This is only a preview of the September 2008 issue of Silicon Chip. You can view 26 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. Items relevant to "Railpower Model Train Controller":
Items relevant to "Versatile LED/Lamp Flasher":
Items relevant to "Ultra-LD Mk.2 200W Power Amplifier Module, Pt.2":
Items relevant to "DSP Musicolour Light Show; Pt.4":
Items relevant to "LED Strobe & Contactless Tachometer, Pt.2":
Purchase a printed copy of this issue for $10.00. |
Basslink
Basslink
the high voltage DC power link
between Victoria & Tasmania –
the longest undersea power
link in the world
by Michael Goebel
Back in February 2000 a tender was granted for building
a high voltage DC power link between Victoria and
Tasmania. It was finally commissioned in April 2006 and
now feeds power in both directions between Victoria and
Tasmania, depending on demand in those states.
siliconchip.com.au
September 2008 13
V
ictoria and Tasmania have different power supply
problems. Victoria has Australia’s most inefficient
thermal power stations, burning brown coal – Loy
Yang power station and mine is shown above. So it makes
sense to use Tasmania’s hydo-electricity when it is available, especially during peak power periods.
On the other hand, when Tasmania suffers droughts they
often do not have enough hydro-electric power capacity
to meet their own demand. When that happens, Tasmania
can draw power from Victoria’s grid.
But there is a problem – the 300km-wide stretch of
often-wild water between the two called Bass Strait. And
connecting them is not quite as easy as erecting a couple
of giant pylons and stringing some cable over the water.
Enter Basslink
Basslink is a 400kV DC bi-directional, undersea electricity interconnector, rated to transmit 500MW continuously
and 630MW peak from Tasmania. The link can operate at
up to 600MW for up to 10 hours, providing that it is ‘precooled’ (six hours at no more than 300MW). In the other
direction, up to 480MW can be sent to the relatively small
Tasmanian grid.
Basslink consists of:
• a 290km-long submarine power cable, the longest of
its type in the world, from McGaurans Beach near Giffard
in Victoria’s Gippsland, to Four Mile Bluff, above George
Town on Tasmania’s north coast.
• a 60.8km overhead power line to the Victorian coast.
14 Silicon Chip
• a 6.6km underground cable in Victoria.
• an 11km overhead line section to the Tasmanian
coast.
• a 1.7km underground cable in Tasmania.
The high voltage cable used for the Basslink is 15cm
thick. The undersea cable alone weighs 17,400 tonnes.
From McGauran Beach it runs for a few kilometres as an
underground cable and finally emerges as an above- ground
line running 70km to Loy Yang power station.
There, the high voltage DC from Tasmania is converted
into AC with the help of thyristor “converter valves” so it
can be fed into the 3-phase power system. These valves have
nothing to do with thermionics or vacuum tubes but are
entirely solid state. The thyristors are made of pure mono
crystalline silicon and are effectively used as switches.
When Victorian power is being fed to Tasmania, a similar
station in Tasmania transforms the DC into AC. So power
transmission can be made in both directions, with similar
conversion equipment for AC to DC and DC to AC at each
end.
Why DC?
In an AC distribution system, voltage conversion is
simple – just use a transformer. However, above certain
power levels and over long distances, for submarine cables, high voltage AC transmission links have significant
disadvantages compared to DC, despite DC having to be
converted twice.
For a long transmission path, the smaller losses and
siliconchip.com.au
reduced construction cost of a DC line can offset the additional cost of converter stations at each end of the line.
Also, at high AC voltages significant amounts of energy
are lost due to corona discharge, the capacitance between
phases or in the case of buried cables, between phases and
the soil or water in which the cable is buried.
Long undersea cables have a high capacitance. While
this has minimal effect for DC transmission, the current
required to charge and discharge the capacitance of the
cable causes additional I2R power losses when the cable
is carrying AC. In addition, some AC power is lost in the
dielectric.
HVDC can carry more power per conductor, because for a
given power rating the constant voltage in a DC line is lower
than the peak voltage in an AC line. This voltage determines
the insulation thickness and conductor spacing.
Therefore DC transmission line corridors can be used to
carry more power into an area of high power consumption,
which can again lower costs.
Another advantage of HVDC systems is that they require
only two cables as opposed to the three needed for three
phase current transmission. As a result, an HVDC overhead
line also requires less space.
So the key advantages of a DC link over an AC link
are:
• It allows power transmission between AC networks
with different frequencies, or networks which cannot be
synchronised. A good example of this is Japan which has
50Hz and 60Hz power grids.
• Inductive and capacitive parameters do not limit the
transmission capacity or the maximum length of a DC
overhead line or cable.
• In addition, the full conductor cross section is utilised
because there is no “skin effect” at DC.
