Vous êtes sur la page 1sur 7

218 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 21, NO.

1, JANUARY 2006

Power-Quality Measurements in an Autonomous


Island Grid With High Wind Penetration
Stavros A. Papathanassiou, Member, IEEE, and Fritz Santjer

AbstractAutonomous power systems, such as those existing in In Greece, more than 50 islands exist with relatively small,
small and medium-size islands, typically consist of medium voltage diesel-powered grids, in many of which significant wind pen-
networks fed by conventional power stations and present charac- etration levels have been reached (exceeding 50% during low
teristics not typical for large interconnected grids. In this paper,
results are presented and analyzed from a power-quality measure- load hours). To obtain field data for the operation of such sys-
ment campaign, conducted in a diesel-powered island system with tems, a two-week measurement campaign was organized on the
significant wind penetration. Emphasis is placed on the voltage island of Samos, where 4 MW of wind turbines were installed
flicker, which was found quite increased under all operating condi- and operating.
tions. It is shown that the diesel engines of the conventional power In this paper, results from the measurements are presented,
station are the main source of flicker in the system, due to the
pulsed variation of their torque. The operation of the wind farms, focusing on the power-quality characteristics at the medium-
on the other hand, does not result in any notable deterioration of voltage (MV) level of the network. The main objective is to
the voltage quality level in the system. demonstrate the effect of the diesel generating units, which are
Index TermsDiesel-driven generators, flicker, power quality, the main factor differentiating the island grid from a weak distri-
wind power generation. bution network in the interconnected system, and to evaluate the
impact from the operation of the wind turbines. Particular em-
phasis is placed on the voltage flicker, the main source of which
I. INTRODUCTION
is shown to be the diesel generating units.

T HE connection of wind turbines (WTs) to the grid is


often constrained by power-quality considerations (i.e.,
by the concern of utilities for the possible deterioration of the
voltage quality of the network). Several investigations have II. DESCRIPTION OF THE SYSTEM
been performed over the years, analyzing the steady-state and
fast voltage variations, flicker emissions, switching transients The autonomous power system of Samos island is fed by a
and harmonics (e.g. [1][7]), which contributed significantly diesel power station (DPS), which, at the time of the measure-
in better understanding the possible effects of the WT connec- ments, had a total installed capacity of approximately 30 MW,
tion. Significant standardization work has also been carried comprising several units from 2.0 MW to 6.3 MW (Fig. 1).
out, with a special reference to IEC Standard 61 400-21 ([8]) Basic generators are two 6.3 MW units , followed by three 4
regarding the power-quality characterization of WTs intended MW units, the main characteristics of which are summarized in
for grid-connected operation. Table I (unit numbering as used by the DPS operators).
Nevertheless, the existing regulations and technical evalua- The distribution network on the island consists of several
tion procedures are focused on interconnected grids, of practi- 15-kV medium voltage (MV) overhead feeders, departing from
cally fixed frequency and relatively high short-circuit capacity, the DPS busbars. An important characteristic of the network is
where the majority of the installed wind capacity is connected. the length of the MV feeders (several tens of km per feeder),
Power quality issues in weak island grids, including the effect which creates voltage regulation problems.
from the WT connection and operation, have been hardly inves- The consumer load during the Summer period, when the mea-
tigated and relatively scarce information exists on such systems surement campaign was conducted, is predominantly touristic,
(e.g. [9][11]). Although in principle the phenomena should be varying approximately from 7 to 23 MW.
of similar nature, important differences exist due to the isolated At the time of the measurements three wind farms were op-
mode of operation of these systems, which are typically fed by erating on the island. One with 9 225 kW pitch controlled,
autonomous diesel power stations. constant speed machines, near the town of Pythagorio, and an-
other two wind farms with stall regulated, constant speed WTs
Manuscript received May 26, 2004. This work was suported by the Research (9 100 kW old units and 750 250 kW recently installed
Project Integration of Wind Turbines into Electricity Supply Networks with ones), at Marathokambos. The wind farms are connected to two
Limited Power Transportation CapacityWiring (JOR3-CT98-0245) by the 15 kV feeders as shown in Fig. 2 (Lines #240 and #260).
European Union. Paper no. TPWRD-00252-2004.
S. A. Papathanassiou is with the School of Electrical and Computer En- To obtain a complete picture of the voltage quality in the is-
gineering, National Technical University of Athens (NTUA), Athens 15780, land system, a 2-week measurement campaign was conducted
Greece (e-mail: st@power.ece.ntua.gr). in mid-Summer, when optimal wind conditions prevail. Mea-
F. Santjer is with the Deutsches Windenergie-Institut (DEWI), Wil-
helmshaven D-26382, Germany (e-mail: f.santjer@dewi.de). surement locations, shown in Fig. 2, include the DPS, the three
Digital Object Identifier 10.1109/TPWRD.2005.848723 wind farms and selected nodes in the network.
0885-8977/$20.00 2006 IEEE
PAPATHANASSIOU AND SANTJER: POWER-QUALITY MEASUREMENTS IN AN AUTONOMOUS ISLAND GRID 219

