Chapter 23 HVDC 980

23.1 HVDC electrical power transmission

Electrical power is generated in ac form and is also usually distributed and consumed in an ac form. Its
long distance transmission between these two stages may be an ac or a dc transmission system. In a
high-voltage dc (HVDC) system the generated 50/60Hz ac is controlled-rectified to dc, transmitted, then
at the receiving end, converted from dc back to 50/60Hz ac. A HVDC system is two ac systems
connected by a dc transmission system, where the ac systems can be totally independent.

The dc-link of a HVDC transmission system is either

CHAPTER 23  a controlled voltage dc-link or

 a controlled current dc-link.

A HVDC controlled current link has the following characteristics.

The link in highly inductive, achieved with series inductance at each end.
The converter/inverter technology is operated in a controlled current mode, thus the
converter/inverter devices require reverse voltage blocking ability. Symmetrical blocking thyristor
HV Direct-Current Transmission devices are applicable, but such devices are restricted to line commutation and phased control.

A HVDC controlled voltage link has the following characteristics.

The link in highly capacitive, achieved with parallel connected capacitance at each end.
The converter/inverter technology is operated in a controlled voltage mode, thus the
converter/inverter devices can be uni-directional voltage blocking IGBT technology. Since
devices are gate commutatable, a switching frequency of kHz’s is possible, thus PWM
techniques can be employed for harmonic minimisation.

Y Y Y Y
Id
Y Y Id Y Y
Originally electrical power generation, transmission, and distribution systems were direct current. The
advent of the three-phase induction motor and the ability of transformers to converter one ac voltage to V
V
another ac voltage level (at the same frequency), saw the unassailable rise to dominance of ac electrical q =3 r =1 s = 2
power systems. But for long distance electrical power overhead line transmission, of just a few hundred p=qxrxs
kilometres of typically about 300 to 550km, a dc transmission system is a viable possibility. For p = 12 (a)
Y Δ
V Δ Y
underwater or underground electrical power transmission, ac may not be viable at just 50km due to high
capacitive charging currents because of the close proximity of the cables (particularly subsea cables). Y Y Y Y
Id Y Y Y Y
Typical application boundaries between the use of HVAC and HVDC are shown in figure 23.1.
V
The 1980km HVDC link between Xiangjiaba and Shanghai is the largest dc overhead line in the world, V
with a capacity of 6.4GW and a ±800kV DC link voltage. The longest overhead HVDC of 2,375 km
(3.15GW ±600kV) is the Rio Madeira link in Brazil.
V
A Brazilian functional HVDC transmission system is rated at ±600kV on 785km and 805km transmission V
lines. Each of the two bipolar dc transmission systems carry 3.15GW. Also involved are three, three- Y Δ Id Δ Y
Y Δ Δ Y q =3 r =1 s = 4
phase 765kV ac lines which are 1GVAr variable capacitor series compensated (FACTS) at two
p=qxrxs
intermediate substations. p = 12
(b) (c)

± 800kV >6.4GW Figure 23.2. HVDC transmission systems: (a) 6 pulse monopole; (b) 12 pulse monopole; and
(c) 12 pulse bipolar, converter bridge configurations.

23.2 HVDC configurations (also see section 23.16)

MW

There are a number of different configurations for transmitting dc power, depending on the number of
System voltage kV

cables employed. Each uses a three-phase fully-controlled thyristor converter (rectifier) coupled through
a dc link to another identical three-phase fully-controlled thyristor converter (inverter). Both converters
System power

have the same modular structure except the converter connections to the dc link are interchanged for
one converter, hence power flow is fully reversible. Since the valves can only conduct current in one
direction, power reversal is achieved by changing the polarity of the dc link terminal voltages through
control of the converter thyristor firing delay angles. The rectification mode (positive dc link voltage) is
achieved with thyristor firing angles of 0 < α < ½π while inversion (negative dc link voltage) is achieved
with firing angles of ½π < α < π. Because one converter terminal connection is reversed, a rectifying
voltage (0 < α < ½π) is opposed by an inverting voltage (½π < α < π) – subtractive not additive voltages.
23.2i - Monopole and earth return
The monopole configurations shown in figure 23.2 parts a and b (6 pulse and 12 pulse respectively) use
just one transmission cable and earth (where resistance is dominated by the earth electrodes each end)
is used as the negative return. Occasionally, a metal earth return may be used, but importantly any
Cable/route length km return is at ground potential thus does not need the full transmission voltage insulation. The converter
>2000km 2 bipolar
output terminals are reversed relative to one another, as indicated by the direction of the thyristors in the
Figure 23.1. HVDC and HVAC transmission system operational boundaries. symbol blocks.

BWW
981 Power Electronics Chapter 23 HVDC 982

Issues involved in using a ground return are Current Source converter HVDC transmission systems
 Electrochemical corrosion of buried metals objects, like pipelines
 Electrode chemical reaction under the sea The thyristor current source converter approach is well established and its reliability has been proven.
 Magnetic field disturbances when the go and return paths become unbalanced Most existing dc transmission systems are based on current source converter technology. HVDC
The monopole system is limited in power handling capability, typically 1.5GW above the ground, and transmission systems based on current source technology can use line-commutated converters LCC or
600MW below the ground or under the sea. capacitor commutated converters CCC. Fig. 23.4 shows the line and capacitor commutated converter
23.2ii - Bipolar HVDC transmission systems. In the CCC converter case, the series commutation capacitors can be on
In the bipolar arrangement two high voltage conductors, at opposite potentials with respect to ground, either side of the line voltage transformers and generate reactive power proportional to the load current.
are used as shown in figure 23.2c and figure 23.3 in more detail. Any pole imbalance uses an earth When series capacitors are used on the line side, before the ac filters, the configuration may include
return, if a low-voltage metal ground return is not used. The bipolar configuration has a number of parallel reactors and TCSC thyristor valves, forming a Controlled Series Capacitor Converter, CSCC.
advantages over the monopole arrangement.
 Normally no earth-current flows which minimises earth losses and any earth related The capacitor commutated converter extends the applications of the HVDC transmission systems based
environmental effects, including minimal corrosion of underground system on current source converters to systems with a low short circuit ratio. This converter improves system
metal components immunity to the risk of commutation failure; replaces the switch capacitors by a combination of fixed
 If a fault develops on one pole, the other pole can continue to operate in a capacitors and reactors for filtering purposes only, and reduces transformer rating and size. As a result,
monopole arrangement, using the earth as the return path the size of the converter station in CCC-HVDC transmission systems is reduced compare to a LCC-
 For a given power rating, each conductor has half the cross-sectional area of the HVDC system with the same power rating. The capacitor-commutating converter can be connected to
monopole line, thus reducing the extra cost of using a second conductor ac systems with a short circuit ratio as low as 1.0 without the need for a synchronous condenser or static
 The same dc transmission line towers can carry two lines with a small additional synchronous compensator (STATCOM) to provide additional reactive power and can maintain constant
capital cost voltage at the point of common coupling.
A homopolar hvdc link is formed if the two high voltage conductors have the same polarity, with,
undesirably, a high ground or metal return current. According to IEEE Standard 1204-1997, the system short circuit ratio (SCR) and effective short circuit
The bipolar system is capable of higher transmission powers than the monopole configuration. Bipolar ratio (ESCR) are defined as:
systems can carry over 3GW at voltages of over ±500kV, over distances well in excess of 1000km.
S
23.2iii - Tripole SCR 
Two of the three conductors of an ac system are used in a bipolar configuration, with the third conductor Pdc
used as a parallel monopole with bidirectional power flow capability. The bidirectional capability of the S  QC
third conductor allows each of the two bipolar conductors to carry higher than rated current when each in ESCR 
Pdc
turn is relieved periodically by the monopole system, such that all three conductors do not exceed rated
2
I R losses. In this way each of the three conductors experience the same thermal losses. This is
 
achieved if the bipolar currents are cycled every few minutes between 0.366pu ½ 3  1 and 1.366 pu
where S is the ac system three-phase symmetrical short circuit level in MVA at the converter terminal
 
½ 3  1 , with ±1 pu being appropriately alternated in the monopole. As a result, 80% more power can
(ac bus) calculated at rated terminal voltage (1 pu), Pdc is the rated dc power of the terminal in MW, and
QC is the three-phase fundamental MVAr at rated Pdc and rated terminal voltage. This includes ac filters
be transmitted compared with the ac equivalent, using the same conductors, towers, etc. Unlike the ac
and shunt connected reactive power capacitors used to compensate for the reactive power absorbed by
equivalent, the dc system can be fully loaded without system instability or need for reactive power
the converter. The system strength is classified as follows:
compensation.
23.2iv - Back-to-back  system is strong if SCR > 3 or ESCR > 2.5
Two different asynchronous ac systems in close proximity, possibly operating at different frequencies,  system is weak if 3 > SCR > 2 or 2.5 > ESCR > 1.5
can be interconnected by either a monopole or bipolar system. Since no dc transmission cables are  system is very weak (very low SCR) if SCR < 2 or ESCR < 1.5
necessary, because of the close proximity of the two systems, the system type and voltage level are not
restrictive.
Smoothing
23.2v - Multi-terminal inductors
More than two converters are connected to the same dc link, where simultaneously, at least one
converter operates in the rectification mode and at least one other converter operates in an inversion
mode. Mechanical switching of the converter terminals is necessary. The control system is more
elaborate than for normal point-to-point transmission and power reversal is affected by current reversal,
(as opposed to the usual voltage reversal used in point-to-point transmission). See section 23.12. AC bus
Y DC filters
Y
Redundant breaker with
Δ
?
V continuous metal neutral connection V

AC
filters

V V Switch capacitors
and reactors (a)

Figure 23.3: Bipolar HVDC system with station optional metallic return.
983 Power Electronics Chapter 23 HVDC 984

Smoothing
inductors 23.4 Twelve-pulse ac line frequency converters

The six-pulse line-frequency fully-controlled thyristor converter was discussed in chapter 14.6. Harmonic
filters are required on both the ac and dc side of the converter as shown in figure 23.7. The ac side
Y Y harmonics occur at 6n±1 the fundamental, while the dc side harmonics are generated at 6n. To reduce
AC bus the filtering requirements, and increase the effectiveness of the filtering, on both the ac and dc sides,
DC filters most high power HVDC systems use 12-pulse, 30° phase-shifted transformer/converter arrangements.
The ac side harmonics now occur at 12n±1 and the dc components are generated at 12n.
Twelve-pulse converter operation is achieved by using the series bridge connection in conjunction with
Y Δ
? Δ-Y and Y-Y compound connected transformers as shown in figure 23.6b. (A delta connection is usually
employed on the lower voltage side of the transformer.) Figure 23.11a shows the arrangement in more
AC detail, with the necessary transformer turns ratio to ensure each converter bridge produces the same
filters output voltage at the same thyristor firing delay angle. Voltage matching between the ac line and
required dc link voltage is achieved with the transformer turns ratio N, shown in figure 23.11a. The
(b) series thyristors in each bridge provide paths which allow both converter currents to be equal.

Figure 23.4: Converter station technologies for the current source based HVDC transmission systems:
(a) line commutated high voltage dc (LCC-HVDC) transmission system and AC DC
side side
(b) series capacitor (valve side) commutated high voltage dc (CCC-HVDC) transmission system.

a
23.3 Typical thyristor HVDC transmission system Y Y
6 pulse valve
A fully modular hardware structure is used. The few watts needed for gate power is harvested from group
thyristor anode level components. b
valves shaded
YΔ
Each 8.8kV symmetrically blocking thyristor is configured in a module with its gate electronics and RC
snubber as shown in figure 23.5. Many thyristors are connected in series to form a valve, with internal
c q =3 r =1 s =2
static voltage sharing resistors and a saturable reactor turn-on snubber as shown in figure 23.5a. The p=qxrxs
Y Y
saturable reactor usually incorporates shunt damping to control turn-on current oscillation at saturation. p=6
(a) (b)
Six valves are needed to form the 6 pulse valve group converter bridge in figure 23.6a, while 12 valves
are used to form the twelve pulse value group converter in figure 23.6b. Each group of four valves in a
single vertical stack form quadrivalves as shown in figure 23.6b. Each quadrivalve may contain
hundreds of series connected thyristors to give the necessary hundreds of kV pole voltage rating.
AC DC
Typical six and twelve pulse valve group configurations are shown in figure 23.6, which form the unipolar Y Δ side side
converters in figure 23.2 parts a and b respectively. A more detailed circuit diagram of the 12-pulse
converter and its MOV voltage protection and dc and ac harmonic filtering circuitry, in a substation
installation, is shown in figure 23.7. The dc inductors Ldc in each pole assist in filtering harmonic currents a
and smooth the dc side current thereby reducing the current level for the onset of discontinuous current YΔ
flow. Because the inductors control the dc link side di/dt, converter commutation is more robust. No dc
filters may be necessary with the back-to-back HVDC converter configuration.
b

Saturable
reactor c Y Y

RC snubber
C1 Valve damping circuits a
electronics 3 Quadrivalves
C2 shaded
C1>>C2 b q =3 r =1 s =4
Static voltage sharing p=qxrxs
resistors p = 12

(a) (b) c
Y Y
(c) (d)
Vf/b Valve
electronics

Figure 23.6. Monopole converter bridges:

Six-pulse valve group (a) converter bridge schematic and (b) six-pulse valve group converter
symbol; Twelve-pulse valve group converter configuration with star-star and star-delta connected
Figure 23.5. Thyristor valve: (a) modules components assembled into a valve and (b) valve symbol. converter transformers: (c) converter schematic and (d) twelve-pulse valve group converter symbol.
985 Power Electronics Chapter 23 HVDC 986

As a result of the transformer configuration, the corresponding upper and lower transformer voltages are 1
2

displaced by 30°, where VaYs-nY leads VaΔs-nΔ by π radians. The dc-link current Id is assumed constant bn  
 i a (t ) sin n t d t
because of the large smoothing inductor Ldc (linear and typically ½H). If source impedance is neglected, 0

then the various circuit current waveforms are constituted from rectangular current blocks as shown in   2 2
2
  2 
2 3 Id 3
sin n t d t   d sin n t d t 
figure 23.11b. Each converter operates with the same firing delay angle α, with respect to the voltage 2I d I
references shown in figure 23.11b. Because the transformer primaries are in parallel, the input current is
  
 0 2 3
sin n t d t 

 3 2 3 (23.8)
the sum of the appropriate two transformer phase currents, namely ia = iYa + iΔa for phase a. The Fourier  
3 2 3
 2

series for each transformer primary phase current is obtained from analysis of the appropriate six-pulse 2I d   2 
converter current, for each converter 
3n   cos n 2  cos n ( 3  2 )  cos n ( 3  2 )  cos n (  2 ) 
 