Thyristor “valves”
The outdoor valves for earlier systems were designed
with oil-immersed thyristors with parallel/series connection of thyristors and an electromagnetic firing system.
Further development went via air-insulated, air-cooled
valves to the air-insulated, water cooled design, which is
the state-of-the-art in HVDC valve design.
The development of thyristors with higher current
and voltage ratings has eliminated the need for parallel
connection and reduced the number of series-connected
thyristors per valve.
Light triggered thyristors
It has long been known that injecting photons into the
gate instead of electrons can turn on thyristors. See Fig 5.
The route of the predominantlyundersea HVDC interconnect between
the Loy Yang power station in
Victoria’s Gippsland and the Tasmanian
hydro grid connection at George Town, near Launceston in
northern Tasmania.
This is known as LTT (Light triggered thyristor) technology
and reduces the number of components in the thyristor
control by up to 80%.
This simplification results in increased reliability and
availability of the transmission system. With LTT technology, the gating light pulse is transmitted via fibre-optic
cable (itself providing high isolation), through the thyristor housing directly to the thyristor wafer. Therefore no
elaborate electronics, including auxiliary supplies, are
needed to control the high potential.
Innovations in almost every other area of HVDC have
been constantly adding to the reliability of this technology
with economic benefits for users throughout the world.
Light Pipe
Cu
Si
Cu
Mo
A graphical diagram of one of the Siemens laser-controlled
LTT thyristors which make up the “valve” used in the AC/
DC and DC/AC conversion in Basslink, with an exploded
photo of the device at right.
siliconchip.com.au
September 2008 15
The high-performance thyristors installed in HVDC
plants today typically have silicon wafer diameters of up
to 125mm (6 inches), blocking voltages up to 8kV and current carrying capacity up to 4kA DC.
Thus no parallel thyristors need to be installed, however, series connection is necessary to handle the HVDC
voltages.
The required optical gate power is just 10mW. The forward overvoltage protection is integrated into the wafer.
Further benefits of direct light triggering are the unlimited
black start capability and operation during system undervoltage or system faults without traditional limitations.
In the case of convential (electrically triggered) thyristors (ETT), this is only possible if sufficient firing energy
is stored long enough in the thyristor electronics.
A simple voltage divider circuit made from standard
off-the-shelf resistors and capacitors allows monitoring of
the thyristors performance.
Monitoring signals are transmitted at very-much-safer
ground potential through another dedicated set of fibre
optic cables as for the LTT.
All electronic circuits needed for the evaluation of performance are now located at ground potential in a protected
environment, further simplifying the system.
The extent of monitoring is the same as for the ETT. It
is expected that this technology will become the industry
standard in HVDC thyristor valves of the 21st century,
paving the way towards maintenance-free thyristor valves.
Laser control
The thyristors in the Basslink project are controlled by
10mW laser flashes, transmitted via glass fibres. These thyristors, which have a diameter of 100mm, were produced by
Infineon and are made of silicon, molybdenum and copper.
To achieve a DC voltage of 400kV, several dozen thyristors
per converter valve are connected in series.
All of these thyristors must trigger within 1µs in order
to ensure that none are overloaded or damaged.
Valve Design
The modular concept of the Siemens thyristor valves
permits different mechanical setups to best suit each application: single, double, quadruple valves or complete
six-pulse bridges, either free standing or suspended from
the building structure.
The standard Siemens valves for long distance transmission are suspended from the ceiling of the valve hall to
allow them to withstand earthquakes, especially important
in areas prone to seismic activity such as Japan and New
Zealand.
The suspension insulators are designed to carry the
weight and additional loads originating for example from
The old and the new: at left is a bank of six mercury-arc rectifiers for a 100kV, 140A high-voltage DC supply from around
1942. At right are the HVDC valve towers in the 3000MW. ±500kV, 1000km HVDC transmission link between Guizhou and
Guangdong in China. Basslink is a smaller distance but is predominantly under water, which brings in a whole new set of
problems to overcome (photo courtesy Siemens).
16 Silicon Chip
siliconchip.com.au
an unbalanced weight distribution due to insulator failure,
an earthquake or during maintenance.
Connections between modules (piping of cooling circuit, fibre optic ducts, buswork and suspension insulator
fixtures) are flexible in order to allow stress-free deflections of the modules inside an MVU (multiple valve unit)
structure.
Each valve is made up of three modules. Four arresters,
each related to one valve, are located on one side of the
valve tower. Ease of access for maintenance purposes, if
required, is another benefit of the Siemens valve design.
By varying the number of thyristors per module and the
number of modules per valve, the same design can be used
for all transmission voltages that may be required.
Thyristor cooling
The thyristors are stacked in the module with a heatsink
on either side. The water connection to the heatsinks can
be designed in parallel or series.
