Aalborg Universitet

Harmonic Analysis of Offshore Wind Farms with Full Converter Wind Turbines

Kocewiak, Lukasz Hubert; Hjerrild, Jesper; Bak, Claus Leth

Published in:
Proceeding of the 8th International Conference on Large-Scale Integration of Wind Power into Power Systems

Publication date:
2009

Document Version
Publisher's PDF, also known as Version of record

Link to publication from Aalborg University

Citation for published version (APA):

Kocewiak, L. H., Hjerrild, J., & Bak, C. L. (2009). Harmonic Analysis of Offshore Wind Farms with Full Converter
Wind Turbines. In Proceeding of the 8th International Conference on Large-Scale Integration of Wind Power into
Power Systems Energynautics GmbH.

General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

- Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
- You may not further distribute the material or use it for any profit-making activity or commercial gain
- You may freely distribute the URL identifying the publication in the public portal -
Take down policy
If you believe that this document breaches copyright please contact us at vbn@aub.aau.dk providing details, and we will remove access to
the work immediately and investigate your claim.

Downloaded from vbn.aau.dk on: July 07, 2022

 Harmonic Analysis of Offshore Wind Farms
with Full Converter Wind Turbines
Łukasz Hubert Kocewiak, Jesper Hjerrild, Claus Leth Bak

 effects of harmonics [18].

Abstract--This paper presents the harmonic analysis
of offshore wind farm (OWF) models with full converters
A. Summary of a problem
represented as harmonic sources and measurement data on the Mainly wind turbines in large offshore wind farms are
point of common coupling (PCC) during normal operation. The equipped with induction generators, connected to the grid
model describes a wind farm (WF) with full rated converters through full-scale frequency converters. The concept of such
installed connected to a shore with a long HV cable. The way configuration is shown in Fig. 1
of propagation and effects of harmonics are presented
Full Scale Converter
for different study cases. Modeling strategies of harmonic
sources for harmonic analysis are described and compared.
AC DC
Different results dependent on applied harmonic models
are shown and discussed in this paper. G
Simulation results are compared with representative DC AC
measurement data obtained from measurement campaign done
in an OWF situated in United Kingdom. The problems Fig. 1 Wind turbine configuration with an induction generator and a full-
scale converter.
and limitations related to present standards and practice will
be discussed based on measurements from offshore wind farm.
Full-rated power converter applied in wind turbines is
Index Terms--full-rating converters, harmonic analysis, used to interface a generator that provides a variable voltage
offshore wind farm, wind turbine, validation with at variable frequency to a supply network operating at fixed
measurements frequency and including features that allow the power
converter to remain connected to the supply network and
I. INTRODUCTION retain control during supply network fault and transient
conditions. Those features allow the converter to find a
T HE tendency to rapid increase in grid-connected,
especially, in onshore located coastal areas, and offshore
located open seas, WFs has made it necessary for
broad application in the wind power industry. This kind of
power converter includes a generator bridge electrically
connected to the stator of the generator and a network
manufacturers and transmission system operators (TSO) to
bridge. A DC link is connected between the generator bridge
investigate in large OWFs from a power grid operation,
and the network bridge. A filter with network terminals is
harmonic emission, stability and control point of view. connected between the network bridge and the supply
Nowadays wind turbines are often grouped in large wind network through the wind turbine transformer as shown in
farms, installed offshore and connected directly to the high Fig. 1 [15].
voltage network via long AC cable lines [9]. The electrical
conditions present in the collection grids for large offshore
wind farms are not similar to any other industrial application
and should be investigated separately.
The number of wind turbines with full converters used in
large offshore wind farms is rapidly increasing. Most
frequently they are connected through a widespread MV
cable network and connected to the transmission system by
long HV cables. This generates advantages such as higher
power system reliability and dispersed generation of
electricity but also represents new challenges to the industry
in relation to understanding the nature, propagation and