2 3 Id  1 1 1 1  n  1, 3, 5, 
iYa  cos   cos 5  cos 7  cos11  cos13 .....  (23.1)
N   5 7 11 13  The Fourier coefficients for the line current are then defined by
and an  ½anY  ½an 
(23.9)
2 3 Id  1 1 1 1  bn  ½bnY  ½bn 
i a  cos   cos 5  cos 7  cos11  cos13 .....  (23.2)
N   5 7 11 13  From which the input line current is defined by
Because of the symmetry of a three-phase system, no triplens exist in each input current. The total line 

current drawn is i a (t )  
n 1,2,3,
2 I an sin(n t  n ) (23.10)
4 3 Id  1 1 1 
i a  iYa  i a  cos   cos11  cos13  cos 23 ..... (23.3)
N   11 13 23  Output voltage
The 12-pulse transformer/converter arrangement cancels harmonic components 6×(2n-1) ±1. The converter outputs are series connected, hence the output voltage is additive for each pole, namely
Vdr = Vdr1 + Vdr2. The converter output voltage, with the constraint that both converters have a trigger
The line current ia rms value is 1  1
3 I d / N and the rms fundamental is 2 6 I d / N  . delay of α, is
6 2 V
The ac line current harmonics occur at 12n±1. The valve side ac line current, shown as N×IYsa (or Vdr  Vdr 1 Vdr 2  VLL cos   2.70 LL cos  (23.11)
N×IΔsa) in figure 23.11b has an rms value of I d 2 / 3 , and once rectified, the valve unipolar current has N N
an rms value of Id /√3. The peak output voltage occurs midway between the peak voltage from each converter, and for α = 0
VLL V
More general equations are for the general case of asymmetrical converter firing when 1  2. Vdr  2 2 cos 15  1.932 2 LL (23.12)
Assuming a turns ratio between the phase windings of the star connected primary and the star N N
connected secondary is 2:1, and the turns ratio between the primary and the delta connected secondary
is 2:√3, then Each converter delivers six current blocks of magnitude Id, comprised of two ⅔π current blocks π
i a  ½iYa  3 2 i a (23.4) radians apart in each converter arm. Since each converter output is the same but shifted by π radians,
provided the two converters have equal delay angles, the resultant 12 current block per cycle results in
Due to waveform symmetry, no dc component or even harmonics exist.
the dc side voltage harmonics in Vdr being of the order 12n.
The Fourier coefficients anY and bnY of iYa (= anY + j bnY) are
If the two delay angles are controlled individually, the two outputs add but with harmonic components
 5 1 11
1 
2
1 6 6
given by two six pulse converter, displaced by π, where
anY   i a (t ) cos n t d t    I d cos n t d t   I d cos n t d t 
1
 0   3V  
1 1 2 cos 21 
Vdr 1  max  cos 1   sin  6n  t  6n 1  
7
 6 1 1
 (23.5)  
 6n  1  6n  1  6n  1  6n  1
6
  n 1
2 2

4I d n  
 sin sin n 1 n  1, 3, 5,  (23.13)
n 3 3V max  
2 cos 2 2 
 5 1 11
 Vdr 2   cos  2   1

1
 sin  6  n  t  
1  
 
   6n  1  6n  1  6n  1  6n  1
1 6

2 2 6n 2
2
16 6 n 1
  i a (t ) sin n t d t    I d sin n t d t   I d sin n t d t   
1
bnY
 0   7 where
 6 1 1  (23.6)
6
 cos  6n  1 1 cos  6n  1 1 
n  

4I d
sin cos n 1 n  1, 3, 5, 
6n 1  n   tan 1 
 6n  1  6n  1 
n 3  sin  6n  1 1 sin  6n  1 1 
  
The Fourier coefficients anΔ and bnΔ of the delta winding current iΔa (= anΔ + j bnΔ) are   6n  1  6n  1 
2
1
an    i a (t ) cos n t d t  cos  6n  1  2 cos  6n  1 2 
 
 0
6n 2  n   tan 
 6n  1 1  6n  1 
 3 2 2
 2   sin  6n  1  2 sin  6n  1 2 
  
3
 Id 2I d 
  3 cos n t d t   3 cos n t d t    6n  1  6n  1 

2  0 2  2 
  3 (23.7) When the two delays angles differ, 1  2, the output voltage harmonics occur at order 6n.
  2

 I 
   3 cos n t d t
d
 The dc link ripple current depends on the total dc link inductance L T and the ripple magnitude varies with
 23 2  the rectifier and inverter converter firing angles, αr and αi. Provided the minimum link current is greater
than the ripple magnitude, each 12-pulse converter contributes a component given by
2I d   2 
  sin n 2  sin n (  2 )  sin n (  2 )  sin n (  2 )  3 2 VLL
3n   3 3  I d min  I d ripple  
0.023 sin 
  LT
 987 Power Electronics Chapter 23 HVDC 988

The total ripple current is the sum of the connected converter component. Each converter can be and the real power transfer, which is inverted into the ac system is given by
operating at a different frequency, ω, be connected to different line voltages VLL, and be operating in 2 6 I
different modes so is operating at different delay angles, α. Pi  Vdi I d  3 V LL I a 1 cos i  3 V LL  I d  cos i  2.7  V LL  d  cos i
N N (23.17)
converter bridge dc bus  Q i / tan i
transformer converter dc filters surge capacitor
Converter unit To maximize the ac system power and minimize the reactive power, the delay angle αi should be large
(→ π). From equation (23.17), to minimise the link I R loss, Id should be minimised (maximise dc link
6 pulse 2

Ldc voltage) and the delay angle αi should be large (→ π). Thus the inversion voltage should be as large as
hv
½H
dc
possible, avoiding commutation failure. Then the maximum αi decreases as current increases since
linear thyristor commutation time increases with current (and temperature).

P-Q operation of the twelve pulse series connected converter

Metallic return The theory of controlled active and reactive power operation of a series connected 12-pulse converter
transfer breaker
ac assumes an ideal input transformer, negligible source impedance, and a constant current from the
filters
converter dc output bus. The output voltage contribution due to each constituent 6-pulse converter is
given by
Midpoint dc
3V
bus arrestor Vdri  max cos i (23.18)
hv
Neutral bus Neutral bus

ac
dc bus arrestor surge capacitor where Vmax is the peak line voltage of the voltage applied to each converter, α i is the firing angle of the
Y Y earth line constituent converter, and i represents either converter 1 or 2.
arrestor and electrode
By multiplying (23.18) by the dc load current Idc, the active power is given by
Qi  Pmax sin i (23.19)
earth return
Y Δ
transfer breaker where
3
½H Pmax  I oV max
linear hv 
Converter unit
12 pulse
Ldc dc
which represents the maximum power delivered from each converter at αi = 0. Similarly, the reactive
power absorbed by the converter is given by
dc reactor and dc bus dc line
Qi  Pmax sin i (23.20)
arrestor arrestor arrestor
The active power Pd and reactive power Qd, drawn by the 12-pulse converter are the sum of the
contribution made by each converter, namely
Figure 23.7. Thyristor HVDC station. Pd  Pmax (cos 1  cos 2 )
(23.21)
Qd  Pmax (sin 1  sin 2 )
23.4.1 Rectifier mode From these two equations, the 12-pulse converter draws active power of 2Pmax and zero reactive power
Figure 23.11b shows that the angle between the input ac voltage and its fundamental current Ia1 is at α1 = α2 = 0. Similarly, at α1 = α2 = ½π, the converter draws reactive power of 2Pmax and zero active
determine by and equals, the phase delay angle αr. The phasor diagram for rectification is shown in power. Therefore, 2Pmax is the base power, 1 pu, and the equations in (23.21) become
figure 23.12a. For a constant link current Id, the fundamental ac input power factor is cos αr while the
input reactive power is given by Ppu  ½(cos 1  cos 2 )
(23.22)
Q r  3 VLL I a 1 sin r  Pr tan r Q pu  ½(sin 1  sin 2 )
2 6 I (23.14)
 3 VLL  I d  sin r  2.7  VLL  d  sin r The variation of Ppu and Qpu with simultaneous variation of α1 and α2 is shown in the three-dimensional
N N plots figure 23.8.
The rms of the fundamental line current Ia1 is 2 6I d / N  .
The real power transfer, which is the rectifier output power, is given by
Pr  Vdr I d  3 VLL I a 1 cos r  Q r / tan r
2 6 Id (23.15) 1

 3 VLL  I d  cos r  2.7 V LL   cos r

N N

Reactive power in pu
0.8
1
Active power in pu
To maximize the power flow and minimize the reactive power, the delay angle αr should be small. From 0.5
0.6

equation (23.15), to minimise the link I R loss, both Id and the delay angle αr should be small. That is,
2
0 0.4
from equation (23.11), the rectifier output voltage should be maximised.
-0.5
A low converter firing angle minimise the reactive power, reduce snubber losses, and reduces the 0.2

harmonic content. -1
180 0
180
150 180
150
23.4.2 Inverter mode 120
150
180
The same basic rectifier mode equations hold in the inversion mode except that αi > ½π. Operational
120 120 150
90 90 120
90
waveforms and the phasor diagram for this mode are shown in figure 23.12b. The reactive power is 60 α1 (degree) α2 (degree)
60 90
α2 (degree) 60 60
30 30
30 30 α1 (degree)
2 6 I 0 0
Q i  3 VLL I a 1 sin i  3 VLL  I  sin i  2.7  VLL  d  sin i 0 0
N d N (23.16)
 Pi tan i
Figure 23.8. Effects of varying of 1 and 2 on: (a) active power and (b) reactive power.
 989 Power Electronics Chapter 23 HVDC 990

Ldc
id(t) = Id
The 12-pulse converter may operate with symmetrical firing of each 6-pulse converter, where α1 = α2 or = constant
with asymmetrical firing where α1  α2 (or a combination of both). Different power loci can be realized by AC ia
+ DC + +
different combinations of firing angles (α1 and α2). Three power loci are shown in part ‘a’ of figure 23.9. side side
iYa iYsa
The symmetrical firing P-Q locus is represented by the outer semicircle with centre ‘0’ and radius of 1 a aYs
pu, and is obtained by the symmetrical firing of α1 and α2 from 0 to π using the parts of equation (23.22). q =3 r =1 s =4
All points in the P-Q plane within the outer semicircle can be achieved by unique combinations of α1 and N:1 p=qxrxs
α2. The figure also shows power loci for asymmetrical firing which are represented by the circumference - Vdr1 p = 12
ia
of the two inner semicircles with radii of 0.5 pu. These are obtained by varying α1 from 0 to π while α2 is b
cYs nY
held at ‘0’ in the case of rectification and by varying α2 from 0 to π while α1 is held at π in the case of bYs
inversion. Part ‘a’ also shows the power locus for constant VAr operation indicated by the line parallel to Id
the Ppu axis. With asymmetrical firing the maximum reactive power is decreased to one-half the case for Y Y
symmetrical firing. Part ‘b’ of figure 23.9 shows the variation of input current THD with varied delay c
Y Y -
angles and reflects the adverse effect of asymmetrical firing on the THD. Vdr
+ ia = iYa + iΔa
Vdr
iΔa Δ Y aΔs iΔsa ia
1
a ib
ic
0.9 60 1=2 √3N : 1 ΔY
Asymmetrical firing
0.8 (a1 = a2) 50 nΔ
cΔs Vdr2
Reactive power in pu

0.7 40 b
0.6 %THD 30
Constant VAr operation (a 1 ? a2) bΔs
0.5
20 (a)
a1 = 180° a1 = 0° to 180°
0.4
a2 = 0° to 180° a2 = 0° 10
c
0.3
- -
0
0.2 Minimum VAr power locus 180
150
0.1 180
120 150
90 120
0 60 90 Vdr
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 a2, degree 60
30 30 a1, degree
Active power in pu
0 0 = Vdr1 + Vdr2

(a) (b)

Vdr2
Figure 23.9: Twelve-pulse fully-controlled converter: (a) power loci and (b) input current THD.
Vdr1
VcYs-nY VaYs-nY
The symmetrical firing power locus is associated with the best THD and the worst supply current power -30° o° αr
factor. On the other hand, asymmetrical firing decreases the reactive power flowing in the system, which
improves the input power factor but deteriorates the input current THD, which limits the power factor
improvement. As a result, asymmetrical firing of the 12-pulse converter offers the possibility of a smaller
reactive power compensator for power factor compensation, at the expense increased input current
harmonics. o° αr

Active power, pu
Active power, pu
1 (α1 = α2 = 0)
α= 0 1
1

N×iYa
cos α ½(cos 1  1) , 2  0 Rectifier quadrant + Id
Sd(pu) Ia 180° 330°
dc output voltage

Rectifier quadrant
cos Φ Sd(pu)
cos Φ
dc output voltage

α=Φ -
Φ
0 30° 90° 150° Ia 0° 30°
(α1 = α2 = π/2)
α= π/2
0

-1 0
0
Reactive -1 0
1
Reactive -Id
1 Sd(pu) power, pu
power, pu ½(cos 1  cos 2 ) , 1  2
N×iΔa
Inverter quadrant
Inverter quadrant ½(cos 2  1) , 1  

Id

2I d 3
-1
(α1 = α2 = π) 
-1 α=π -1 3
Control voltage, pu
α Control voltage, pu α1, α2 (b)
-1 0 1 -1 0 1
N×ia
α2 , (α1 = π)
α1 = α2
π/2 π/2
α1 , (α2 = 0)  1 3 Id
π π
  1 3  1 I d Ia1
  2 3  1 I d fundamental of Ia

Figure 23.10: Power locus of 6-pulse and 12-pulse converters and per unit output voltage. Figure 23.11. Twelve-pulse valve group converter configuration with star-star and star-delta connected
converter transformers: (a) converter schematic and (b) twelve-pulse valve group waveforms.
991 Power Electronics Chapter 23 HVDC 992

23.5 Twelve-pulse ac line frequency converter operation control

P P
Rectification and inversion modes of converter operation and the line (or natural) commutation process
Q q =3 r =1 s =4 Q of the three-phase fully-controlled thyristor converter have been considered in chapter 14.4 for
Id p=qxrxs Id rectification, with overlap in 14.5, and 14.6 for inversion with overlap.
p = 12 One converter operates as a rectifier (power flow from ac to dc) and the other dc link converter operates
Y Y Y Y
+ - as an inverter (power flow from dc to ac). Either terminal converter can operate as an inverter or rectifier,
Vdr1 Vdr since the delay angle determines the mode (voltage) of operation. The power flow between the two ac
Vdr1
ia = iYa + iΔa ia = iYa + iΔa = Vdr1+Vdr2 systems connected to the HVDC link is controlled by controlling the delay angle of each converter.
- Vdr + Current only flows in one direction in the dc link, from the rectifier to the inverter. A simple system model
ia + = Vdr1+Vdr2 ia - is shown figure 23.13 where the link dc resistance is represented by Rdc and source reactance, hence
ib ib Vdr2 overlap effects have been neglected. Let the transformer turns ratio factor N equal 1. The dc current Id is
ic Vdr2 ic
- + V Vdi
ΔY ΔY I d  dr (23.23)
Rdc
where
3
Vdr  2 VLL cos r for 0  r  ½
Vdr2  (23.24)
½Vdr 3
Vdi  2 VLL cos i for ½  i  
Vdr1 
Id
VcYs-nY
VcYs-nY
+ Rdc -
-30° o° αi
Vdr Vdi
VaYs-nY VaYs-nY P
o° αr o° αi (a)
Vdr1 - +
½Vdr
-30° o° αr Vdr2
N×iYa rectification inversion
N×iYa αr < 90° αi > 90°
Id Id
+ +
Ia 180° 300° Ia 180°
- - 0 30° 90° 150°
Vd
0 30° 90° 150° Ia 360° Ia
-Id -Id
N×iΔa
N×iΔa Vdi

Id Id
 
3 2I d 3 operating
 
2I d
3 3 point maximum αi
Ia1 decreases with
N×ia increased Id
N×ia Ia1 (b)

 1 I
3 d  1 I
3 d

  1 3  1 I d   1 3  1 I d
Idref  Vdr Id

  2 3  1 I d   2 3  1 I d

Figure 23.13. Basic HVDC transmission system: (a) circuit diagram and (b) load line characteristics.