Fig. 1. Diesel engine at the Samos power station.

TABLE I
DATA FOR THE DIESEL GENERATOR UNITS OF THE POWER STATION

Fig. 2. Electrical system of Samos island (only two MV feeders shown).

III. POWER-QUALITY MEASUREMENT RESULTS


A. Steady-State Measurements
The power produced by the diesel generators (active and reac-
tive) is shown in Fig. 3 for a period of about 2 weeks. The daily
load profiles with peaks at 13:00 and 21:00 hours and minima
around 6:00 h are clearly visible. The increase in the DG daily Fig. 3. Active and reactive power variation of the diesel generating units over
the measurement period (1-h values).
peak after the middle of the measurement period is due to the
decrease of the wind in that interval, resulting in the loss of
34 MW of wind power. In Fig. 3 the very low power factor of [12], during 95% of the time). The statistics of the voltage devia-
the DPS is readily observed. Values as low as 0.7 are observed tions (10 min average values) at Pythagorio are shown in Fig. 4
during the first days, increasing to 0.800.85 during the subse- for the whole measurement period. Although the operation of
quent low wind period. The low power factor is characteristic the WTs should raise the voltage levels ([13]), the opposite ef-
of the island load (to a large extent due to the air-conditioning fect is noted because the compensating capacitors of the 9
units), but in the first days it is further decreased due to the op- 225 kW wind farm were out of operation. Remedial action with
eration of the wind farms. installation of capacitors in the network, at the wind farms and
As it was observed on site, the poor power factor very often possibly at the DPS busbars is urgently needed for the island
resulted in the synchronous generators of the DPS reaching their system.
excitation limit. Thus, they were unable to maintain an increased The three phase voltages of the system exhibit an unbal-
voltage level at the DPS busbars (typically 15.75 kV, i.e. 1.05 ance under all operating conditions, which is attributed to the
p.u.), to compensate the voltage drop on the long distribution nonuniform distribution of the single-phase consumer loads.
feeders. In general, the voltage regulation on the island grid is The voltage unbalance factor (VUF) (e.g., IEC 61 000-2-1,
very poor, even at the power station busbars, where the voltage [14]) is generally acceptable (lower than 2%), although higher
drops below 0.95 p.u. in certain cases. The voltage deviation at values reaching 5% also exist during certain intervals. The
remote nodes often exceeds 10%, which is the typical steady- operation of the wind turbines generally reduces the voltage
state under-voltage limit in MV networks ( 10% according to unbalance.
220 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 21, NO. 1, JANUARY 2006

Fig. 6. Short-term flicker index during the low wind days of the measurement
period, as a function of the active power of feeders 240 and 260.

Fig. 4. Statistics of the steady-state voltage deviations at Pythagorio node over


the whole measurement period.

Fig. 5. System frequency.


