Koutiva 2006
Koutiva 2006
Abstract—This paper studies the behavior of a high-voltage di- Classical HVdc links are not suitable for this type of trans-
rect current link based on voltage-source converters, which feeds mission application, as they require a high short-circuit ratio
a weak ac network with power produced from an offshore wind between the two ends. On the contrary, the technology of HVdc
farm (WF) of induction generators. Its control system, which is
based on adaptive fuzzy controllers, manages to offer very satis- based on voltage-source converters (VSCs) can be a feasible
factory performance, without the need for a detailed mathemat- solution [1]–[3]. It presents the advantages of dc transmis-
ical model, but just prior knowledge of the behavior of the elec- sion and due to the high switching capability of the VSCs’
trical system. Using the simulation program PSCAD/EMTDC, the insulated-gate bipolar transistors (IGBTs), it can instantly reg-
study was performed under both steady-state and transient condi- ulate the reactive power and, consequently, the ac voltage,
tions. The results show that the link supplies the variable power of
the WF to a weak grid, keeping the ac voltage fluctuations in the independent of the real power flow. So the short-circuit ratio
point of common coupling at an acceptable level. Moreover, due to between the two ends of the link does not have to be high.
the ability of the control system to adjust the stator frequency of Furthermore, the power quality and system stability can be
the induction generators in relation to the wind velocity, maximum improved via continuously adjustable reactive power support
wind power acquisition is achieved. with ac voltage feedback control.
Index Terms—DC–AC power conversion, fuzzy control, HVdc While some control methods for HVdc links based on VSCs
transmission, induction machines, voltage control, wind power connecting WFs of synchronous generators to strong ac sys-
generation. tems have been relatively discussed in the literature [4], [5],
the topology regarded in this paper has not been studied so far.
I. INTRODUCTION Aiming to cope with a more demanding case, the connection of
a WF of induction generators to a weak ac grid via an HVdc link
III. CONTROL PRINCIPLES OF A VSC-BASED HVDC LINK IV. CONTROL SYSTEM DESCRIPTION
The active power P flowing between each converter and its The control system is based on fuzzy logic (FL). This type of
adjacent ac network is controlled by changing the phase angle control, approaching the human reasoning that makes use of the
, between the fundamental frequency voltage generated by the tolerance, uncertainty, imprecision, and fuzziness in the deci-
converter, Vg, and the ac bus voltage, Vn, (Fig. 1). Neglecting sion-making process, manages to offer a very satisfactory per-
the losses at the transformer, the active power is calculated formance, without the need of a detailed mathematical model
according to (1) of the system, just by incorporating the experts’ knowledge into
fuzzy rules. In addition, it has inherent abilities to deal with im-
(1) precise or noisy data; thus, it is able to extend its control ca-
pability even to those operating conditions where linear control
where is the transformer leakage reactance. techniques fail (i.e., large parameter variations).
The control of power angle is achieved by regulating the an- The main drawback of an FL-based control system is that the
gular frequency of the ac voltage created by the VSC. For ex- tuning of its membership functions (MFs) needs too much “trial
ample, when the wind power increases, the angular frequency at and error.” In order to reduce the time-consuming process of
the point tends to increase. This is detected by , which the MFs tuning or to ameliorate the performance when it does
gives an order to increase the phase angle of the PWM sinu- not satisfy the specification, we can apply an online-tuned adap-
soidal reference signal and, consequently, the phase angle tive fuzzy control system (AFCS). AFCSs can adapt to their en-
of the fundamental ac voltage of , with respect to vironment and acquire new knowledge by themselves through
the ac voltage at the bus of the WF . According to (1), by learning. A possible arrangement of such a system is the imple-
increasing the phase angle , more power is driven to the dc mentation of a fuzzy controller (FC) to adjust the parameters of
system and, consequently, the angular frequency returns to its another FC. This adjustment is accomplished online. The main
reference value. FCs MFs are tuned online through the supervisor-FC, which fol-
In normal operation mode, the active power, which is supplied lows the reasoning of an expert which would manually tune the
to the HVdc link, must be equal to the active power received by MFs. This type of control offers, except for the automatic MFs
the network at the other side of the link plus the losses. This adjustment, an adapting tuning of the MFs, according to the be-
power balance is achieved by keeping constant the voltage at havior of the system. AFCs are thus very suitable for the con-
the point G (Fig. 1). Therefore, is designated the role of trol of systems which are strongly fluctuating, such as systems
the dc voltage regulator at this point. For example, when the WF of large WFs of induction generators.
KOUTIVA et al.: OPTIMAL INTEGRATION OF AN OFFSHORE WIND FARM TO A WEAK AC GRID 989
Fig. 3. HVDC link based on VSCs and its control circuit block diagram.
Fig. 5. Membership functions for the fuzzy set ! 0! . Fig. 7. Membership functions for the fuzzy set phase angle .
TABLE I
FUZZY RULES FOR FC
TABLE II
FUZZY RULES FOR FC
(5)
(6)
(7)
where
wind turbine power;
wind density ;
Fig. 10. Wind velocity.
tip speed ratio;
pitch angle;
blade impact area;
wind velocity;
performance coefficient;
blade semidiameter;
angular velocity of the turbine;
constants.