  I a1  VaYs-nY   I a1  o
VaYs-nY
The rectifier output power is
o

αr
Pr  Vdr I d (23.25)
αi
while the power supplied to the inverter is
  I a1  Pi  Vdi I d (23.26)
  I a1 
where
Ia1 Ia1
Pr Vdr I d Vdi I d  I d2Rdc  Pi  I d2Rdc (23.27)
(a) (b)
If transformer per phase leakage inductance Ls, referred to the converter side, is accounted for, then the
resultant overlap at commutation reduces the output voltage for each six-pulse converter.
3Ls
Figure 23.12. Twelve-pulse valve group converter configuration with star-star and star-delta
3
Vdr  2  

2 VLL cos r 
 6
I   2 VLL cos r  Ls I d
 d  
 (23.28) 
connected converter transformers operating in: (a) a rectifying mode and (b) an inverting mode.
993 Power Electronics Chapter 23 HVDC 994

droop due to
Both the rectifier and inverter dc output voltages can be compensated for leakage reactance Vdr link resistance
RECTIFIER Vdi minimum extinction
INVERTER
commutation overlap. Remember the overlap voltage component is not a loss element in the resistor angle characteristic
R
sense. It represents a ‘lossless’ loss of voltage which increases with current. ref droop
S Id
If the link voltage is controlled by the inverter and the dc current controlled by the rectifier, then the load 
 D
constant current A
line characteristic in figure 23.13c results. The inverter voltage is kept slightly below the rectifier voltage. operation °
minimum rectifier E
As the load current increases, the inverter terminal voltage is reduced. This is because the time to safely B
delay angle
commutate the inverter thyristors increases with current, hence αi must decrease, as shown by the characteristic
Vdr constant Id
current operation° γ
droop in the output characteristics in figure 23.13b. The more detailed practical approach to HVDC greater than minimum
power transport control is considered in section 23.5.1. αr = 90° Vdi
extinction angle
Vdr = 0
T αi =γ= 90°
23.5.1 Control and protection F Vdi = 0
ref ref
Id Id (a) (b) I I < Id Id
HVDC transmission systems must operate under tightly controlled conditions. Dc-link current and the
two terminal voltages are precisely controlled to affect the desired power transfer. Accurate system
quantity measurements are required, which include at each converter bridge, the dc-link current, the dc- (c) (d)
minimum rectifier Operating
side voltages, and the delay angle α for each converter/inverter. Vd R delay angle point Vd
Two terminal dc transmission systems have a preferred control mode during normal operation. characteristic
S D
D C Vdr
B A
Inverter Vdi
R S
Under steady-state conditions, the inverter controls the dc voltage by one of two methods. RECTIFIER
E H E
 The inverter maintains a constant delay angle αi > 90°, or extinction angle γ, where γ = π - αi, minimum extinction
Operating
Vdi Vdr angle characteristic INVERTER
This constant angle maintenance causes the dc voltage Vd to droop 
with increasing dc droop point
current Id, as shown in the minimum constant extinction angle  characteristic in figure Y 
Imargin  INVERTER RECTIFIER
23.14b and labelled A-B-C-D in figure 23.14c. The weaker the ac system at the inverter, the X Imargin
VDCDL F T F
steeper the droop characteristic.
ref
 Alternatively, the inverter may operate in a dc voltage controlling mode which is the constant Vd II Id Id II ref
Id Id
characteristic shown dashed in figure 23.14b and labelled B-H-E in figure 23.14c. To achieve
Figure 23.14. Steady-state Vd - Id characteristics for a two terminal HVDC system.
this, the extinction

angle γ must increase beyond its minimum value characterised in figure
23.14b as  .
Current margin
Rectifier The rectifier and inverter controllers both receive the dc current demand I dref but the inverter current
If the inverter is operated in either a minimum constant γ or constant Vd mode, then the rectifier is used demand is decreased by an amount termed the current margin Imargin, as shown in Figure 23.14c. This
to control the dc link current Id. This is achievable provided the delay angle αr is not at its minimum limit. current margin is usually a constant magnitude of about 10% of rated current. The inverter current
The steady-state constant current characteristic of the rectifier is shown in figure 23.14a as the vertical controller endeavours to control the dc link current to II = I dref – Imargin but the rectifier current controller
section of the characteristic S-T and is labelled S-C-H-T in figure 23.14c. The rectifier delay angle is dominates and maintains the dc current at I dref. In steady-state the rectifier controller overrides the
increased toward αr = 90° if the link current attempts to increase beyond the reference level I dref . During inverter controller which is not able maintain a dc current I dref - Imargin. The inverter current controller only
an attempted short circuit fault, the rectifier delay angle reaches 90° which sets the rectifier output becomes active when the rectifier current controller has reduced its delay angle αr to the minimum limit.
voltage to zero, as shown by equation (23.24), and controls the fault current to I dref as shown by This rectifier minimum delay angle limit is characterised by R-S in figure 23.14c.
trajectory S-T in figure 23.14a. Control characteristic performance
Variations in the ac voltages change the control operating point of the system as follows.
The operating point of the HVDC system is where the rectifier characteristic intersects the inverter i. When the rectifier side ac voltage decreases and/or the inverter side ac voltage increases, then the
characteristics, either at point C or point H on figure 23.14c, depending on which of the two inverter transformer tap adjustment mechanisms on both the rectifier and inverter transformers should attempt to
control methods is being employed. The operating point is tuned over a period of tens of seconds by
remedy this ac voltage regulation problem. If the disturbance is large enough, the new stable operating
adjusting the line-side tap changers of the converter transformers.
The inverter controls the dc-link voltage as follows. point is shown in figure 23.14d. The R-S characteristic falls below points D or E, the operating point will
 The inverter establishes the desired dc voltage Vd by tap changer adjustment, if it is operated in shift from point H to somewhere on the vertical characteristic D-E-F where it is intersected by the
a constant minimum γ mode. lowered minimum αr R-S characteristic as shown in figure 23.14d. The inverter converts to current
 If the inverter is operated in a constant Vd mode, the tap changer is adjusted to produce control, controlling the dc current Id to the value II = I dref - Imargin, approximately 10% of rated current and

constant Vd with an extinction angle slightly larger than the minimum  . the rectifier effectively controls the dc voltage provided it is operating at its minimum delay angle
characteristic R-S. The dc power flow is relatively unaffected and safely returns to the normal operating
The ac-side tap changers on the rectifier end transformers are adjusted so that the delay angle α is condition shown in figure 23.14c, once the ac disturbances have subsided.
small but with a 5° working range, whilst maintain the constant current I dref. If the inverter is constant dc ii. If the rectifier ac voltage increases, the reference current, which is set by the rectifier, is unaffected.
voltage operated at the operating point H, and if the dc current I dref is increased so that the operating Since the link voltage is set by the inverter, any increase in rectifier ac voltage does not affect the power
point H moves towards A, the inverter control mode reverts to constant extinction angle γ control when flow. This can be seen in figure 23.14c where increasing the rectifier characteristic R-S does not affect
operating in the droop region A-B. The voltage Vd droops to less than the desired value, so the inverter the intersection of the operating point C, whence power flow is unaffected by an increase in the rectifier
end transformer tap changer progressively boosts the dc-side voltage until dc voltage control ac-side voltage.
recommences, in region B-C-D. iii. If the inverter side ac voltage is decreased, the link voltage decreases proportionally, but the current
is unaffected because it is set by the rectifier. The power flow is therefore decreased in line with the
Not all HVDC transmission systems use constant dc voltage control, which is the horizontal inverter ac voltage decrease. This can be seen in figure 23.14c where the inverter side ac voltage
characteristic B-H-E in figure 23.14c. Instead, the tap changer in conjunction with the constant extinction decrease will lower the inverter characteristic A-B-C-D. The operating point C, the operating voltage,
angle γ control characteristic A-B-C-D in figure 23.14c, provides dc voltage control. decreases but the current is unaffected. Thus the power transferred is decreased in-line with the inverter
ac-side voltage.
995 Power Electronics Chapter 23 HVDC 996

vi. Worst case conditions are a dc-link short circuit. As seen in figure 23.14a, the rectifier maximum  to maintain adequate commutation margin in inverter operation so that the valves can recover
current is I dref, while figure 23.14b shows that the maximum inverter current is II. The maximum fault their forward blocking capability after conduction before their voltage polarity reverses;
current is therefore limited to the operational current margin Imargin. As seen in figures 23.14a and  to control higher-level parameters such as frequency in isolated mode or provide power
23.14b, the control angle of both converters moves to 90°, which produces 0V converter output oscillation damping to help stabilize the ac network;
voltages, as shown by equation (23.24). Hence the power associated with any short circuit fault is low,  to compensate for loss of a pole, a generator, or an ac transmission circuit by rapid
unlike short circuit faults in ac transmission systems. readjustment of power;
 to ensure stable operation with reliable commutation in the presence of system disturbances;
Voltage dependant current demand limit (VDCDL)  to minimize system losses and converter reactive power consumption; and
If the ac voltages sag significantly because of weak ac systems, it may not be possible to maintain full  to ensure proper operation with fast and stable recovery from ac system faults and
load current. In such a case the dc-link voltage is decreased, and the controller characteristics are disturbances.
dictated by the trajectories X and Y in figure 23.14c. A controller which reduces the maximum current
demand in such conditions, is termed a voltage dependant current demand limiter, or VDCDL. The
23.6 Delta/Delta/Double polygon 18 pulse converter
current is not reduced to zero so that recovery response is faster once the dc-link voltage has sufficiently
recovered.
Six-pulse bridge circuits can be in series or parallel to create an 18-pulse converter. In series, as in
figure 23.16, power source harmonic voltages can be eliminated. The series connection does not need
Power flow reversal inter-phase transformers, unlike parallel connected converters.
The controllers can be designed such that the transition from the rectifier controlling current to the Transformer leakage inductances are an important practical feature in the design. An alternative method
inverter controlling current, is automatic and smooth. That is, seamless automatic power flow reversal is using physically separate phase-shifting transformers is also feasible. In principle this connection can be
achieved by interchanging the inverter and rectifier functions, as seen in figure 23.15. This is realised by extended to higher pulse numbers by increasing the number of polygon windings and modifying the
appropriate control of the delay angles, hence terminal polarities, but the dc link current direction does phase shift accordingly. Practical limitations occur due to reactance effects, efficiency, and the limit on
not reverse. Such a bi-directional power flow requirement may be necessary when two ac systems are available turns ratios. The shift angle is 360° divided by the pulse number.
required to bi-directionally interchange power. The rectifier delay angle is progressively increased while ac input
the inverter delay angle is decreased, such that the rectifier and inverter voltage difference is control to
be virtually constant. This is achieved if αr + αi ≈ 180° is maintained.
Id Id

+ P + - P -
Vdr Vdi Vdr Vdi
Vd x Id -Vd x Id
- - + +
constant
Vdr > Vdi Vdr < Vdi dc current
Rectification inversion inversion Rectification
α < 90° α > 90° α > 90° α < 90°

minimum rectifier minimum rectifier

Operating
+Vd delay angle
point +Vd delay angle
R characteristic R characteristic
S Vd x Id S
A A
D RECTIFIER D
RECTIFIER C
+Pdc minimum extinction +Pdc minimum extinction
Vdi Vdr Vdi Vdr
angle characteristic angle characteristic
Imargin droop Imargin droop
INVERTER INVERTER
 

dc load
F T
 -Pdc T F
 -Pdc
ref
II Id Id IR ref
Id Id
RECTIFIER RECTIFIER
INVERTER +Pdc INVERTER +Pdc
-Pdc -Pdc C

L M  L Current for dc output equivalent to
J
K M
 J  straight-through connection

K
- Vd


-Vd  Operating
point -Vd x Id
(a) (b)

Figure 23.15. Power reversal in HVDC systems by voltage polarity reversal, not current reversal.

23.5.2 HVDC Control objectives

The fundamental objectives of a CSI-based HVDC control system are as follows:

 to control basic system quantities such as dc line current, dc voltage, and transmitted power
accurately and with sufficient speed of response; Figure 23.16. Delta/delta/double polygon for 18 pulse rectifiers.
997 Power Electronics Chapter 23 HVDC 998

Appropriate zig-zag and forked connections could also be employed to provide similar 18 pulse Polygon winding currents
characteristics. However, polygons are easier to use for low-voltage outputs because the higher number i 1m  i 3m  i 1
of turns facilitates selection of the turns ratios.
i 2m  i 1m  i 2
23.6.1 Analysis of Double-Wound Polygon Subtracting these equations gives
Analysis of closed polygon connections tends to be more cumbersome than that for open (fork) i 1m  i 3m  i 2m  i 1m  i 1  i 2
connections. Thus the procedure for the general polygon case is presented. Incorporate i 1m  i 2m  i 3m  0
Let Φ be the phase shift between polygon output line-to-line voltage VLs and the delta input line-to-line i1  i 2
voltage VLin. Let the turns/phase on the delta winding be Np and polygon windings have Nl turns for the i 1m 
long winding and Ns turns for the short winding. Figure 23.17 shows the arrangement for reference. 3
Also
i2  i3 i 3  i1
i 2m  and i 3m 
3 3

In general, the converter input currents consist of current pulses with amplitude Id. They incorporate
harmonics of the form (6k ± 1), which include positive-and negative-sequence currents. To calculate the
rms current in the winding, these harmonics are considered individually then summed (√Σi ) to
2

determine the total rms value. The fundamental and rms currents in each converter line are displaced by
±⅔π depending on their sequence. Thus, each component, hence the total, is multiplied by
½
1  cos  2 3  2   sin 2 3  2 
 
that is, √3, as it flows into the secondary windings. The idealized converter line current is a 120°
conduction square wave with rms value of Id⅔ . Substituting these results into the equation for i1m,
½

yields
2
secondary winding rms current 3 1 I  0.471I d
3 3 d
When load currents; i1, i2, and i3 which are characteristic of a three-phase bridge rectifier, are present,
the line input current to the delta winding is calculated as follows. The sum of ampere-turns on the
transformer core must be zero. Thus
i AN p  i 1mN l  i 3mN s  0
from which
Nl Ns
i A  i 1m i
N p 3m N p
and, similarly
Nl Ns
i C  i 3m i
N p 2m N p
The delta winding line input current is given by
i 1in  i A  i C
Figure 23.17. Double-wound polygon connection phasor diagram. Thus
Nl N
i 1in  i  i   s i 2m  i 3m 
N p 1m 3m Np
Polygon voltages and phase shift
From Figure 23.17
From the polygon vector drawing in figure 23.17 i 1m  i 3m  i 1 and i 2m  i 1m  i 2
VLs V VNl Thus
 Ns  i 2m  i 3m  i 1  i 2
sin120 sin  sin 60  
Thus Therefore
2  Nl Ns  Ns
polygon long winding voltage: VN 1  VLs sin  60    i 1in  i 1   i
3  N p N p  2 N p
and When the voltage ratio output to input is 1:1,
2 Nl N 3 1
polygon short winding voltage: VNs  VLs sin  (4.2)  2 and s 
3 Np 3 Np 6
From the figure 23.17
The phase shift is governed by the ratio of the secondary turns and is determined to be 15°.
VNs sin 60
tan   Using these ratios, the line input current is
VNl VNs cos 60 i 1in  1.1153i 1  0.299i 2 for 15
Thus for a turns ratio given by n = VNl /VNs Extending this analysis, the polygon winding kVA can be determined in respect to the dc load power.
3 The result is
  tan1
 1.209  sin  60     sin  
D.W.polygon winding VA
1  2n
Vdo I d
999 Power Electronics Chapter 23 HVDC 1000

