1 110-04 PDF
1 110-04 PDF
1, 2010, 29–36
The recent developments in semiconductors and control equipment have made the voltage source converter based high
voltage direct current (VSC-HVDC) feasible. This new DC transmission is known as “HVDC Light or “HVDC Plus by
leading vendors. Due to the use of VSC technology and pulse width modulation (PWM) the VSC-HVDC has a number of
potential advantages as compared with classic HVDC. In this paper, the scenario of back-to-back VSC-HVDC link connecting
two adjacent asynchronous AC networks is studied. Control strategy is implemented and its dynamic performances during
disturbances are investigated in MATLAB/Simulink program. The simulation results have shown good performance of the
proposed system under balanced and unbalanced fault conditions.
K e y w o r d s: HVDC, voltage source converter (VSC), IGBT, SPWM, control design
1 NOMENCLATURE are: the STATCOM, the SSSC, the UPFC, and the volt-
age source converter based HVDC transmission (VSC-
UL the sinusoidal AC voltage in the AC network HVDC) [2, 3].
UV (1) the fundamental line to line voltage (valve The VSC-HVDC system is the most recent HVDC
side) technology. It consists of two VSC, one of which oper-
XL the leakage reactance of the transformer ates as a rectifier and the other as an inverter. The two
δ phase shift between UL and UV (1) converters are connected either back-to-back or joined by
IV source current a DC cable, depending on the application. Its main func-
L, R phase reactor inductance and resistance
tion is to transmit a constant DC power from the rectifier
C DC side capacitance
station to the inverter station, with high controllability.
ω source voltage angular frequency
m modulation index The VSC-HVDC has several main advantages against the
P, Q AC active, reactive power inputs conventional HVDC based on thyristors. They are [4, 5]:
Ud , Id , Pd DC side voltage, current, power 1. Independent control of the active and reactive power
α, β stationary α – β axis output from each terminal.
d, q synchronous d– q axis 2. Reduced requirements for harmonic filters.
p, n positive, negative components
∗ reference value for controller 3. Improvements of the power quality and system stabil-
ity.
4. Elimination of the requirement for a local power gen-
eration.
2 INTRODUCTION
These features make the VSC-HVDC attractive for con-
nection of weak AC system, island networks, and renew-
The availability of the modern semiconductor devices able energy sources, which may be located in remote area,
such as the Insulated Gate Bipolar Transistor (IGBT) [1] to a main grid.
has led to the development of a new generation of power
electric converters. These devices, unlike the conventional This paper presents the elements of back-to-back VSC-
thyristors which have no intrinsic turn-off ability, are of HVDC where the converter stations are located at the
the fully controlled type. The most common convert- same site and transmission line or cable is not needed.
ers, which employ the self commutating, high voltage, The paper wills first give a brief description about the
high current, and high switching frequency power elec- VSC-HVDC transmission system and its terminal control
tronic devices, are the Voltage Source Converters (VSC). functions. Following that, typical operating contingency
A number of FACTS controllers which use VSC as their scenarios are simulated in order to evaluate transient per-
basic building block have been already in operation in formance. The simulation results confirm that the control
various parts of the world. The most popular controllers strategy has fast response and strong stability.
∗ Electrical Engineering Department Intelligent Control & Electrical Power Systems Laboratory (ICEPS) Djillali Liabes University, Sidi
Bel-Abbes, 22000, Algeria, med [email protected], [email protected], [email protected], [email protected]
The fundamentals of VSC transmission operation may A 200 MW (±100 kV DC) asynchronous interconnec-
be explained by considering each terminal as a voltage tion is used to transmit power from a 230 kV, 2000 MVA,
source connected to the AC transmission network via a 60 Hz network having a SCR of 10 to 230 kV, 1000 MVA,
three-phase reactor. The two terminals are interconnected 50 Hz network having a SCR of 5. As shown in Fig. 2, it
either back-to-back or through a DC link, as schemati- consists of DC capacitors, two converters, passive high-
cally shown in Fig. 1. pass filters, phase reactors, and transformers.
