This is only a preview of the September 1991 issue of Silicon Chip. You can view 44 of the 96 pages in the full issue, including the advertisments. For full access, purchase the issue for $10.00 or subscribe for access to the latest issues. Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
|
Each day of eve.zy year, there are an average
44,000 thunderstorms and 8 million lightning
flashes. A big storm will have many
thousands of lightning strikes and these can
cause great damage. Hence, it is important to
be able to plot the course of thunderstorms.
Plotting the course of
As you read this article, a lightning
monitoring system for the states of
New South Wales and Victoria is in
the process of being set up, in time for
the next "lightning season" in the coming summer. Lightning monitoring re-
By LEO SIMPSON
TrT2 = a
•
R2
Fig.1: LPATS receivers record the
precise time that they detect a
lightning strike. When two receivers
detect a lightning strike
simultaneously, the time difference
between the two is zero and the
lightning can be assumed to be
anywhere on a straight line
equidistant from the two receivers.
When receiver 1 detects the strike
before receiver 2, the strike has
occurred on a hyperbola which passes
around receiver 1.
12
SrucoN CHIP
ceivers placed up to 500km apart will
be able to plot each lightning strike,
virtually as it happens, with an accuracy of as little as 200 metres.
The information provided by the
lightning monitoring system will be
of great importance to government
bodies such as the various electricity
commissions, Telecom, state railway
authorities and defence establishments, as well oil refineries, large
chemical plants and organisers of
sporting events where large crowds of
people are expected.
There is already some monitoring
of thunderstorms being done in Australia apart from that provided by
weather radar. For example, the Northern Territory Power and Water Authority uses the Stormscope system to
monitor the progress of thunderstorms
in the Katherine region. If a large storm
threatens the area and the 132kV transmission line in particular, additional
gas turbines in the Katherine power
station are brought on line, so that an
interruption to power transmission
from Darwin will not cause blackouts .
And dui;ing the particularly damaging thunderstorm which hit Sydney last summer, Prospect Electricity
(previously known as Prospect County
Council) had prior warning of the
storm's extent from a direction finding storm location system and thus
had alerted all its line crews. Sydney
Electricity (previously known as Sydney County Council) did not have
this warning and thus did not learn
the full extent of the storm damage
until many hours later.
The system to be installed for monitoring NSW and Victoria is a great
deal more involved than the
Stormscope system which works in
conjunction with weather radar. The
problem with weather radar is that
while it can give a good indication of
cumulo nimbus clouds and heavy
rain, it does not detect lightning.
Stormscope does, but not with any
great accuracy.
The new system is known as LPATS,
which stands for Lightning Positioning & Tracking System. It is a Time of
Arrival (TOA) system whereby the
time when a lightning stroke is detected at a number of remote radio
receivers is precisely recorded. Then,
with the position of each the remote
receiver being fixed and known, the
position of the lightning strike can be
calculated.
Basic principle
The principle of the Time of Arrival system is illustrated in Fig.1.
Here we see two radio receivers which
are lo,cated a considerable distance
apart which may be up to 500 kilometres. Now consider a lightning strike
which is recorded at exactly the same
time by the two receivers. A moment's
thought will reveal that the strike must
have been somewhere on a straight
line equidistant between the two receivers. This line is depicted in Fig.1
as (Ti-Tzl = O.
Now consider another lightning
stroke which is somewhat closer to
receiver 1 than to receiver 2. Receiver
1 will detect the stroke at a time before receiver 2. Again, by a similar
process of deduction, the lightning
stroke must have occurred somewhere
along a curved line shown as (T 1 -T 2 )
= a. Other lightning strikes can be
shown to have occurred anywhere
along a hyperbolic line which circles
receiver 1 or receiver 2.
Now if we add another receiver as
shown in Fig.2, we get more information about the possible location of a
lightning stroke. Receivers 1 and 2
give a "time difference line" of T1 -T 2
while receivers 2 and 3 give a time
difference line of T 2 -T 3 . Receivers 1
and 3 give a third time difference
line. The location where all three intersect is the position of the lightning
strike. Or is it? In fact, there are some
thunderstorms
situations where three receivers are
not enough to give a clear result so
four receivers is the practical minimum in an LPATS network. And in
practice, to give a degree of redundancy, five or six receivers are used.
