//INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM Wind Power in Power Systems Edited by Thomas Ackermann Royal Institute of Technology Stockholm, Sweden //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM Wind Power in Power Systems KTH VETENSKAP OCH KONST ROYAL INSTITUTE OF TECHNOLOGY Electric Power Systems http://www.ets.kth.se/ees //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM Wind Power in Power Systems Edited by Thomas Ackermann Royal Institute of Technology Stockholm, Sweden //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM Copyright Ó 2005 John Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England Telephone (þ44) 1243 779777 Email (for orders and customer service enquiries): cs-books@wiley.co.uk Visit our Home Page on www.wiley.com All Rights Reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1T 4LP, UK, without the permission in writing of the Publisher Requests to the Publisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed to permreq@wiley.co.uk, or faxed to (þ44) 1243 770571 This publication is designed to provide accurate and authoritative information in regard to the subject matter covered It is sold on the understanding that the Publisher is not engaged in rendering professional services If professional advice or other expert assistance is required, the services of a competent professional should be sought Other Wiley Editorial Offices John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA Wiley-VCH Verlag GmbH, Boschstr 12, D-69469 Weinheim, Germany John Wiley & Sons Australia Ltd, 33 Park Road, Milton, Queensland 4064, Australia John Wiley & Sons (Asia) Pte Ltd, Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809 John Wiley & Sons Canada Ltd, 22 Worcester Road, Etobicoke, Ontario, Canada M9W 1L1 Library of Congress Cataloging in Publication Data Wind power in power systems / edited by Thomas Ackermann p cm Includes bibliographical references and index ISBN 0-470-85508-8 (cloth : alk paper) Wind power plants Wind power I Ackermann, Thomas II Title TK1541.W558 2005 621.310 2136—dc22 2004018711 British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 0-470-85508-8 Typeset in 10/12pt Times by Integra Software Services Pvt Ltd, Pondicherry, India Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire This book is printed on acid-free paper responsibly manufactured from sustainable forestry in which at least two trees are planted for each one used for paper production //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM To Moana, Jonas and Nora //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM Contents Contributors Abbreviations Notation Units xx xxix xxxvi xlvi Introduction Thomas Ackermann Part A Theoretical Background and Technical Regulations Historical Development and Current Status of Wind Power Thomas Ackermann 2.1 Introduction 2.2 Historical Background 2.2.1 Mechanical power generation 2.2.2 Electrical power generation 2.3 Current Status of Wind Power Worldwide 2.3.1 Overview of grid-connected wind power generation 2.3.2 Europe 2.3.3 North America 2.3.4 South and Central America 2.3.5 Asia and Pacific 2.3.6 Middle East and Africa 2.3.7 Overview of stand-alone generation 2.3.8 Wind power economics 2.3.9 Environmental issues 2.4 Status of Wind Turbine Technology 2.4.1 Design approaches 2.5 Conclusions Acknowledgements References 8 11 11 11 13 16 16 17 18 18 20 21 22 23 23 23 //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_01_PREA01.3D – – [1–48/48] 20.12.2004 7:23PM Contents viii Wind Power in Power Systems: An Introduction Lennart So¨der and Thomas Ackermann 3.1 3.2 3.3 3.4 3.5 3.6 Introduction Power System History Current Status of Wind Power in Power Systems Network Integration Issues for Wind Power Basic Electrical Engineering Characteristics of Wind Power Generation 3.6.1 The wind 3.6.2 The physics 3.6.3 Wind power production 3.7 Basic Integration Issues Related to Wind Power 3.7.1 Consumer requirements 3.7.2 Requirements from wind farm operators 3.7.3 The integration issues 3.8 Conclusions Appendix: A Mechanical Equivalent to Power System Operation with Wind Power Introduction Active power balance Reactive power balance References Generators and Power Electronics for Wind Turbines Anca D Hansen 4.1 Introduction 4.2 State-of-the-art Technologies 4.2.1 Overview of wind turbine topologies 4.2.2 Overview of power control concepts 4.2.3 State-of-the-art generators 4.2.4 State-of-the-art power electronics 4.2.5 State-of-the-art market penetration 4.3 Generator Concepts 4.3.1 Asynchronous (induction) generator 4.3.2 The synchronous generator 4.3.3 Other types of generators 4.4 Power Electronic Concepts 4.4.1 Soft-starter 4.4.2 Capacitor bank 4.4.3 Rectifiers and inverters 4.4.4 Frequency converters 4.5 Power Electronic Solutions in Wind Farms 4.6 Conclusions References Power Quality Standards for Wind Turbines John Olav Tande 5.1 Introduction 5.2 Power Quality Characteristics of Wind Turbines 25 25 25 26 28 29 32 32 33 34 40 40 41 41 46 47 47 48 49 50 53 53 53 53 55 55 59 62 65 66 69 70 72 72 72 73 74 75 77 77 79 79 80 //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_28_CHA27.3D – 621 – [603–628/26] 17.12.2004 10:55PM Wind Power in Power Systems 621 At the time of reconnection, T2, switch SW 01 was closed and section with wind turbine WT 01 was reconnected to the entire power system During the experiment, the voltage and the current at the low-voltage side of the 0.7/10 kV transformer of wind turbine WT 01 were measured, as shown in Figure 27.9 The sampling frequency of the measuring equipment was kHz (Raben et al., 2003) and any delay in the measured signals produced by the equipment can be neglected During islanded operation, the no-load capacitor of wind turbine generator WT 01 was kept grid-connected This was done to reduce the possibility of a voltage drop at the terminals of wind turbine WT 01 during islanded operation (Raben et al., 2003) 27.3.2 Measured behaviour 2000 1800 1600 1400 1200 1000 800 600 400 200 –200 –400 –600 –800 –1000 (a) 1200 Phase current/500 ms 1000 800 600 