and
R 1
−L ω
0
d ivdn ivdn L uvdn
= −
dt ivqn ω −RL
ivqn 0 L1 uvqn
1
0
uLdn
+ L 1 . (10)
0 L uLqn
d
uvdp = uLdp − Rivdp + ωLivqp − L ivdp , (11)
dt
d
uvqp = uLqp − Rivqp − ωLivdp − L ivqp , (12)
dt
and
d
uvdn = uLdn − Rivdn − ωLivqn − L ivdn , (13)
Fig. 5. Inner Current Controller
dt
d
uvqn = uLqn − Rivqn − ωLivdn − L ivqn . (14)
dt
5.3 Inner current loop
The mean voltages over the sample period k to k + 1
For each of the phases we can write (Fig. 2) are derived by integrating (11), (12), (13) and (14) from
kTs to (k + 1)Ts and dividing by Ts (where Ts is the
div
UL − u v = L + Riv . (6) sampling time).
dt
During unbalanced operation [8] the expression for the uvdp = uLdp − Rivdp +
voltage drop over the reactor (R, L ) holds for positive as L
well as for negative-sequence voltages and currents. The ωLivqp − {ivdp (k + 1) − ivdp (k)}, , (15)
Ts
voltages drops are described by the differential equation
uvqp = uLqp − Rivqp −
R
−L 0 0
ivaX ivaX L
d ωLivqp − {ivqp (k + 1) − ivqp (k)} , (16)
ivbX = 0 −R 0 ivbX −
L Ts
dt
ivcX 0 0 −R ivcX
L
1
0 0
1
0 0
and
L uvaX L uLaX
0 1 0 uvbX + 0 1 0 uLbX . (7)
L L
0 0 L1 uvcX 0 0 L1 uLcX uvdn = uLdn − Rivdn −
L
Where X = (p) for positive sequence and (n) for negative ωLivqn − {ivdn (k + 1) − ivdn (k)} , (17)
Ts
sequence. Equation (7) can be transformed to the α β -
uvqn = uLqn − Rivqn +
frame. This gives for the voltages and currents:
L
R 1 ωLivqn − {ivqn (k + 1) − ivqn (k)} , (18)
d ivαX −L 0 ivαX L 0 uvαX Ts
= −
dt ivβX 0 −RL
ivβX 0 L1 uvβX
1 By assuming linear current and constant network voltage
0 uLαX
+ L 1 . (8) (the network voltage varies very little during a switching
0 L uLβX
time period) during one sample period Ts we obtain from
(15) through (18):
Equation (8) can be further transferred into the rotating
dq -frame:
R
uvdp (k + 1) = uLdp (k) − {ivdp (k + 1) + ivdp (k)}
d ivdp
−R ω
ivdp
1
0
uvdp
2
= L − L 1 ωL
dt ivqp ω −RL
ivqp 0 L uvqp + {ivqp (k + 1) + ivqp (k)}
1 2
0 uLdp L
+ L 1 . (9) − {ivdp (k + 1) − ivdp (k)} , (19)
0 L uLqp Ts
Journal of ELECTRICAL ENGINEERING 61, NO. 1, 2010 33
d d
Id0 = −(Id1 + Id2 ) = −2C (Ud1 − Ud2 ) = −2C Ud0 .
dt dt
(23)
By changing the conduction time of the switches in a
pole it is possible to change the average of the DC mid-
point current Id0 and thereby control the difference volt-
age Ud0 . For example, a positive difference (Ud0 ≥ 0 )
can be decreased to zero if the amplitude of the reference
voltage which generates a positive midpoint current is in-
creased at the same time as the amplitude of the reference
voltage which generates a negative DC midpoint current
is decreased. This is done by the addition of an offset com-
Fig. 7. Positive-Sequence impedances of the two AC networks
ponent to the sinusoidal reference voltage. Consequently,
the bridge voltage becomes distorted, and to limit the
distortion effect, the control has to be slow. Finally, for
R
uvqp (k + 1) = uLqp (k) − {ivqp (k + 1) + ivqp (k)} better performance this function should be activated in
2
the station controlling the DC voltage.