Fig. 7. Typical measurement of the three-phase voltages at the DPS busbars
(minimum loadno wind conditions).
The system frequency, shown in Fig. 5, is well controlled.
Maximum excursions do not exceed 0.5 Hz, even during the
to wind farms after subtracting the existing background flicker
strong wind period of the first days, which is well below the
from the planning levels, would be quite restrictive, especially
deviation limit set in [12] for noninterconnected systems ( 1
if it is taken into account that the island network is very weak,
Hz during 95% of the time and 7.5 Hz during 100% of the
with short circuit capacities as low as 10 MVA at many nodes.
time). This is important because stability concerns are often a
These observations warranted a further investigation of the in-
barrier to increasing wind penetration in island systems.
creased voltage variability that resulted in the high flicker index
values. In Fig. 7, a typical 20-min recording is shown of the
B. Flicker
three phase-to-earth voltages at the 15-kV busbars of the DPS.
Flicker measurements along the whole period of the measure- A significant high frequency ripple can be observed in all phase
ment campaign (approximately two weeks in the midsummer voltages, which is shown in more detail for a 5-s period. Such a
period) revealed that the voltage flicker levels are relatively high ripple was always present in the voltages of all nodes in the net-
at all nodes of the network, irrespective of the load level and the work and was not considerably affected by the operation of the
output power of the wind turbines. Actual values recorded at wind turbines. Its magnitude and frequency spectrum character-
medium-voltage (15 kV) level for the short-term flicker index istics, however, depend on the diesel generators in operation, as
range between 0.3 and 0.6, with similar values for the long- it is analyzed in more detail in the next section, pointing toward
term flicker index . In Fig. 6, the measured values are the internal combustion engines of the DPS as the most prob-
plotted for the low wind period of Fig. 3, during which the WTs able source of the high flicker levels in the network.
were mostly out of operation, as a function of the active power
of feeders 240 and 260. IV. DIESEL GENERATOR CONTRIBUTION TO FLICKER
Flicker values around 0.6 for and particularly for are
quite high, although they are still lower than the compatibility A. Torque Pulsations of a Diesel Engine
levels (1.0 and 0.8, for and respectively, according to It is a known fact that generating units driven by internal
IEC 61 000-3-7, [15]). Nevertheless, applying the EMC coor- combustion engines may create voltage flicker problems, as it
dination principles of the IEC 61 000 series of standards, [15], has been demonstrated by field measurements in [17], where
[16], the flicker emission limits of individual users should be the basic relations for the quantitative analysis of the phenom-
such that the overall resulting disturbance does not exceed the enon have been presented. This is due to the pulsed nature of
planning levels of the network. Therefore, the limit available the engine shaft torque, which is developed by the discreet
PAPATHANASSIOU AND SANTJER: POWER-QUALITY MEASUREMENTS IN AN AUTONOMOUS ISLAND GRID 221

Fig. 8. Shaft torque due to one cylinder of the engine.

firing process of the individual cylinders. The problem becomes


exacerbated and may raise flicker concerns when firing asym-
metries exist between individual cylinders (due to fuel quality,
fuel/air ratio deviations, ignition asymmetries, etc.), giving
rise to significant low frequency harmonics (pulsations) in the
shaft torque. Depending on the total inertia of the rotating parts
(engine, flywheel, generator rotor) and the electrical damping
provided by the generator, these harmonics are reflected on
the stator current and hence on the terminal voltage of the
generator. The resulting voltage fluctuations are determined by
Fig. 9. Total shaft torque of a 12-cylinder four-stroke engine for (a)
the impedance of the system components, by the number and symmetrical firing conditions and (b) misfiring of cylinder #6.
size of the generators in operation, as well as by the action of
the AVRs, in the case of fast acting regulators. The respective period (i.e., the time interval between consec-
The shaft torque of a diesel engine is the superposition of the utive power strokes of the engine cylinders), will be
individual torques produced by each cylinder, the form of which
is illustrated in Fig. 8 ([17]), with the initial negative section cor- (6)
responding to the compression phase. In a four-stroke engine,
the firing process of a specific cylinder is repeated every two Using (4) and (5), the cylinder and engine firing frequencies
shaft revolutions, whereas in a 2-stroke one, in every shaft rev- and are calculated and included in Table I for the diesel
olution. The speed of rotation of a diesel generator with poles generators of the Samos power station.
and electrical frequency is given by For ideal (perfectly symmetrical) operation of the diesel en-
gine, the power spectrum of the shaft torque would reveal spec-
(1) tral components (harmonics) at multiples of the basic firing fre-
quency
and the respective revolution period
(7)
(2)
On the other hand, if the torques developed by the engine
cylinders are not identical, the shaft torque will present pulsa-
Hence, the period between successive power strokes (firings)
tions at the cylinder power stroke frequency . This is illus-
of one engine cylinder is
trated in Fig. 9(b), where a misfire occurs in cylinder #6 of the
12-cylinder engine. In such a case, the base frequency of the
(3) shaft torque waveform becomes equal to and its spectral
analysis would reveal components at harmonic frequencies
and the respective power stroke frequency
(8)
(4) In such a case, the engine firing frequency and its multi-
ples are also included in the harmonics. For instance,
where for a 2-stroke engine and for a 4-stroke is the harmonic of the fundamental cylinder firing fre-
one. quency .
With perfect symmetry in the firing of all cylinders, the total
shaft torque is as shown in Fig. 9(a), presenting a base pulsa- B. Validation With Measurements
tion frequency equal to the engine firing frequency , which is
The theoretical analysis of the previous section is fully con-
the product of the cylinder power stroke frequency by the
firmed by the measurements obtained in Samos, which reveals
number of cylinders
that the spectral content of the recorded voltage ripple coincides
with the expected harmonic frequencies of the operating diesel
(5) generators.
222 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 21, NO. 1, JANUARY 2006

Fig. 11. Simulation results showing the coupling between the engine shaft
torque pulsations due to cylinder firing asymmetries [diagram (a)] and the
voltage at the DPS busbars [diagram (b)].