Phase–phase voltage/500 ms 400 Voltage (V) Current (A) Figure 27.10 presents the measured behaviour of the phase current, IL, and the phase-tophase voltage, ULL During islanded operation, the current of the no-load compensated induction generator of wind turbine WT 01 was not zero, because the induction motors in section absorbed a certain amount of electric power produced by the induction generator of wind turbine WT 01; also, some reactive power was exchanged via the internal power network of section 1, as there was the no-load impedance of the transformers and the cable charging During islanded operation, and shortly after reconnection, higher harmonics of the fundamental frequency of 50 Hz were seen in the measured current The higher harmonics were presumably caused by the induction machines This coincides with the results of previous studies (Pedersen et al., 2000) The measured voltage showed no (significant) drop during islanded operation At tripping, t ¼ T1, there was no DC offset in the measured phase current However, there was a noticeable DC offset in the measured phase current at reconnection, t ¼ T2 There was no DC offset during tripping because the opening of switch SW 01 was an unbalanced three-phased event (Section 27.2.1 explains the elimination of the DC offset in the phase current during tripping.) Switch SW 01 was closed at the same time, t ¼ T2, 200 –200 –400 –600 –800 –1000 2.2 2.4 2.6 2.8 Time (s) 3.2 3.4 3.6 3.8 –1200 (b) 2.2 2.4 2.6 2.8 3.2 Time (s) 3.4 3.6 3.8 Figure 27.10 Measured behaviour of (a) phase-current and (b) phase-to-phase voltage Reprinted from Akhmatov, V Analysis of Dynamic Behaviour of Electric Power Systems with Large Amount of Wind Power, Ph.D thesis, Technical University of Denmark, Kgs Lyngby, Denmark, copyright 2003, with permission from the copyright holder //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_28_CHA27.3D – 622 – [603–628/26] 17.12.2004 10:55PM 622 Full-scale Verification of Dynamic Wind Turbine Models in all the three phases The reconnection was therefore a balanced transient event initiating the DC offset in the phase current During and after islanded operation, the voltage magnitude was almost unchanged The measured current showed a fluctuating behaviour after reconnecting The natural frequency of the current fluctuations was about Hz and could not be explained by the dynamic behaviour of the induction generator only The value of the natural frequency indicated that the current fluctuations were related to the torsional oscillations of the shaft system of wind turbine WT 01 27.3.3 Modelling case The positive-sequence equivalent of the experimental network is implemented in the tool PSS/ETM that is used for the validation It is essential that the network representation and its load flow solution are in agreement with the factual conditions of the experiment SEAS provided the data of the internal power network of the wind farm, including for the cables and the transformers, and the manufacturer Bonus Energy supplied the data of the fixed-speed wind turbines (Akhmatov, 2003a) The values of the phase current and phase-to-phase voltage, which were measured just before opening switch SW 01, are used for initialising wind turbine generator WT 01 The numeric value of the apparent power, S, of the no-load compensated induction generator of wind turbine WT 01 is computed from: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffi S ¼ P2 þ Q2 ¼ 3ULL IL ; ð27:1Þ where P denotes the electric power of the induction generator of wind turbine WT 01 (PG), and Q ¼ QG À QC, where QG is the reactive power absorbed by the induction generator and QC is the reactive power of the no-load capacitor From the measured behaviour of the current and the voltage shown in Figure 27.10, the numeric value of the apparent power of the no-load compensated induction generator is estimated to be 80 kVA This is the value before the islanded operation Using this information and the data of the generator and the no-load capacitor, and applying iterations, we arrive at the initial operational point of the induction generator of wind turbine WT 01 Assuming a regular wind distribution over the wind farm, other gridconnected wind turbines (in three other sections) were initialised to the same operational points The operational points of the induction motors of the disconnected wind turbine generators are not given but are estimated from the measured current and voltage during islanded operation During islanded operation, the measured value of the phase current peak is in the range of 60–65 A, and the voltage magnitude does not change significantly This current represents mainly the electric power absorbed by the induction motors in section of the wind farm Using the above-described procedure for initialising wind turbine generator WT 01, the operational points of the induction motors are estimated The induction motors absorb approximately 10 kW per (disconnected) wind turbine Despite our efforts carefully to represent the experimental conditions in the simulation case, there will always be a small number of uncertainties and missing data The //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_28_CHA27.3D – 623 – [603–628/26] 17.12.2004 10:55PM Wind Power in Power Systems 623 manufacturer did not provide the damping coefficients that are required for the twomass model equations, for instance The representation is simplified and the damping coefficients are set to zero in the simulations In dynamic simulations, the wind turbine generator is computed with the transient fifth-order model For the induction motors, however, the standardised PSS/ETM model CIMTR4 is used – that is, the common third-order model of induction motors (i.e the current transients in the induction motors during reconnection are neglected) The no-load capacitor of wind turbine WT 01 is modelled in accordance with the dynamic interface of the tool PSS/ETM Therefore, the transients in the capacitor current are also neglected, even though there are probably such current transients in the