This work is made as part of an Industrial Ph.D. project supported by Fig. 2 Wind turbines from Burbo Bank Offshore Wind Farm
the Danish Ministry of Science, Technology and Innovation, project (daylife.com).
number 08-044839.
Ł. H. Kocewiak, J. Hjerrild, are with DONG Energy, Denmark (e-mail:,
lukko@dongenergy.dk, jeshj@dongenergy.dk). In order to investigate harmonic emission of offshore
C. L. Bak is with the Institute of Energy Technology, Aalborg wind farms, measurement campaigns have been conducted.
University, Denmark (e-mail: clb@iet.aau.dk).
In this paper Burbo Bank Offshore Wind Farm (BBOWF) is
taken into consideration as an example, where measuring
B. Proposed analysis approach
systems were installed and used for simultaneous
measurement at different locations within the wind farm. The analysed WT conversion system is to be rated at 3.6
The BBOWF is located on Burbo Flats in Liverpool Bay MVA output power focusing on a full converter solution for
(see Fig. 3). At its closest point, the site is approximately 6.4 different kinds of generators in order to avoid dealing with
km from the Selfton coastline and 7.2 km from North Wirral. slip rings, which is extremely important if offshore wind
The wind farm consists of 25 SWT-3.6-107 Variable Speed farms are exposed to truculent environment, and any low
wind turbines [7], each with rated capacity of 3.6 MW. voltage ride through (LVRT) difficulties characteristic for
BBOWF is therefore capable of providing a maximum demanding grid codes [7].
output of 90MW of electricity. The WT has a fully rated Modeling strategies of harmonic sources such as power
power converter IGBT-based power electronic converter. converters for harmonic analysis sometimes give different
results which will be shown in this paper.
Resonances may be excited by a relatively small distortion
source in the system or by an imbalance in the converter
components or control. The resulting amplification of the
small source by the resonant characteristics of the system
can compromise the normal operation or even lead to
instability. That is why power converters from WTs
harmonic emission point of view should be deeply
investigated [2, 17]. The schematic clarification of this
behavior is shown in Fig. 5.

PCC Cc

TPT
vS
ZSR Series
resonance

Parallel
resoance T2

Fig. 3 Location of Burbo Bank Offshore Wind Farm in Liverpool Bay.

T1 L1

The wind turbines are arranged as it is described in Fig. 3. C2

Three 33 kV radials go back to the shore-based substation Fig. 5 Simple representation of a WF connected to the network with
adjacent to the SP Manweb BSP substation and they are possible resonances excitation.
located near the consumption centre. 9, 8 and 8 wind
turbines are connected to each radial respectively as An electrical system, which contains capacitance in the
presented in Fig. 4. The substation consists of the wind park form of cables, overhead lines or capacitor banks, will have
transformer, a capacitor bank and earthing transformer with some frequencies where the reactance of the capacitors and
is also used for substation supplies. the reactance of the system are equal, opposite and in
parallel. This situation becomes very important if large
offshore wind farms are connected with a long cable to the
network. High long HV cable or capacitor banks
capacitances create resonances in a low frequency range, up
to 1000 Hz, where harmonic content exists very often. When
a parallel resonance appears that harmonic current will be
excited to oscillate between the energy storage in the
inductance and the energy storage in the capacitance. It
affects every WF component negatively and finally it may
damage the system [3].

