamos.
Hydro-Quebec's transmission lines stretch
for over
1000 kilometres.
Power lines, telephone lines and even railway
lines are all
potential conduits for "geomagnetically
induced
currents" (GICs) of
hundreds of amperes.
Power companies
are vulnerable because their power lines guide
the GICs
towards sensitive components such as transformers
at
power stations
and substations. A transformer changes a
high
voltage
supply of alternating current into
a low voltage
supply or vice
versa. It consists of a giant doughnut of
iron
with two sets
of windings on each side of the structure.
The
voltage in one set
of windings induces a magnetic field in the
iron core,
which in turn induces a voltage in the second
set
of windings.
The ratio of the number of windings in the
two
coils determines the
change in voltage.
High-performance
transformers are delicate machines. They
are
designed to
cope with voltages within a specific
range of
amplitudes
and frequencies. Outside
these bounds, the
transformer behaves
unpredictably.
The trouble
with GICs is that the voltages associated
with
them change
this delicate balance. In particular, they set
up
voltages at
harmonic frequencies to the ordinary load.
These
frequencies
are transformed but in a way
that can rapidly
spiral
out of control. The result is
wildly fluctuating
voltages called
voltage asymmetries. If the power is not
shut
down, these
can create enough heat to damage the iron
core
beyond repair.
Worse, these fluctuations pass rapidly
through
the
network so that neighbouring
transformers also become
affected. Within
seconds an entire network can collapse as one
transformer after
another fails.
Exactly
this happened to Hydro-Quebec's
power system that
fateful
morning. "Voltage regulations need to be within
5 to
10 per cent of a
nominal value. If you fall outside that,
you
generally see
a system collapse and the start of a
domino
effect," says
John Kappenman, an expert in the
effects of
geomagnetic
storms at the Metatech Corporation,
based in
Goleta,
California.
Many
other electricity utilities
around the world also
suffered the
effects of GICs that morning. Further south,
the
iron core of a
transformer at a New Jersey power station
burnt
out and
had to be replaced at a cost
of several million
dollars.
Later, researchers at
the Oak Ridge National
Laboratory in
Tennessee predicted the potential effects of
a
geomagnetic
storm only slightly more severe than the
one in
1989. They
concluded that the ensuing blackouts
and chaos
could cost the
US economy up to $6 billion dollars in
lost
business.
Astronomers
are forecasting storms just as big
as those in
1989 for the
next solar maximum, if not bigger. As the
Sun
passes through
its 11-year cycle, solar astronomers
measure
the activity on its
surface by counting the number of sunspots
and the number
of groups of sunspots they can see during
a
predetermined
period, usually a month or a year.
Together,
these
numbers allow them to calculate
an index of solar
activity known as
the International Sunspot Number. During the
solar minimum, the
sunspot number can be as low as 10. In July
1989, during the
last solar maximum, it peaked at 159. And
in
March 1958, it
reached 201, the highest level ever
recorded
(see Figure).
Cycle 23 "will be
one of the largest on record, and comparable
to the last
two solar cycles", says a panel of
international
experts chaired
by Jo Ann Joselyn of the US National
Oceanic
and
Atmospheric Administration's Space Environment
Center in
Boulder,
Colorado. They warn that the sunspot
number could
reach 190, peaking
sometime between June this year and January
2001.
Can
anything be done to avert disaster?
Leonard Bolduc, a
researcher
at Hydro-Quebec's Institute
of Research in
Electricity of
Quebec, who was working on the night of
the
failure, has
studied the network's breakdown. There is
little
that Hydro-Quebec
can do to prevent GICs. Instead, Bolduc says
the company's
strategy is to design grids
that can cope.
"Hydro-Quebec
has spent a lot of money trying to
understand
the phenomena
and to evaluate all its equipment during a
GIC
storm," he
says.
Its solution has
been to fit its power lines with
capacitors,
known as
transmission line series capacitors, that prevent
the
flow of direct
current without affecting alternating current.
The company has
spent more than C$1.2 billion fitting the
new
capacitors. It has
also set up monitoring equipment that spots
voltage
asymmetries and warns operators to
redistribute the
load to other
parts of the network, by bringing online
other
generators in
different areas. "We are confident
that our
network could now
support such a big storm," says Bolduc.
Currents and electrojets
Another
approach is to predict the severity
of geomagnetic
storms before they
hit the Earth so that preventive action can
be taken. But
this isn't easy. The interaction between
the
Earth's magnetic
field and the particles in the hot plasma
is
extremely
complex. Ari Viljanen and Risto
Pirjola of the
Finnish
Meteorological Institute have been
studying this
process.
They began by modelling the
interaction between
plasma from
the Sun and the Earth's magnetic field, and
the
way this
generates currents in ionised regions of the
Earth's
upper
atmosphere. These
currents--called the
auroral
electrojet--have
an electric field associated with them.