The parallel cooling circuit provides all thyristors with
the same water temperature. This allows a better utilisation of the thyristor capability and offers the additional
advantage that electrolytic currents through the heatsinks
– the cause for electrolytic corrosion – can be avoided by
placing grading electrodes at strategic locations in the
water circuit.
The parallel wafer cooling principle has been in use
for more than 25 years, with no corrosion problems ever
encountered. Water cooling also does not require any deoxygenising equipment.
Fire!
With such enormous power involved, arcing and fire is
a constant risk. This has been minimised through many
steps:
• All oil has been eliminated from the valve and its
components. Snubber capacitors and grading capacitors
use SF6 as a replacement for impregnating oil.
• Only flame-retardant and self-extinguishing plastic
materials are used.
• A wide separation between the modular units ensures
that any local overheating will not affect neighbouring units.
• Careful design of the electrical connections avoids
loose contacts.
The past has shown that Siemens HVDC installations
have never been exposed to a hazardous fire risk. The
tests performed on actual components and samples in the
actual configuration as used in the valve indicate that the
improved design indeed is flame retardant and the risk
of a major fire following a fault is extremely low or even
non existent.
REPLACE
One end of BassLink – the 400kV DC valve hall near the Loy Yang power station in Victoria’s Gippsland. The valves hang
from the ceiling, predominantly to allow them to move, protecting them from damage in case of earthquake.
siliconchip.com.au
September 2008 17
High Voltage DC: how it works
As noted elsewhere in this article,
the basic element in all HVDC conversion is the light triggered thyristor
(LTT) or SCR (silicon controlled rectifier). These are connected in series
banks of a hundred or more SCRs, to
give a total rating which may be 500kV
DC or more and hundreds of Amps.
When connected in such a way, they
are referred to as “valves” and each
valve can be regarded as a single
device. In reality, to function in this
way, all the SCRs in a valve must be
triggered on within one microsecond
of each other. This is achieved by a
laser light pulse fed to the every SCR
in the valve.
Once an SCR is triggered, it breaks
into conduction and it continues to
conduct until the load current falls
to a very low value (below the “holding” current) or the voltage across it
is reversed in polarity. So in effect,
once an SCR or valve is triggered
into conduction, it behaves just like
a conventional diode, albeit a very
large diode!
So how are these banks of LTTs,
or valves, connected to perform the
conversion from AC to DC or DC to
AC? And how does the same setup
perform either AC/DC or DC/AC without altering the connections?
To keep it simple, let’s first consider
the conversion of AC to DC. Consider
that all power grids are 3-phase systems, with the difference between
respective AC phases being 120°.
Fig.1 shows a 3-phase transformer
connected to a 6-SCR bridge rectifier.
Incidentally, this is exactly the same
connection as used in a standard
car alternator (it has six large diodes
pressed into its casing.)
Each diode (or LTT) conducts when
+
+DC
THREE-PHASE TRANSFORMER (STAR/STAR)
φ1
RED
R
φ3
φ2
BLUE
W
WHITE
0V
PHASE
2
18 Silicon Chip
D3
B
it has a forward voltage across it and so
each diode conducts for a maximum of
60°. The result is a DC voltage with a
superimposed 300Hz AC ripple (for a
50Hz grid), as shown in Fig.2.
At any one time, only two diodes will
be conducting, for example, D1 & D5 or
D6, D2 & D4 or D6 and so on. The result
is that there are six diode-conduction
periods providing the AC to DC conversion which is more familiarly known as
rectification.
Exactly the same setup can be used
to convert the DC back to AC except that
the power flow is in the other direction.
In DC/AC conversion, only two SCRs
conduct at any one time (as in rectification) but they need to be triggered into
conduction at the right time to energise
the particular transformer windings.
However, the resultant waveforms on
the output side of 3-phase transformers are anything but a clean sine wave.
Hence quite heavy filtering is required
in order to prevent large harmonics in
the distribution grid.
PHASE
3
T
–
D2
Fig.1: AC to DC conversion in an HVDC system is D4
D6
D5
essentially the same as the rectification process
0V
in a 3-phase power system, whether it is the alternator of a car or in a country’s electric power grid. Six diodes (or SCRs) are
required and two diodes will be conducting at any one time. Note that while
we show a star-connected transformer in this case, it could just as easily be a
delta-connected system.
DC OUTPUT
PHASE
1
D1
Fig.2: This set of waveforms
shows how a 3-phase
system is converted to DC
using the schematic of
Fig.1. The three voltage
phases are separated
by a phase difference of
120° and the resultant DC
from the rectifiers has a
superimposed ripple at six
times the mains frequency,
ie, 300Hz for a 50Hz grid.
Because there are six periods (or
combinations) of SCR conduction in
this DC/AC conversion process, this
is referred to as a 6-pulse converter;
six trigger pulses are required for one
cycle of AC which will be 50Hz or 60Hz,
depending on the country where it is
used.