measurements (Larsson and Thiringer, 1995) 27.3.4 Model validation The main target is to validate the user-written dynamic model of Type A wind turbines, which was implemented in the tool PSS/ETM This model contains the fifth-order model of induction generators and a shaft system representation with the two-mass model For reasons of comparison, the same simulations are carried out using of the following models: A model with the fifth-order model of induction generators and the lumped-mass model (i.e assuming a very stiff shaft system); a model containing the common third-order model of induction generators and the two-mass model of the shaft system The simulation tool PSS/ETM operates with positive-sequence equivalents of the network models and computes the phasor values of voltages and currents Figure 27.11 shows the simulated behaviour, in phasor values, of the voltage and the current For the validation, we therefore compare the measured phase values with the computed phasor values As demonstrated in Section 27.2.1, the current phasor follows the magnitude behaviour of the phase current with developed DC offset The behaviour simulated with the dynamic wind turbine model is in agreement with the measured behaviour The simulated current phasor is correctly initiated, drops to around 60 A (peak-phase) during islanded operation, contains the fundamental frequency transients during reconnection and shows an oscillating behaviour with the natural frequency of around Hz The wind turbine model containing the lumped-mass model of the mechanical system underpredicts values for the current transients at reconnection Furthermore, it does not predict oscillations of the current after reconnecting The wind turbine model with the common third-order model of induction generators underpredicts values of the current at reconnection, because the fundamental frequency transients in the machine current are neglected This model does, however, predict current oscillations with the natural frequency of around Hz The oscillating behaviour of the current phasor is related to the shaft torsional mode and to the natural frequency of the generator itself The simulated generator rotor speed shown in Figure 27.12 illustrates this When representing the shaft system with the //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_28_CHA27.3D – 624 – [603–628/26] 17.12.2004 10:55PM Full-scale Verification of Dynamic Wind Turbine Models 624 2.0 1200 1.8 TM PSS/E 1100 fifth-order, two-mass model 1000 1.6 900 Voltage (V) Current (kA) 1.4 1.2 1.0 0.8 0.4 0 0.2 0.4 0.6 0.8 1.0 1.2 Time (s) 1.4 1.6 1.8 2.0 (b) 2.0 0.2 0.4 0.6 0.8 1.0 1.2 Time (s) TM 1.8 TM PSS/E 1.6 fifth-order, two-mass model 1.4 1.6 1.8 2.0 PSS/E fifth-order, two-mass model 1.6 1.4 1.4 Current (kA) Current (kA) 2.0 1.8 1.2 1.0 0.8 1.2 1.0 0.8 0.6 0.6 0.4 0.4 0.2 0.2 (c) fifth-order, two-mass model 600 500 200 100 0.2 TM PSS/E 700 400 300 0.6 (a) 800 0.2 0.4 0.6 0.8 1.0 1.2 Time (s) 1.4 1.6 1.8 2.0 (d) 0.2 0.4 0.6 0.8 1.0 1.2 Time (s) 1.4 1.6 1.8 2.0 Figure 27.11 Computed behaviour of (a) current phasor and (b) voltage phasor, both computed with the dynamic wind turbine model containing the fifth-order generator model and the two-mass shaft model Current-phasor computed with the model containing (c) the fifth-order generator model and the lumped-mass model of the mechanical system and (d) the third-order generator model and the two-mass shaft model Reprinted from Akhmatov, V Analysis of Dynamic Behaviour of Electric Power Systems with Large Amount of Wind Power, Ph.D thesis, Technical University of Denmark, Kgs Lyngby, Denmark, copyright 2003, with permission from the copyright holder 0.010 005 PSS/E TM two-mass model 0.008 003 Speed deviation (p.u.) Speed deviation (p.u.) 004 002 001 –001 –002 0.006 0.004 0.002 –0.002 –0.004 –003 –0.006 –004 –0.008 –005 (a) 0.2 0.4 0.6 0.8 1.2 Time (s) 1.4 1.6 1.8 –0.01 (b) PSS/E 0.2 0.4 0.6 TM 0.8 1.2 Time (s) lumped-mass model 1.4 1.6 1.8 Figure 27.12 Computed behaviour of the generator rotor speed deviation (minus slip) for (a) the two-mass model and (b) the lumped-mass model Note: The curves are to different scales Reprinted from Akhmatov, V Analysis of Dynamic Behaviour of Electric Power Systems with Large Amount of Wind Power, Ph.D thesis, Technical University of Denmark, Kgs Lyngby, Denmark, copyright 2003, with permission from the copyright holder //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_28_CHA27.3D – 625 – [603–628/26] 17.12.2004 10:55PM Wind Power in Power Systems 625 two-mass model, the behaviour of the generator rotor speed shows oscillations with the natural frequency of around Hz This is similar to the current phasor behaviour The reason is the strong coupling between electrical and mechanical parameters in induction generators (Akhmatov, Knudsen and Nielsen 2000) When using the lumped-mass model, the generator rotor speed does not show such oscillating behaviour Consequently, there are no oscillations in the simulated current phasor To summarise this discussion, only the dynamic wind turbine model containing the fifth-order model of induction generators and the two-mass representation of the shaft system gives sufficiently accurate results This full-scale validation also demonstrates that the links between the various parts of the dynamic wind turbine model have been accurately implemented 27.3.5 Discrepancies between model and measurements Having shown that the dynamic wind turbine model with the fifth-order induction generator model and the two-mass model of the shaft system gives sufficiently accurate results, we will now evaluate and explain possible discrepancies between the simulations carried out with this model and the measurements First, the simulated results not include higher harmonics of the fundamental electric frequency, which are there in the measurements The reason is that this user-written model is implemented in the tool PSS/ETM, which is a fundamental frequency tool This is one of the restrictions of this dynamic wind turbine model and of the simulation tool Second, during