I R Ltr

UR UL
UC Cc

Fig. 6 Series resonance circuit where reactance of the capacitor and the
inductor are equal.
Fig. 4 Layout of Burbo Bank Offshore Wind Farm in Liverpool Bay.
 A series resonance can also occur where the reactance of in the 33 kV collection grid the main modeled components
the capacitors and the reactance of the system are equal but are:
in series and this would cause a low-impedance path for o three cable radials with different cross-sections,
harmonic current as shown in Fig. 7. Therefore, the o earthing transformer with connected loads,
amplitude of the current during the resonance can reach very o capacitor bank (PFC) for reactive grid code
high value. Series resonance can result in high voltage compliance;
distortion levels between the inductance and the capacitor in in the 0.69 kV for each WT the main modeled components
the series circuit. During the resonance the inductive are:
reactance of system components such as transformers Ltr is o 4 MVA wind turbine transformer,
equal to the capacitive reactance of cables or capacitor o full scale PWM converter,
banks Cc as shown in (1): o high frequency filter designed to cut-off power
converter switching frequency,
1 o grid converter reactor.
Ltr  (1)
Cc A. External network configuration
BBOWF is connected to SP Manweb’s Network which is
This condition is fulfilled at the resonant frequency designed to operate substations in interconnected groups
described by (2): with standard transformer and cable sizes. In order to limit
effects of distortion of the system voltage waveform, the
1 harmonic content of any connected load complies with the
f  (2)
2 Ltr Cc limits set out in Engineering Recommendation G4/5 [10].
The 132 kV and 33 kV networks comprise sections of
underground cable or overhead lines or combinations of
When system configuration creates harmonic resonance each. The BBOWF is connected to the 132 kV Wallasey
and harmonic excitation exists as well, large current of the Circuit 1.
resonant frequency will flow through the circuit causing The cables and overhead lines in the network could create
large voltage drop [4, 5]. parallel harmonic resonances dependent on the system
configuration characterised by the system impedance. As
both factors are difficult to identify from measurement data,
the network model configuration for simulation purpose has
been chosen based on load profiles from ScottishPower
Long Term Development Statement [14].
UL UC
B. Subsea cables and transformers
The submarine cables connect the wind turbines to each
UR other and to the submarine export cables, which in turn
I connects the wind farm to the onshore substation in
Fig. 7 Series resonance phasor diagram. Wallasey (34/132 kV). All land and submarine cables have
been modeled as long-lines with skin effect approximated as
In case of series resonance high voltages drop will have a square-root function against harmonic order [8, 9].
opposite signs. Therefore, the sum of the two voltages will For a precise modeling of high frequency effects of
be zero, but each of them will have high amplitude as transformers, additional capacitances need to be considered.
presented in Fig. 6. This kind of problem may occur in The high frequency model provides an accurate frequency
OWFs especially from the resonant circuits of cable response with respect to voltages and currents at the
connections [6]. transformer terminals as it is implemented in Power
Factory [16].
II. SYSTEM CONFIGURATION C. Wind turbine representation
In the simulated system three voltage levels have been A voltage sourced converter defines the voltage waveform
taken into consideration. The 0.69 kV on each wind turbine, at the bus-bar to which it is connected, that is why the most
33 kV collection grid and extended 132 kV power grid have accurate harmonic model is a harmonic voltage source. The
been included in the model created in DIgSILENT Power series reactance in the model represents the coupling
Factory 14.0. reactance that is modeled internally in the full scale PWM
In the 132 kV export system the main modeled converter model. The harmonic voltage is defined by a
components are: frequency series used as a look-up table.
o SP Manweb grid, In many applications, harmonic injections are given as
o overhead line (OHL) with 1.83 km length and 5.78 harmonic current injections at the output of the converter.
km cable connection to the substation, Therefore to represent the converter in a more realistic way,
o 33 kV consumption centre, a harmonic current source has been defined and the
o 90 MVA park transformer; amplitude and angle of the harmonic currents have been
defined. This approach is only valid if it can be assumed that
the coupling reactance is very high compared to other observe interaction between power converter DC side
network impedances. In this case, the equivalent voltage capacitor with the smoothing reactor and the AC system
source can be transformed into a pure current-source, impedance.
without internal admittance, with sufficient accuracy. In Results presented in Fig. 9 show that relations between
cases, in which this cannot be assumed, the actual level of harmonics are similar in comparison to measurement data
harmonic current injections is depends on the network but harmonic distortion is much lower than observed in
impedance. BBOWF. It shows that much lower impedance in case of
For both current source and voltage source the unbalanced voltage source causes more impedance damping and shifting
representation has been chosen [16]. when harmonic current source is used for simulations.
D. Measuring equipment
The measurements were carried out with a PC equipped
with National Instruments data acquisition card, running by a
programme developed in LabVIEW. Voltage and currents
were sampled at 44.1 kHz, using NI PCI-4472 8-Channel
Dynamic Acquisition Board. The dynamic signal acquisition
board has analog filter to remove any signal components
beyond the range of the analog to digital converters (ADCs).
However, in order to cut frequency components above half
programmed sampling rate digital anti-aliasing filters are
implemented.