Fig. 10. Measured power spectral density of the voltage at the DPS busbars.
Generating units in operation: (a) DG #7,8,9,12 and (b) DG #7,8,9. have been performed, to verify the coupling of the torque pulsa-
tions with the voltage of the system, without attempting to repro-
In Fig. 10, the power spectral density of the rms phase voltage duce the exact form of the actual voltage variations. Indicative
at the DPS busbars is illustrated for two characteristic operating results are shown in Fig. 11, where one 6 MW engine is oper-
conditions (one-cycle rms values have been used resulting in a ating , with the torque of cylinder #1 reduced
cut-off frequency of 25 Hz). In Fig. 10(a), where DGs #7, 8, at 80% of the normal value. It is observed that the voltage at the
9 and 12 are operating, the spectral content is located at fre- station busbars presents an oscillation of approximately 20 V
quencies , including thus the char- peak-to-peak, with a fundamental frequency equal to the engine
acteristic harmonics of both engine types according to Table I value (2.08 Hz).
(multiples of the cylinder power stroke frequency, ). Upon
disconnection of DG #12, the power spectral density of the bus V. WT EFFECT ON VOLTAGE FLICKER
voltage is modified as shown in Fig. 10(b), where the harmonics WTs are a well-known source of flicker emissions during
appear now at , i.e. at the characteristic frequencies their normal operation, due to the highly variable nature of their
of DGs #7, 8 and 9. This excellent agreement is observed using input power, to the rotor torque harmonics, as well as to
any voltage measurements, either at the DPS or at the network, other resonance modes excited by the variability of the wind.
both at MV and LV level. Therefore, their connection to the grid is expected to increase
It is therefore confirmed that the recorded voltage fluctuations the existing levels of voltage flicker, to a degree which depends
are due to firing asymmetries of the diesel power station engines. on the characteristics of the specific machines (e.g. the flicker
For the power stroke frequencies shown in Table I, the resulting index , according to [8]) and on the network parame-
voltage fluctuations occur in the frequency range between 220 ters (mainly the short circuit capacity and impedance angle,
Hz, including thus the range 510 Hz, which contributes most and , at the point of common coupling).
significantly to the flicker levels ([18]). Had the engine operation Yet, during the measurement campaign it was observed that
been ideal, only the frequencies (12.5 Hz and 33.33 Hz) the connection of the wind turbines typically resulted in a
and their multiples would have been observed. In such a case, noticable decrease of the overall flicker level. Such a situation
the voltage variations and flicker problem would be much less is illustrated in Fig. 12, corresponding to a 2-h interval, during
severe, due to the shift to higher frequencies, as well as because which the 9 225-kW wind farm is successively cut out and in
of the more effective filtering of the torque pulsations by the for measuring purposes. In the diagram, the wind farm output
mechanical inertia of the generating sets. power is shown, together with the active power of one diesel
Determination of the exact firing pattern for each engine, to generator (representative of the variation of the total DPS
quantitatively account for the time-domain voltage variations power) and the measured voltage flicker at the 15-kV busbars
and their frequency spectrum, is beyond the scope of this inves- of the DPS. It is apparent that each time the WTs are cut in, the
tigation. Nevertheless, simulations of the engine-generator-net- flicker is significantly reduced, which is a rather unexpected
work combination using a suitable transient stability code ([19]) finding.
PAPATHANASSIOU AND SANTJER: POWER-QUALITY MEASUREMENTS IN AN AUTONOMOUS ISLAND GRID 223

Fig. 14. Spectrum of the active power of feeder #240 before and after cut-in
2
of the 9 225-kW wind farm.

Fig. 12. Flicker variation at the DPS busbars during a 2-h interval with
2
successive connection and disconnection of the 9 225-kW wind farm.