reconnection, the fundamental frequency transients of the simulated current phasor are somewhat lower than in the measured phase current This can be explained by the fact that the induction motors and the no-load capacitor are modelled according to the interface of the tool PSS/ETM That means that during balanced transient events the fundamental frequency transients are neglected in the simulated current However, such transients have probably influenced the measured phase current during reconnection Last, the damping of the current oscillations is lower in the simulated behaviour than it is in the measured behaviour The reason is that the simulations are carried out with the damping coefficients set to zero However, there is always finite damping in the real wind turbine construction and its generator This validation shows only relatively small discrepancies between simulated and measured behaviour The discrepancies were minimised by accurately implementing the network model equivalent into the simulation tool and by accurate modelling 27.4 Conclusions Validation is an important and indispensable part of modelling electricity-producing wind turbines in dynamic simulation tools The dynamic wind turbine models are to be used for analysing power system stability in the context of connecting large amounts of wind power to the grid Therefore, the dynamic wind turbine models have to be based //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_28_CHA27.3D – 626 – [603–628/26] 17.12.2004 10:55PM 626 Full-scale Verification of Dynamic Wind Turbine Models on sufficiently accurate assumptions and to predict accurate responses of the wind farms to transient events in power networks There is no point in applying models that have not been validated and that are inaccurate The results presented here demonstrate that it is possible to develop and implement in existing simulation tools dynamic wind turbine models of sufficient accuracy and complexity The focus here was on the validation of the user-written dynamic model of Type A wind turbines implemented in the simulation tool PSS/ETM, for use by the Danish power company NESA Type A wind turbines may be considered the simplest among the existing wind turbine concepts Even for this concept, though, modelling and validation has been challenging The dynamic wind turbine model contains (a) the transient fifth-order model of induction generators, (b) the two-mass representation of the shaft system and (c) the aerodynamic rotor model and the model of blade-angle control The common thirdorder model of induction generators should not be used to investigate balanced transient events with a significant voltage drop The lumped-mass model of the shaft system does not predict accurately the interaction between the wind turbines and the electric power network during grid disturbances It is also important to choose the details of the aerodynamic rotor model in accordance with the target of the investigation The dynamic wind turbine model was validated against measurements and also against simulations Here, the partial validation has demonstrated that the main parts of the dynamic wind turbine model are developed and implemented in the simulation tool with sufficient accuracy The full-scale validation showed that all the parts of the complete model and also the links between them are represented with sufficient accuracy There were only relatively small discrepancies between the results of the dynamic wind turbine model and measurements There will always be such discrepancies because of uncertainties in the data of the wind turbine and the power network and restrictions of the simulation tool and the dynamic wind turbine model These discrepancies can be minimised by careful modelling Dynamic wind turbine models (of different wind turbine concepts) should be implemented, in the near future, in the existing simulation tools that are applied to analyse power system stability Individual validations are expensive and therefore simulation tool suppliers should commercialise and standardise the dynamic wind turbine models whenever possible In this way such models would become accessible to every user, which would be generally advantageous These standardised models should be validated in cooperation with wind turbine manufacturers, organisations involved in large wind farm projects or independent research organisations; the validation reports should also be published References [1] Akhmatov, V (2001) ‘Note Concerning the Mutual Effects of Grid and Wind Turbine Voltage Stability Control’, Wind Engineering 25(6) 367–371 [2] Akhmatov, V (2003a) Analysis of Dynamic Behaviour of Electric Power Systems with Large Amount of Wind Power, PhD thesis, Technical University of Denmark, Kgs Lyngby, Denmark, available at http://www.oersted.dtu.dk/eltek/res/phd/00-05/20030403-va.html //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_28_CHA27.3D – 627 – [603–628/26] 17.12.2004 10:55PM Wind Power in Power Systems 627 [3] Akhmatov, V (2003b) ‘On Mechanical Excitation of Electricity-producing Wind Turbines at Grid Faults’, Wind Engineering 27(4) 257–272 [4] Akhmatov, V., Knudsen, H., Nielsen, A H (2000) ‘Advanced Simulation of Windmills in the Electrical Power Supply’ International Journal of Electrical Power and Energy Systems 22(6) 421–434 [5] Akhmatov, V., Knudsen, H., Nielsen, A H., Pedersen, J K., Poulsen, N K (2003) ‘Modelling and Transient Stability of Large Wind Farms’, Electrical Power and Energy Systems, 25(2) 123–144 [6] Eltra (2000) ‘Specifications for Connecting Wind Farms to the Transmission Network’, ELT1999-411a, Eltra Transmission System Planning, Denmark [7] Hinrichsen, E N., Nolan, P J (1982) ‘Dynamics and Stability of Wind Turbine Generators’, IEEE Transactions on Power Apparatus and Systems, 101(8) 2640–2648 [8] Kazachkov, Y A., Feltes, J W., Zavadil, R (2003) ‘Modeling Wind Farms for Power System Stability Studies’ presented at IEEE Power Engineering Society General Meeting, Toronto, Canada [9] Knudsen, H., Akhmatov, V (1999) ‘Induction Generator Models in Dynamic