ωL
− {ivdp (k + 1) + ivdp (k)}
2
L
− {ivqp (k + 1) − ivqp (k)} , (20) 6 SIMULATIONS RESULTS
Ts
6.2 Step responses troller. The AC voltages at the filter bus (UL1 and UL2 ),
can be kept constant except for some transients that oc-
In order to test the dynamic responses of the VSC- cur when both the step changes are applied. The active
HVDC regulators, three test cases have been studied: power flow is adjusted to the new setting within 30 ms.
– At t = 1.3 s the reference active power of converter At t = 2.0 s the reactive power step change is applied,
1 (P1∗ ) is changed from −0.5 pu to +0.5 pu, and at this step change cause transients on the DC voltage, but,
t = 2.5 s from +0.5 pu to +1.0 pu as expected, the step change of the active power causes a
– From t = 2.0 s to t = 2.7 s the reference reactive power much higher transient than that with the change in reac-
of converter 1 (Q∗1 ) is changed from 0.0 pu to −0.4 pu. tive power. At t = 2.5 s the active power change to 1 pu,
and later another step is also applied on the reference DC
– At t = 3.0 s the reference DC voltage of converter 2 is
voltage of the inverter at t = 3.0 s. The reactive power
changed from +1.0 pu to +0.095 pu.
and the DC voltage can track the references of the reac-
First the converter 1 active power reference value is tive power and the DC voltage. The active and reactive
changed from from −0.5 pu to +0.5 pu and then set to power responses are decoupled by the control design.
1.0 pu at t = 2.5 s. As can be seen from Fig. 8 the active
power can track the reference of the active power. 6.3 AC side perturbations
The transferred active powers at both sides change the
direction which causes transients on the DC voltage then A remote three phase to ground fault was first applied
returns to the reference value due to the DC voltage con- at t = 1.5 s during 0.12 s (5 cycles on 60 Hz system) at
Journal of ELECTRICAL ENGINEERING 61, NO. 1, 2010 35
station 1. A second perturbation follows. A single phase ing the fault. All oscillations in voltages and currents at
to ground fault was applied at t = 1.8 s during 0.10 s both systems, means that the phase voltages and cur-
(5 cycles on 50 Hz system) at station 2 in order to in- rents at both systems are unbalanced. During the severe
vestigate the behavior of VSC-HVDC during unbalanced
three-phase fault at station 2 at t = 2.1 s, the AC voltage
faults. Then, a three-phase to ground fault is applied at
station 2 at t = 2.1 s and is cleared at 5 cycles after the at station 2 side is decreased to 0.1 pu during the fault
fault, ie, at t = 2.2 s. Figure 9 presents the simulations and recovers fast and successfully to 1.0 pu voltage after
results. From the simulation, it can be noted that during clearing the fault. The transmitted power flow is reduced
the two first faults (t = 1.5 s at station 1 and t = 1.8 s to very low value during the fault and recovers to 1.0 pu
at station 2), the active power flow is 1.0 pu, transmit- after the fault. The DC voltage, which can be controlled
ted from converter 1 to converter 2, and is kept constant to 1.0 pu during the fault, has some oscillations at the
during the fault. The DC voltage drops and it contains
beginning of the fault and at clearing the fault, and its
an oscillation during the faults.
maximum transient value is about 1.2 pu. On the other
Consequently the transferred DC power contains also
the oscillation. During these two faults the transmitted hand, the phase currents at station 1 side decrease to low
power can be kept constant except a small oscillation dur- values to reduce the power flow.
36 M. Khatir — S.A. Zidi — S. Hadjeri — M.K. Fellah: DYNAMIC PERFORMANCE OF A BACK-TO-BACK HVDC STATION . . .