The
horizontal
component of this field at the
Earth's surface
together with
the conductance of the surface are the
crucial
factors that
determine the strength of GICs. So Viljanen
and
Pirjola have had to
model the conductivity too.
By
combining their models of the auroral
electrojet and the
conductivity
of the Earth's surface, they
have created a
formidable
tool. Their overall model produces data that
come
within 20
per cent of the values of GICs
measured in the
Finnish power
system.
The Finnish
researchers now want to turn the
model on its
head. They say that
by using data from GICs in Finland,
their
model can
throw light on the processes at work in the
upper
atmosphere. In
effect, they hope to turn the entire
Finnish
power grid
system into a giant instrument for
studying the
interaction between
the magnetosphere and the solar wind.
Kappenman also
has ambitious plans. He is division manager
for
Metatech's Applied
Power Solutions Division, and the architect
of a new
computer model called PowerCast designed to
predict
the effects of
GICs before they occur. His first customer
is
National
Grid--the company that operates
Britain's power
transmission
network--which is installing his
system this
month.
PowerCast
uses a model of the
interaction between solar
plasma,
the magnetosphere and the geology
of the Earth's
surface. But
it links this to a model of
the power grid
itself. The
National Grid uses about 900
transformers and
PowerCast
takes into account all of them when deciding
which
might be damaged by
impending magnetic storms.
The data that
PowerCast uses to make its predictions come
from
a small
observatory called the Advanced Composition
Explorer
(ACE), a
spacecraft that sits in the solar wind
approximately
1.5 million
kilometres upstream of Earth at the point
where
the
gravitational forces from the Earth and
the Sun balance
one another.
ACE measures the composition of the solar
wind
and gives
roughly one hour's warning of an impending
solar
storm.
Using these
data, PowerCast will give the
National Grid a
minute-by-minute
update of the threat that GICs pose to
the
network. Operators
can then take appropriate steps to mitigate
the
effects of any impending storm
while maintaining the
supply. It's a
difficult job, says Kappenman. "Power
companies
are not like phone
companies where, if it gets too busy,
they
can give you a
busy signal." Metatech claims that the
system
works with
"reasonable accuracy". Just how it it will
perform
during the
forthcoming solar maximum remains to be seen.
Satellites
are also at risk during solar
storms. The US
Department
of Defense has estimated that
disruptions to
government
satellites from space weather cost
about $100
million a
year, and that even when the
Sun is relatively
placid, as it
was in 1994 and 1995, about 150
malfunctions
occur
annually.
When the
communications satellite Galaxy IV failed last
May,
it brought
down communications networks and put 45
million
pagers out
of action. The satellite's manufacturer,
Hughes
Electronics
Corporation, says an on-board processor failed
as
a result
of a random event. Others believe
a more likely
culprit is the
Sun.
Killer electrons
Dan Baker,
director of the Laboratory for
Atmospheric and
Space Physics
at the University of Colorado in Boulder,
and
his colleagues
have studied the solar weather conditions
that
existed at the
time of the failure. They found that a
large
number of
high-energy electrons had become
trapped in the
Earth's
magnetosphere in the two-week
period before the
satellite
failed, as a result of exceptionally
stormy solar
weather. According
to Baker, it was "one of the most
intense
periods that we've
seen for the last two or three years".
Electrons
with kinetic energies greater
than a million
electronvolts have
been dubbed "killer electrons". They
smash
through the
skin of spacecraft and lodge inside
dielectric
materials such as
thermal blankets, electronic boards, coaxial
cables and
electrical insulation. If more electrons
arrive
than can leak
away, the buildup of charge can create
strong
electric
fields inside the spacecraft, a process called
deep
dielectric
charging. Eventally, arcing occurs as
electrons
jump between
areas at different potentials. It is these
tiny
bolts of
lightning that destroy spacecrafts' electronics.
"We
don't know for
sure if this caused the Galaxy IV
failure,"
says Baker,
but he says several other spacecraft
also had
problems during the
same period.
And nobody
knows whether networks that have been designed
in
the 10
years since the last solar maximum
will cope. "As
technologies change,
new vulnerabilities to solar events crop
up," says Lou
Lanzerotti, a geophysicist at Bell Labs, the
R&D
arm
of Lucent Technologies.
He points to the
huge
proliferation of
wireless networks. "We find that there are
a
few solar radio
bursts every solar maximum that are larger
in
amplitude at Earth
than the noise level in a cellular system."
Could these
bursts drown wireless networks in a sea of
noise,
putting
millions of cellphones and other
wireless devices
around
the world out of action? "It's
something that we
haven't
thought about before because
we didn't have the
technology
and didn't need to
think about it," says
Lanzerotti.
With the solar maximum approaching, time is
fast
running
out.
David Appell is a science journalist based in Gilford, New Hampshire
Further reading:
* The
Space Weather Bureau at
Marshall Space Sciences
Laboratory, Huntsville, Alabama
---------------------------------------
From New Scientist, 27 February 1999
© Copyright New Scientist, RBI Limited
1999