23.7 Filtering and power factor correction Solution

As shown in figure 23.7, both ac and dc side filtering is used to reduce radiated EMC on the dc link and Because of the transformer turns ratio, the transformer dc-side ac voltages are double the ac-side
voltages.
conducted EMC on the ac sides which causes power losses and interference.
i. The inverter delay angle is derived from equation (23.24)
On the dc-side, the large link inductance at each converter (typically between H and H) is 3
supplemented with LC filters, tuned to eliminate selected 12n current harmonics in a 12-pulse system. A Vdi  2 VLL cos i
th 
filter is notch tuned to eliminate (shunt) one specific harmonic, usually the most dominate 12 . At
3
50/60Hz, these filters are capacitive thus provide reactive power absorbed by the converters. 500kV  2 2  220kV  cos i
Harmonics occur at 12n±1 on the ac-side. Again, tuned LC filters eliminate (shunt) specific low order 
harmonics and a general high pass shunt filter is used for components above the 11 and 13 .
th th that is i  147.3
Generally, higher pulse order (>12) transformer/converter arrangements are not attractive in HVDC
because of the difficulties in producing high-voltage transformers (auto-transformers tend to be used). ii. From Pi = Vdi × Id, the link current is
Additional to the VAr compensation provided by the ac harmonic filters, pure capacitance may also be 500  10 6 W
used. In order to avoid overcompensation voltage regulation problems which can occur at low power Id   1000 A
500  10 3 V
transmission levels, the extra capacitance tends to be switched in-circuit as needed.
The rectifier output voltage is given by equation (23.23), rearranged
The main transformers may be provided with ac-side voltage taps to adjust the secondary voltage, as
Vdr  Vdi  I d Rdc
considered in section 23.5.1. The taps are switched automatically by motorised tap-changing drives,
which only operate when large voltage changes occur for prolonged periods of time. If the transformers =  -500kV + 1000A×1Ω = 508kV
have only Y-Y winding configurations, they may also have a low-voltage delta tertiary winding for VAr The rectifier delay angle is derived from equation (23.24)
compensation, provide ancillary supplies, and suppression of transformer core triplen harmonic fluxes. 3
Vdr  2 VLL cos i

Feature summary of Line Commutated Thyristor Systems 3
508kV  2  2  230kV  cos r

With the exception of multi-terminal HVDC the basic steady state operation required of FACTS and yields r  35
HVDC systems may be achieved through the use of thyristor based systems. There are however limits
on the performance and functionality that may be achieved.
iii. The input power is the output plus dc-link resistive losses, that is
Advantages Pr  I d2  Rdc  Pi
a) High voltage thyristor devices give low losses and are available in robust high-current capacity = 10002×8Ω + 500MW = 508MW
single-wafer capsules that fail short circuit. The input VAr from the ac side is
b) Line commutated HVDC and FACTS have an established track record at transmission voltage Qr  Pr tan r
and power levels.
= 508MW  tan35° = 355.7MVAr
Similarly, the inverter VAr into the ac side (indicated by the negative sign) is
Disadvantages
Qi  Pi tan ir
a) Line commutated systems inject significant low frequency harmonics which must be eliminated
by large (physically and electrically) passive filter arrangements. The presence of these filters = 500MW  tan148° = -312.4MVAr
may lead to circulating harmonic currents which must be mitigated by damping networks. The efficiency is
Filters and damping networks may have to be designed specifically for each location and may Pi

not be optimal for all operating conditions. Po
b) Line commutated systems are inherently limited in response time (limited to line frequency 500MW
switching) and may suffer control capability limits. For example thyristor based HVDC: = ×100 = 98.43%
508MW
i. cannot decouple the real and reactive power injected into the network.
ii. may have limited operating range depending on the source impedance provided by the Figure 23.18. HVDC Example 23.1. Id = 1kA
ac network at the point of connection.
c) Line commutated circuits tend to require large passive components leading to large footprint
systems. + VRdc = 8kV +
460kV 440kV
230kV Vdr = 508kV 220kV
50Hz Vdi = 500kV
60Hz
Example 23.1: Basic six-pulse converter based hvdc transmission
Δ:Y Y:Δ
√3:2 - - 2:√3
The basic six-pulse converter dc transmission system represented by figure 23.18 connects a 230kV ac
rms, 50Hz system to a 220kV, 60Hz system. The 6-pulse converters at each transmission line end are
508MW
interface by a Δac-Ydc transformers of turns ratio √3:2, as shown, such that the transformer dc-side line 8MW
500MW
voltage is double the ac side line voltage. The power transmission is 500MW to the inverter which is Rectification inversion
356MVAr αr = 35° αi = 148°
maintained at a dc voltage level of 500kV. The total dc-line resistance is 8Ω. -312MVAr
♣
Determine
i. The inverter delay angle, αi.
ii. The dc-link current, hence rectifier output voltage, thence rectifier delay angle, αr.
The following example is based on example 23.1.
iii. The rectifier input power and VAr, and inverter VAr, thence system efficiency.
1001 Power Electronics Chapter 23 HVDC 1002

Example 23.2: 12-pulse hvdc transmission This is 182MVAr to the ac side of the inverter.

The dc transmission line represented by figure 23.13 connects a 230kV ac rms, 50Hz system to a ii. At 500kV and 600MW:
220kV, 60Hz system. The 12-pulse bipolar converters at each transmission line end are interface by a Pi 600MW
Y-Y transformer of turns ratio 1:1 and a Δ-Y transformer of turns ratio 1:√3, each with a converter side Id    1200A
inductance of 1mH. The rectifier delay angle is α = 30° for 500MW power transmission and the inverter
Vdi 500kV
advance angle is α = 160° (in order to avoid any reactive power increase), which maintains the dc Accounting for the link resistive voltage drop
voltage level at 500kV at the inverter end. Vdr  Vdi  Rdc I d

The total line resistance is 8Ω and the dc link smoothing inductance is large enough to initially consider
 500kV  8  1200A  509.6kV
The efficiency is
the dc current to be ripple free.
Pi VI
Determine   100  i d  100
2 Pr Vr I d
i. The transformer tap ratios at each end, the dc link efficiency, I R losses, and both terminal VAr
ii. If the rectifier tap ratio of 0.866 results in the transmission current limit giving a power of 500kV
600MW, find the delay angle and line efficiency for 500kV at the inverter.   100  98.1%
509.6kV
iii. The value of the dc link inductance Ldc such that the link peak to peak current is 0.1pu the
The necessary rectifier angle, accounting for transformer reactive inductance, is
average load current at full load (1200A), assuming the normalised magnitude of the dc side
3 3Ls 
harmonic V12 is 0.15pu maximum (with respect to the 50Hz supply). Vdr  2   2 ai VLL cos r  I
  d 
Solution
3 3  2    50Hz  1  103 H 
i. From Pi = Vdi × Id, the link current is 509.6  103 V  2   2  0.866  230  103 V cos r   1200A 
  
500  10 6 W
Id   1000 A which gives a rectifier delay angle of αr = 18.5°.
500  10 3 V
The inverter voltage, accounting for the transformer tapping ai is given by an equation similar to iii. The maximum link voltage from the 50Hz rectifier, accounting for leakage at maximum
equation (23.28), that is current is
3 3Ls  3 3Ls 
Vdi  2   2 ai VLL cos r  I Vdr  2   2 VLL cos r  I
  d    d 
3 3  2    60Hz  1  103 H  3 3  2 50Hz  1  103 H 
500  103 V  2   2 ai 220  103 V cos 180  160   1000A   2 2 230kV  cos 0  1200A 
      
which gives a transformer tap ratio at the inverter end of ai = 0.896.  620.5 kV
From equation (23.23), the rectifier voltage is The magnitude of the 600Hz component (12×50Hz) is 15% of 620.5kV, namely 93.1kV, which
Vdr  Vdi  Rdc I d produces a ripple current of 10% of rated current, 1200A, namely 120A. Thus from v = Ldi/dt
t 12
 500kV  8  1000A  508kV Ldc  v 12
i 12
and the necessary transformer tap ratio ar is derive from
1
3Ls
 93.1kV   1.3H
3  12  50Hz  120A
Vdr  2   2 ar VLL cos r  Id 
  
Figure 23.19. HVDC Example 23.2.
3 3  2    50Hz  1  103 H  Id = 1000A
508  10 V  2  
3
2 ar 230  10 V cos 30 
3
1000A 
    Y:Y Y:Y
which gives a transformer tap ratio at the rectifier end of ar = 0. 945. 1:1 + VRdc = + 1:1
230kV 8kV 220kV
50Hz 60Hz
The link efficiency is Vdr = Vdi =
Δ:Y Y:Δ
Pi VI √3:1
508kV 500kV 1:√3
  100  i d  100
Pr Vr I d
- -
500kV
  100  98.4% 508MW 500MW
508kV 8MW
Rectification inversion
293MVAr
The I R losses are 1000A ×8Ω = 8MW or (508-500) /8 Ω, dissipated, distributed along the line. αr = 30° αi = 160°
2 2 2 -182MVAr
W ♣
The rectifier reactive power is given by equation (23.14), that is
Qr  Pr tan r
23.8 VSC-based HVDC
= (500MW + 1000A 2×8Ω)×tan30° = 293.3MVAr
This is 293.3MVAr from to rectifier ac side. Voltage source converters (VSC) for dc-transmission are one of two-level, neutral point clamped or
modular multi-level converters.
The inverter reactive power is given by equation (23.16), that is Voltage source converter-based (VSC) dc-transmission consists of a bipolar two-wire HVDC system with
Qi  Pi tan i self-commutatable converters connected pole-to-pole, as shown in figure 23.20. DC capacitors are used
at each VSC dc-side to provide a stiff dc voltage source. The dc capacitors are grounded at their
= 500MW × tan160° = - 182MVAr
1003 Power Electronics Chapter 23 HVDC 1004

electrical centre point to establish the earth reference potential for filtering and the transmission system. modes can be selected independently at either end of the dc link. Figure 23.20 shows the characteristic
The VSC is effectively mid-point grounded and dc filters and a zero-sequence blocking inductor are ac voltage phasors including the controlled variables Vdc, Id, Q, and VL.
used to mitigate interference on any metallic communication circuits adjacent to the dc cables. There is
no earth return operation. The converters are coupled to the ac system through ac phase inductors and 23.8.2 Power control concept
power transformers, with harmonic filters located between the phase inductor and the transformer. The
ac filters are tuned to multiples of the switching frequency, as shown in figure 23.24. This arrangement dc-link power
minimizes harmonic currents and avoids dc voltage stresses in the transformer, which allows use of a The dc-link power flow concepts are not complicated by ac phasor considerations. No reactive power is
standard ac power transformer for matching the 50/60Hz AC network voltage to the converter ac voltage involved with dc, only the real power flows. Consider the HVDC configuration depicted in figure 23.21
necessary to produce the desired dc transmission voltage. involving converter #1 and converter #2.
Converter DC-link Converter
valves DC cables DC valves Idc
Rdc
capacitors capacitors
AC AC phase AC phase AC
transformer inductors inductors transformer
VL1 XC1 I VL2
C1 IC2 XC2
Vdc1 Vdc2
IL1 IL2

Vac 1 Vac 2 VC1 VC2

Iac 1 Iac 2

Vdc-1 Vdc-2 (a)

AC AC
harmonic harmonic
filters filters

1 2

Vac-ref1 Vac-ref2 
 VC 1I L 1  VC 2I L 2
* *
 I 2

dc
Rdc
Vdc-ref1 Vdc-ref2 jXR1 jXR2

IC1 IC2
ac dc dc ac VL1-n VL2-n
voltage voltage voltage voltage
control PWM control control PWM control VC1-n VC2-n (b)

current current
control Pref 1 Pref 2 control
Qref 1 Qref 2
Figure 23.21. VSC HVDC transmission dc-side, using two-level voltage source inverters.
Figure 23.20. Two level symmetrical mono-polar VSC HVDC transmission.

If converter #1 produces an ac voltage represented by

The IGBT valves used in VSC converters are comprised of series-connected IGBT cells. Present VC 1  VC 1  cos 1  j cos 1 
technology uses 2.5kV IGBT die, with a 4.5kV objectives, parallel connected on a common electrically- (23.29)
 VC 1 1
conducting substrate in 2500A sub-modules, with 30 series connected sub-modules in a string cell.
Strings are then series connected to produce the required link valve voltage requirement. The IGBT while the other, converter #2, is represented by the voltage source
switching frequency is limited to about 2 kHz. The valves are cooled with circulating water and water to VC 2  VC 2  cos  2  j cos  2 
air heat exchangers. (23.30)
 VC 2  2
The structure is constructed to shield electromagnetic interference (EMI) radiation.
Then if the power transmitted equals the power received, then
23.8.1 VSC-Based HVDC control VC 1I L*1 VC 2I L*2   0 (23.31)
Power flow between the VSC and ac network can be controlled by changing the phase angle of the
If the dc link resistive losses are incorporated equations (23.31) becomes
converter ac voltage VS with respect to the filter bus ac voltage VT, whereas the reactive power can be
controlled by changing the magnitude of the fundamental component of the converter ac voltage VT with VC 1I L*1 VC 2I L*2   I dc2 Rdc (23.32)
respect to the filter bus ac voltage VS. By independently controlling these two aspects of the converter
voltage, operation in all four quadrants is possible. This means that the converter can be operated in the ac-side powers
middle of its reactive power range near unity power factor to maintain dynamic reactive power reserve The fundamental base apparent power ST at the filter bus between the converter reactor and the AC
for contingency voltage support similar to a static VAr compensator. This also means that the real power filter is defined as follows (see figure 23.22):
transfer can be changed rapidly without altering the reactive power exchange with the ac network or
S T  PT  jQT  3 VT  I R (23.33)
waiting for switching of shunt compensation.
The active and reactive power components on the grid-side are defined as (see section 25.3 of ac power
Reactive power control can be used for dynamic voltage regulation to support the ac interconnection, by transmission):
synthesising a balanced set of three phase voltages. Black start (restoring system operation without an V V S  sin  V V S  sin 
external energy source) capability is also a feature. P  T  T
L XR
(23.34)
Independent control of the VSC ac voltage magnitude and phase relative to the system voltage V S  cos  VT
decouples the active and reactive power control loops for HVDC system regulation. The active power QT  VT
XR
control loop can be set to control either the active power (dc-link current) or the dc-side voltage. In a dc
link, one station is selected to control the active power while the other controls the dc-side voltage. The where: δ = phase angle between the filtered voltage VT and the converter output voltage VS
reactive power control loop controls either the reactive power or the ac-side voltage. Either of these two L = inductance of the converter ac line inductance
1005 Power Electronics Chapter 23 HVDC 1006

jXR Transformer Idc thermal limit

P(φ)
cable
IT Idc pu
1pu 1pu Rectifier
Idc mode
VXR Power IGBT/diode
transformer valve Vdc
Vdc Power +½
VS-n VT-n VL-n transformer limit Vac
IR XR IT IL Phase reactor Vdc 0.9 pu
XR

Active power
IL +½
1:N
{Im} Q(φ) -½ Vac Q(φ)
VS VT VL 1:N pu pu
1.0 pu
dc
IT VS-n Harmonic VL
ITq capacitance Vac
VXR -½
δ filter 1.1 pu
Harmonic
ITp VT-n {Re} (a) filter Idc cable
(c) limit 1pu Inverter
(b) mode
Figure 23.22. VSC HVDC transmission ac-side. Q consumption reactive power Q generation
Inductive Capacitive
‘under-excited’ ‘over-excited’
Changing the phase angle δ controls the active power flow P between the converter and the filter bus
and consequently between the converter and the ac network. The apparent power is S=P+jQT. Figure 23.24. P-Q active and reactive power control locus, showing varying voltage limits.