III. HARMONIC ASSESSMENT

A range of harmonic analysis studies has been undertaken Fig. 9 Harmonic analysis of the BBOWF Power Factory model at the PCC
with harmonic voltage source applied. Harmonic voltages are presented in
over a range of network operation conditions. It determines % against harmonic order.
the harmonic impedance which the BBOWF is likely to see
and the typical WF harmonic emission in the PCC. The In the next step harmonic current source has been used to
frequency range of interest is 0-1250 Hz. analyse BBOWF harmonic emission in the PCC. As it is
Analysed measurement data is presented in Fig. 8. During presented in Fig. 10, results are different in comparison to
measurements study process it has been observed that harmonic voltage source. The first noticeable difference is
mainly 3rd, 5th, 7th, 11th and 13th harmonics have changed 7th harmonic domination when for voltage source 5th
depending on production and measurement time. Presented harmonic is the highest. It has been observed during
in the paper measurement data frequency analysis has been measurement data analysis that this happens also in real life.
assumed to be the most common and representative. This observation raises the question whether the set of
measurement data is representative and appropriate to this
case study. From impedance plot it will be seen that high
resonance peak close to 7th harmonic is the main reason of
this amplification.

Fig. 8 Harmonic analysis of the voltage waveform measured with 44.1 kH

sampling ratio at the PCC. Harmonic voltages are presented in % with
harmonic order on the abscissa axis.
Fig. 10 Harmonics at the PCC obtained from the BBOWF Power Factory
Basis of daily and annual load profiles from Distribution model simulation with harmonic current source applied. Harmonic voltages
Long Term Development Statement [14] typical network are presented in % with harmonic order on the abscissa axis.
configuration has been chosen for simulation purposes.
Other possibilities have been taken into consideration as From the harmonic emission analysis in the PCC it can be
well, but no significant changes have been observed. observed that both models are not able to give satisfactory
Firstly harmonic voltage source has been applied and proper results. Real harmonics level is somewhere between
analysed during simulation process. It should be emphasised both obtained from calculations. From measurement data
that no DC link capacitive impedances have been taken into analysis of different periods during day and night it has been
consideration [11]. This generalization makes impossible to observed that harmonic level, especially in case of 5th and
7th, had never been so high as in case of current source and
as low as it was calculated from simulation with voltage
source.
Different power converter modeling for harmonic load
flow analysis has also shown that different models have an
influence on harmonic impedance in the PCC. This affects
harmonic levels and in consequence WF harmonic emission
assessment. It has been shown that can play a crucial role in
analysis process and power quality assessment.

Fig. 14 Angle of the network impedance calculated at the PCC with

current source applied.

As it was described previously impedance plot proves (see

Fig. 11, 12) that voltage source behaves in a different way
for low frequencies in comparison to current source.
Unfortunately both models show inaccurate results. In case
of harmonic voltage source there is too high damping around
5th and 7th harmonics while for harmonic current source it is
too low. The results show that real measurement waveform
Fig. 11 Magnitude of the network impedance calculated at the PCC with spectrum is somewhere between two different simulated
voltage source applied.
study cases.