Simulation Tools’, International Conference on Power System Transients IPST’99, Budapest, Hungary, pp 253–259 [10] Larsson, A., Thiringer, T (1995) ‘Measurements on and Modelling of Capacitor-connecting Transients on a Low-voltage Grid Equipped with Two Wind Turbines’, International Conference on Power System Transients IPST’95, Lisbon, Portugal, pp 184–188 [11] Øye, S (1986) ‘Unsteady Wake Effects Caused by Pitch-angle Changes’, in IEA R&D WECS Joint Action on Aerodynamics of Wind Turbines, 1st Symposium, London, U.K., pp 58–79 [12] Pedersen, J K., Akke, M., Poulsen, N K., Pedersen, K O H (2000) ‘Analysis of Wind Farm Islanding Experiment’, IEEE Transactions on Energy Conversion 15(1) 110–115 [13] Pedersen, J K., Pedersen, K O H Poulsen, N K., Akhmatov, V., Nielsen, A H (2003) ‘Contribution to a Dynamic Wind Turbine Model Validation from a Wind Farm Islanding Experiment’, Electric Power Systems Research 64(2) 41–51 [14] Raben, N., Donovan, M H., Jørgensen, E., Thisted, J., Akhmatov, V (2003) ‘Grid Tripping and Re-connection: Full-scale Experimental Validation of a Dynamic Wind Turbine Model’, Wind Engineering 27(3) 205–213 [15] Snel, H., Schepers, J G (Eds) (1995) ‘Joint Investigation of Dynamic Inflow Effects and Implementation of an Engineering Method’, ECN-C-94–107, Netherlands Energy Research Foundation, ECN, Petten, The Netherlands //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_28_CHA27.3D – 628 – [603–628/26] 17.12.2004 10:55PM //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_29_CHA28.3D – 629 – [629–652/24] 17.12.2004 10:51PM 28 Impacts of Wind Power on Power System Dynamics J G Slootweg and W L Kling 28.1 Introduction In most countries, the amount of wind power generation integrated into large-scale electrical power systems covers only a small part of the total power system load However, the amount of electricity generated by wind turbines is increasing continuously Therefore, wind power penetration in electrical power systems will increase in the future and will start to replace the output of conventional synchronous generators As a result, it may also begin to influence overall power system behaviour The impact of wind power on the dynamics of power systems should therefore be studied thoroughly in order to identify potential problems and to develop measures to mitigate those problems The dynamic behaviour of a power system is determined mainly by the generators Until now, nearly all power has been generated with conventional directly grid-coupled synchronous generators The behaviour of the grid-coupled synchronous generator under various circumstances has been studied for decades and much of what is to be known is known However, although this generator type used to be applied in wind turbines in the past, this is no longer the case Instead, wind turbines use other types of generators, such as squirrel cage induction generators or generators that are grid-coupled via power electronic converters The interaction of these generator types with the power system is different from that of a conventional synchronous generator As a consequence, wind turbines affect the dynamic behaviour of the power system in a way that might be different from synchronous generators Further, there are also differences in the interaction with the power system between the various wind turbine types presently applied, so that the various wind turbine types must be treated separately This also applies to the various wind park connection schemes that can be found discussed in the literature Wind Power in Power Systems Edited by T Ackermann Ó 2005 John Wiley & Sons, Ltd ISBN: 0-470-85508-8 (HB) //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_29_CHA28.3D – 630 – [629–652/24] 17.12.2004 10:51PM 630 Impacts on Power System Dynamics This chapter discusses the impact of wind power on power system dynamics First, we present the concepts of power system dynamics and of transient and small signal stability, as well as the most important currently used wind turbine types Then, we will study the impact of these wind turbine types and of various wind farm interconnection schemes on the transient stability of the power system For this, we will analyse the response to disturbances, such as voltage and frequency changes Finally, we will deal with the impact of wind power on the small signal stability of power systems and use eigenvalue analysis for this purpose 28.2 Power System Dynamics Power system dynamics investigates how a power system responds to disturbances that change the system’s operating point Examples of such disturbances are frequency changes because a generator trips or a load is switched in or disconnected; voltage drops due to a fault; changes in prime mover mechanical power or exciter voltage, and so on A disturbance triggers a response in the power system, which means that various properties of the power system, such as node voltages, branch currents, machine speeds and so on, start to change The power system is considered stable if the system reaches a new steady state and all generators and loads that were connected to the system before the disturbance are still connected The original power system is considered unstable if, in the new steady state, loads or generators are disconnected Two remarks must be made at this point First, when a system is stable, the new steady state can either be identical to or different from the steady state in which the system resided before the disturbance occurred This depends on the type of disturbance, the topology of the system and the controllers of the generators Second, that a power system is unstable does not necessarily mean that a disturbance leads to a complete blackout of the system Rather, the system’s topology is changed by protection devices that disconnect branches, loads and/or generators during the transient phenomenon, in order to protect these In most cases the changed system will be able to reach a new steady state, thus preventing a complete blackout However, although the ‘new’ system that results after the changes is