As shown in figure 23.23a, for active power flows:

 if the VS phase-lags VT, active power P flows from the ac to the dc side (rectifier) The capacitive reactive power capability increases with decreasing voltage when it is needed most.
 if the VS phase-leads VT, the active power P flows from the dc to the ac side (inverter) Similarly, the inductive reactive power capability increases with increasing network voltage when it is
Changing the amplitude difference between the filter voltage VT, and the converter voltage VS controls needed most. For a given ac system voltage the converter can be operated at any point within its
the reactive power flow between the converter and the ac network. respective ‘circle’. At low ac voltages the VA limit dominates. For high ac voltages, the dc-voltage limit is
restrictive but it is not desirable to inject reactive power when ac voltage already is high.
As shown in figure 23.23b, for reactive power flows:
 if VT > VS, there is reactive power consumed from the ac network
 if VS > VT, there is reactive power generated into the ac network 23.9 HVDC components
Rectifier Inverter Reactive Power
Consumed generated VX The basic structure of VSC based HVDC system is shown in Figure 23.25. The function and design of
VX VX the major power components in this and an LCC system follows.
DC
Reactor
VX VSC
VT VS VS VT Power
PCC Transformer AC Reactor
DC Link DC DC cable or
VS VT VS VT overhead line
δ δ CB Capacitor Filter

IR AC Harmonic
Filter
IR IR
Active power flow IR Reactive power flow
(a) (b)
P Q

Figure 23.23. Active and reactive power phasor diagrams.

Figure 23.25. Symmetrical mono-polar two-level VSC-HVDC transmission system.

With the PWM (Pulse-Width-Modulation, see Section 17.2.3) controlled VSC it is possible to create any
phase angle and voltage amplitude (within limits set by the dc-link voltage magnitude) by changing the i. AC Circuit Breaker, CB
PWM modulation depth and the relative phase displacement respectively, by using phase-locked-loop
grid synchronised displacement. This allows independent control of the active and reactive power. AC circuit breakers are used to connect and disconnect the HVDC system during normal and fault
The typical P-Q diagram (which depends on the ac system voltage), which is valid within the whole conditions. During a dc fault, the ac circuit breaker disconnects the circuit to eliminate feeding the fault
steady-state ac network voltage, is shown in the figure 23.24. This figure illustrates the grid real power, from the ac side. There is no special design or requirements compared to normal ac circuit breakers
used in power systems. Unlike LCC HVDC, a VSC-based system has no inherent capability to clear dc
P, and reactive power, Q, capability of the HVDC VSC converter terminal, measured at the
line faults. A pole-to-pole or pole-to-ground fault will cause the VSC dc capacitor to be discharged and
interconnection point (point of common coupling, PCC), as a function of ac system voltage. Normally
fault current will be maintained via the VSC free-wheeling diodes until the VSC substation ac circuit
converters are sized based on the specific P-Q envelope, Q-V envelope, and dc link voltage.
breakers trip. With a temporary fault, power transmission continues after a normal start-up of the
The 1st and 2nd quadrants represent rectification and the 3rd and 4th inversion. A positive Q indicates
scheme. This can be achieved in the order of 10 seconds.
delivery of reactive power to the ac network. Because the dc-link decouples the two converters, reactive
power can be controlled independently at each station. There are dc-link voltage (Q generation
ii. AC-filters
restrictions) and current (inverter - igbt - power) limitations that have been taken into account in this
Two basic shunting filter characteristics are applicable to HVDC, namely high Q tuned filters (band pass)
typical P-Q diagram. Depending on the cable design, the cable dc current maximum may restrict the
for filtering (shunting) of a specific harmonic and low Q damping filters (high pass) for attenuation
rectifier power limit. There is also a steady state minimum dc-voltage level limit, which may prevent
(shunting) of more than one high order harmonic.
continuous absorption of large amounts of reactive power.
1007 Power Electronics Chapter 23 HVDC 1008

In a typical HVDC scheme, ac filters contain two or three grounded and/or ungrounded tuned filter and should be minimized. The ac filter side of the inductor is effectively at ground at high frequencies
branches. A typical second order filter and its frequency characteristics are shown in figure 23.26. AC and the capacitance across the inductor should therefore be low. To achieve low capacitance the
filters for VSC HVDC converters have lower ratings than those for conventional HVDC converters and converter inductors have air coils (with a magnetic return circuit) rather than magnetic cores.
are not required for reactive power compensation. These filters are permanently connected to the
converter bus and not switched with transmission loading, as in conventional HVDC. iv. Power transformer
Because of the use of the converter ac inductors and the use of VSC PWM, the current in the
Voltage source converters can be operated with different control schemes, most of which use pulse transformer windings contains minimal harmonics and is not exposed to any dc voltage. The transformer
width modulation to control the ratios (magnitude and phase) between dc and ac side fundamental is a number of configured 50/60Hz, two-winding limb, single transformers (because of transportation
frequency voltages. At the ac side converter terminal, the voltage to ground is not sinusoidal. The most limitations (size and weight) and redundancy) or a three phase power transformer, with taps and a tap-
frequent application of voltage source converters is as a machine drive (in industrial applications) where changer. The secondary voltage, the filter bus voltage, is controlled by the tap changer to achieve the
this is of less concern. However, connecting a voltage source converter to a transmission or distribution maximum active and reactive power from the VSC, both consumed and generated. The tap changer is
system requires the voltage to be sinusoidal. This necessary filtering is achieved by means of the located on the secondary side, which has the largest voltage swing, and also to ensure that the ratio
converter inductor and ac filters. between the outer radial line winding and a possible tertiary winding is fixed. This tertiary winding feeds
the station auxiliary power system and if delta connected suppresses any core triplen fluxes.
Ω In order to maximize the active power transfer, the converter generates a low frequency zero-sequence
C1 voltage (<0.2pu) which is blocked by the ungrounded transformer secondary winding.
500 The power transformer provides the following features:
 Reactance is located between the VSC and the ac system, which decreases the ac filter size.
C2 C1 L1
400  Match the voltage between the ac system and the VSC AC terminals, resulting in optimum switch
300 rating utilization.
 Eliminate the 5 and 7 harmonics by connecting the transformer in 12-pulse Yyd configuration.
th th
Rd
L1
200  Prevents zero sequence current flow between the ac system and the VSC.
L2 Rd L2 C2
100
Transformer should sustain:
0  The fundamental current stresses
0 10 20 30 40 50  Saturation characteristics of the transformer magnetic field
Order of harmonic
 Low and middle frequency harmonics stresses
 Dielectric stresses due to normal/transient operating voltage, especially with VSC PWM, and
Figure 23.26. Single phase equivalent circuit diagram and harmonic impedance characteristic of a
harmonics. The extending composite value-side bushing is tested to higher dc stress levels.
double tuned filter.
v. DC-link Capacitors
The distorted waveform of the converter terminal voltage can be described as a series of harmonic The dc capacitor is the stored energy element for a VSC, providing it with a stiff dc voltage which is
voltages essential presumption with VSC topologies. The primary functions of the dc link side capacitor are:
 Provide a low-inductance path for switch turn-off current. Due to the stray inductance, the turn-off
V S  V h cos h ot  h 
h 1
commutating current results in transient voltage stresses on the switching devices. These stresses
th
where Vh is the h harmonic voltage. are minimised by reducing connection lengths between the dc link capacitor and the switching
devices.
 Temporary energy store, as it stabilises connection operation and facilitates closed loop control
The magnitude of the harmonic voltages vary with the dc voltage, the switching frequency (or pulse
adjustment.
number) of the converter and converter PWM technique.
 Reduce harmonic dc link voltage ripple. HVDC capacitors use a dry, self-healing, metallised film
Harmonic cancellation PWM or optimal PWM utilizes a sinusoidal PWM with 3rd harmonic (triplen) design, which have advantages over oil filled technologies. A dry capacitor design offers long life,
injection. A 3rd harmonic is added to the fundamental frequency to increase the voltage, hence power high capacity, and low inductance, in a non-corrosion, non-radiating plastic housing.
rating of the converter, as considered in chapter 17.1.3vii.  Decreases harmonic coupling between different VSC substations connected to the same dc link.
iii. Converter line inductors The dc link capacitor design parameters are:
There is one converter oil filled inductor per phase, with EMC shields to eliminate magnetic fields
 Continuous dc operation voltage;
outside the windings. The converter line inductor is a key components in a voltage source converter with
which continuous and independent control of active and reactive power is possible.  Limits of dc voltage ripple under transient conditions, such as ac system faults;
The main purposes of the converter ac inductors are:  Harmonic currents conducted to the dc side; and
 Peak discharge current during dc faults.
 to provide low-pass filtering of the PWM voltage to give the desired fundamental
frequency voltage. The converter generates harmonics related to the switching fre- A small voltage ripple requirement implies large capacitance: while small capacitance offers fast active
quency. The harmonic currents are blocked by the converter inductor and the harmonic power control. DC capacitor sizing is a trade-off between voltage ripple, lifetime, and fast dc link control.
content on the ac bus voltage is reduced by an ac shunt filter, which diverts the Thus dc link capacitor design bounds is a function of two inequalities:
harmonic currents; 2 S S
 to provide active and reactive power control. The fundamental frequency voltage across  C dc 
the inductor defines the power flow (both active and reactive) between the ac and dc Vdc2 2Vdc Vdc
sides. The P-Q diagram in figure 23.23 shows the active and reactive power definitions;
 to limit line short-circuit currents. The short-circuit voltage of the converter inductor is where CDC dc-link capacitance, S apparent converter power, VDC rated dc-link voltage, ω electrical
typically 15%; frequency, ΔVDC allowed voltage ripple (peak-peak), and τ is the time constant of dc capacitor charging.
 to prevent dc circuit resonance; and The time constant τ is selected to be less than 10 ms to satisfy low dc-link voltage ripple.
 minimises intermittent dc-link current flow.
vi. DC-filters
The stray capacitance across the line inductor is minimised in order to minimize the harmonics coupled Instead of over-sizing the dc link capacitor, a series LC filter can be used across the link capacitor to
into the ac filter side of the inductor. The high dv/dt on the bridge terminals at switching results in current eliminate targeted harmonics. It is parallel connected to the dc link capacitor to decrease the total
pulses (i=C dv/dt) through all stray capacitances to ground. As these currents pass through the valve, equivalent impedance of the dc circuit. This reduces harmonic currents injected into the dc line.
1009 Power Electronics Chapter 23 HVDC 1010

vii. DC-series reactor The outer sheath on land cables is normally a thermoplastic polyethylene (PE) sheath
In VSC-HVDC systems, a dc reactor (air-core coils with magnetic return paths of iron-core coils with air or an extruded halogen PVC flame-retardant sheath. PE is a harder material offering
slots) may be connected after the dc capacitor to: better mechanical protection.
 Reduce the dc fault current gradient in order to protect VSC free wheeling diodes; For long term physical cable integrity, the surface of the outer sheath may be provided
 Harmonic current reduction in the dc overhead line or cable; and with a thin conductive layer, which is simultaneously extruded with, and thus strongly
 Critical resonance tuning within the dc circuit. bonded to, the non-conductive underlying jacket.

A dc reactor in a VSC-HVDC system is considerably smaller than that used in an LCC HVDC scheme Additionally optic fibre cable for communication purposes, may be laid with the bundled.
and typically less than 10mH. To limit dc fault current levels, link inductances approaching those found
in LCC HVDC systems tend to be specified. The most common cable technologies include:
 Mass-impregnated (MI) insulation cables
viii. Overhead and subsea/underground HVDC cables  Low-pressure oil-filled (LPOF) cables
To transmit electric energy using direct current over a distance, both cables and overhead transmission  Extruded cross-linked polyethylene (XLPE) cables
lines can be used. There is also a small difference between underground and submarine dc cables. The
cost of cables is considerably larger than overhead lines but there are also other aspects associated a) Mass-impregnated single core mass cables
with the basic principle of VSC transmission that may influence the choice: MI insulation is made by winding kraft processed paper around the conductor until the required
insulation thickness is achieved. The cable is then placed in a vacuum chamber and heated to remove
 Cables often have less impact on the environment than overhead transmission lines. any air and moisture. A viscous oil is introduced which impregnates the paper, increasing the electrical
 It is considerably simpler to obtain right of way for new cables. strength of the insulation. A recent development uses polypropylene laminated papers, where thin layers
 Since a VSC produces only one dc voltage polarity, the cable does not need to be designed for of plastic are laminated to the paper layers, allowing the cable to operate at higher temperatures and
voltage polarity reversal. This allows less expensive and simpler cables, such as extruded XLPE therefore higher current ratings. HVDC,
dc cables, to be used in long distance VSC transmission systems.
 Cables are less prone to faults than overhead transmission lines. Since overhead transmission For bulk power transmission, mass impregnated cables prove to be the most suitable solution because
2
lines are exposed to lightning strikes and pollution, faults along them are likely. Most line outages of their capacity to operate up to ±600 kV dc (1600A, 2500mm ) and transmission capacity over 1000
are transient, with transmission recommencing once air insulation is restored. Permanent faults on MW for monopole solutions. Therefore an HVDC in bipolar connection is able to transmit up to 2GW with
overhead lines are easier to repair. In the event of a cable fault, the outage would be permanent MI cables. These cables can be installed at depths to 1000m under the sea level and with nearly
and the repair time may be significant. unlimited transmission length. The system capacity is limited by the conductor temperature, which can
 VSC HVDC has difficulties coping with dc faults and most existing links operate with dc cables. reduce overload capability. Most submarine cables are the mass-impregnated type. Figure 23.27a
Manufacturers have been reluctant to offer VSC HVDC with overhead lines, and the 2010 Caprivi shows a photo of mass impregnated cable manufactured by ABB.
VSC HVDC link is the only such installation. Also these cables tolerate fast dc voltage polarity reversal making them suitable for LCC HVDC.