IV. CONCLUSIONS
This paper describes large offshore wind farm harmonic
emission assessment obtained basis of different models of
full scale power converter. Voltage and current harmonic
sources have been taken into consideration in analysis
process. Burbo Bank wind farm situated in Liverpool Bay
has been used as example. Simulation results have been
compared with measurements. Results were presented up to
25th harmonic.
The measurement results have been used for verification
Fig. 12 Angle of the network impedance calculated at the PCC with of simulation models of the wind farm thereby making it
voltage source applied. possible to have a more accurate determination of harmonic
emission and propagation in a wind farm mainly during
Frequency sweep impedance plots show that for different steady state operation.
power converter modeling scenarios different impedances It was shown that different harmonic sources in offshore
characteristics appear in the PCC as well. It also is reflected wind farms modeling techniques give different results. Both
in presented above harmonic emission analysis. harmonic load flow and sweep frequency response analysis
in the point of common coupling (PCC) in BBOWF give
different results for different models. Analysis has shown
that the results obtained in Power Factory are similar but not
identical to measurement data. This fact implies that
simulation techniques in the frequency, time and harmonic
domains and modeling of the wind turbines as harmonic
sources should be extended. It is necessary to find a better
agreement between theory and experiment.
The comparison of power converter represented as a
harmonic voltage source or current source shows that in case
of voltage source the impedance is noticeably lower and
hence has larger impact on the system. It causes more
Fig. 13 Magnitude of the network impedance calculated at the PCC with impedance damping and shifting than when harmonic
current source applied. current source is investigated.
From a general investigation of the external network, it
seems not to have any significant resonant conditions that
rise suspicions.
The need for accurate simulations is major for OWFs as
consequences of faults are more severe in terms of repair
costs and lost revenue than for onshore based WFs. The [16] DIgSILENT Power Factory manual, v.14.0, DIgSILENT GmbH,
2008, Gomaringen, Germany.
result of simulations can always be questioned depending on
[17] P.W. Lehn, “Direct harmonic analysis of the voltage source
the accuracy of the component models used in the simulation converter”, IEEE Transactions on Power Delivery, vol. 18, no. 3,
programme, and validation of models and simulations with July 2003, p. 1034-1042.
reliable measurements performed in a real large WF, makes [18] Sokratis T. Tentzerakis; Stavros A. Papathanassiou, “An Investigation
it possible to verify and improve the simulations to give of the Harmonic Emissions of Wind Turbines”, IEEE Transactions
on Energy Conversion, Volume 22, Issue 1, March 2007 p. 150-158.
more reliable results [18]. [19] S. Tentzerakis, N. Paraskevopoulou, S. Papathanassiou, P.
Insufficient agreement between simulations and Papadopoulos, “Measurement of wind farm harmonic emissions”, in
measurements is a premise for future work on models Proc. IEEE Power Electronics Specialists Conference, 15-19 June
development. Both IEEE and IEC standards consider 2008, p. 1769-1775.
harmonics in a general sense, without regard to characteristic
harmonics generated by certain types of equipment or VI. BIOGRAPHIES
special operation modes [12, 13]. The above presented
analysis shows the need to extend harmonic sources Łukasz Kocewiak was born in Grójec, Poland, in 1983. He received B.Sc.
description in standards and to define more precisely power and M.Sc. degrees in electrical engineering from Warsaw University
of Technology.
converters more precisely and other wind farm components Currently he is an Industrial PhD student in cooperation with DONG