stable, the ‘old’ system was unstable and stability has been regained by changing the system’s topology There are two different methods to investigate the dynamics of a power system in order to determine whether the system is stable or not The first method is time domain analysis The type of time domain analysis that we use here is also referred to as dynamics simulation, fundamental frequency simulation or electromechanical transient simulation (see also Chapter 24) This approach subjects the system to a disturbance after which its response (i.e the quantitative evolution of the system’s properties over time) is simulated In this way, it can be decided whether the system is stable or not In the case of instability, strategies can be designed to change the system’s topology in such a way that stability is regained with minimum consequences to loads and generators The second method is frequency domain analysis, also referred to as an analysis of the small signal properties of the system or as eigenvalue analysis Frequency domain analysis studies a linearised representation of the power system in a certain state The linearised representation makes it possible to draw conclusions as to how the power //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_29_CHA28.3D – 631 – [629–652/24] 17.12.2004 10:51PM Wind Power in Power Systems 631 system will respond in the analysed state to incremental changes in the state variables The results of the frequency domain analysis consist of the following: An overview of the system’s eigenvalues; in case of complex eigenvalues, the damping and frequency of the corresponding oscillation can be calculated from the eigenvalue, whereas in case of real eigenvalues the associated time constant can be calculated An overview of the participation factors for each of the eigenvalues or at least those eigenvalues that are considered of importance, because their damping is low, for instance; the participation factors contain information on the relationship between the calculated eigenvalues and the system’s state variables This information can be used to identify the state variables that can be used to affect a certain eigenvalue For more detailed information on the frequency domain analysis of a power system, including the calculation and the use of participation factors, see, for example, Kundur (1994) The two methods mentioned are complementary Dynamics simulation can analyse the complete response of a system to a disturbance However, the simulation contains information on the system’s response only to the specific disturbance that is studied For other disturbances new simulations have to be carried out Frequency domain analysis, in contrast, yields a complete overview of the response to an incremental change in any of the system’s state variables for the system in its current operating state However, owing to the nonlinearity of a power system, the results are valid only for the investigated power system topology and for small changes in the state variables If the topology of the power system is changed or the state variables change significantly (e.g by connecting or disconnecting a load or a line or by changing the operating points of the generators) different eigenvalues will result, because if the changes in state variables are large the linearised representation of the system will no longer be valid 28.3 Actual Wind Turbine Types The vast majority of wind turbines that are currently being installed use one of three main types of electromechanical conversion system The first type is known as the Danish concept In Chapter 4, the Danish concept is introduced as Type A.(1) An (asynchronous) squirrel cage induction generator is used to convert the mechanical energy into electricity Owing to the different operating speeds of the wind turbine rotor and the generator, a gearbox is necessary to match these speeds The generator slip slightly varies with the amount of generated power and is therefore not entirely constant However, because these speed variations are in the order of 1%, this wind turbine type is normally referred to as a constant-speed or fixed-speed wind turbine The Danish, or constant-speed, design is nowadays nearly always combined with stall control of the aerodynamic power, although pitch-controlled constant-speed wind turbine types have been built, too (1) For definitions of wind turbine Types A–D, see Section 4.2.3 //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_29_CHA28.3D – 632 – [629–652/24] 17.12.2004 10:51PM 632 Impacts on Power System Dynamics The second type uses a doubly fed induction generator instead of a squirrel cage induction generator, introduced in Chapter as Type C Similar to the previous type, it needs a gearbox The stator winding of the generator is coupled to the grid and the rotor winding to a power electronic converter, nowadays usually a back-toback voltage source converter with current control loops In this way, the electrical and mechanical rotor frequencies are decoupled, because the power electronic converter compensates the difference between mechanical and electrical frequency by injecting a rotor current with variable frequency Variable-speed operation becomes possible That means that the mechanical rotor speed can be controlled according to a certain goal function, such as energy yield maximisation or noise minimisation The rotor speed is controlled by changing the generator power in such a way that it equals the value derived from the goal function In this type of conversion system, the control of the aerodynamic power is usually performed by pitch control The third type is called the direct-drive wind turbine because it does not need a gearbox It corresponds to Type D in Chapter A low-speed multipole synchronous ring generator with the same rotational speed as the wind turbine rotor converts the mechanical energy into electricity The generator can have a wound rotor or a rotor with permanent magnets The stator is not coupled directly to the grid but to a power electronic converter This may consist of a