The wind model used in this application is a three- component
model [7] and is defined by the following equation:
(8)
where is the base wind velocity, is the ramp wind Fig. 11. Power at the sending and the receiving end of the VSC-based HVdc
component, and is the noise wind component. link.
The 20-kV ac network has a short-circuit capacity (SCC)
equal to 45 MVA. The dc cables are coaxial and 100 km long; dc
capacitors of 500 F are used at each converter. AC capacitors
of a total capacity equal to 10 F are placed at the side of the
WF in order to supply a constant amount of the reactive power
that is necessary for the operation of the inductive generators.
A detailed, nonlinear simulation of the system has been used in
order to test the system under steady-state and transient condi-
tions. Each machine of the WF was represented separately in
the simulation in order to take into account the interaction be-
tween the generators of the WF. The basic simulation program
used is PSCAD/EMTDC. Because the control system was based
on fuzzy logic, the fuzzy logic toolbox of MATLAB was also Fig. 12. Electrical frequency and its reference.
used. Through C++ programming, the design of AFCs was ac-
complished, converting the simple fuzzy controllers of the fuzzy
reach its optimal value for the specific wind velocity. Before the
logic toolbox into AFCs, which can be self-tuned online.
system manages to reach the steady-state, which corresponds
to the wind velocity of 8 m/s, the wind speed changes again,
VI. EVALUATION OF THE SYSTEM PERFORMANCE
rising up to 11 m/s. Then, the real power at both ends increases,
This paragraph presents the simulated response of the system, causing a respective increment of . Even in this disturbance,
under steady-state and two disturbances, a sudden decrease, and stays close to and increases until it reaches its optimal
a sudden increase of the wind speed. value. When this happens, , and reach a
The system is simulated with a wind velocity which fluctu- new steady-state, which corresponds to the maximum absorp-
ates with a mean value near 9 m/s and at 200 s, following an tion of real power from the WF.
almost step change, drops to a mean value near 8 m/s (Fig. 10). Fig. 13 shows the dc voltage at the receiving end of the link
After a few seconds, the wind velocity rises suddenly up to and its reference . As decreases, the
11 m/s. decreases too. The control system detects the deviation of the dc
Figs. 11–14 show the response of the system at the above voltage from its reference and gives an order to adjust the phase
mentioned disturbances. When the wind velocity decreases, the angle to push less power to the ac system and to increase the dc
values of the real power at the two ends and voltage. Finally, the dc voltage reaches its reference value again.
start to decrease (Fig. 11). Then, the control system gives an Respectively, when the dc voltage rises, the control system gives
order to decrease (Fig. 12). We can observe that is an order to push more power to the ac system, so the dc voltage
continuously attached to . keeps decreasing, trying to returns to its reference value.
KOUTIVA et al.: OPTIMAL INTEGRATION OF AN OFFSHORE WIND FARM TO A WEAK AC GRID 993
VII. CONCLUSION
In this paper, the connection of a big offshore WF of induc-
tion generators to a long-distance weak ac grid is studied. The
only feasible solution for this connection is an HVdc link based
on VSCs, which offers the possibility to supply the variable real
power of the WF to the low short-circuit capacity bus at the PCC
with an acceptable power quality. This system has a very un-
stable and fluctuating behavior by nature, so it demands a quick,
flexible, and adaptive control. The proposed control system is
based on adaptive fuzzy control. The AFCs that are presented in
this paper have the ability to be self-tuned online in order to have
Fig. 13. DC voltage at the grid. an optimum performance in every operating condition. They can
easily adapt to the quickly changing and fluctuating values of
the wind velocity, so they offer the possibility to achieve, apart
from an optimum integration of the WF to the weak ac grid,
maximum wind power acquisition, driving the WFs to the max-
imum aerodynamic efficiency.
REFERENCES
[1] G. Venkataramanan and B. K. Johnson, “A superconucting dc trans-
mission system based on VSC transmission technologies,” IEEE Trans.
Appl. Superconduct., pt. 2, vol. 13, no. 2, pp. 1922–1925, Jun. 2003.
[2] W. Lu and B.-T. Ooi, “DC overvoltage control during loss of converter
in multy-terminal voltage-source converter-based HVDC,” IEEE Trans.
Power Del., vol. 18, no. 3, pp. 915–920, Jul. 2003.
[3] B. R. Andersen, L. Xu, and P. J. Horton, “Topologies for VSC transmis-
sion,” Power Eng. J., vol. 16, pp. 142–150, Jun. 2002.
[4] L. Xu, V. G. Agelidis, and E. Acha, “Steady state operation of HVDC
Fig. 14. RMS ac voltage at the grid. power transmission system with voltage-source converters and simulta-
neous VAR compensation,” in Proc. EPE, Graz, Austira, 2001.
[5] X. Koutiva, T. Vrionis, N. A. Vovos, and G. B. Giannakopoulos, “A
simulation model of an HVDC link using VSC’s and a control system
exploiting its capabilities,” in Proc. CSIT, 2002.
[6] S. Heier, Wind Energy Conversion Systems, NY: Wiley, 1998, p. 35.
[7] P. M. Anderson and A. Bose, “Stability simulation of wind turbine
systems,” IEEE Trans. Power App. Syst., vol. PAS-102, no. 12, pp.
3791–3795, Dec. 1983.