Fig. 15. Active and reactive power spectrum of one 225-kW WT.

ples of the rotor rotational speed, (where or 43


rpm for the specific WT), and the turbine torsional, aeroelastic
and electro-mechanical eigenfrequencies. These frequencies
are identified on the diagram of Fig. 15, showing the spectrum
of the output power (active and reactive) of one 225-kW WT.
The identification of specific modes is possible, as it has already
Fig. 13. Harmonics of the voltage at the 15-kV busbars of the DPS (rms
2
phase-to-earth) (a) before and (b) after cut-in of the 9 225-kW wind farm.
been done in [2], where the output power spectrum of the same
WT type is given for operation in the interconnected system.
The basic differentiation here is the presence of the diesel en-
In the previous section it was shown that the voltage flicker
gine-related voltage harmonics (at , ),
in the power system of Samos is created by the torque pulsa-
which are coupled to the WT output power via the dynamics of
tions of the diesel engines. For a given system load, the cut in
its induction generator. Notably, the transfer function from the
of the wind farm results in a reduction of the power output of
terminal voltage to output power presents a resonance around
the diesel engines, with a corresponding reduction in the mag-
the rotor electromechanical poles frequency, in the vicinity of
nitude of the engine torque pulsations and hence the resulting
8 Hz.
voltage oscillations. This is visible in the diagrams of Fig. 13,
where the spectrum of the DPS voltage is shown, before and
after cut-in of the WTs. The most important voltage harmonic VI. CONCLUSION
at 12.51 Hz is reduced by about 15% after cut-in of the WTs, This paper presents results from a power-quality measure-
causing a similar reduction in the overall flicker levels when the ment campaign, conducted in the autonomous, diesel-powered
WTs are operating. The same applies for other components at system of the Greek island of Samos, where significant wind
frequencies around 10 Hz, which mostly contribute to the flicker penetration also exists. Phenomena addressed include the
effect. steady-state voltage and frequency variations, with the em-
Variations in the output of the wind turbines certainly exist, as phasis placed on the voltage flicker problem, which was found
it is evident in the power spectral density of the active power of to be quite severe under all operating conditions. The spectral
Feeder #240, shown in Fig. 14 for the interval before and after analysis of the recorded voltage variations revealed that the
cut-in of the WTs. Nevertheless, these variations are masked, diesel generators of the conventional power station are the main
regarding their flicker effect, by the aforementioned significant source of flicker in the system.
reduction in the diesel engine contribution to the flicker. The operation of the WTs was not found to create any fre-
The variability of the WT output power is concentrated quency regulation problems. Voltage regulation issues exist due
around characterstic frequencies, including the integer multi- to malfunctioning central compensating capacitors of the wind
224 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 21, NO. 1, JANUARY 2006

farms, as well as due to the lack of sufficient compensation for [12] Voltage Characteristics of Electricity Supplied by Public Distribution
the reactive power demand of the load. An important finding of Systems, European Std. EN 50160, 1999.
[13] N. G. Boulaxis, S. A. Papathanassiou, and M. P. Papadopoulos, Wind
the campaign was the actually positive effect of the WT connec- turbine effect on the voltage profile of distribution networks, Renew.
tion and operation on the fast voltage variations and the resulting Energy, vol. 25, no. 3, pp. 401415, Mar. 2002.
flicker. [14] Electromagnetic Compatibility (EMC)Part 2: EnvironmentSection
1: Description of the EnvironmentElectromagnetic Environment for
In total, although the penetration of wind power to nonin- Low-Frequency Conducted Disturbances and Signalling in Public
terconnected island systems is dealt with severe conservatism, Power Supply Systems, IEC Std. 61 000-2-1, 1990.
among other things for fears of power quality and frequency reg- [15] Electromagnetic Compatibility (EMC)Part 3: LimitsSection 7: As-
sessment of Emission Limits for Fluctuating Loads in MV and HV Power
ulation issues, the field measurements performed demonstrate Systems, IEC Std. 61 000-3-7, 1996.
that such concerns are largely unfounded. [16] Electromagnetic Compatibility (EMC)Part 1: GeneralSection 1:
Application and Interpretation of Fundamental Definitions and Terms,
IEC Std. 61 000-1-1, 1992.
ACKNOWLEDGMENT [17] P. M. Anderson and M. Mirheydar, Analysis of a diesel-engine driven
generating unit and the possibility for voltage flicker, IEEE Trans. En-
The authors gratefully acknowledge the contribution of ergy Convers., vol. 10, no. 1, pp. 3747, Mar. 1995.
CRES (M. Koulouvari and S. Tentzerakis) and PPC (Dr. D. [18] Electromagnetic Compatibility (EMC)Part 4: Testing and Measure-
ment TechniquesSection 15: Flickermeter-Functional and Design
Conidaris and DPS personnel) to the measurement campaign, Specifications, IEC Std. 61 000-4-15, 2003.
as well as the valuable input of Prof. M. Papadopoulos and Dr. [19] S. A. Papathanassiou, N. G. Boulaxis, and M. P. Papadopoulos, A sim-
N. Boulaxis of NTUA, both to the campaign and the analysis ulation code for power systems with high wind penetration, in Proc.
of the measurements. European Wind Energy Conf., Copenhagen, Denmark, Jul. 2001.