In reality, the setup outlined in Fi.g.1 is
not used now. Instead, we have a more
complicated system, shown in Fig.3.
This uses a 3-phase transformer with
star and delta-connected secondaries
feeding a set of 12 SCRs or valves.
This makes use of the fact that while
the three phases are separated by 120°,
when the outputs of a star and delta
system are connected together, you
effectively have six phases separated
by 60°.
When used for AC/DC conversion
(ie, rectification), the resultant DC has
600Hz ripple (for a 50Hz grid) and there
are 12 combinations of SCR conduction.
To consider just one 30° conduction
period, when D1 conducts, so will D5,
D8 & D11 or D6, D9 & D12.
Since there are 12 conduction periods, this is referred to as a 12-pulse
converter. The six voltage waveforms
from the star and delta windings are
depicted in Fig.4.
Note that we have shown the Star
and Delta winding outputs with their
normal amplitude relationship, whereby
the phase-to-phase (delta) voltage is √3
times the phase to neutral (star) voltage,
to highlight how the 6-phase system
comes about.
siliconchip.com.au
+DC
STAR
SECONDARY
φ1
PRIMARY
WINDING
THREEPHASE φ2
AC
φ3
D1
D2
D3
D4
D5
D6
D7
DELTA
SECONDARY
D10
D8
D9
D11
D12
HIGHVOLTAGE
DC
0V
Fig.3: the schematic of a 12-pulse AC to DC converter. While this circuit is
ostensibly an AC to DC converter, it can just as easily be run in the opposite
direction, converting DC to 3-phase AC by triggering the SCRs at appropriate
times to energise the relevant transformer windings.
In reality, we assume that their amplitudes will be the same.
For DC/AC conversion, the process is
exactly the same as for a 6-pulse system
except that we now have 12 SCRs (or
valves) which are triggered at 12 points
in the waveform to energise the star and
delta windings. Because the harmonics
in a 12-pulse converter are twice as high
Fig.4: the phase relationship in a
+
3-phase system using star and
delta waveforms together, to drive
the converter circuit of Fig.3.
The resultant phase waveforms
are separated by only 60° and the
0V
resultant DC from rectification
has ripple at 12 times the mains
frequency, ie, 600Hz for a 50Hz
system. Note that the harmonics
produced in conversion can run
as high as the 30th or higher and
–
require elaborate filtering.
in frequency and reducing in amplitude
by the same amount, it is much easier
to filter them out and prevent them being
fed into the distribution grid.
Note that the schematic of Fig.3 is
much simplified compared to reality and
does not show all the filtering, protection
and ancillary components.
Note also that the DC/AC conversion
process assumes that there is already
3-phase power available on the AC grid,
to provide the essential triggering and
synchonisation of the SCR valves.
In a complete HVDC system there is
DELTA
φ1
STAR
φ4
DELTA
φ2
STAR
φ5
DELTA
φ3
STAR
φ6
T
an AC/DC converter at each end of the
transmission line and in the case of the
Basslink system and others throughout
the world, they can transfer power in
both directions. See Fig.5.
So was
Edison
right all
along?
Anyone familiar with the history
of electrical energy will know that
Thomas Edison was a very strong
proponent of DC and many early
systems in towns and cities were
based on DC.
Ultimately though, the sheer
advantages of the AC system, as
promoted by Westinghouse and
largely invented by Nicola Tesla,
won out and now AC generation and
transmission of power are universal,
usually at 50Hz or 60Hz.
So why is high voltage DC now
being promoted? In fact, the advantages of AC over DC are still
manifest. However, for power transmission over very long distances,
say 800km or more or for distances
of 50km or more undersea, DC has
advantages in that usually only
one cable instead of three, in the
case of 3-phase AC transmission,
is involved.
By transmitting electrical power
at very high voltages, eg, 400kV or
higher, the resistance losses are
reduced, as are the weight of the
cable, its supporting structure and
so on.
Furthermore, DC transmission
makes it possible to transfer power
between two electrical grids that operate at different frequencies (50Hz
& 60Hz), as happens in Japan.
So Edison did not get it right.
High voltage DC power transmission is the solution to problems
that Edison is unlikely to ever have
foreseen.
SC
+DC
φ1
AC
GRID1
φ2
φ1
AC–DC
CONVERSION
HIGH-VOLTAGE
DC TRANSMISSION
φ3
AC–DC
CONVERSION
φ2
AC
GRID2
φ3
0V
Fig.5: a complete HVDC system in schematic form. In very long distance systems, as used in China or Brazil, power
transmission is usually in one direction only but a complete AC to DC and DC to AC converter is required at the start
and finish of the transmission line.
siliconchip.com.au
September 2008 19
|