back-to-back voltage source converter or a diode rectifier with a single voltage source converter The electronic converter makes it possible to operate the wind turbine at variable speed Similar to the case for Type C wind turbines, pitch control limits the mechanical power input The three main wind turbine types are illustrated in Figure 19.2 (page 421) Apart from the wind turbine types depicted in Figure 19.2, other types have been developed and used They include wind turbines with directly grid-coupled synchronous generators and with conventional synchronous as well as squirrel cage induction generators combined with a gearbox and coupled to the grid with a full-scale power electronic converter Currently, there are hardly any of these wind turbine types on the market, though, and we will not take them into consideration here 28.4 Impact of Wind Power on Transient Stability 28.4.1 Dynamic behaviour of wind turbine types 28.4.1.1 Constant-speed wind turbines Constant-speed wind turbines (Type A) use a directly grid-coupled squirrel cage induction generator to convert mechanical into electrical power The behaviour of a constantspeed wind turbine is determined by the intrinsic relationship between active power, reactive power, terminal voltage and the rotor speed of the squirrel cage induction generator This relationship can be studied using the network equivalent, depicted in Figure 28.1 (Kundur, 1994) In this figure, U is the voltage, I is current, s the slip, R the resistance and L the reactance The indices s, t, s, m and r stand for leakage, terminal, stator, mutual and rotor, respectively The values of the generator parameters are given in Table 25.3 (page 567) //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_29_CHA28.3D – 633 – [629–652/24] 17.12.2004 10:51PM Wind Power in Power Systems lt Ut Ls 633 jLsσ jLrσ lr jLm Rr Rr (1 – s) s Figure 28.1 Network p equivalent of squirrel cage induction generator Note: U ¼ voltage; ffiffiffiffiffiffiffi I ¼ current; s ¼ slip; j ¼ À1; L ¼ reactance; R ¼ resistance; subscripts s, t, s, m and r refer to leakage, terminal, stator, mutual and rotor, respectively Squirrel cage induction generators are likely to become unstable after a voltage drop (Van Cutsem and Vournas, 1998) The explanation for this is as follows Figure 19.3 (page 422) illustrates the relationship between active power output and rotor slip as well as between reactive power consumption and rotor slip with the terminal voltage Ut as a parameter It shows that: The lower the terminal voltage, the larger the absolute value of the rotor slip that corresponds to a certain amount of active power generation The larger the rotor slip, the larger the reactive power consumption If the generator terminal voltage drops (e.g because of a fault) only a small amount of electrical power can be fed into the grid because the generated electrical power is proportional to the terminal voltage However, the wind continues to supply mechanical power Owing to the resulting imbalance between supplied mechanical power and generated electrical power, the generator speeds up This results in a decreasing slip Once the fault is cleared, the squirrel cage induction generator draws a large amount of reactive power from the grid because of its high rotational speed, as can be seen in Figure 19.3(b) Owing to this reactive power consumption, it can happen that the terminal voltage recovers only relatively slowly after the fault is cleared However, if the generator terminal voltage is low, the electrical power generated at a given slip is lower than that at nominal terminal voltage, as shown in Figure 19.3(a) (page 422) If the rotor accelerates faster than the terminal voltage is restored, the reactive power consumption continues to increase This leads to a decrease in the terminal voltage and thus to a further deterioration of the balance between mechanical and electrical power and to a further acceleration of the rotor Eventually, the voltage at the wind turbine will collapse It may then be necessary to disconnect the turbine from the grid to allow the grid voltage to restore Depending on the design and the settings of its protection system, the wind turbine will either be stopped by its undervoltage protection or accelerate further and be stopped and disconnected by its overspeed protection It can only be reconnected after //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_29_CHA28.3D – 634 – [629–652/24] 17.12.2004 10:51PM 634 Impacts on Power System Dynamics restoration of the grid voltage in the affected parts of the network That may take several minutes, particularly if other protection systems were activated during the disturbance too The exact quantitative behaviour of the terminal voltage and the required restoration time depend on the actual wind speed, wind turbine characteristics, network topology and protection system settings Wherever possible, a fault should be removed from the system before the wind turbine becomes unstable because of the mechanism pointed out above Otherwise, a large amount of generation may be lost The fault should be cleared quickly in order to limit the overspeeding and, consequently, to limit the amount of reactive power that is consumed for restoring the voltage The time that is available to clear the fault before it leads to voltage and rotor speed instability is called the critical clearing time Akhmatov et al (2003) propose a number of countermeasures to prevent instability of constant speed wind turbines (see also Chapter 29): Constant-speed wind turbines, which are usually stall-controlled, can be equipped with pitch drives that quickly increase the pitch angle when an acceleration of the rotor is detected This reduces the mechanical power and consequently limits the rotor speed and the reactive power consumption after the fault, thus reducing the risk of instability The wind turbines can be equipped with a controllable source of reactive power [e.g a static condenser (STATCON) or static VAR compensator (SVC)] to deliver the reactive power required to increase