Timing
In order to give precise location of
lightning strikes, all of the receivers
in the system must have the same
very precise time reference. In the
USA, the LORAN navigational system
has been used but now the Navstar
Global Positioning Satellite (GPS) system is the preferred reference. LPATS
makes use of the civilian access standard positioning service of GPS which
has an accuracy of up to 100 metres.
This is used to establish the location
of the receivers at installation. After
that, in normal operation, the GPS
signal is used to continually synchronise the 10MHz time clock.
The receiver itself uses a simple
whip antenna to pick up the lightning
signal. An enclosed helix antenna is
used to receive the GPS satellite timing signals. The detection receiver has
a bandwidth of 2kHz to 500kHz and
apparently uses an AM detector although the manufacturers, Atmospheric Research Systems, Inc. are coy
about giving any details. However,
the detection process is good enough
to produce a good approximation of
the waveform of the lightning strike.
T1-T2 • a
Lightning and thunderstorms can do tremendous damage in Australia. With a
precise system for plotting thunderstorms and lightning strikes, the hazards can
be minimised and any damage more quickly repaired.
Fig.2: with three receivers, three "time
difference" hyperbolas can be plotted
although only two are shown here.
'l\vo hyperbolas will intersect at two
points or (rarely) touch at one point.
Three hyperbolas are needed to plot a
single unambiguous location for every
lightning strike.
SEPTEMBER1991
13
Fig.3: how lightning is located by
time of arrival:
(1). The signal will be detected at
each receiver at a different time
relative to the event, dependent
on the distance from the event.
(2). Time is measured at each
site with a resolution of 100
nanoseconds (±50 nanoseconds).
(3). Each receiver has a 10MHz
timebase which is typically
synchronised 20 times a second
from a precise source such as
the Navstar GPS satellites.
(4). A minimum of three
receivers are required for a
solution. Achievable accuracy is
1 microsecond and within 200
metres, dependent on the size of
the network.
This is fed to an 8-bit analog to digital
converter with a 200 nanosecond sample rate. The waveform is then stored
in memory with 100 microseconds of
storage.
Two characteristics of the strike
waveform are important - the peak
current of the strike and its risetime.
The exact peak of the waveform is
crucial because that is used tu define
the time of the strike. If a preset signal
threshold was used to define the time
of strike, there would be timing errors
because of the large range of magnitude of lightning strikes - they can
range from a peak current of less than
1000 amps to more than 100,000 amps
and they can have a duration of 500
milliseconds.
Since the peak of the lightning strike
is timed with an accuracy of ±0.1
microseconds, any inaccuracies which
could occur due to the differing
magnitudes of lightning strikes are
eliminated.
The digitising and storage process
also allows other information such as
the stroke polarity and total stroke
energy to be determined. All this information about the time of the strike
and its amplitude is sent by a serial
14
SILICON CHIP
data link such as a phone or radio
modem to a central computer which
calculates the exact location of the
strike.
This information is stored for later
analysis and is also available for immediate display on area maps by the
users of the LPATS service. They can
plot the progress of storms as they
develop and, with experience, they
can predict where they are heading
and the likely amount of lightning
damage.
If necessary, vital equipment can be
shut down or otherwise protected,
sporting events can be cancelled,
crowds evacuated from golf courses
and so on.
Noise rejection
You might wonder how a receiver
with a bandwidth of 2kHz to 500kHz
would be able to discriminate between
local radio interference noise and a
lightning strike. After all, a lightning
strike which may be 200km or more
away from the receiver will not produce a very strong signal. Local radio
interference can easily be much
stronger.
The answer is that the detection
receiver really does not have to perform the discrimination process. Why?
Because the only naturally occurring
electromagnetic event that can be
simultaneously detected by four or
more LPATS receivers which are many
hundreds of kilometres apart is a lightning strike. Hence, if less than four
receivers in an LPATS system detect
an electromagnetic discharge, it is not
recorded as a lightning discharge by
the central computer. This method
has led to a high degree of detection
accuracy.
Naturally, each receiver needs to be
sited away from strong sources of radio interference but apart from that,
the installations are quite uncritical.