The choice between cables and overhead lines is influenced by environmental constraints as well as an b) Low pressure oil filled (LPOF) cables
overall optimization that considers total capital cost and transmission system reliability. LPOF cables are insulated with paper impregnated with low viscosity oil under an overpressure of a few
bars. This ensures voltages up to 500 kV and capabilities up to 1000 MW for underground installations.
DC cable types It can be used for both ac and dc transmission applications. Since oil flow is required along the cable,
HVDC cables are constructed as follows: the cable length is limited to around 80 km. Environmentally, the risk of oil leakage must be taken into
 Conductor account.
The shape of the conductor is round and built up of compacted stranded round (copper
or aluminium) wires or, for large cross-sections, concentric layers of keystone shaped c) Extruded Cross linked polyethylene (XLPE) cables
wires. XLPE cables offer a high working temperature, and use a lapped non-impregnated thin polyethylene film
 Insulation system insulator extruded over the cable core, to allow long and deep submarine installation, associated with
The HVDC polymeric insulation system consists of: increased electrical operational stresses. A bonded metal-polyethylene laminate can be bonded can
- Conductor screen help ensure cable physical integrity.
- Insulation
- Insulation screen XLPE cables are available up to ±320kV dc, giving 1000MW per bipolar, with development aimed at
the insulation system is triple-extruded and dry-cured. (thickness is operating voltage increasing the voltage rating of cable technology and reducing its cost. This will make VSC transmission
dependant). more attractive for a number of applications. Most underground cable insulators are made from extruded
 Metallic screen insulation. Figure 23.27b shows a XLPE cable. XLPE cables are unsuitable for use in CSC applications
Copper wire screen, for land cables, with cross-section design for fault currents. where, in order to reverse the direction of power flow, the polarity must be reversed.
 Metallic sheath
A lead alloy sheath is provided for submarine cables, while a metal-polyethylene
laminate may be provided for land cables. The laminate is bonded to the polyethylene, Underground and submarine cables
which gives mechanical properties. Three types of cables are used for underground and submarine cables as shown in Table 23.1. The
 Inner jacket (for submarine cables) different between the underground and submarine cables are the conductor material and the armour
A polyethylene sheath is extruded over the lead sheath to provide mechanical and layer, which is used in submarine cables to withstand the axial mechanical tension during laying and
corrosion protection for the lead sheath. operation.
 Tensile armour (for submarine cables)
The tensile armour consists of galvanized round steel wires close to each other helically Cables with copper conductors are used for submarine applications and aluminium conductors for land.
twisted round the cable, which is flooded with bitumen for corrosion protection. The Cable ratings for bipolar VSC-HVDC are double those shown in table 23.1. Cu has high electrical
tensile armour is needed when the cable is laid in the sea. The tensile armour also conductivity and mechanical properties; being easy to create strong joints. However it is heavy and more
offers mechanical protection against impacts and abrasion for a cable that is sufficiently expensive, thus is used when its mechanical properties are mandatory, such as submarine cables. Aℓ
buried in the seabed. has low conductivity and low mechanical properties; being difficult to implement joints. It is lighter and
 Outer sheath cheaper than Cu.
The outer serving for submarine cables consists of two layers of polypropylene yarn, the
inner one impregnated with bitumen. The polypropylene yarn is a semi-wet covering.
1011 Power Electronics Chapter 23 HVDC 1012

Table 23.1. Cables for underground and submarine application DC cable environmental factors
Electro Magnetic Fields (EMF)
Type Mass-impregnated LPOF XLPE DC power transmission, via underground cable, generates low frequency magnetic fields in the
range of 50μT, which is the same order as the background static magnetic field of the earth,
Conductor Cu/Aℓ Cu/Aℓ Cu/Aℓ
which are 8,000 times lower than the 400mT deemed the accepted threshold assessed as no
Insulation Paper and mass Paper and fluid Cross linked PE risk to human health.
Voltage kV 600 500 320 Magnetic field
HVDC cables are laid in pairs, close together, carrying the same current levels in opposite
Maximum conductor size mm2 3000 3000 2400 directions, so the overlapping static magnetic fields tend to cancel each other out. The resulting
Maximum current A 2000 2000 2700 magnetic field decreases rapidly to zero with distance from the cables.
Acoustic noise
Capacity per Cable MW 1000 600 1000
Unlike ac transmission lines, HVDC cables do not make noise, unlike all converter stations.
Weight kg/m 30-60 40-80 15-61 Ground heating
Converter type LCC or VSC LCC or VSC VSC or unidirectional LCC During operation, the temperature of an underground cable will rise dependent on the current
carried and load factor. Heat distribution to surrounding soil depends on the backfill material.
The impact of heat release on soil temperature is strictly local and very limited. Typically 1°C
above the line at 1m depth.

Conductor ix. HVDC dc circuit breaker

copper or aluminium High-current high-voltage dc circuit breakers are problematic. Semiconductor approaches involve series
connection of many devices, resulting voltage sharing problems and high on-state losses. Hybrid
Conductor screen
semi-conductive polymer
approaches use a combination of semiconductors for turn-off speed and vacuum circuit breakers for low
losses. The utilisation of dc circuit breakers is problematic in a multi-terminal hvdc system since clearing
Insulation and isolation are required both sides of the dc fault. Polymer or porcelain housed metal oxide (ZnO)
dry cured hvdc polymer surge arresters are vital in all hvdc systems, and are extensively used in all dc circuit breakers.
insulation screen
semi-conductive polymer
Example 23.3: HVDC transmission with voltage source controlled dc-link #1

Metallic screen swelling tape A 200kV, 1000A, HVDC is being considered as interconnection between two ac systems as shown in
copper wires Figure 23.29. As shown the HVDC converter ratings are 200MVA. The 50Hz load inverter grid draws a
total of 150MVA at a 0.75 power factor, while the rectifier VSC draws the power requirement from the
swelling tape lead alloy sheath
60Hz network at unit power factor. Sinusoidal Pulse Width Modulation (SPWM) is used with triangular
2
carrier wave with a frequency 1350 Hz. The copper dc cable is 1000 mm , with total distance 100 km
aluminium inner jacket
Assuming copper has specific resistance of R=1.39×10 Ω /km, determine:
-2
laminate polyethylene
i. Transformer turns ratio for the rectifier and inverter transformers.
tensile armour ii. DC cable losses.
outer covering galvanised iii. Active and reactive power delivered at the inverter side.
polyethylene steel wire iv. Active and reactive power drawn at the rectifier side.
outer covering
polyethylene Rectifier
200kV dc Inverter
Converter Station #1 Converter Station #2
(a) (b) (c) PCC PCC
#1 DC cable / overhead #2
line
Figure 12.27. ABB HVDC cables: VDC
(a) mass impregnated, MI, (b) XLPE land and (c) XLPE submarine. VDC
200 MVA 200 MVA
16 % 10 %
230 kV 320 kV
DC cable modelling 60 Hz
Figure 23.29. VSC HVDC system for Example 23.3. 50 Hz
In basic low frequency dynamic studies, dc cables are represented as lumped T or π models as shown
in Figure 23.28. The cable parameters are a function of the cable length. For example, the parameters
of a 500 kV, 2000 A dc cable are: R=0.05 Ω/km, C= 0.26 µF/km. Solution
In more detailed studies, for example, mid cable dc faults, a distributed parameter model is used.
i. On the inverter side:
R ½R
½R ½R
½R Vdc 50Hz  200kV
R
Vdc 50Hz
VLN 50Hz   70.721kV
½C
½C ½C
½C C
C 2 2
VLN 50Hz at star side  3VLN 50Hz  122.347kV
(a) (b)

Figure 23.28. Basic cable: (a) T model and (b) π model.

1013 Power Electronics Chapter 23 HVDC 1014

V pr2 320kV 2
Ltr  X tr  0.10  51.5 Solution
St 200MVA
S 50Hz 200MVA i. The total active 50Hz power at the load end is
I L 50Hz    361.271A
P50Hz  S 50Hz  pf 50Hz
3 VL 50Hz 3  320kV
 100MVA  0.75  75MW
VL N conv50Hz est  VL N grid 50Hz  j X L I L The 50Hz reactive power flow into the dc link is
Q  VAr 2  P 2
 320 / 3  j 51.5  361.271  185kV  j 18.497kV
 203.497kV  1002  752  66.1MVAr
The transformer turns ratio is 203.497kV/122.347kV = 1.663 The 50Hz ac line current is
S 50Hz
I L 50Hz 
On the rectifier side: 3 VL 50Hz
Vdc 60Hz  Vdc 50Hz  I dc Rdc  200kV  1000A  1.39  201.39kV
100MVA
V   441A
VLN 60Hz  dc 60Hz  71.212kV 3  130kV
2 2 The necessary phasor diagram is referenced with respect to neutral.
VLN 60Hz _ at _ star _ side  3VLN 60Hz  123.198kV
The reference 50Hz phase voltage is
VL 50Hz
V2 230kV 2 VL N 50Hz 
Ltr  X tr pr  0.16  42.32 3
St 200MVA
130kV
S 60Hz 200MVA   75.06kV  0
I L 60Hz    502.63A 3
3 VL 60Hz 3  230kV The angle between the line current and the phase voltage is
VL N conv 60Hz est  VL Ngrid 60Hz  jX L I L   cos1 0.75  41.4
 230kV / 3  j 42.32  502.63A  133kV  j 21.271kV  111.729kV The line reactance at the 50Hz end is
X L 50Hz  2 f  L
The transformer turns ratio is 111.729kV/123.198kV = 0.9
 2  50Hz  25mH = 7.85
ii. DC cable losses The 50 Hz converter line to neutral voltage (and load angle) is
Rdc  2Ro Cu Cu  2  1.39  102  100km  2.78 VL N conv50Hz  VL N 50Hz  jX L I L
Pdc loss  I dc2 Rdc  1000A 2  2.78  2.78MW  1.39% of 200MW  75kV  j 7.85  441A  41.4
 75kV + 3.4548.6  77.3kV + j2.59
iii. The total active 50Hz power at the inverter is  77.34kV  1.92
P50Hz  S 50Hz  pf 50Hz The line-to-line voltage at the 50Hz VSC converter is
 150MVA  0.75  112.5MW VL  3VL N
The 50Hz reactive power flow into the DC link is  3  77.34kV = 134kV
Q  VAr 2  P 2
ii. The ±80kV dc-link has a pole-to-pole voltage of 160kV. From Pi = Vdi × Id, the link current is
 1502  112.52  99.21MVAr
75MW
Id   468.75A
160kV
iv. At rectifier side: Prec = 112.5MW + 2.78MW = 115.28MW The cable voltage drop is
Qrec = 0
V dc  I d  Rdc
♣
 468.25A  2  0.9375kV
The VSC converter dc voltage at the 60Hz transmitting end is 160kV + 0.938kV = 160.938kV such
Example 23.4: HVDC transmission with voltage source controlled dc-link #2 the pole voltages are ±80.469kV. These results are shown in figure 23.30.

The VSC dc-voltage transmission line represented by figure 23.26 connects a 130kV ac rms, 50Hz The VSC power delivered into the link is
three-phase system to a 120kV, 60Hz three-phase system, some 100km apart. The voltage-controlled Pdc 60Hz  I dVdc 60Hz
converter at each nominal ±80kV transmission line end are interface by a Δ-Y transformer with 25mH of  468A  2  80.47kV  75.36MW
line and leakage reactance referred to the grid side, and with a turns ratio to match the converter voltage With unity power at the 60Hz sending end P = S, the line current is
2
requirements for the ±80kV dc link. The cable is rated at 625A with 300mm of copper giving a total line S 60Hz P
resistance is 2Ω. I L 60Hz   60Hz
3VL 3VL
The 50Hz load end grid draws a total of 100MVA at a 0.75 power factor, while source end VSC draws 75.36MW
  362.6A
the power requirement from the 60Hz network at unit power factor. 3  120kV

Determine the phasor diagram for: For the 60Hz transmission end phase or diagram, the line to neutral voltage is required from the
i. the 50Hz receiving end 120kV line voltage at the 60Hz end, thus
ii. the 60Hz transmitting end.
1015 Power Electronics Chapter 23 HVDC 1016
±500MW
±100MVAr
V L 60Hz
V L N 60Hz 
3
120kV
  69.3kV
3
The 60Hz line reactance is
Figure 20.31. VSC HVDC terminal in example 23.5.
X L 60Hz  2 f  L60Hz
 2  60Hz  25mH = 9.45 Solution
The 60Hz VSC converter side-line to neutral voltage is therefore
VL Nconv 60Hz  VL N 60Hz  jX L 60Hz I L 60Hz i. The total converter (apparent) power:
 69.3kV  j 9.45  362.6A0 S c  Pc2  Qc2
 69.3kV  3.426kV   90
500  10   100  10 
2 2
 6 6
 510MVA
 69.385kV  2.83
whence the converter current is:
The line voltage at the sending converter is
Sc
VLconv 60Hz  3VL Nconv 60Hz Ic 
3Vc
 3  69.385kV = 120.18kV
510MVA
  765A
75.36MW 75.36MW 75.36MW 75MW 75MW 75MW 3  385  103

800kW ii. From the SCR expression

0MVAr 1.24MVAr 1.53MVAr 66.1MVAr
Vs2 275, 0002
z sm    6.3
120kV
130kV SCR  Pdc 24  500  106
50Hz
60Hz
The ac grid resistance is
363A 470A 2Ω 441A
a 1 zg
Rg 
1   X /R 
+ + 2
120kV 120.2kV 134kV 130kV
161.2kV
b 2 6.3
  0.5234
25mH + 160kV + 25mH 1  122
9.45Ω 7.85Ω
and the ac grid reactance is
c 3
X g  Rg  X / R
 0.5234  12  6.28
100km
69.3kV 75kV Hence the ac grid impedance, referred to grid side, is:
69.4kV 77.3kV
z s  Rg  ( X g  X t )  j X i / nt2
N
100MVA 275kV 2
pf = ¾  0.5234  (6.28  0.11 )  j 2 50  0.03  ( 275kV / 385kV )2
600MVA
 6.302
IL =363A VL-N = 77.3kV
69.3kV
VXL =
Note that the dc side inductor has no impact on the fault current.
-2.83°
1.92° 3.45kV
Converter ac fault current is:
VXL = -41.4° VL-N =
75kV
Vg / 3
69.4kV
3.43kV
IL =441A I ac fault   ( 275kV / 385kV )
zs
275kV / 3
60Hz transmitting end phasor diagram   (275 kV/ 385kV )  4544A
6.302
50Hz receiving end phasor diagram
Figure 23.30. HVDC VSC Example 23.4. This fault current is almost 6 times the rated current, 765A. Typical IGBT reverse diodes would take this
♣ current for less than 20ms. This might be too short time for ac CB to clear the fault. In this HVDC system
the phase reactors Li would need to be increased, if possible.
Example 23.5: HVDC transmission with voltage source controlled dc-link #3
iii. The dc fault current is:
Figure 23.31 shows one terminal of a 500MW VSC HVDC system. If the converter is rated at Pdc = 
500MW and Q = ±100MVAr, determine: I dc fault  I ac fault
6
i. the rms current in switches,
ii. the rms current in switches for a dc fault near dc bus, as indicated. 
 4544A  5828A
Comment on the result, considering the converter switches. 6
iii. Hence the dc fault current in the dc cable for this RMS fault current. ♣
1017 Power Electronics Chapter 23 HVDC 1018