REFERENCES
[1] G. Gerdes and F. Santjer, Power quality of wind turbines and their in- Stavros A. Papathanassiou (S93M98) was born
teraction with the grid, in Proc. Eur. Wind Energy Conf., Thessaloniki, in Greece in 1968. He received the Diploma in
Greece, Oct. 1994. Electrical Engineering and the Ph.D. degree from the
[2] T. Thiringer, Power quality measurements performed on a National Technical University of Athens (NTUA),
lowvoltage grid equipped with two wind turbines, IEEE Trans. Athens, Greece, in 1991 and 1997, respectively.
Energy Convers., vol. 11, no. 3, pp. 601606, Sep. 1996. He was with the Distribution Division of the Public
[3] , Power quality impact of a sea located hybrid wind park, IEEE Power Corporation of Greece in power-quality and
Trans. Energy Convers., vol. 16, no. 2, pp. 123127, Jun. 2001. dispersed generation studies, responsible for the elab-
[4] A. Larsson, Flicker emission of wind turbines during continuous oper- oration of DG interconnection guidelines. In 2002, he
ation, IEEE Trans. Energy Convers., vol. 17, no. 1, pp. 114118, Mar. was a Lecturer with the Electric Power Division of
2002. NTUA. His research mainly deals with wind turbine
[5] , Flicker emission of wind turbines caused by switching opera- technology and the integration of distributed generation (DG) in distribution
tions, IEEE Trans. Energy Convers., vol. 17, no. 1, pp. 119123, Mar. networks.
2002. Dr. Papathanassiou is a member of CIGRE and a registered professional en-
[6] M. P. Papadopoulos, S. A. Papathanassiou, S. T. Tentzerakis, and N. gineer and member of the Technical Chamber of Greece.
G. Boulaxis, Investigation of the flicker emission by grid connected
wind turbines, in Proc. Int. Conf. Harmonics Quality of Power, Athens,
Greece, Oct. 1998.
[7] N. G. Boulaxis, S. A. Papathanassiou, and M. P. Papadopoulos, Wind
turbine effect on the voltage profile of distribution networks, Renew. Fritz Santjer was born in Germany in 1960. He re-
Energy, vol. 25, no. 3, pp. 401415, Mar. 2002. ceived the Diploma in Electrical Engineering from
[8] Wind Turbine Generator SystemsPart 21: Measurement and Assess- the University of Siegen, Siegen, Germany, in 1989.
ment of Power Quality Characteristics of Grid Connected Wind Tur- In 1990, he joined the German Wind Energy In-
bines, IEC Std. 61 400-21, 2001. stitute (DEWI) where he works with grid connection
[9] G. W. Scott, V. F. Wilreker, and R. K. Shaltens, Wind turbine generator and power quality of wind turbines and wind farms
interaction with diesel generators on an isolated power system, IEEE and with stand-alone systems. In 2000, he became
Trans. Power App. Syst., vol. PAS-103, no. 5, pp. 933937, May 1984. Head of the Electrical Systems group at DEWI.
[10] G. A. Vokas and A. V. Machias, Harmonic voltages and currents on two He is an assessor for the MEASNET power-quality
Greek islands with photovoltaic stations: study and field measurements, procedure and involved in national and international
IEEE Trans. Energy Convers., vol. 10, no. 2, pp. 302306, Jun. 1995. working groups regarding guidelines on power
[11] M. P. Papadopoulos, S. A. Papathanassiou, and S. T. Tentzerakis, Mod- quality and the grid connection of wind turbines. He was involved in various
eling of induction machine main flux saturation in the arbitrary reference European research projects concerning grid connection and power quality
framea case study on wind turbine self-excitation, in Proc. Int. Conf. of wind turbines, stand-alone systems, and simulations of wind turbines and
Electric Machines, Vigo, Spain, Sep. 1996. networks.

Vous aimerez peut-être aussi