In fact, if a receiver site does become
noisy, its threshold of detection is
automatically adjusted, under software ~ontrol.
Accuracy
In practice, the LPATS system can
locate lightning strikes to within 200
metres at the centre of the network
(depending on its overall size), ranging out to a kilometre for strikes well
outside the region covered by the reC!')ivers. You might wonder if the ac-
curacy could be improved, down to
say 50 metres or less. In practice, the
answer is no. For a start, the Navstar
GPS enables positioning only within
100 metres; although the military capability of GPS enables targets to be
located to within less than 10 metres!
Second, there are inevitable errors,
both random and systematic, which
add up to give the ultimate positional
accuracy for lightning strikes of within
200 metres.
But there is a third reason why lightning strikes cannot be located with
better accuracy and that has to do
with the path of the strike itself. This
is usually several thousand metres
long and is rarely over a straight vertical path. So while•LPATS could perhaps locate the centre of a discharge
to within better than 200 metres, the
exact point where it hit the ground
would still be unpredictable.
In practice, where a lightning strike
causes substantial damage, it will usually be fairly easy to locate the exact
point, once the LPATS system has
done its work. In practice too, the
users of the information provided by
an LPATS network will know precisely where any damage prone installation is, given the locality of a
lightning strike.
Suffice to say that the information
on lightning strikes and thunderstorms from LPATS is far more precise than from any other lightning
detection system previously developed, especially those based on direction finding antennas.
A very good example of the efficacy
of an LPATS system was given during
a thunderstorm in the USA, on 13th
June 1991. This took place during the
1991 US Open Golf Tournament at
the Hazeltine National Golf Club. A
spectator was killed by lightning during this storm and several people were
injured.
A subsequent inquiry into this tragedy was able to obtain archived data
which showed the initial development
of the storm, its path and even the
strike which killed the player. Had
the event's organisers had access to
this information during the storm, it
is likely that no-one would have been
hurt.
Acknowledgement
Our thanks to Ken Ticehurst of
Kattron Pty Ltd, Ourimbah NSW and
to Dr Rodney Bent of Atmospheric
40.273N
9'J.675W
132 mi
170°
_
10381 Strokes in 1:07
18:21
USA NATIONAL LIGHTNING DATA
-
□ 12Clx120
Magnified
Magnified
-
Normal
Comm Connected
Fig.4: this is a screen display from an LPATS network covering the United
States. There is a facility to zoom in on thunderstorms & lightning strikes are
plotted virtually as they occur. The system can locate lightning strikes to within
200 metres at the centre of the network, depending on its overall size.
11 5. 0 0 119 .00 12J.00 127 . 00 1J 1 .00 1J5.00 1J9,00 14J.00 147 .00
*
-12.00
- 16 . 00
-12.00
Darwin
~1.0~
1.0
1.0
-20.00
*~
Da~ r
w
C
::::,
Ij::
<
...I
-24 . 00
-28 . 00
- J 2.00
*Alice
l. 0
~
~~rt
1.5
~
1.0_1.o
-28.00
_J
-J6.00
*
-20.00
-24.00
1.0
Springs
1.0
L~
-16.00
-]2 .00
Adelaide
-]6.00
- 40. 00 ~ ~ ~ ~~ ~ ~ ' - - ' - - ' - ~ - ' - - ~ ~ ~ ' - - ' - - ' - ~ . , . , _ ~ L . . . . L . ~' - ' - - ' - ~ ~ -40. 00
115. 00 11 9. 00 123.00 127 . 00 1]1 . 00 135 . 0 0 139.00 14J.00 147.00
LONGITUDE
Fig.5: this is a plot of location accuracy for an LPATS network covering the
whole of Australia. In this notional system, lightning receivers are located at
Dampier, Darwin, Townsville, Adelaide, Perth and Alice Springs. Note that in
spite of the huge distance between the receivers, there is a large area for which
lightning strikes could be pinpointed to within a kilometre or better.
Research Systems Inc for their assistance in preparing this article. Thanks
also to Michael Nott of the Northern
Territory Power and Water Authority
for information on their Stormscope
warning system.
SC
SEPTEMBER
1991
15
|