Example 23.6: HVDC transmission with voltage source controlled dc-link #4 Vc max  Vcd2 Vcq2  60kV 2  287kV 2  293.3kV
Consider the same VSC HVDC terminal as in the previous example (figure 23.31), assuming the PCC Vac Vc max 275kV  293.3kV
bus voltage is controlled to 275kV. Vmax %  100  100  3.2%
Vac 275kV
Determine: This voltage swing, 1.7% - -3.2%=4.9%, is relatively small.
i. the converter AC voltage for maximum power demand
a) Pc=+500MW, Qc=+100MVAr ii. Repeating the same procedure with Li =0.12H:
b) Pc= -500MW, Qc= -100MVAr Vmin %  1.9%
Refer the converter voltage to grid side and determine the percentage voltage swing. Vmax %  6.6%
ii. the maximum ac voltage swing, assuming the phase reactance is increased 4 times to contain
The total voltage swing is 1.9% - -6.6%=8.5%, which is acceptable, but should account for when
dc faults.
dimensioning equipment.
Solution ♣
Assuming the reference coordinate frame is linked to the PCC voltage, then the converter current is:
1 Vac Vc 1 Vac Vcx  jVcy 23.10 Twelve-pulse transformer based HVDC
Ic  
3 j X t  X i  3 j X t  X i 
Figure 23.32 show a HVDC configuration that takes advantage of a star-delta, phase-shifting,
1 Vcy 1 Vac Vcx transformer secondary to halve the series IGBT connection problems associate with high-voltage
I cx  jI cx  j th th
3 Xt  Xi 3 Xt  Xi applications. The twelve-pulse transformer arrangement also reduces the 5 and 7 ac harmonics due
The PCC power is: to the cancellation effect because of the 30° phase shift between the two secondary windings. These
advantages are traded against the complexity associated with an extra high-voltage transformer
Pc  jQc  3Vac I cx  j 3V ac I cx winding. A VSC variation is to use multilevel configurations involving chained cells, called a capacitor
VacVcy
2
V Vcx clamped multilevel M C converter. The cell comprise is a capacitor of large capacitance, which is
  jVac ac switched in and out of the series string circuit, as introduced in chapter 17.3.4 and considered in 23.13.
Xt  Xi Xt  Xi
Therefore the voltage components are 3φ AC Converter DC-link Converter 3φ AC
Q X  X i  transformer valves DC cables DC valves transformer
Vc d   ac t Vac AC phase capacitors capacitors AC phase
Vac AC inductors inductors AC
dc inductor dc inductor
P X  X i 
harmonic harmonic
and dc filter and dc filter
filters filters
Vc q   ac t
Vac

i. a)
Qac  X t  X i 
Vcd   Vac
Vac
 275kV 2 
100MVAr  0.11  2 50  0.03  ( 275kV / 385kV )2 
  600MVA   275kV
275kV
-12kV  275kV  263kV
P X  X i 
Vcq   ac t
Vac
Figure 23.32. HVDC NPC VSC using twelve-pulse transformer arrangement.
 275kV 2 
500MW  0.11  2 50  0.03  ( 275kV / 385kV )2 
  600MVA  2
23.11 VSC-HVDC transmission systems based on the modular multilevel converter, M C
275kV
 60kV 2
The M C approach exploits the potential of a multilevel structure and pulse width modulation. The
filtering requirements are reduced due to generation of high quality ac voltage (small filters are required).
Vc min  Vcd2 Vcq2  600002  2630002  269.76kV The use of a large number of levels with small voltage steps, results in low dv/dt and reduced voltage
stress on the insulation of the interfacing transformers. This allows the use of standard 50/60Hz
Vac Vc min 275kV  269.76kV transformers without need to withstand dc link voltage or harmonic currents. The effective switching
Vmin %  100  100  1.7%
Vac 275kV frequency per device is low, which reduces the switching losses and overall conversion losses, and
does not produce significant high frequency noise. Fig. 23.33 shows one-phase of a 7-level modular
i. b) With Pac and Qac reversed converter, which relies on the cell capacitors to create a multilevel voltage waveform at the converter
Q ac  X t  X i  terminal. As the number of levels increases, it can generate high voltage with low harmonic content,
Vcd   Vac using medium voltage devices (2.5kV or 4.5kV). DC link capacitors are not required, which significantly
Vac reduces the space requirement for the converter station. As the converter depends on phase voltage
12kV  275kV  287kV redundancy to maintain approximately constant voltage across the cell capacitors, the modular
Pac  X t  X i  multilevel converter performs better than the NPC converter during unbalanced operation, and
Vcq   symmetrical and asymmetrical faults, without increasing the risk of device failure or system collapse.
Vac The ability of the modular converter to ride through different types of ac faults makes it attractive in
 60kV power system applications where restrictive grid codes must be met. The absence of the dc link
1019 Power Electronics Chapter 23 HVDC 1020

capacitors in VSC-HVDC transmission system based on the modular converter, make sizing the cell Table 23.2: Comparison between different HVDC technologies
capacitors critical. They are sized to be able to store enough energy to support the converter dc link
during transient events, otherwise the system may fail to meet transient ride-through requirements. The
Current source converter based Voltage source converter based
first commercial modular multilevel converter based HVDC transmission system project is the Trans Bay
LCC CCC Neutral-point clamped Modular (M2C)
cable project in USA, rated at 400MW active power transfer and ±170MVAr STATCOM functionality with line commutated converter capacitor commutated converter
Two-level
(3 level) (half bridge)
a bipolar dc link voltage of ±200kV. The line is 85km of submarine cable.
Switching device Thyristor Thyristor IGBT IGBT IGBT

For this type of multi-level converter, the cell capacitors experience the full load current, unlike the rail switching losses negligible negligible high moderate low
capacitor of a two level system. High capacitance and large physical size result. This capacitance is the
On-state losses low low moderate moderate moderate
dominant factor in the size of this type of converter. But the main dc link capacitors in both ends of the
line are eliminated. There is dc current component in the leg current and no dc link fault protection is Reduced due elimination of
Significantly reduced Significantly reduced Significantly reduced
Station size large switch capacitors and their
offered. circuit breakers
(30% to 50% of LCC) (30% to 50% of LCC) (30% to 50% of LCC)

2
Active continuous from ±10% to continuous from ±10% to Continuous Continuous Continuous
The essential features of the m c are power control ±100% ±100% From 0 to ±100% From 0 to ±100% From 0 to ±100%
 A highly modular construction both in the power section and in control and protection, allowing a Active
DC voltage polarity must be DC voltage polarity must be Instantaneous and no Instantaneous and no Instantaneous and no
changed and not changed and not change of DC voltage change of DC voltage change of DC voltage
converter hall with a small footprint or a building with a low profile. power reversal
instantaneous (0.5-1s) instantaneous (0.5-1s) polarity polarity polarity
 In normal operation, only one level per converter arm switches at any given time. Thus the ac Independent
voltages can be adjusted in small increments and a dc voltage with little ripple can be achieved, control of active no no yes yes yes
which minimizes the level of generated harmonics and in most cases eliminates the need for ac and reactive power
filters. Additionally, the small output voltage steps produce minimal radiant or conducted high- Reactive power 20% to 50%, with additional
50% to 60% no no no
demand series capacitance
frequency interference. Limited (lagging VAr only) Continuous and Continuous and inherent Continuous and
 The low switching frequency of the individual semiconductors results in low switching losses, thus Reactive power and discontinuous using inherent within the within the converter inherent within the
Limited (lagging VAr only)
resulting in high system efficiency. capability switch shunt capacitors for converter control at no control at no additional converter control at
leading VAr additional cost cost no additional cost
 Uses standard proven components.
 The produced voltages and currents (low di/dt and dv/dt) allow the use of standard ac transformers. Power levels Up to 6400MW Up to 6400MW Up to 1200MW Up to 1200MW Up to 1200MW
 Fast system response to faults, particularly ac faults. Controllability - Low Low High High High
Response time (0.1-0.2s) (0.1-0.2s) (0.03-0.05s) (0.03-0.05s) (0.03-0.05s)
AC filters large large small < 2-level converter no
+½Vdc +½Vdc DC filter yes yes no no no
1 C Expensive with high Expensive with high
Expensive with high insulation requirement insulation requirement to
Expensive with high Cheap with standard
insulation requirement to to withstand switching withstand switching of
Converter insulation requirement to insulation.
C 2 withstand voltage stresses of large voltage steps large voltage steps with
transformer withstand voltage stresses Additional large AC
during power reversal, but with high frequency. high frequency.
during power reversal inductor is needed.
with reduced MVA rating Additional large AC Additional large AC
C 3 inductor is needed. inductor is needed.
Expensive with high Expensive with high
VAC DC cable
insulation requirement to insulation requirement to Cheap and light weight Cheap and light weight Cheap and light
C 4 withstand voltage stresses withstand voltage stresses extruded cable extruded cable weight extruded cable
during power reversal during power reversal
Due to voltage distortion
C 5 Commutation
and lack or reactive power Significantly reduced no no no
failure
during AC fault
Lp
PCC
ωt Connection of strong
Vdc VAC systems only (SCR>2.5), Independent of
Independent of systems Independent of systems
Lp connection of weak system Connection of strong and systems strength and
Applications strength and network strength and network
is possible but at additional weak systems with SCR≥1 network without
without generation without generation
6 cost by using STATCOM or generation
synchronous condenser
Very good, but its ability
7 to cope with
AC fault
AC filters Possible with high risk of Possible with reduced risk asymmetrical faults
ride-through excellent excellent
commutation failure of commutation failure depends on capacitor
8 capability
voltage balancing
strategy used
9
Possible with
Possible with combinations of
Possible with increasing
increasing the inherent inductances
DC fault ride- Possible with inherent Possible with inherent the interfacing reactor
10 interfacing reactor within each arm and
through capability inductance in the DC side inductance in the DC side XL>0.25pu and current
XL>0.25pu and current interfacing reactors
higher rated diodes
-½Vdc higher rated diodes XL>0.25pu or bypass
-½Vdc thyristors
limited to 3 due to increase limited to 3 due to increase extendable to any extendable to any extendable to any
Figure 23.33: One-phase of 7-level modular converter used as basic HVDC building block and complexity of the control complexity of the control number of terminals number of terminals number of terminals
Multi-terminal
components of the ac output voltage before filtering. system to achieve current system to achieve current practically feasible. practically feasible. practically feasible.
configuration
balancing and power balancing and power DC side isolation is DC side isolation is With single protection
reversal reversal difficult. difficult. zone.
Delivery time 36 month 36 month 24 month 24 month Not yet known

Redundancy yes yes yes yes yes

Manufacturers ABB, SIEMENS and AREVA ABB ABB ABB SIEMENS

 1021 Power Electronics Chapter 23 HVDC 1022

The cell capacitors are used as energy storage tanks to construct fundamental voltage from predefined
nd
discrete steps, and must be sized to handle the fundamental and the 2 harmonic currents in the
converter arm, with minimum cell capacitor voltage ripple. There are several approaches for sizing of the 2×600MVA
Ii1p Ii3p
2 n1p Idc1 CB1p 150km CB5p
ndc1p
CB6p 150km CB3
n3p
M C cell capacitance, but can be based on the concept of inertia. The inertia H is defined as the ratio of 600MVA
400kV/300kV
VSC1 2×600MVA 600MVA
B1 400kV/300kV
the electrostatic energy stored in all converter cells to the converter nominal apparent power rating, ~ ~
= =
Snom. F1
CB9p
±640kV G3
2 G1
V  20000MVA ~ Pac3 20000MVA
½C  dc  400kV ~ = 400kV

 N c   3 C Vdc
2 SCR=3 = 600MVA B3 SCR=20

H  6  Nc  (23.35)
CB1n 150km CB5n ndc1n CB6n 150km CB3
Idc3 VSC3 400kV/300kV
600MVA
S nom Nc S nom 400kV/300kV n1n n3n

N c S nom H
C  (23.36)
3 Vdc2 75km
75km
For real power transfer, as with HVDC, an inertia H between 30ms to 40ms provides satisfactory (b)
performance. Nc is the number of cells per arm, C is the capacitance per cell, and Vdc is the pole-to-pole
dc link voltage. CB10p
2×600MVA n2p CB2p CB7p CB8p CB4p n4p
600MVA Idc2 2×600MVA
VSC2 VSC4 600MVA
400kV/300kV 150km ndc2p
Table 23.2 provides a comparison between high-voltage dc transmission system technologies. The ~
150km
~
400kV/300kV

=
comparison focuses on issues such as: control flexibility, fault ride-through capability, conversion losses, =

electromagnetic compatibility (EMC) issues, and provision of auxiliary functionality such as voltage, G2 ±640kV
G4

20000MVA ~ Pac4 20000MVA

frequency, and damping support. The table does not highlight that the MMC converter is most suited to 400kV ~ = 400kV
SCR=3 = CB10n 600MVA B4 SCR=20
high modulation index, near fixed frequency operation. It cannot function if producing dc output currents, B2 600MVA 400kV/300kV
CB2n CB7n CB8n CB4n
as with ac machine control operating around zero speed (see chapters 10.6.5 and 13.3.4). 400kV/300kV

150km ndc2n n4n

150km

23.12 Multi-terminal VSC HVDC Figure 23.34. Four-terminal VSC-HVDC networks: (a) mono-polar link and (b) bipolar link.

Multi-terminal HVDC, as shown in figure 23.34, uses VSC topologies where power flow control is
achieved with bidirectional converter currents. Figure 23.34 shows a four terminal, mono-polar HVDC
topology, while figure 23.34b shows a four terminal, bipolar link HVDC system with the 20 necessary dc 23.13 HVDC Earth Electrodes
breakers/isolators shown as the crossed square symbols. Ten dc breakers/isolators are necessary for
the mono-polar system. Earth electrodes are essential component in a monopolar HVDC transmission s ystem, since they
Exploitation of multi-terminal VSC HVDC is hampered by dc fault protection limitations. Even with high carry continuous operating current. Earth electrodes provide a significantly cheaper solution than a
transformer leakage inductance of 0.3pu, typical ac currents feeding a dc fault exceed 10pu in less than scheme using a second conductor (with half the nominal voltage).
5ms, which are experience by the converter freewheel diodes. A purely semiconductor dc circuit breaker Earth electrodes are also used in all bipolar HVDC systems and in HVDC m ulti-point systems. All high
can isolated the fault within tens of microseconds, limiting inrush currents to safe levels, but losses are voltage systems, including HVDC schemes require a reference for voltage measurement, insulation
high since the semiconductors are rated for continuous full load current. A hybrid approach of a coordination, and overvoltage protection. In a bipolar HVDC system, the station neutral point can be
electromechanical circuit breaker (plus a series IGBT) in parallel with a semiconductor string, is not connected to the ground mat of the HVDC station to which the line-side star points of the converter
viable, since electromechanical breaker operation requires a few millisecond, during which time the transformers are also connected. Since the direct currents in the two HVDC poles never exactly
shunting semiconductors (IGBTs) will unrealistically have to conduct, then commutate, 5 to 10 times cancel, in spite of current balancing control, a differential current flows continuously f rom the station
rated current. neutral point to ground. It is common practice to locate the grounding of the station neutral point at
The alternative is to combine the converter and dc circuit breaker functions by using a converter with dc some distance (10km to 50km) from the HVDC station by means of special earth electrodes.
block capability, as with cascaded H-bridge topologies, at the expense of increased losses. Earth electrodes for HVDC systems may be land, coastal or submarine electrodes. In monopolar HVDC
At a converter, the need for dc circuit breakers is avoided by using the ac side breakers, in conjunction systems, which exist almost exclusively in the form of sub marine cable transmission systems, there are
with over-current rated converter diodes and a dc isolator/disconnector. Isolating the other side of the dc fundamental differences between the design of anode and cathode electrodes.
side fault is problematic if the remaining parts of the multi-terminal system are to continue functioning.
i. The Horizontal Land Electrode
600MVA 600MVA If a sufficiently large area of flat land with relatively homogeneous ground characteristics is available, the
400kV/300kV 300kV/400kV horizontal ground electrode is the most economical form of a land electrode. The electrode conductor,
B1 (0.0002+j0.32)pu (0.0002+j0.32)pu B3
j0.18pu
j0.18pu which is generally of iron, is laid horizontally at a 2m depth, embedded in coke filled trench of 0.5 x
G1 ~ ~ G3 2
= 150km 150km = 0.5m cross section. The advantage of this design is in anodic operation, where the passage of the
Vdc5 Vdc3
current from the electrode conductor into the coke bed is carried primarily by electrons, and is thus not
VSC1 VSC3 associated with loss of the material (ions).
600MVA 600MVA