harmonic models. It has been shown that different modes Energy and Aalborg University. The main direction of his research
give different result and in consequence different harmonic is related with harmonics and nonlinear dynamics in power electronics and
emission assessment of OWFs what implies problems with power systems.
agreement with standards and restricted grid codes. Jesper Hjerrild was born in 1971. He received the M.Sc. and Ph.D.
degrees in electrical engineering from the Technical University of
V. REFERENCES Denmark, Lyngby, in 1999 and 2002, respectively.
Currently he has been employed at Dong Energy. His main technical
[1] T. Ackerman, “Wind Power in Power Systems”, John Wiley and
interest is electrical power systems in general, involving a variety of
Sons, I edition, 2005, p. 56.
technical disciplines including modelling of power system including wind
[2] N. R. Watson, J. Arrillaga, “Power System Harmonics”, John Wiley
power and power system control, stability and harmonics. Furthermore, he
and Sons, 2003.
also works with designing of the wind farm
[3] A. Baggini, “Handbook of Power Quality”, John Wiley and Sons,
From 2002 until 2004 Jesper Hjerrild was employed at DEFU (The
2008.
Association of Danish Energy Companies R&D).
[4] M. Madrigal, E. Acha, “Power Systems Harmonics: Computer
Modelling and Analysis”, John Wiley and Sons, 2001.
[5] George J. Wakileh, “Power Systems Harmonics: Fundamentals, Claus Leth Bak was born in Djursland, Denmark, in 1965. He received B.
Analysis, and Filter Design”, Springer, 2001. Sc. in Electrical Power Engineering from the engineering college in Århus
[6] W. Wiechowski, J.C. Hygebjerg, P.B. Eriksen, “Higher Frequency in 1992, he received M.Sc. in Electrical Power Engineering in 1994.
Performance of AC Cable Connections of Offshore Wind Farms – He is an Associate Professor at Aalborg University with experience on
Studies of the Danish TSO”, in Proc. 7th International Workshop on high voltage engineering, relay protection for transmission systems and
Large Scale Integration of Wind Power and on Transmission substation automation and dynamic analysis (PSCAD/EMTDC) of large
Networks for Offshore Wind Farms, 26-27 May, 2008, p. 211-217. power systems.
[7] V. Akhmatov, J. Nygaard Nielsen, J. Thisted, E. Grøndahl, P. Egedal,
M. Nørtoft Frydensbjerg, K. Høj Jensen, “Siemens Windpower 3.6
MW Wind Turbines for Large Offshore Windfarms”, in Proc. 7th
International Workshop on Large Scale Integration of Wind Power
and on Transmission Networks for Offshore Wind Farms, 26-27
May, 2008, p. 494-497.
[8] C. H. Chien, R. W. G. Bucknall, “Theoretical Aspects of the
Harmonic Performance of Subsea AC Transmission Systems for
Offshore Power Generation Schemes”, in Proc. Generation
Transmission and Distribution, 2006, p. 599-609.
[9] W. Wiechowski, P. B. Eriksen, “Selected studies on offshore wind
farm cable connections – challenges and experience of the Danish
TSO”, in Proc. Power and Energy Society General Meeting –
Conversion and Delivery of Electrical Energy in the 21st Century,
20-24 July 2008, p. 1-8.
[10] Planning levels for harmonic voltage distortion and connection of
non-linear equipment to transmission and distribution networks in
the UK, Engineering Recommendations (ER) G5/4-1.
[11] Measurements and Assessment of Power Quality Characteristics of
Grid Connected Wind Turbines, IEC Standard 61400-21,
International Electrotechnical Commission, 2008.
[12] Recommended Practice for Harmonic Control in Electric Power
Systems, IEEE Standard 519-1992, Institute of Electrical and
Electronics Engineers, 1992.
[13] Assessment of Emission Limits for Distorting Loads in MV and HV
Power Systems, IEC Standard 61000-3-6, International
Electrotechnical Commission, 1996.
[14] Long Term Development Statement, SP Manweb for the years
2008/09 to 2012/13, ScottishPower, [Online] Available:
http://www.scottishpower.com/OtherDocuments.htm
[15] R. Jones, P. B. Brogan, E. Grondahl, H. Stiesdal, “Power
Converters”, U.S. Patent 7 372 174 B2, May 13, 2008.