the speed at which the voltage is restored Mechanical and/or electrical parameters of the wind turbine and the generator can be changed, but this often has the disadvantage of increased cost, reduced efficiency and a more complicated mechanical construction These measures aim either at reducing the amount of overspeeding during the fault or at supplying reactive power to accelerate voltage restoration after the fault Although the measures mitigate the problem, they not completely solve it: it originates from the working principle of an induction generator, which is not principally changed by the above measures It should be noted that constant-speed wind turbines may become unstable at times other than after a fault The above sequence of events may also be initiated by a relatively small drop in terminal voltage This may be the result of a nearby synchronous generator tripping or a highly inductive load switching in, for instance When the wind turbine delivers its nominal power and the terminal voltage drops slightly, rotor speed will increase, because a larger slip is required to deliver nominal power at a terminal voltage below nominal This leads to an increase in the reactive power consumption, which in turn results in a further decrease in terminal voltage This can lead to a voltage collapse that is not preceded by a short circuit This is an example of voltage instability The response of a squirrel cage induction generator to changes in grid frequency is similar to that of synchronous generators The frequency of the stator field is identical to the grid frequency divided by the number of pole pairs of the generator If this frequency changes, the mechanical rotor frequency changes as well The change in energy stored in the rotating mass, which is caused by the rotor speed change is either fed into the system (in the case of a drop in grid frequency) or drawn from the system (in the case of an increase in grid frequency) //INTEGRAS/KCG/PAGINATION/WILEY/WPS/FINALS_14-12-04/0470855088_29_CHA28.3D – 635 – [629–652/24] 17.12.2004 10:51PM Wind Power in Power Systems 635 However, there is an important difference in the responses of synchronous generators in power plants and of squirrel cage induction generators that are used in constant-speed wind turbines In conventional power plants, a controllable prime mover is used If there is a frequency drop, the mechanical power applied to the generator can be adjusted in order to counteract this frequency drop When the frequency increases, the prime mover power is reduced, whereas when the frequency decreases, the prime mover power is increased In wind turbines, this is not possible because the wind cannot be controlled Thus, although constant-speed wind turbines tend to damp frequency deviations by either releasing energy from or storing energy in the rotating mass, the effect is weaker than in the case of synchronous generators in power plants We would like to stress here that the different responses are not due to the different generator types but to the fact that in power plants the prime mover can be adjusted to counteract frequency changes This is normally not possible in wind turbines 28.4.1.2 Variable-speed wind turbines The dynamic behaviour of variable-speed wind turbines is fundamentally different from that of constant-speed wind turbines Variable-speed wind turbines use a power electronic converter to decouple mechanical frequency and electrical grid frequency This decoupling takes place not only during normal operation but also during and after disturbances Power electronic components are very sensitive to overcurrents because of their very short thermal time constants, as mentioned in Chapter 25 When a drop in terminal voltage occurs, the current through the semiconductors increases very quickly The controller of the power electronic converter samples many quantities, such as terminal voltage, converter current, grid frequency and so on, at a high sampling frequency, in the order of kilohertz A fault is therefore noticed instantly by the power electronic converter The variable-speed wind turbine is then quickly disconnected in order to prevent damage to the converter If a power system has a high penetration of wind power from variable-speed turbines, this effect is naturally undesirable If the variable-speed wind turbines are disconnected at a relatively small voltage drop, a large amount of generation might be lost Such a situation may arise when a fault in the high-voltage transmission grid causes a voltage drop in a large geographical region This would lead to severe problems with the power balance in the associated control area or even on the system level Therefore, grid companies with large amounts of wind power presently revise their connection requirements for wind turbines, and are starting to require wind turbines to remain connected to the grid during a fault (E.On Netz, 2001; see also Chapter 7) This does not seem to be a major problem, though The literature presents approaches where the semiconductors are controlled in such a way that during voltage drops the current is limited to the nominal value (Petterson, 2003; Saccomando, Svensson and Sannino, 2002) Presently, most variable-speed wind turbines are not equipped with current controllers for a continued operation during voltage drops However, if system interaction were to require variable-speed wind turbines to have current controllers this would not be a very complicated issue ... 3.5 3.6 Introduction Power System History Current Status of Wind Power in Power Systems Network Integration Issues for Wind Power Basic Electrical Engineering Characteristics of Wind Power Generation... Modelling Issues 10.3.1 Future development of wind power 10.3.2 Wind regime 10.3.3 Wind power forecast models 10.3.4 Grid connection 10.3.5 Modelling of power systems with large-scale wind power. .. value 9.2.5 Grid investment value 9.3 The Value of Wind Power 9.3.1 The operating cost value of wind power 9.3.2 The capacity credit of wind power 9.3.3 The control value of wind power 9.3.4 The