75km
ii. Vertical Land Electrode
±320kV
For high specific resistance surface layers, a vertical deep hole electrode may be viable.
600MVA 600MVA
400kV/200kV 200kV/400kV iii. Cathodic Submarine Electrodes
(0.0002+j0.32)pu
j0.18pu 150km 150km
(0.0002+j0.32)pu B4 Cathodic submarine electrodes of a monopolar HVDC system with submarine power transmission
j0.18pu
G2 ~ ~ G4 cable do not present problems. Since there is no material corrosion, a copper cable laid on the bottom
= = suffices, but the cable length should ensure that the surface current density produces an electrical
B2 Vdc4
VSC2 Vdc6 VSC4 field of less than 3V/m in the surrounding water.
600MVA 600MVA
iv. Anodic Submarine Electrodes

(a)
1023 Power Electronics Chapter 23 HVDC 1024

Prefabricated electrode modules are lowered to the ocean floor and then connected to the feed cable.
When the submarine electrodes are divided into sections which are connected to the HVDC station by Disadvantages
means of separate feed cables, the electrode can be remotely monitored.  Improved control is achieved at the expense of increased losses in the power converter.
Increased losses are the result of:
v. Anodic Coastal Electrode i. A higher frequency switching leads to increased switching loss.
The coastal electrode is similar in design to the vertical land electrode. Graphite rods surrounded by a ii. IGBT devices exhibit higher on-state voltage drop than thyristors with similar voltage
coke bed are installed in boreholes which are sunk along the coastline. Such coastal electrodes are ratings. For a given application, this leads to increased conduction loss.
readily accessibility for inspection, maintenance and regeneration.  IGBT modules have lower power capability than available thyristor packages, leading to
A coastal electrode can also be a horizontal land electrode if the ground has the necessary increased power component count. Footprint is larger than that of thyristors.
conductivity or if low conductivity can be achieved by irrigating the trench with salt water. With salt  IGBT devices have lower current overload capability than thyristor based systems.
water, current flow to the opposite electrode occurs through the water.
 High dv/dt transients may be present at the output. This can cause problems with cable systems
and also electromagnetic interference.
 Higher costs than LCC HVDC.
23.14 HVDC VSC features  DC faults are a serious issue since a VSC behaves like an uncontrolled rectifier bridge during a
DC fault. Diode overrating is typically required and the fault is cleared by AC circuit
breakers. Typical recovery time for dc faults (assuming AC circuit breaker tripping) is around
IGBT based systems used in transmission are based around the self-commutating voltage source
10s.
inverter (VSI) principle; high frequency switching allows converter synthesised controlled voltage at its
 The development of multi-terminal dc and dc grids requires very fast dc circuit breakers.
output. This voltage appears as a power frequency fundamental with harmonics at the switching
 IGBT failure in short circuit is possible but it is much more complex than with thyristors, which
frequency. With appropriate control, the voltage source inverter may appear as a controlled current
naturally fail in short circuit due to disc packaging.
source.
 Higher power losses than LCC HVDC
The attributes of HVDC voltage source converters allow for simpler ac system integration, plus the  Lower power ratings than LCC HVDC
following system technological features.
The progression and expansion to multi-terminal HVDC systems is hampered by the fact that VSC
 Independent, continuous control of active and reactive power at each terminal, that is, four- technology cannot inherently block the ac side from the effects of dc-side faults.
quadrant operational control, down to low power levels.
 Power flow can be reversed instantaneously (50-100ms) without the need to reverse the dc link
voltage polarity (only dc current direction reverses), so XLPE (or polarised) cables can be used. 23.15 HVDC LCC features
 Steady-state reactive power capability offers voltage control for weak ac grids.
 Independent control of reactive power at each terminal while maintaining full dc voltage for The main advantages of LCC HVDC include:
efficient transmission operation. The converter station can be used to provide voltage support to  Thyristor devices have low losses and are available in robust high voltage (6-10kV), high current (3-
the ac network while transmitting any level of active power, at no additional cost. VSC-HVDC 4kA) capacity single wafer capsules
can be configured to provide fast frequency or damping support to the ac networks through  Thyristors have excellent overcurrent capability
active power modulation  Thyristors, in disc packaging, naturally fail in short circuit which is desired for high-voltage valves
 If there is no need to transmit active power, VSC-HVDC can operate in a sleep mode (zero  Thyristor costs are lower than IGBT costs
active power exchange) and both converter stations operate as two independent STATCOMs to  Line commutated HVDC has an established track record at transmission voltage and power levels.
regulate ac network voltages.  Resilience to dc side short circuit faults. The system can operate well at low dc voltage and dc fault
 Dynamic reactive power reserve capability from the voltage source converters for contingency recovery is well controlled.
voltage support of the interconnected ac system. The VSC can be connected to blacked-out  Provides ac system frequency stabilisation in most installations, but performance is limited because
networks and re-energize them. This feature eliminates the need for a start-up generator in of reactive power issues. AC voltage control can also be implemented.
applications where space is critical or expensive such as with offshore wind farms.
 Less filtering requirements, about 15 to 20% of rated power, due to minimal injection of low The disadvantages of LCC HVDC are:
frequency harmonics, due to PWM modulation. This results in a reduced system installation  Inject significant low order harmonics which must be eliminated using large (physically and
size. electrically) passive filter arrangements. The presence of these filters may lead to circulating
 No requirement for switched filters or shunt capacitor banks for reactive power compensation harmonic currents which must be mitigated by damping networks. Filters and damping networks may
with changes in power transfer. have to be designed specifically for each location and may not be optimum for all operating
 No inherent fault current contribution to increase circuit breaker interrupting duty. Fault current conditions.
contribution is naturally limited to maximum load current but can be reduced during faults by fast  Inherently limited in their response time (limited to line frequency switching) and may face limits to
acting VSC control. their control ability. For example thyristor based HVDC:
 Robust with respect to ac network faults, thereby allowing continuity of power transmission and o Cannot decouple the real and reactive power injected into the network.
limitation of spread of system disturbances. o Requires large reactive power compensation. This reactive power is supplied using passive
 VSC HVDC is short-circuit proof for line-to-line-to-ground and line-to-line faults on the dc-side. filters and switched capacitors and reactors. The switched banks are required since the amount
Good response to ac faults. Since the VSC converter actively controls the ac current, the VSC- of reactive power needed varies with the active power magnitude and operating mode (inversion
HVDC contribution to the ac fault current during is limited to rated current or less, and the or rectification). As a result, the HVDC converter station based on line commutated converters is
converter can remain in operation to provide voltage support to the ac networks provided the ac bulky and has a large footprint.
network remains stable.  Power reversal necessitates the dc link voltage polarity be reversed. The XLPE cables have a range
 Can operate in ac grids with extremely low short-circuit levels or with passive loads. Can of cost and performance advantages over traditional paper/oil insulated of dc cables, but they can not
operate in an unsymmetrical network (e.g. during ac network faults) and provide unbalance take fast voltage polarity reversal. Also, the time required to reverse the power flow is limited to
control to compensate asymmetrical loads. The faster converter response allows counteraction around 0.5s, which imposes operational constraints.
of network distortion and transients.  There is a theoretical minimum power flow of 10% in the dc link to guarantee satisfactory operation.
 Facilitate the development of smart grids, especially during transition from the grid mode to an  It is only suitable for connection of the ac networks with a high short circuit ratio (SCR > 2.5).
intended/or unintended islanding mode.  Suffers from commutation failure (dc-side short circuit) during ac faults. Typically a sudden ac voltage
drop of 5-10% will cause commutation failure. A commutation failure causes transient (0.5-1s)
interruption in power transfer.
1025 Power Electronics Chapter 23 HVDC 1026

 Vulnerable to dc side open circuit conditions. This is an unlikely event and causes overvoltage which Reading list
is readily limited by surge arresters.
 Multi-terminal topologies can be built but with performance and cost penalties. DC grids are
rd
Mohan, N., Power Electronics, 3 Edition,
challenging. Wiley International, 2003.

Acha, E., et al., Power Electronic Control in Electrical Systems,

23.16 Features of conventional HVDC and HVAC transmission Newes, 2002.
The following are comparison features between ac and conventional dc electrical power transmission: HVDC Connecting to the future, Alstom Grid, 2010
 Increased power system stability since dc power flow is controlled by converter delay angles.
Converter response times are of the order of ms, while the dc link-time constant is
significantly longer. Thus, the dc-link short circuit current can be limited. Problems
 No distance effect on stability (voltage angle), with dc transmission, plus no intermediate stations
needed every few hundred kilometres. 23.1 Show that if a three-phase ac transmission system is converted to a tripole hvdc system with the
 The fast converter response times means power reversal can be achieved in ms, much faster third conductor transmitting a square-wave current of ±1 pu, that the other two phases can alternately
than within ac systems.    
conduct ½ 3  1 and ½ 3  1 , yet not exceed their thermal rating.
 Steady-state power flow is related to dc-link resistance, as opposed to ac-line reactance.
 Transmission line inductance has zero impedance to dc, whereas inductive reactance is relatively
large in ac systems. A minimal ac skin effect occurs with ac transmission. 23.17 Appendix: HVDC topology classification
 AC power is lost due to dielectric losses. Capacitance between conductors is open circuit to dc,
2
but in ac systems capacitive reactance current paths exist which increase I R losses. Figure 23.35a shows the symmetrical mono-polar HVDC link configuration, which has the following
 Under sea cables have high capacitance which has minimal effect on dc transmission advantages and disadvantages:
 DC back-to-back converters (with minimal link inductance, tens of milliHenry’s) allow  Since the virtual supply mid-point is established by dc cable stray capacitances, the converter
synchronisation between 50Hz and 60Hz systems or same frequency systems at different transformers are not exposed to dc voltage stresses;
voltage phase angles.  no ground dc current;
 DC transmission line and its towers are cheaper per unit length than for ac, but dc terminating  no ac current contribution from the grid during a pole-to-ground dc side fault; and
stations are more complicated and expensive than ac terminating stations  requires two fully insulated dc cables and no redundancy; so tends to be an expensive solution.
 DC converters have limited overload capability and are less reliable because of the ancillaries Figure 23.35b shows an asymmetrical mono-polar VSC-HVDC link with a metallic return. The main
(filters, cooling management, etc.). advantages and disadvantages of this configuration are:
 DC transmission can have a lower visual impact, plus a narrower corridor, hence lower  Metallic ground cable is not fully insulated, hence reduced expenditure on the dc cables;
environmental impact than ac systems  no dc ground current;
 Electric and magnetic fields for dc transmission have lower environmental implication than ac  permits extension to a bi-polar configuration;
fields  exposes the interfacing transformer to dc voltage stresses; and
 Corona effects tend to be less significant with dc that for ac conductors  no redundancy.
 DC transmission is more robust to lightning effects, since more effective surge protection is used. Figure 23.35c shows asymmetrical mono-polar VSC-HVDC link with ground return, where the main
 Polymer cables age slower, giving a longer lifetime, with dc features and drawbacks of this approach are:
 Maximum voltage gives lower I R losses for a given transmitted power level
2
 Uses a single dc cable, which reduces cost and transmission losses;
 No unnecessary energy transfer since no reactive VAr (S=VI*=P+jQ), with dc  allows extension to a bi-polar link;
 no redundancy;
The general advantages of ac transmission, over dc transmission, are  exposes transformer to dc voltage stresses;
 No costs associated with ac-dc-ac conversion equipment  with ac current in-feed from the ac side, a pole-to-ground fault exposes the converter switches to
 Transformer (and autotransformer) voltage matching increased current stresses; and
 Reactive power and harmonics readily compensated  causes environmental concerns due to the use of grounded electrode and continuous operation
 Not restricted to only point-to-point connection, as is HVDC (there is no fully functioning exception) with ground current, hence requires special permit.
 Established system control methods Figure 23.35d shows symmetrical bipolar VSC-HVDC link with ground return. The main benefits and
 No ac transformer dc voltage stressing due to asymmetrical phase control alignment, and no I R
2 shortcomings are:
and core losses due to high harmonic currents  Does not expose converter transformer to dc stresses;
 ac switch gear and breakers, (and particularly vacuum circuit breakers up to 33kV) are very  no ac current in-feed from the grid during pole-to-ground dc network faults;
effective - compared with the difficulties in breaking dc current  dc link capacitors and cable insulation must be rated for the full dc voltage; and
 Lower current harmonics  requires permission to operate with ground return during outage of one dc cable (ac current in-
feed exist).
Figure 23.35e shows a symmetrical bi-polar VSC-HVDC topology with metallic return. This configuration
has the following advantages and disadvantages:
 50% redundancy of total rating;
 suffers from ac current in-feed from the grid during a pole-to-ground dc fault;
 transformer must be designed withstand dc voltage stresses; and
 use of electrodes causes environmental concerns, therefore requires permission.
Figure 23.35f is a symmetrical bi-polar HVDC link with metallic return with the following features:
 50% redundancy of total rating;
 no ground current;
 increases transmission capacity, without the need to raise converter dc link voltage; and
 exposes transformer to dc voltage stress.
 1027 Power Electronics Chapter 23 HVDC 1028

B1n T1p T2p B2p

L1 L2
B1 B2 Vdc1p Vdc2p
T1 T2
L1 L2
Vdc1 Vdc2

VSC1p VSC2p
VSC1 VSC2
(a)
LV DC

B1n T1n T2n B2n

L1 L2
Vdc1n Vdc2n

B1 T1 T2 B2
L1 L2 VSC1n VSC2n
Vdc1 Vdc2
(f)
LV DC
VSC1 VSC2
(b)
Figure 23.35. Different mono-polar and bipolar VSC-HVDC topologies:
(a) symmetrical mono-polar link, (b) asymmetrical mono-polar link with metallic return,
(c) asymmetrical mono-polar link with ground return, (d) symmetrical bipolar link with
floating ground return, (e) symmetrical bipolar with ground return, and (f) symmetrical bi-
polar link with metallic return.
B1 T1 T2 B2
L1 L2
Vdc1 Vdc2

VSC1 electrodes VSC2

(c)

B1 T1 T2 B2
L1 L2
Vdc1 Vdc2

VSC1 VSC2

(d)

B1n T1p T2p B2p

L1 L2
Vdc1p Vdc2p

VSC1p VSC2p

electrodes

B1n T1n T2n B2n

L1 L2
Vdc1n Vdc2n

VSC1n VSC2n
(e)